Carview!

CARVIEW

MOTORHOMES

Select Language

HTTP/2 200 server: nginx date: Sat, 27 Dec 2025 16:21:47 GMT content-type: text/xml last-modified: Sun, 27 Apr 2025 09:51:47 GMT etag: W/"680dfe33-7a292" content-encoding: gzip still don't have a titlehttps://venthur.de/2025-01-12T16:00:00+01:00Bastian Venthur's BlogInvestigating the popularity of Python build backends over time (II)2025-01-12T16:00:00+01:002025-01-12T16:00:00+01:00Bastian Venthurtag:venthur.de,2025-01-12:/2025-01-12-build-backends.html

Last year, I analyzed the popularity of build backends used in pyproject.toml files over time. This post is the update for 2024.

<a href="2024-01-26-build-backends.html">Last year</a>, I analyzed the popularity of build backends used in <code>pyproject.toml</code> files over time. This post is the update for 2024. <h2>Analysis</h2> Like last year, I’m using <a href="https://tomforb.es/">Tom Forbes’</a> fantastic <a href="https://py-code.org/datasets#metadata">dataset</a> containing information about every file within every release uploaded to <a href="https://pypi.org">PyPI</a>. To get the current dataset, I followed the <a href="https://github.com/venthur/pypi-analyze">same process</a> as in <a href="2024-01-26-build-backends.html">last year’s analysis</a>, so I won’t repeat all the details here. Instead, I’ll highlight the main steps: <ul> <li>Download the parquet files from the <a href="https://py-code.org/datasets#metadata">dataset</a></li> <li>Use <a href="https://duckdb.org/">DuckDB</a> to query the parquet files, extracting the project name, upload date, the <code>pyproject.toml</code> file, and its hash for each upload</li> <li>Download each <code>pyproject.toml</code> file and extract the build backend. To avoid redundant downloads, I stored a mapping of the file hash and their respective build backend</li> </ul> Downloading all the parquet files took roughly a week due to <a href="https://github.com/">GitHub’s</a> rate limiting. <a href="https://tomforb.es/">Tom</a> suggested leveraging the Git v2 protocol to <a href="https://git-scm.com/docs/protocol-v2#_fetch">fetch</a> the data directly. This approach could bypass rate limiting and complete the download of all <code>pyproject.toml</code> files in just 20 minutes(!). However, I couldn’t find sufficient documentation that would help me to implement this method, so this will have to wait until next year’s analysis. Once all the data is downloaded, I perform some preprocessing: <ul> <li>Grouped the top 4 build backends by their absolute number of uploads and categorized the remaining ones as “other”</li> <li>Binned upload dates into quarters to reduce clutter in the resulting graphs</li> </ul> <h2>Results</h2> I modified the plots a bit from last year to make them easier to read. Most notably, I binned the data into quarters to make the plots less noisy, and secondly, I stopped stacking the relative distribution plots to make the percentages directly readable. The first plot shows the absolute number of uploads (in thousands) by quarter and build backend. <a href="images/2025-absolute.png"><img alt="Absolute distribution of build backends by quarter" src="images/2025-absolute.png" title="Absolute distribution of build backends by quarter"></a> The second plot shows the relative distribution of build backends by quarter. <a href="images/2025-relative.png"><img alt="Relative distribution of build backends by quarter" src="images/2025-relative.png" title="Relative distribution of build backends by quarter"></a> In 2024, we observe that: <ul> <li><a href="https://setuptools.pypa.io/">Setuptools</a> continues to grow in absolute numbers and remains around the 50% mark in relative distribution</li> <li><a href="https://python-poetry.org/">Poetry</a> maintains a 30% relative distribution, but the trend has been declining since 2024-Q3. Preliminary data for 2025-Q1 (not shown here) supports this, suggesting that Poetry might be surpassed by <a href="https://hatch.pypa.io/">Hatch</a> in 2025, which showed a remarkable growth last year.</li> <li><a href="https://flit.pypa.io/">Flit</a> is the only build backend in this analysis whose absolute and relative numbers decreased in 2024. With a 5% relative distribution, it underlines the dominance of Setuptools, Poetry, and Hatch over the remaining build backends.</li> </ul> The script for downloading and analyzing the data is available in my <a href="https://github.com/venthur/pypi-analyze">GitHub repository</a>. If someone has insights or examples on implementing the Git v2 protocol to download the <code>pyproject.toml</code> file given the repository URL and its hash, I’d love to hear from you!New python-debianbts in experimental2024-05-08T15:00:00+02:002024-05-08T15:00:00+02:00Bastian Venthurtag:venthur.de,2024-05-08:/2024-05-08-new-python-debianbts-in-experimental.html

Thanks to Peter De Wachter, python-debianbts no longer depends on an external SOAP library

<a href="2024-04-20-wsdl-for-debbugs.html">Last month</a>, I asked for help migrating <a href="https://github.com/venthur/python-debianbts">python-debianbts</a> to zeep, a SOAP library that would replace the now unmaintained pysimplesoap. The main blocker was the lack of a proper WSDL file provided by Debian’s BTS software, debbugs. Peter De Wachter pointed out other issues with debbugs’ SOAP implementation and provided a <a href="https://github.com/venthur/python-debianbts/pull/60">patch</a> that solves the whole issue by removing the dependency on an external SOAP library altogether by implementing the required bits of the SOAP protocol directly in python-debianbts. The new version is completely backwards compatible, and the test suites of python-debianbts and reportbug are still passing. Apparently, while working on this patch, Peter also uncovered an issue with the type hints defined in Python’s <code>xml.etree</code> module, for which he also provided a <a href="https://github.com/python/typeshed/pull/11841">patch</a>. Really great work! I’ve uploaded the new version to Debian/experimental for now to get some exposure and feedback, before uploading it to unstable. Peter, thank you very much for your support, I really appreciate it!Help needed: creating a WSDL file to interact with debbugs2024-04-20T14:00:00+02:002024-04-20T14:00:00+02:00Bastian Venthurtag:venthur.de,2024-04-20:/2024-04-20-wsdl-for-debbugs.html

I need help from someone with experience in SOAP and debbugs to create a WSDL file for Debian's Bug Tracking System

I am upstream and <a href="https://packages.debian.org/sid/python3-debianbts">Debian package</a> maintainer of <a href="https://github.com/venthur/python-debianbts">python-debianbts</a>, which is a Python library that allows for querying Debian’s Bug Tracking System (BTS). python-debianbts is used by reportbug, the standard tool to report bugs in Debian, and therefore the glue between the reportbug and the BTS. <a href="https://en.wikipedia.org/wiki/Debbugs">debbugs</a>, the software that powers Debian’s BTS, provides a <a href="https://en.wikipedia.org/wiki/SOAP">SOAP</a> interface for querying the BTS. Unfortunately, SOAP is not a very popular protocol anymore, and I’m facing the second migration to another underlying SOAP library as they continue to become unmaintained over time. <a href="https://github.com/mvantellingen/python-zeep">Zeep</a>, the library I’m currently considering, requires a <a href="https://en.wikipedia.org/wiki/Web_Services_Description_Language">WSDL</a> file in order to work with a SOAP service, however, debbugs does not provide one. Since I’m not familiar with WSDL, I need help from someone who can create a WSDL file for debbugs, so I can migrate python-debianbts away from <a href="https://github.com/pysimplesoap/pysimplesoap">pysimplesoap</a> to zeep. <h2>How did we get here?</h2> Back in the olden days, reportbug was querying the BTS by <a href="2007-03-01-decoupling-presentation-from-logic.html">parsing its HTML output</a>. While this worked, it tightly coupled the user-facing presentation of the BTS with critical functionality of the bug reporting tool. The setup was fragile, prone to breakage, and did not allow changing anything in the BTS frontend for fear of breaking reportbug itself. In 2007, I started to work on <a href="2007-03-07-a-screencast-is-worth-a-thousand-pictures.html">reportbug-ng</a>, a user-friendly alternative to reportbug, targeted at users not comfortable using the command line. Early on, I decided to use the BTS’ SOAP interface instead of parsing HTML like reportbug did. 2008, I extracted the code that dealt with the BTS into a separate Python library, and after some collaboration with the reportbug maintainers, reportbug adopted python-debianbts in 2011 and has used it ever since. 2015, I was working on <a href="2015-08-18-please-help-to-port-python-debianbts-to-python3.html">porting python-debianbts to Python 3</a>. During that process, it turned out that its major dependency, SoapPy was pretty much unmaintained for years and blocking the Python3 transition. <a href="2015-10-13-thanks.html">Thanks to the help of Gaetano Guerriero</a>, who ported python-debianbts to pysimplesoap, the migration was unblocked and could proceed. In 2024, almost ten years later, pysimplesoap seems to be unmaintained as well, and I have to look again for alternatives. The most promising one right now seems to be <a href="https://github.com/mvantellingen/python-zeep">zeep</a>. Unfortunately, zeep requires a WSDL file for working with a SOAP service, which debbugs does not provide. <h2>How can you help?</h2> reportbug (and thus python-debianbts) is used by thousands of users and I have a certain responsibility to keep things working properly. Since I simply don’t know enough about WSDL to create such a file for debbugs myself, I’m looking for someone who can help me with this task. If you’re familiar with SOAP, WSDL and optionally debbugs, please <a href="mailto:venthur@debian.org">get in touch</a> with me. I don’t speak Pearl, so I’m not really able to read debbugs code, but I do know some things about the SOAP requests and replies due to my work on python-debianbts, so I’m sure we can work something out. There is a <a href="https://git.savannah.gnu.org/cgit/emacs/elpa.git/plain/Debbugs.wsdl?h=externals/debbugs">WSDL file</a> for a debbugs version used by <a href="https://www.gnu.org/">GNU</a>, but I don’t think it’s official and it currently does not work with zeep. It may be a good starting point, though. <h2>The future of debbugs’ API</h2> While we can probably continue to support debbugs’ SOAP interface for a while, I don’t think it’s very sustainable in the long run. A simpler, well documented REST API that returns JSON seems more appropriate nowadays. The queries and replies that debbugs currently supports are simple enough to design a REST API with JSON around it. The benefit would be less complex libraries on the client side and probably easier maintainability on the server side as well. debbugs’ maintainer seemed to be in agreement with this idea <a href="https://lists.debian.org/debian-debbugs/2018/12/msg00001.html">back in 2018</a>. I created an attempt to define a <a href="https://github.com/venthur/debbugs-proposal/blob/master/yaml-unresolved/swagger.yaml">new API</a> (<a href="https://app.swaggerhub.com/apis-docs/venthur/debbugs/1.0.0">HTML render</a>), but somehow we got stuck and no progress has been made since then. I’m still happy to help shaping such an API for debbugs, but I can’t really implement anything in debbugs itself, as it is written in Perl, which I’m not familiar with.Investigating popularity of Python build backends over time2024-01-26T19:00:00+01:002024-01-26T19:00:00+01:00Bastian Venthurtag:venthur.de,2024-01-26:/2024-01-26-build-backends.html

Inspired by a Mastodon post, I wanted to investigate how the popularity of build backends used in pyproject.toml files evolved since the introduction of PEP-0517 in 2015.

Inspired by a <a href="https://framapiaf.org/@fcodvpt/111533565837454708">Mastodon post</a> by Françoise Conil, who investigated the current popularity of build backends used in <code>pyproject.toml</code> files, I wanted to investigate how the popularity of build backends used in <code>pyproject.toml</code> files evolved over the years since the introduction of <a href="https://peps.python.org/pep-0517/">PEP-0517</a> in 2015. <h2>Getting the data</h2> <a href="https://tomforb.es/">Tom Forbes</a> provides a huge <a href="https://py-code.org/datasets#metadata">dataset</a> that contains information about every file within every release uploaded to <a href="https://pypi.org">PyPI</a>. To get the current dataset, we can use: <div class="codehilite"><pre><code>curl -L --remote-name-all $(curl -L "https://github.com/pypi-data/data/raw/main/links/dataset.txt") </code></pre></div> This will download approximately 30GB of parquet files, providing detailed information about each file included in a PyPI upload, including: <ol> <li>project name, version and release date</li> <li>file path, size and line count</li> <li>hash of the file</li> </ol> The dataset does not contain the actual files themselves though, more on that in a moment. <h2>Querying the dataset using duckdb</h2> We can now use <a href="https://duckdb.org/">duckdb</a> to query the parquet files directly. Let’s look into the schema first: <div class="codehilite"><pre><code>describe select * from '*.parquet'; ┌─────────────────┬─────────────┬─────────┐ │ column_name │ column_type │ null │ │ varchar │ varchar │ varchar │ ├─────────────────┼─────────────┼─────────┤ │ project_name │ VARCHAR │ YES │ │ project_version │ VARCHAR │ YES │ │ project_release │ VARCHAR │ YES │ │ uploaded_on │ TIMESTAMP │ YES │ │ path │ VARCHAR │ YES │ │ archive_path │ VARCHAR │ YES │ │ size │ UBIGINT │ YES │ │ hash │ BLOB │ YES │ │ skip_reason │ VARCHAR │ YES │ │ lines │ UBIGINT │ YES │ │ repository │ UINTEGER │ YES │ ├─────────────────┴─────────────┴─────────┤ │ 11 rows 6 columns │ └─────────────────────────────────────────┘ </code></pre></div> From all files mentioned in the dataset, we only care about <code>pyproject.toml</code> files that are in the project’s root directory. Since we’ll still have to download the actual files, we need to get the <code>path</code> and the <code>repository</code> to construct the corresponding URL to the mirror that contains all files in a bunch of huge git repositories. Some files are not available on the mirrors; to skip these, we only take files where the <code>skip_reason</code> is empty. We also care about the timestamp of the upload (<code>uploaded_on</code>) and the <code>hash</code> to avoid processing identical files twice: <div class="codehilite"><pre><code>select path, hash, uploaded_on, repository from '*.parquet' where skip_reason == '' and lower(string_split(path, '/')[-1]) == 'pyproject.toml' and len(string_split(path, '/')) == 5 order by uploaded_on desc </code></pre></div> This query runs for a few minutes on my laptop and returns ~1.2M rows. <h2>Getting the actual files</h2> Using the <code>repository</code> and <code>path</code>, we can now construct an URL from which we can fetch the actual file for further processing: <div class="codehilite"><pre><code>url = f"https://raw.githubusercontent.com/pypi-data/pypi-mirror-{repository}/code/{path}" </code></pre></div> We can download the individual <code>pyproject.toml</code> files and parse them to read the <code>build-backend</code> into a dictionary mapping the file-<code>hash</code> to the build backend. Downloads on <a href="https://github.com/">GitHub</a> are rate-limited, so downloading 1.2M files will take a couple of days. By skipping files with a hash we’ve already processed, we can avoid downloading the same file more than once, cutting the required downloads by circa 50%. <h2>Results</h2> Assuming the data is complete and my analysis is sound, these are the findings: There is a surprising amount of build backends in use, but the overall amount of uploads per build backend decreases quickly, with a long tail of single uploads: <div class="codehilite"><pre><code>>>> results.backend.value_counts() backend setuptools 701550 poetry 380830 hatchling 56917 flit 36223 pdm 11437 maturin 9796 jupyter 1707 mesonpy 625 scikit 556 ... postry 1 tree 1 setuptoos 1 neuron 1 avalon 1 maturimaturinn 1 jsonpath 1 ha 1 pyo3 1 Name: count, Length: 73, dtype: int64 </code></pre></div> We pick only the top 4 build backends, and group the remaining ones (including <a href="https://pdm-project.org/latest/">PDM</a> and <a href="https://www.maturin.rs/">Maturin</a>) into “other” so they are accounted for as well. The following plot shows the relative distribution of build backends over time. Each bin represents a time span of 28 days. I chose 28 days to reduce visual clutter. Within each bin, the height of the bars corresponds to the relative proportion of uploads during that time interval: <a href="images/2024-relative.png"><img alt="Relative distribution of build backends over time" src="images/2024-relative.png" title="Relative distribution of build backends over time"></a> Looking at the right side of the plot, we see the current distribution. It confirms Françoise’s findings about the current popularity of build backends: <ul> <li>Setuptools: ~50%</li> <li>Poetry: ~33%</li> <li>Hatch: ~10%</li> <li>Flit: ~3%</li> <li>Other: ~4%</li> </ul> Between 2018 and 2020 the graph exhibits significant fluctuations, due to the relatively low amount uploads utizing <code>pyproject.toml</code> files. During that early period, <a href="https://flit.pypa.io/">Flit</a> started as the most popular build backend, but was eventually displaced by <a href="https://setuptools.pypa.io/">Setuptools</a> and <a href="https://python-poetry.org/">Poetry</a>. Between 2020 and 2020, the overall usage of <code>pyproject.toml</code> files increased significantly. By the end of 2022, the share of <a href="https://setuptools.pypa.io/">Setuptools</a> peaked at 70%. After 2020, other build backends experienced a gradual rise in popularity. Amongh these, <a href="https://hatch.pypa.io/">Hatch</a> emerged as a notable contender, steadily gaining traction and ultimately stabilizing at 10%. We can also look into the absolute distribution of build backends over time: <a href="images/2024-absolute.png"><img alt="Absolute distribution of build backends over time" src="images/2024-absolute.png" title="Absolute distribution of build backends over time"></a> The plot shows that Setuptools has the strongest growth trajectory, surpassing all other build backends. Poetry and Hatch are growing at a comparable rate, but since Hatch started roughly 4 years after Poetry, it’s lagging behind in popularity. Despite not being among the most widely used backends anymore, Flit maintains a steady and consistent growth pattern, indicating its enduring relevance in the Python packaging landscape. The script for downloading and analyzing the data can be found in my <a href="https://github.com/venthur/pypi-analyze">GitHub repository</a>. It contains the results of the duckb query (so you don’t have to download the full dataset) and the pickled dictionary, mapping the file hashes to the build backends, saving you days for downloading and analyzing the <code>pyproject.toml</code> files yourself.dotenv-cli update2023-07-01T15:00:00+02:002023-07-01T15:00:00+02:00Bastian Venthurtag:venthur.de,2023-07-01:/2023-07-01-dotenv-cli-update.html

dotenv-cli now ues exec on POSIX systems to create new processes

Thanks to <a href="https://github.com/venthur/dotenv-cli/pull/31">Nicholas Guriev</a>, <a href="https://github.com/venthur/dotenv-cli">dotenv-cli</a> now uses <code>exec</code> instead of <code>popen</code> to create the new process on <code>POSIX</code> systems. As a <a href="2019-05-12-dotenv-cli.html">refresher</a>, dotenv-cli is a package that provides the <code>dotenv</code> command. <code>dotenv</code> reads the <code>.env</code> file from the current directory, puts the contents in the environment variables, and executes the given command with the extra environment variables set. <code>dotenv</code> comes in handy if you follow the <a href="https://12factor.net/config">12 factor app</a> methodology or just need to run a program locally with specific environment variables set. With this new change, when you call <div class="codehilite"><pre><code>dotenv my_awesome_tool </code></pre></div> instead of forking a new process for <code>my_awesome_tool</code>, effectively creating a child process of <code>dotenv</code>, <code>dotenv</code> now uses <code>exec</code> to “become” the new process. This is a bit cleaner, as there is no longer a <code>dotenv</code>-process running that you don’t actually care about, and less-error prone when trying to send signals such as <code>SIGTERM</code> to the processes that runs <code>my_awesome_tool</code>. Unfortunately, <code>exec</code> does not work properly under Windows, so here we still fall back to using <code>popen</code>. This new feature is available in version 3.2.0 which is available on <a href="https://pypi.org/project/dotenv-cli/">PyPI</a> and <a href="https://packages.debian.org/sid/python3-dotenv-cli">debian/unstable</a>.Blag 2.0 released2023-06-17T16:00:00+02:002023-06-17T16:00:00+02:00Bastian Venthurtag:venthur.de,2023-06-17:/2023-06-17-blag.html

I released a major update on blag, which introduces a few backwards incompatible changes and many improvements.

A few days ago, I released a major update on <a href="https://github.com/venthur/blag">blag</a>, my blog-aware static-site generator, which introduces a few backwards-incompatible changes and many improvements over the old version. <h2>Good-looking default theme</h2> The old bare-bones default theme has been replaced with a good-looking one, based on the one used on <a href="https://venthur.de">this blog</a>: <img alt="Blag Screenshot" src="images/blag.png"> It comes with a light- and dark theme that switches automatically based on the browser setting, as well as fitting light- and dark syntax highlighting themes for code blocks. <h2>Improved <code>quickstart</code></h2> The <code>blag quickstart</code> command has been improved. Additionally to generating the configuration, it now also populates the working directory with the <code>templates</code>-, <code>static</code>- and <code>content</code> directories, containing the updated default theme and a few content pages to get you started. <h2>No internal fallback templates anymore</h2> Related to the changes in <code>quickstart</code>, the internal fallback template has been removed, and blag now completely relies on the templates in the local <code>templates</code> directory. This makes it more transparent for the user what is happening while simplifying blag’s internal logic. However, this is a backwards incompatible change! In the case of a missing template, the user will be warned with a hint on how to obtain the missing template. <h2>Index and archive are now separate</h2> Previously, the front-page would always show the archive of all articles. This is not very useful when your blog contains more than a few dozen articles. With blag 2.0, the previous <code>archive</code> has been split into <code>index</code> and <code>archive</code>, where <code>index</code> is the front-page showing only the most recent 15 articles by default and linking to the <code>archive</code> which shows all articles. There’s also two corresponding templates in the <code>templates</code> directory. <h2>Miscellaneous</h2> <ul> <li>Various dependencies have been updated.</li> <li>blag’s <a href="https://blag.readthedocs.io">documentation</a> has been migrated from <a href="https://www.sphinx-doc.org/">Sphinx</a> to <a href="https://www.mkdocs.org/">MkDocs</a>, which is a bit more lightweight and easier to maintain</li> <li>A packaging <a href="https://github.com/venthur/blag/issues/103">issue</a> has been fixed, where the <code>tests/conftest.py</code> was missing in the source distribution</li> </ul> Blag 2.0 is available on <a href="https://pypi.org/project/blag/">pypi</a>, <a href="https://packages.debian.org/sid/blag">debian/unstable</a> and <a href="https://github.com/venthur/blag">github</a>The State of Python Packaging in 20222022-12-18T20:00:00+01:002022-12-18T20:00:00+01:00Bastian Venthurtag:venthur.de,2022-12-18:/2022-12-18-python-packaging.html

Every year, I revisit the current best practices for Python packaging. This is my summary for 2022.

Every year or so, I revisit the current best practices for Python packaging. <a href="2021-06-26-python-packaging.html">This</a> was my summary for 2021 – here’s the update for 2022. <h2>PyPA</h2> <a href="https://www.pypa.io/en/latest/">PyPA</a> is still the place to go for information, best practices and tutorials for packaging Python projects. My only criticism from last year, namely that PyPA was heavily biased towards their own tooling (e.g. <code>pipenv</code>), has been addressed: the <a href="https://packaging.python.org/guides/tool-recommendations">tool recommendations</a> section lists now several tools for the same purpose with their own ones not necessarily being the first anymore. <h2>setup.py, setup.cfg, requirements.txt, Pipfile, pyproject.toml – oh my!</h2> This is the reason why I’m revisiting the documentation every year, to see what’s the current way to go. Good progress has been made since last year: <h3>Bye setup.py and setup.cfg – hello pyproject.toml</h3> <code>pyproject.toml</code> finally got mature enough to replace <code>setup.py</code> and <code>setup.cfg</code> in most cases. Recent versions of <code>setuptools</code> and <code>pip</code> now fully support <code>pyproject.toml</code> and even PyPA’s packaging tutorial completely switched their example project from away <code>setup.py</code> towards <code>pyproject.toml</code>, making it an official recommendation. So, now you can replace your <code>setup.py</code> with <code>pyproject.toml</code>. If you had already some kind of declarative configuration in <code>setup.cfg</code> you can move that as well into <code>pyproject.toml</code>. Most tools, like <code>pypy</code> or <code>pytest</code> also support configuration in <code>pyproject.toml</code> (flake8 being a notable exception…) so there’s no reason to keep <code>setup.cfg</code> around anymore. Actually, if you migrate to <code>pyproject.toml</code> it is best to do it properly and remove <code>setup.py</code> and <code>setup.cfg</code> as <code>setuptools</code> behaves a bit buggy when building a package that has either of them and the <code>pyproject.toml</code>. <h3>requirements.txt</h3> <code>requirements.txt</code> are still needed if you develop a “deployable” application (vs. a library) and want to provide pinned dependencies, i.e. with the specific versions that you’ve tested your application with. Usually, the list of requirements in <code>requirements.txt</code> is the same as defined in <code>pyproject.toml</code>, but with pinned versions. <h3>Pipfile + Pipfile.lock</h3> I still do completely ignore <code>Pipfile</code> and <code>Pipfile.lock</code> as they are only supported by <code>pipenv</code> and not backed by any standard. <h2>Summary</h2> The major change this year was the proper support of <code>pyproject.toml</code>. I am slowly replacing all <code>setup.py</code> and <code>setup.cfg</code> in my projects with <code>pyproject.toml</code> and haven’t discovered any issues yet. Even packaging those packages as Debian packages is well-supported by Debian’s tooling. I’m still running a quite boring stack based on <code>pyproject.toml</code> and <code>requirements.txt</code>, ignoring more advanced tools like <code>poetry</code> or such for dependency management. My build-system defined in <code>pyproject.toml</code> requires <code>setuptools</code> and I’m using <a href="https://github.com/pypa/build">build</a> for building and <a href="https://github.com/pypa/twine">twine</a> for uploading. Since PyPA changed the packaging tutorial towards <code>pyproject.toml</code> and away from <code>setup.py</code>, I think we will slowly see <code>setup.py</code> and <code>setup.cfg</code> go away over the years. Speaking of PyPA, I’m happy that they changed their attitude towards a more unbiased recommendation of tooling.Mastodon2022-11-12T17:00:00+01:002022-11-12T17:00:00+01:00Bastian Venthurtag:venthur.de,2022-11-12:/2022-11-12-mastodon.html

Due to recent events around Twitter, I finally decided to give Mastodon a try.

Due to recent events around Twitter, I finally decided to give <a href="https://joinmastodon.org">Mastodon</a> a try. Naturally, I find the idea of an open and decentralized platform much more appealing than the privately owned walled gardens that became so hugely popular in the past two decades. I’m curious whether Mastodon can keep up the momentum of the last two weeks and eventually establish itself as an alternative to Twitter. On that note, I think it will be interesting to see how well moderation of hate speech etc. works- and scales on Mastodon. I believe that Twitter, Facebook and the likes, spend a significant amount of resources on content moderation, so this may become a huge headache for Mastodon instance admins when it gets more popular. <h3>Choosing an Instance</h3> For no particular reason, I went with the <a href="https://mastodon.social">mastodon.social</a> instance, so my handle is <a href="https://mastodon.social/@venthur">@venthur@mastodon.social</a>. After my first steps, I realized that choosing the instance does have an impact, particularly if you follow the local timeline. <a href="https://mastodon.social">mastodon.social</a> is currently one of the biggest instances and therefore the local timeline is very busy and the topics naturally very random. Maybe I’ll try out a more specialized instance such as <a href="https://fosstodon.org/about">fosstodon</a> at some point – one of the awesome features of Mastodon is that it actually supports the migration of accounts between instances! I wonder if there are plans to have an official Debian instance? <h3>First Impressions</h3> So far I’m quite happy with Mastodon. There’s good quality content and I already found a few people that I was following on Twitter here on Mastodon too. Interestingly, some of them actually are more active on Mastodon than on Twitter. But truth to be told: many of them do cross-post on both mediums, and most of the people I follow on Twitter are not on Mastodon yet. I do like the concepts of the local- and federated timelines, although they are quite busy. I like that you can set the language of your (individual) toots which allows users to filter their timelines for languages. In practice, this does not work so well yet, and I still see a lot of different languages in my local and federated timelines. I assume the problem is that people just forget to set the language of their toots so the default language is used. This problem should be easily solvable in the frontend by “guessing” the language once a few words have been typed. I also like the idea that you can follow hashtags, although for me the results were mixed. I tried to follow <code>#debian</code> and <code>#python</code> and got a lot of toots that were not really relevant for me in the case of <code>#python</code> I got quite spammed with job ads so I had to unfollow it again. Unfollwing a hashtag is not very intuitive, as the tags are not in your list of people you follow, so you have to find the page of the hashtag itself (e.g. <a href="https://mastodon.social/tags/debian">#debian</a>) and unfollow from there. You see, there are some rough edges here and there, but I find the overall experience much more enjoyable than Twitter. <h3>Debian Folks in the Fediverse?</h3> Which brings me to: are there any Debian Developers or -Maintainers out there to follow in the <a href="https://en.wikipedia.org/wiki/Fediverse">fediverse</a>? I found most of the ones I’m following on Twitter, but I’m sure there’s more hiding out there.Keychron keyboards fixed on Linux2022-08-01T22:00:00+02:002022-08-01T22:00:00+02:00Bastian Venthurtag:venthur.de,2022-08-01:/2022-08-01-keychron-keyboards-fixed-on-linux.html

Since Linux 5.19, the Function keys of Keychron keyboards are finally working out of the box.

<a href="2021-04-30-keychron-c1-on-linux.html">Last year</a>, I wrote about on how to get my buggy Keychron C1 keyboard working properly on Linux by setting a kernel module parameter. Afterwards, I <a href="https://lore.kernel.org/all/f82dd7a1-a5c6-b651-846c-29f6df9436af@redhat.com/">contacted</a> Hans de Goede since he was the last one that contributed a major patch to the relevant kernel module. After some debugging, it turned out that the Keychron keyboards are indeed misbehaving when set to Windows mode. Almost a year later, <a href="https://lore.kernel.org/all/20220505191221.36172-2-bryancain3@gmail.com/">Bryan Cain</a> provided a patch fixing the behavior, which has now been <a href="https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/diff/drivers/hid/hid-apple.c?id=v5.19&id2=v5.18">merged</a> to the Linux kernel in 5.19. Thank you, Hans and Bryan!blag is now available in Debian2022-06-18T18:00:00+02:002022-06-18T18:00:00+02:00Bastian Venthurtag:venthur.de,2022-06-18:/2022-06-18-blag-in-debian.html

After a year I finally got around creating a Debian package for blag.

<a href="2021-03-24-blogging.html">Last year</a>, I wrote my own blog-aware static site generator in Python. I called it “<a href="https://github.com/venthur/blag">blag</a>” – named after the <a href="https://blog.xkcd.com/">blag</a> of the webcomic xkcd. Now I finally got around packaging- and uploading blag to Debian. It passed the <a href="https://ftp-master.debian.org/new.html">NEW queue</a> and is now part of the distribution. That means if you’re using Debian, you can install it via: <div class="codehilite"><pre><code>sudo aptitude install blag </code></pre></div> Ubuntu will probably follow soon. For every other system, blag is also <a href="https://pypi.org/project/blag/">available on PyPI</a>: <div class="codehilite"><pre><code>pip install blag </code></pre></div> To get started, you can <div class="codehilite"><pre><code>mkdir blog && cd blog blag quickstart # fill out some info nvim content/hello-world.md # write some content blag build # build the website </code></pre></div> Blag is aware of articles and pages: the difference is that articles are part of the blog and will be added to the atom feed, the archive and aggregated in the tag pages. Pages are just rendered out to HTML. Articles and pages can be freely mixed in the <code>content</code> directory, what differentiates an article from a page is the existence of the <code>dade</code> metadata element: <div class="codehilite"><pre><code>title: My first article description: Short description of the article date: 2022-06-18 23:00 tags: blogging, markdown ## Hello World! Lorem ipsum. [...] </code></pre></div> blag also comes with a dev-server that rebuilds the website automatically on every change detected, you can start it using: <div class="codehilite"><pre><code>blag serve </code></pre></div> The default theme looks quite ugly, and you probably want to create your own styling to make it more beautiful. The process is not very difficult if you’re familiar with <a href="https://jinja.palletsprojects.com">jinja</a> templating. Help on that can be found in the “Templating” section of the online <a href="https://blag.readthedocs.io">documentation</a>, the offline version in the <code>blag-doc</code> package, or the man page, respectively. Speaking of the <code>blag-doc</code> package: packaging it was surprisingly tricky, and it also took me a lot of trial and error to realize that <code>dh_sphinxdocs</code> alone does not automatically put the generated html output into the appropriate package, you rather have to list them in the <code>package.docs</code>-file (i.e. <code>blag-doc.docs</code>) so <code>dh_installdocs</code> can install them properly.Managing dotfiles with GNU Stow2021-12-19T14:00:00+01:002021-12-19T14:00:00+01:00Bastian Venthurtag:venthur.de,2021-12-19:/2021-12-19-managing-dotfiles-with-stow.html

How to use GNU Stow to manage the symlinks to the dotfile repository.

Many developers manage their user-specific application configuration – also known as dotfiles – in a version control system such as <a href="https://git-scm.com/">git</a>. This allows for keeping track of changes and synchronizing the dotfiles across different machines. Searching on github, you’ll find <a href="https://github.com/search?q=dotfiles">thousands of dotfile repositories</a>. As your dotfiles are sprinkled all over your home directory, managing them in a single repository is not trivial, i.e. how do you make sure that your <code>.bashrc</code>, <code>.tmux.conf</code>, etc. that life in your dotfile repository appear in the proper places in your home directory? The most common solution is to use <a href="https://en.wikipedia.org/wiki/Symbolic_link">symlinks</a> so that the <code>.tmux.conf</code> in your home directory is just a symlink pointing to the appropriate file in your dotfile repository: <div class="codehilite"><pre><code>$ ls -l ~/.tmux.conf lrwxrwxrwx 1 venthur venthur 34 18. Dez 22:53 /home/venthur/.tmux.conf -> git/dotfiles/tmux/.tmux.conf </code></pre></div> This leads immediately to another problem: how do you manage the symlinks? For the longest time I just manually maintained the symlinks on the various machines, but this approach does not scale well with the number of dotfiles and machines you’re using this repository on. Often, people write their own shell scripts that help them with the maintenance of the symlinks, but at least the solutions I’ve seen so far did not convince me. Last year I stumbled upon <a href="https://www.gnu.org/software/stow/">GNU Stow</a>, an unpretentious little tool that does not reveal at first sight how useful it would be for the job. The description on the website says: <blockquote> GNU Stow is a symlink farm manager which takes distinct packages of software and/or data located in separate directories on the filesystem, and makes them appear to be installed in the same place. </blockquote> Right. <h2>How does it work?</h2> In <code>stow</code>’s terminology, a package is a set of files and directories that need to be “installed” in a particular directory structure. The target directory is the root of the tree in which the package appear to be installed. When you “stow” a package, <code>stow</code> creates symlinks in the target directory that point into the package. Let’s say I have my dotfiles repository in <code>~/git/dotfiles/</code>. Within this repository, I have a <code>tmux</code> package, containing the <code>.tmux.conf</code> dotfile: <div class="codehilite"><pre><code>$ pwd /home/venthur/git/dotfiles $ find tmux tmux # the package tmux/.tmux.conf # the dotfile </code></pre></div> The target directory is my home directory, as this is where the symlinks need to be created. I can now stow the <code>tmux</code> package into the target directory like so: <div class="codehilite"><pre><code>$ stow --target=/home/venthur tmux </code></pre></div> and <code>stow</code> will create the appropriate symlinks to the contents of the package into the target directory: <div class="codehilite"><pre><code>$ ls -l ~/.tmux.conf lrwxrwxrwx 1 venthur venthur 34 2. Jun 2021 /home/venthur/.tmux.conf -> git/dotfiles/tmux/.tmux.conf </code></pre></div> Note that the name of the package (i.e. the name of the directory) does not matter as <code>stow</code> points the symlinks into the package, so you can choose it freely. I usually use the name of the program that the configuration belongs to as the package name. Your package can also contain several files or even a complex directory structure. Let’s look at the configuration for neovim, which lives below <code>~/.config/nvim/</code>: <div class="codehilite"><pre><code>$ pwd /home/venthur/git/dotfiles $ find neovim neovim neovim/.config neovim/.config/nvim neovim/.config/nvim/init.vim $ stow --target=/home/venthur neovim $ ls -l ~/.config/nvim lrwxrwxrwx 1 venthur venthur 41 2. Jun 2021 /home/venthur/.config/nvim -> ../git/dotfiles/neovim/.config/nvim </code></pre></div> At this point we should mention that the target directory for my dotfiles will always be my home directory, so the contents of the packages are either the configuration files or the directory structure as they live in my home directory. <h2>Deleting a package from the parent directory</h2> You can also remove (unstow) a package from the target directory again, using the <code>--delete</code> parameter: <div class="codehilite"><pre><code>$ ls -l ~/.tmux.conf lrwxrwxrwx 1 venthur venthur 34 18. Dez 22:53 /home/venthur/.tmux.conf -> git/dotfiles/tmux/.tmux.conf $ stow --target=/home/venthur --delete tmux/ $ ls -l ~/.tmux.conf ls: cannot access '/home/venthur/.tmux.conf': No such file or directory </code></pre></div> <h2>Stowing several packages at once</h2> Since your dotfile repository will likely contain more than one package, it makes sense to combine the individual <code>stow</code> commands into one, so instead of stowing everything individually, <div class="codehilite"><pre><code>$ stow --target=/home/venthur tmux $ stow --target=/home/venthur vim $ stow --target=/home/venthur neovim </code></pre></div> you can stow everything at once: <div class="codehilite"><pre><code>$ stow --target=/home/venthur */ </code></pre></div> Note that I use <code>*/</code> instead of <code>*</code> to match all directories (i.e. packages), since my dotfiles repository also contains a <code>README.md</code> and a <code>makefile</code>. <h2>Putting it all together</h2> <a href="https://github.com/venthur/dotfiles">My dotfiles</a> repository contains a <code>makefile</code> that allows me to create/update or delete all symlinks at once: <div class="codehilite"><pre><code>all: stow --verbose --target=$$HOME --restow */ delete: stow --verbose --target=$$HOME --delete */ </code></pre></div> The <code>--restow</code> parameter tells <code>stow</code> to unstow the packages first before stowing them again, which is useful for pruning obsolete symlinks from the target directory. Et voilà! Whenever I make a change in my dotfiles repository that involves creating or deleting a dotfile (or a package), I simply call: <div class="codehilite"><pre><code>$ make </code></pre></div> and everything is updated. To delete all dotfile-related symlinks from this machine, I simply run: <div class="codehilite"><pre><code>$ make delete </code></pre></div>The State of Python Packaging in 20212021-06-26T14:00:00+02:002021-06-26T14:00:00+02:00Bastian Venthurtag:venthur.de,2021-06-26:/2021-06-26-python-packaging.html

Every year, I revisit the current best practices for Python packaging. This is my 2021 edition.

Every year or so, I revisit the current best practices for Python packaging. I.e. the way you’re supposed to distribute your Python packages. The main source is <a href="https://packaging.python.org/">packaging.python.org</a> where the official packaging guidelines are. It is worth noting that the way you’re supposed to package your Python applications is not defined by Python or its maintainers, but rather delegated to a separate entity, the <a href="https://www.pypa.io/en/latest/">Python Packaging Authority (PyPA)</a>. <h2>PyPA</h2> PyPA does an excellent job providing us with information, best practices and tutorials regarding Python packaging. However, there’s one thing that irritates me every single time I revisit the page and that is the misleading recommendation of their own tool <a href="https://pipenv.pypa.io">pipenv</a>. Quoting from the <a href="https://packaging.python.org/guides/tool-recommendations">tool recommendations</a> section of the packaging guidelines: <blockquote> Use Pipenv to manage library dependencies when developing Python applications. See Managing Application Dependencies for more details on using pipenv. </blockquote> <a href="https://www.pypa.io/en/latest/">PyPA</a> recommends <a href="https://pipenv.pypa.io">pipenv</a> as the standard tool for dependency management, at least <a href="https://web.archive.org/web/20180504100940/https://packaging.python.org/guides/tool-recommendations/">since 2018</a>. A bold statement, given that <code>pipenv</code> only started in 2017, so the Python community cannot have had not enough time to standardize on the workflow around that tool. There have been no releases of <code>pipenv</code> between 2018-11 and 2020-04, that’s 1.5 years for the standard tool. In the past, <code>pipenv</code> also hasn’t been shy in pushing breaking changes in a fast-paced manner. PyPA still advertises <code>pipenv</code> all over the place and only mentions <a href="https://python-poetry.org">poetry</a> a couple of times, although <code>poetry</code> seems to be the more mature product. I understand that <code>pipenv</code> lives under the umbrella of PyPA, but I still expect objectiveness when it comes to tool recommendation. Instead of making such claims, they should provide a list of competing tools and provide a fair feature comparison. <h2>Distributions</h2> You would expect exactly one distribution for Python packages, but here in Python land, we have several ones. The most popular ones being <a href="https://pypi.org/">PyPI</a> – the official one – and <a href="https://www.anaconda.com/">Anaconda</a>. Anaconda is more geared towards data-scientists. The main selling point for Anaconda back then was that it provided pre-compiled binaries. This was especially useful for data-science related packages which depend on libatlas, -lapack, -openblas, etc. and need to be compiled for the target system. This problem has mostly been solved with the wide adoption of <a href="https://pythonwheels.com/">wheels</a>, but you still encounter some source-only uploads to PyPI that require you to build stuff locally on <code>pip install</code>. Of course there’s also the Python packages distributed by the Operating System, Debian in my case. While I was a firm believer in only using those packages provided by the OS in the very past, I moved to the opposite end of the spectrum throughout the years, and am only using the minimal packages provided by Debian to bootstrap my virtual environments (i.e. <code>pip</code>, <code>setuptools</code> and <code>wheel</code>). The main reason is outdated or missing libraries, which is expected – Debian cannot hope to keep up with all the upstream changes in the ecosystem, and that is by design and fine. However, with the recent upgrade of <a href="https://www.python.org/dev/peps/pep-0513/">manylinux</a>, even the <code>pip</code> provided by Debian/unstable was too outdated, so you basically had to <code>pip install --upgrade pip</code> for a while otherwise you’d end up compiling every package you’d try to install via <code>pip</code>. So I’m sticking to the official PyPI distribution wherever possible. However, compared to the Debian distribution it feels immature. In my opinion, there should be compiled wheels for all packages available that need it, built and provided by PyPI. Currently, the wheels provided are the ones uploaded by the upstream maintainers. This is not enough, as they usually build wheels only for one platform. Sometimes they don’t upload wheels in the first place, relying on the users to compile during install. Then you have <a href="https://www.python.org/dev/peps/pep-0513/">manylinux</a>, an excellent idea to create some common ground for a portable Linux build distribution. However, sometimes when a new version of manylinux is released some upstream maintainers immediately start supporting only that version, breaking a lot of systems. A setup similar to Debian’s where the authors only do a source-upload and the wheels are compiled on PyPI infrastructure for all available platforms, is probably the way to go. <h2>setup.py, setup.cfg, requirements.txt, Pipfile, pyproject.toml – oh my!</h2> This is the part I’m revisiting the documentation every year, to see what’s the current way to go. The main point of packaging your Python application is to define the package’s meta data and (build-) dependencies. <h3>setup.py + requirements.txt</h3> For the longest time, the <code>setup.py</code> and <code>requirements.txt</code> were (and, spoiler alert: still is) the backbone of your packaging efforts. In <code>setup.py</code> you define the meta data of your package, including its dependencies. If your project is a deployable application (vs. a library) you’ll very often provide an additional <code>requirements.txt</code> with pinned dependencies. Usually the list of requirements is the same as defined in <code>setup.py</code> but with pinned versions. The reason why you avoid version pinning in <code>setup.py</code> is that it would interfere with other pinned dependencies from other dependencies you try to install. <h3>setup.cfg</h3> <code>setup.cfg</code> is a configuration file that is used by many standard tools in the Python ecosystem. Its format is ini-style and each tools’ configuration lives in its own stanza. Since 2016 <a href="https://setuptools.readthedocs.io/en/latest/userguide/declarative_config.html">setuptools supports</a> configuring <code>setup()</code> using <code>setup.cfg</code> files. This was exciting news back then, however, it does not completely replace the <code>setup.py</code> file. While you can move most of the <code>setup.py</code> configuration into <code>setup.cfg</code>, you’ll still have to provide that file with an empty <code>setup()</code> in order to allow for editable <code>pip</code> installs. In my opinion, that makes this feature useless and I rather stick to <code>setup.py</code> with a properly populated <code>setup()</code> until that file can be completely replaced with something else. <h3>Pipfile + Pipflie.lock</h3> <a href="https://github.com/pypa/pipfile"><code>Pipfile</code> and <code>Pipfile.lock</code></a> are supposed to replace <code>requirements.txt</code> some day. So far they are not supported by <code>pip</code> or mentioned in any <a href="https://www.python.org/dev/peps/">PEP</a>. I think only <code>pipenv</code> supports them, so I’d ignore them for now. <h3>pyproject.toml</h3> <a href="https://www.python.org/dev/peps/pep-0518/">PEP 518</a> introduces the <code>pyproject.toml</code> file as a way to specify build requirements for your project. <a href="https://www.python.org/dev/peps/pep-0621/">PEP 621</a> defines how to store project meta data in it. <code>pip</code> and <code>setuptools</code> support <code>pyproject.toml</code> to some extent, but not to a point where it completely replaces <code>setup.py</code> yet. Many of Python’s standard tools allow already for configuration in <code>pyproject.toml</code> so it seems this file will slowly replace the <code>setup.cfg</code> and probably <code>setup.py</code> and <code>requirements.txt</code> as well. But we’re not there yet. <code>poetry</code> has an interesting approach: it will allow you to write everything into <code>pyproject.toml</code> and generate a <code>setup.py</code> for you at build-time, so it can be uploaded to PyPI. Ironically, Python settled for the TOML file format here, although there is currently no support for reading TOML files in Python’s standard library. <h2>Summary</h2> While some alternatives exist, in 2021 I still stick to <code>setup.py</code> and <code>requirements.txt</code> to define the meta data and dependencies of my projects. Regarding the tooling, <code>pip</code> and <code>twine</code> are sufficient and do their job just fine. Alternatives like <code>pipenv</code> and <code>poetry</code> exist. The scope of <code>poetry</code> seems to be better aligned with my expectations, and it seems the more mature project compared to <code>pipenv</code> but in any case I’ll ignore both of them until I revisit this issue in 2022. <h2>Closing Thoughts</h2> While the packaging in Python has improved a lot in the last years, I’m still somewhat put off how such a core aspect of a programming language is treated within Python. With some jealousy, I look over to the folks at Rust and how they seemed to get this aspect right from the start. What would in my opinion improve the situation? <ul> <li>Source only uploads and building of weels for all platforms on PyPI infrastructure – this way we could have wheels everywhere and remove the need to compile anything in <code>pip install</code></li> <li>Standard tooling: <code>pip</code> has been and still is the tool of choice for packaging in Python. For some time now, you also need <code>twine</code> in order to upload your packages. <code>setup.py upload</code> still exists, but hasn’t worked for months on my machines. It would be great to have something that improves the virtualenv handling and dependency management. We do have some tools with overlapping use-cases, like <code>poetry</code> and <code>pipenv</code>. <code>pipenv</code> is heavily advertised and an actual PyPA project, but it feels immature in terms of scope, release history (and emojis!) compared to <code>poetry</code>. <code>poetry</code> is gaining a lot of traction, but it is apparently not receiving much love from PyPA, which brings me to:</li> <li>Unbiased tool recommendations. I don’t understand why PyPA is trying so hard to make us believe <code>pipenv</code> would be the standard tool for Python packaging. Instead of making such claims, please provide a list of competitors and provide a fair feature comparison. PyPA is providing great packaging tutorials and is a valuable source of information around this topic. But when it comes to the tool recommendations, I do challenge PyPA’s objectiveness.</li> </ul>Getting the Function keys of a Keychron working on Linux2021-04-30T18:00:00+02:002021-04-30T18:00:00+02:00Bastian Venthurtag:venthur.de,2021-04-30:/2021-04-30-keychron-c1-on-linux.html

The Keychron C1 (and other Keychron models) do not work properly on Linux: the Function keys are not accessible at all, here's how to fix them.

Having destroyed the third Cherry Stream keyboard in 4 years, I wanted to try a more substantial keyboard for a change. After some research I decided that I want a mechanical, wired, tenkeyless keyboard without any fancy LEDs. At the end I settled for a <a href="https://www.keychron.com/products/keychron-c1-wired-mechanical-keyboard">Keychron C1</a> with red switches. It meets all requirements, looks very nice and the price is reasonable. <h2>Surprise!</h2> After the keyboard was delivered, I connected it to my Debian machine and was unpleasantly surprised to notice that the Function-keys did not work at all. The keyboard shares the Multimedia keys with the F-keys and you have an <code>fn</code> key that supposedly switches between the two modes, like you’re used to on a laptop. On Linux, however you cannot access the F-keys at all: pressing <code>fn + F1</code> or <code>F1</code> makes no difference, you’ll always get the Multimedia key. Switching the keyboard between “Windows” and “Mac” mode makes no difference either, in both modes the F-keys are not accessible. Apparently Keychron is aware of the problem, because the quick start manual tells you: <blockquote> “We have a Linux user group on facebook. Please search “Keychron Linux Group” on facebook. So you can better experience with our keyboard.” </blockquote> Customer support at its finest! So at this point you should probably just send the keyboard back, get the refund and buy a proper one with functioning F-keys. <h2>The fix</h2> Test if this command fixes the issue and enables the <code>Fn +</code> F-key-combos: <div class="codehilite"><pre><code># as root: echo 2 > /sys/module/hid_apple/parameters/fnmode </code></pre></div> Depending on the mode the keyboard is in, you should now be able to use the F-keys by simply pressing them, and the Multimedia keys by pressing <code>fn +</code> F-key (or the other way round). To switch the default mode of the F-keys to Function- or Multimedia-mode, press and hold <code>fn + X + L</code> for 4 seconds. If everything works as expected, you can make the change permanent by creating the file <code>/etc/modprobe.d/hid_apple.conf</code> and adding the line: <div class="codehilite"><pre><code>options hid_apple fnmode=2 </code></pre></div> and running <div class="codehilite"><pre><code># as root update-initramfs -u -k all </code></pre></div> This works regardless if the keyboard is in Windows- or Mac-mode, and that might hint at the problem: in both cases the Linux thinks you’re connecting a Mac keyboard. <h2>The rant</h2> Although the fix was not very hard to find and apply, this experience still leaves a foul taste. I naively assumed the problem of having a properly functioning keyboard that simply works when you plug it in, has been thoroughly solved by 2021. To make it worse, I assume Keychron must be aware of the problem because the other Keychron models have the <a href="https://www.google.com/search?q=keychron+linux+function+keys">same issue</a>! But instead of fixing it on their end, they forward you to a facebook “community” and expect you to somehow fix it yourself. So dear Keychron, you do make really beautiful keyboards! But before you release your next model with the same bug, maybe invest a bit on fixing the basics? I see that your keyboards support firmware updates for Windows and Mac – maybe you can talk to the folks over at the <a href="https://fwupd.org/">Linux Vendor Firmware Service</a> and support Linux as well? Maybe you can even fix this annoying bug for the keyboards you’ve already sold? I found it really cute that you sent different keycaps for Windows and Mac setups – a few disappointed Linux users might accept an apology in form of a Linux cap… <h2>Update (2022-08)</h2> Starting with version 5.19, Linux comes with a <a href="https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/diff/drivers/hid/hid-apple.c?id=v5.19&id2=v5.18">fix</a> for the Keychron keyboards, so all models should work as expected out of the box. The relevant <a href="https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=fa33382c7f74a1444f90f324007da1431d7180b2">commit message</a> says: <blockquote> Keychron’s C-series and K-series of keyboards copy the vendor and product IDs of an Apple keyboard, but only behave like that device when set to “Mac” mode. In “Windows” mode, the Fn key doesn’t generate a scancode, so it’s impossible to use the F1-F12 keys when fnmode is set to its default value of 1. To fix this, make fnmode default to the new value of 3, which behaves like fnmode=2 for Keychron keyboards and like fnmode=1 for actual Apple keyboards. This way, Keychron devices are fully usable in both “Windows” and “Mac” modes, while behavior is unchanged for everything else. </blockquote>Writing Makefiles for Python Projects2021-03-31T16:00:00+02:002021-03-31T16:00:00+02:00Bastian Venthurtag:venthur.de,2021-03-31:/2021-03-31-python-makefiles.html

Exploring two different strategies for using Makefiles in Python projects

I’m a big fan of <a href="https://www.gnu.org/software/make/manual/make.html">Makefiles</a>. Almost all <a href="https://github.com/venthur">my side projects</a> are using them, and I’ve been advocating their usage at work too. <code>Makefiles</code> give your contributors an entry point on how to do certain things like, building, testing, deploying. And if done correctly, they can massively simplify your CI/CD pipeline scripts as they can often just stupidly call the respective <code>make</code> targets. Most importantly, they are a very convenient shortcut for you as a developer as well. For Python projects, where I’m almost always using <a href="https://docs.python.org/3/tutorial/venv.html">virtual environments</a>, I’ve been using two different strategies for <code>Makefiles</code>: <ol> <li>assuming that <code>make</code> is executed inside the virtual environment</li> <li>wrapping all virtual environment calls inside <code>make</code></li> </ol> Both strategies have their pros and cons. <h2>Assuming <code>make</code> is executed inside the venv</h2> Let’s have a look at a very simple <code>Makefile</code> that allows for building, testing and releasing a Python project: <div class="codehilite"><pre><code>all: lint test .PHONY: test test: pytest .PHONY: lint lint: flake8 .PHONY: release release: python3 setup.py sdist bdist_wheel upload clean: find . -type f -name *.pyc -delete find . -type d -name __pycache__ -delete </code></pre></div> This is straightforward and a potential contributor immediately knows the entry points to your project. Assuming there is a venv installed already, you have to activate it first and run the <code>make</code> commands afterwards: <div class="codehilite"><pre><code>$ . venv/bin/activate $ make test </code></pre></div> The downside is of course, that you have to activate the venv for every new shell. Which can get a bit annoying when you spawn a new terminal in <code>tmux</code> or put <code>vim</code> into the background to run <code>make</code>. Activating the venv inside make will not work, as each recipe runs in its own shell, moreover each command in each recipe runs in its own shell too. There are workarounds for the latter, i.e. using the <code>.ONESHELL</code> flag, but that does not solve the issue of a new shell in each recipe. <h2>Wrapping the venv calls inside <code>make</code></h2> The second approach mitigates for the issue of activating the venv altogether. I’ve borrowed this idea mostly from <a href="https://github.com/sio/Makefile.venv">makefile.venv</a> and simplified it a lot for my needs. <div class="codehilite"><pre><code># system python interpreter. used only to create virtual environment PY = python3 VENV = venv BIN=$(VENV)/bin # make it work on windows too ifeq ($(OS), Windows_NT) BIN=$(VENV)/Scripts PY=python endif all: lint test $(VENV): requirements.txt requirements-dev.txt setup.py $(PY) -m venv $(VENV) $(BIN)/pip install --upgrade -r requirements.txt $(BIN)/pip install --upgrade -r requirements-dev.txt $(BIN)/pip install -e . touch $(VENV) .PHONY: test test: $(VENV) $(BIN)/pytest .PHONY: lint lint: $(VENV) $(BIN)/flake8 .PHONY: release release: $(VENV) $(BIN)/python setup.py sdist bdist_wheel upload clean: rm -rf $(VENV) find . -type f -name *.pyc -delete find . -type d -name __pycache__ -delete </code></pre></div> The equivalent <code>Makefile</code> now looks immediately more complicated. So let’s break it down. Instead of calling just calling <code>pytest</code>, <code>flake8</code> and <code>python</code> assuming the venv is already activated or all dependencies are installed on the system directly, we explicitly call the ones from the venv by prefixing the command with the path to the venv’s <code>bin</code> directory. To ensure the venv exists, each of the recipes is depending on the <code>$(VENV)</code> target, which ensures we always have an up-to-date venv installed. This works, because the <code>. venv/bin/activate</code>-script basically just does the same: it puts the venv before anything else in your <code>PATH</code>, therefore each call to <code>python</code>, etc. will find the one installed in the venv first. While the <code>Makefile</code> is a bit more complicated, we now can just call <div class="codehilite"><pre><code>$ make test </code></pre></div> and don’t deal with venvs directly any more (well, for those simple cases at least…). If you don’t need to support Windows, you can remove the appropriate block and the <code>Makefile</code> looks relatively tame, even for people that don’t use <code>make</code> very often. <h2>Which one is better?</h2> I think the second approach is more convenient. I’ve used the first approach happily for years and only learned quite recently about the second one. I haven’t really noticed any downsides yet, but I do realize that almost all Python projects with <code>Makefiles</code> I’ve checked out, seem to prefer the first approach. I wonder why that is?Perfection is Achieved When There is Nothing Left to Take Away2021-03-24T23:00:00+01:002021-03-24T23:00:00+01:00Bastian Venthurtag:venthur.de,2021-03-24:/2021-03-24-blogging.html

In 14 years since starting this blog, I changed the software several times, with each iteration simplifying requirements and setup.

In 14 years since starting this blog, I changed the software several times, with each iteration simplifying requirements and setup. <h2>Wordpress</h2> In 2007, I <a href="2007-01-23-hello-world-2.html">started blogging</a>. Back then I used a self-hosted <a href="https://wordpress.org/">Wordpress</a> instance, running on my server. I was never really comfortable with the technology stack involved: a <a href="https://www.php.net/">programming language</a> I didn’t speak, a <a href="https://www.mysql.com/">DB</a> I had little experience with, a relatively heavy setup and of course, the occasional security issues related to that setup. All that just to have a somewhat dynamic website with <a href="https://en.wikipedia.org/wiki/Pingback">pingbacks</a> and comments generated by your visitors… totally worth it and exciting times! You wrote something to the “lazyweb” department and people would actually comment with useful advice! Then came the spammers. First, it was just a little, and you could easily fight it off with the <a href="https://wordpress.org/plugins/akismet/">askimet</a> plugin. Over the years, however the spam-to-useful-comments-ratio shifted to ridiculous levels and the comment section itself became a burden. Hosting the comments on a separate platform like <a href="https://disqus.com/">disqus</a> was not an option for me, so I finally turned the comments off. Without the need for a comment section, the whole idea of storing mostly textual content in a DB just to have it served as a website, suddenly seemed like overkill and an unnecessary security risk. So I started looking for alternatives. I wanted to maintain my content as plain text files. Text files are cool because they are future proof, portable and can be managed in git. Consequently, I was searching for a static site generator. <h2>Pelican</h2> Around 2016 I finally migrated away from Wordpress to <a href="https://getpelican.com/">Pelican</a>. Pelican met all my requirements. I could store my content as plain Markdown files in a git repository. Pelican is written in Python, a language with which I’m very comfortable with. Pelican deals with blog posts and “pages” (i.e. non blog articles) and most importantly: Pelican just generates static HTML. So all that’s required on the server side is a simple web server. Conveniently, it also comes with an importer that allowed me to export the content of my old Wordpress instance straight into Markdown files. Migrating to Pelican was a huge relieve: I finally had a all my content as plain text in a git repository, my server was a lot more secure without the Wordpress instance running and removing MySQL freed up a lot of resources on that machine. <h2>blag</h2> Fast forward to 2021. While I was quite happy with Pelican, it still had too many features I don’t need. My blogging experience can be stripped down to a bunch of Markdown files that serve either as articles or pages, some static files, some templating for a decent design and an Atom feed. Particularly, I do not need pages/articles in multiple languages, a plugin system, support for multiple authors and tons of settings. So I did what every self-respecting nerd would do: I wrote my own site generator. Not to write something better than Pelican, but rather something lesser. Something that does less while still being functionally adequate. I called the result “<a href="https://github.com/venthur/blag">blag</a>”, named after the <a href="https://blog.xkcd.com/">blag of the webcomic xkcd</a>. blag is a blog-aware, static site generator written in Python. It supports the basic features you’d expect from a blog, like an archive, tags and an <a href="https://tools.ietf.org/html/rfc4287">Atom</a> feed. It converts Markdown (with fenced code blocks) and supports <a href="https://jinja.palletsprojects.com">Jinja2</a> templating. With this setup I’m quite happy for now. Maybe sometime there will be a simpler RSS/Atom-successor that does not require site-specific metadata (i.e. only stuff that can be generated from the articles themselves), so I could get rid of blag’s only required configuration.Installing Debian on a Thinkpad T14s2021-02-10T21:00:00+01:002021-02-10T21:00:00+01:00Bastian Venthurtag:venthur.de,2021-02-10:/2021-02-10-installing-debian-on-a-thinkpad-t14s.html

Due to non-free firmware it is still difficult to install Debian on recent hardware.

Recently, I got a new work laptop. I opted for Lenovo’s Thinkpad T14s that comes with AMD’s Ryzen 7 PRO 4750U Processor. Once everything is installed, it works like a charm: all hardware is supported and works out of the box with Debian. However, the tricky part is actually installing Debian onto that machine: The laptop lacks a standard Ethernet port and comes with Intel’s Wi-Fi 6 AX200 module. So if you don’t happen to have a docking station or an Ethernet adapter available during install, you’ll have to install everything over WiFi. The WiFi module, however requires non-free firmware and this is where the fun starts. First, I downloaded an official <a href="https://www.debian.org/distrib/netinst">netinst</a> image and copied it onto a USB drive. Halfway through the installation, it complained that firmware for the WiFi module was missing, and I was stuck as I couldn’t continue the installation without network access. Ok, then – missing non-free firmware it is. The <a href="https://wiki.debian.org/Firmware">wiki</a> suggests using an unofficial image instead, as it supposedly contains “all non-free firmware packages directly”. So I tried an unofficial <a href="https://cdimage.debian.org/images/unofficial/non-free/images-including-firmware/">netinst image with non-free firmware</a>. That also did not work, with the same error as above: the required firmware was missing. I checked the image later and actually couldn’t find the non-free firmware either. Hum. In the end, I had to prepare a second USB drive with the firmware downloaded from <a href="https://cdimage.debian.org/cdimage/unofficial/non-free/firmware/testing/current/">here</a>. I unpacked the firmware into <code>/firmware</code> on the second USB. The installer checks at some point during the installation for firmware installed on other removable media and (hopefully) finds it. In my case it did, and I finally could install everything. I’m quite happy with the laptop, but I have to admit how incredibly difficult it still is to install Debian on recent hardware. As a Debian Developer, I do understand Debian’s position on non-free firmware, on the other hand however, a less technical person would probably have given up at some point and just installed some other Operating System. Update (2022-01-10): As someone suggested to me, an easier solution to ensure internet access is to simply connect your cellphone via USB to the laptop and use tethering.Dear Apple,2021-01-11T22:00:00+01:002021-01-11T22:00:00+01:00Bastian Venthurtag:venthur.de,2021-01-11:/2021-01-11-dear-apple.html

Like Android, Apple should allow other apps like Signal to become the default SMS/MMS app on iOS.

In the light of WhatsApp’s recent move to <a href="https://www.whatsapp.com/legal/updates/key-updates-eea">enforce new Privacy Agreements</a> onto its users, alternative messenger services like <a href="https://www.signal.org">Signal</a> are currently gaining some more momentum. While this sounds good, it is hard to believe that this will be more than a dent in WhatsApp’s user base. WhatsApp is way too ubiquitous, and the whole point of using such a service for most users is to use the one that everyone is using. Unfortunately. Convincing WhatsApp users to additionally install Signal is hard: they already have SMS for the few people that are not using WhatsApp, now expecting them to install a third app for the same purpose seems ridiculous. Android mitigates this problem a lot by <a href="https://support.signal.org/hc/en-us/articles/360007321171-Can-I-send-SMS-MMS-with-Signal-#android_sms">allowing</a> to make other apps – like Signal – the default SMS/MMS app on the phone. Suddenly people are able to use Signal for SMS/MMS and Signal messages transparently. Signal is smart enough to figure out if the conversation partner is using Signal and enables encryption, video calls and other features. If not, it just falls back to plain old SMS. All in the same app, very convenient for the user! I don’t really get why the same thing is <a href="https://support.signal.org/hc/en-us/articles/360007321171-Can-I-send-SMS-MMS-with-Signal-#ios_sms">not possible</a> on iOS? Apple is well known for taking things like privacy and security for its users seriously, and this seems like a low-hanging fruit. So dear Apple, wouldn’t now be a good time to team up with WhatsApp-alternatives like Signal to help the users to make the right choice?Creating Beautiful Github Streaks2020-12-30T00:00:00+01:002020-12-30T00:00:00+01:00Bastian Venthurtag:venthur.de,2020-12-30:/2020-12-30-streak.html

How to generate git commits in the past to create beautiful github streaks.

Github streaks are a pretty fun and harmless way of visualizing your contributions on github or gitlab. Some people use them to brag a little, some companies use them to check out potential candidates. It’s all good as long as everyone is aware how easily you can manipulate them. In order to “fake” a commit date in the past, you have to set the <code>GIT_AUTHOR_DATE</code> and <code>GIT_COMMITTER_DATE</code> environment variables on <code>git commit</code> – both, RFC 2822 and ISO 8601 formats are accepted (for more info see <a href="https://git-scm.com/book/en/v2/Git-Basics-Viewing-the-Commit-History">here</a>). When generating hundreds of commits, the <code>--allow-empty</code> flag is very useful, as we don’t have to actually change the source tree in order to generate a commit. So the command to generate a fake commit looks like this: <div class="codehilite"><pre><code>GIT_AUTHOR_DATE="$TIMESTAMP" GIT_COMMITTER_DATE="$TIMESTAMP" git commit --allow-empty -m "$TIMESTAMP" </code></pre></div> execute it on an empty git repository and push to github or gitlab and you’ll se the result after some time. To automate this, I wrote me a little python script (<a href="https://github.com/venthur/streak">github</a>, <a href="https://pypi.org/project/python-streak/">pypi</a>). You can simply install it via: <div class="codehilite"><pre><code>pip install python-streak </code></pre></div> and run it with: <div class="codehilite"><pre><code>streak </code></pre></div> in an empty git repository. By default, streak will create 3 commits per day with a probability of ~71% per commit for 100 years, starting from 1980-05-09. These parameters can be tuned with: <ul> <li><code>--start START</code>: Beginning of the commits (Format: YYYY-MM-DD)</li> <li><code>--end END</code>: End of the commits (Format: YYYY-MM-DD)</li> <li><code>--max MAX</code>: Maximum number of commits per day.</li> <li><code>--proba PROBA</code>: Probability for a commit to happen.</li> </ul> Generating a few commits per day for the range of 10 years, took roughly 2 minutes on my computer, and man was I a busy beaver in the 80s! <img alt="Streak 1989" src="/images/streak.png" title="The Great Streak of 1989"> I know, I’m for sure not the first one to discover this kind of thing, but it was a fun afternoon project anyways!Introducing Noir2019-08-24T00:00:00+02:002019-08-24T00:00:00+02:00Bastian Venthurtag:venthur.de,2019-08-24:/2019-08-24-introducing-noir.html

noir is a drop-in replacement for black, with the default line length set to 79 characters.

<h2>tl;dr</h2> <a href="https://github.com/venthur/noir">Noir</a> is a drop-in replacement for <a href="https://github.com/psf/black">Black</a> (the uncompromising code formatter), with the default line length set to <a href="https://www.python.org/dev/peps/pep-0008/">PEP-8</a>’s preferred 79 characters. If you want to use it, just replace <code>black</code> with <code>noir</code> in your <code>requirements.txt</code> and/or <code>setup.py</code> and you’re good to go. Black is a Python code formatter that reformats your code to make it more <a href="https://www.python.org/dev/peps/pep-0008/">PEP-8</a> compliant. It implements a subset of PEP-8, most notably it deliberately ignores PEP-8’s suggestion for a line length of 79 characters and defaults to a length of 88. I find the decision and the <a href="https://github.com/psf/black#line-length">reasoning</a> behind that somewhat arbitrary. PEP-8 is a good standard and there’s a lot of value in having a style guide that is generally accepted and has a lot of tooling to support it. When <a href="https://github.com/psf/black/issues/498">people</a> <a href="https://github.com/psf/black/issues/855">ask</a> to change Black’s default line length to 79, the issue is usually closed with a reference to the reasoning in the <code>README</code>. But Black’s developers are at least aware of this controversial decision, as Black’s only option that allows to configure the (otherwise uncompromising) code formatter, is in fact the line length. Apart from that, Black is a good formatter that’s gaining more and more popularity. And, of course, the developers have every right to follow their own taste. However, since Black is licensed under the terms of the <a href="https://opensource.org/licenses/MIT">MIT license</a>, I tried to see what needs to be done in order to fix the line length issue. <h2>Step 1: Changing the Default</h2> <a href="https://github.com/venthur/noir/commit/f322b60072774a098c4b84eaebaa7e349fc7d251">This</a> is the easiest part. You only have to change the <code>DEFAULT_LINE_LENGTH</code> value in <code>black.py</code> from <code>88</code> to <code>79</code>, and black works as expected. Bonus points for doing the same in <code>black.vim</code> and <code>pyproject.toml</code>, but not strictly necessary. <h2>Step 2a: Fixing the Tests</h2> Now comes the <a href="https://github.com/venthur/noir/commit/b5f6ecf51adfa90751e8a7fbc139081e7f676a48">fun part</a>. Black has an extensive test suite and suddenly a lot of tests are failing because the fixtures that compare the unformatted input with the expected, formatted output were written with a line length of 88 characters in mind. To make it more interesting the expected output comes in two forms: (1) as normal reformatted Python code (which is rather easy to fix) and (2) as a <a href="https://en.wikipedia.org/wiki/Diff">diff</a> between the input and the expected output. The latter was really painful to fix – although I’m very much used to reading diffs, I don’t usually write them. <h2>Step 2b: Fixing the Tests</h2> After all fixtures were updated, some tests were still failing. And it turned out that Black is running itself on its own source code as part of its test suite, making the tests fail if Black’s code does not conform to Black’s coding standards. While this is a genius idea, it meant that I had to reformat Black’s code to match the new 79 characters line length, generating a <a href="https://github.com/venthur/noir/commit/b87e2b66833f39aa109e92daab5f859021125ab6">giant diff</a>, that is functionally unrelated to the fix I wanted to make but now part of the fix anyway. This of course makes the whole patch horrible to maintain if you plan to follow along upstream’s <code>master</code> branch. <h2>Step 3: Publish</h2> Since we already got this far, why not publish the fixed version of Black? To my surprise the name <a href="https://pypi.org/project/noir/">noir</a> was still available on PyPi, so I <a href="https://github.com/venthur/noir/commit/2d854112a2d11d34b12ab7fe65361826dc01bc65">renamed</a> my version of Black to Noir and uploaded it to PyPi. You can install it via: <div class="codehilite"><pre><code>$ pip install noir </code></pre></div> Since I didn’t change anything else, this is literally a drop-in replacement for Black. All you have to do is replace <code>black</code> with <code>noir</code> in your <code>requirements.txt</code> and/or <code>setup.py</code> and you’re good to go. The script that executes Black is still called <code>black</code> and the server is still called <code>blackd</code>. <h2>Outlook</h2> While this was a fun exercise, the question remains what to do with it. I’ll try to follow upstream and update my patch whenever a new version will come out. As new versions of Black are released only a handful of times a year, this might be feasible. Depending on how painful it is to maintain the patch for the tests, I might either drop the tests altogether, relying on upstream’s tests passing on their side and just maintaining the trivial patch from Step 1: Changing the <code>DEFAULT_LINE_LENGTH</code>. The latter can probably be automated somehow using github actions – and I’ll probably look into that at some point. Best case scenario, of course, would be if Python changes its recommended line length to 88 and I wouldn’t have to maintain noir in the first place :)Dotenv CLI2019-05-12T00:00:00+02:002019-05-12T00:00:00+02:00Bastian Venthurtag:venthur.de,2019-05-12:/2019-05-12-dotenv-cli.html

dotenv-cli is a Python package that provides the dotenv command.

As a small weekend project I wrote my own <a href="https://github.com/venthur/dotenv-cli"><code>dotenv</code> CLI</a>. Dotenv-CLI is a simple package that provides the <code>dotenv</code> command. It reads the <code>.env</code> file from the current directory puts the contents in the environment and executes the given command. Example <code>.env</code> file: <div class="codehilite"><pre><code>BASIC=basic basic export EXPORT=foo EMPTY= INNER_QUOTES=this 'is' a test INNER_QUOTES2=this "is" a test TRIM_WHITESPACE= foo KEEP_WHITESPACE=" foo " MULTILINE="multi\nline" # some comment </code></pre></div> becomes: <div class="codehilite"><pre><code>$ dotenv env BASIC=basic basic EXPORT=foo EMPTY= INNER_QUOTES=this 'is' a test INNER_QUOTES2=this "is" a test TRIM_WHITESPACE=foo KEEP_WHITESPACE= foo MULTILINE=multi line </code></pre></div> where <code>env</code> is a simple command that outputs the current environment variables. A <code>dotenv</code> CLI comes in handy if you follow the <a href="https://12factor.net/config">12 factor app</a> methodology or just need to run a program locally with specific environment variables set. While <a href="https://github.com/venthur/dotenv-cli">dotenv-cli</a> is certainly not the first <code>dotenv</code> implementation, this one is written in Python and has no external dependencies except Python itself. It also provides a bash completion, so you can prefix any command with <code>dotenv</code> while still be able to use completion: <div class="codehilite"><pre><code>$ dotenv make <TAB> all clean docs lint release test </code></pre></div> Since there’s also some other popular shells out there and I already struggled writing the simple bash completion script, it would be very nice if some more experienced <code>zsh</code>-, <code>fish</code>-, etc.- users could help me out in providing completions for them as well. <code>dotenv-cli</code> is available on Debian and Ubuntu: <div class="codehilite"><pre><code>$ sudo apt-get install python3-dotenv-cli </code></pre></div> and on PyPi as well: <div class="codehilite"><pre><code>$ pip install dotenv-cli </code></pre></div>Orphaning reportbug-ng2018-12-28T00:00:00+01:002018-12-28T00:00:00+01:00Bastian Venthurtag:venthur.de,2018-12-28:/2018-12-28-orphaning-rng.html

After several years without development, it is time to say goodbye to reportbug-ng.

After more than two years without an update, I’ve decided to discontinue developing <a href="https://github.com/venthur/reportbug-ng">reportbug-ng</a> and <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=917557">orphaned</a> the Debian package. Since reportbug is a more than adequate alternative and reportbug-ng’s popcon value is relatively low, it’s probably the best to just remove the package from the archive. I <a href="2007-03-07-a-screencast-is-worth-a-thousand-pictures.html">started developing</a> reportbug-ng 11 years ago as a user-friendly alternative to Debian’s standard reportbug. Not only did it come with a nice GUI that allowed for easy searching and filtering of existing bugs, it also queried our bug tracking system (BTS) using its SOAP interface. Back then, reportbug used to parse the human readable HTML from the website to get the same information. A couple of years later Debian’s reportbug <a href="2009-03-25-reportbug-finally-has-a-gui.html">got a GUI too</a> and became a lot more novice-friendly. At some point <a href="https://alioth-lists-archive.debian.net/pipermail/reportbug-maint/2008-July/000277.html">I asked the reportbug team</a> if they’d be interested in moving away from parsing the BTS’ HTML and using the SOAP interface too. They agreed and I started separating the BTS related part of the code into a separate Python module called <a href="https://github.com/venthur/python-debianbts">python-debianbts</a>. A few years later, in August 2011, reportbug 6.0 started to use python-debianbts for querying the BTS. It has been a fun ride, but it’s time to acknowledge that reportbug-ng has outlived its usefulness and I’m not willing to maintain that Qt based GUI anymore. I still continue to maintain python-debianbts, though.Introducing Litestats2018-10-13T00:00:00+02:002018-10-13T00:00:00+02:00Bastian Venthurtag:venthur.de,2018-10-13:/2018-10-13-litestats.html

litestats is a tool that converts Python's profiler output into a sqlite3 database for easier exploration and analysis.

Profiling in Python has always been easy, however, analyzing the profiler’s output not so much. After the profile has been created you can use Python’s <code>pstats</code> module but it feels quite clumsy and not really empowering. For Python 2 there has been <a href="https://www.vrplumber.com/programming/runsnakerun/">RunSnakeRun</a>, a very convenient tool for analyzing the profiler output, unfortunately that tool hasn’t been updated since 2014. I recently <a href="https://github.com/venthur/snakerunner">ported</a> it to Python3 and wxPython4 but I’m probably not going to maintain that code properly as I’m not very comfortable with wxPython. I still wanted something nicer than <code>pstats</code> for profiling so I wrote <a href="https://github.com/venthur/litestats">litestats</a>. Litestats is a simple command line tool that takes the output of the Python profiler and transforms the data into a sqlite3 database. You can now easily analyze the profiler output using <code>sqlite</code> on the command line, the <code>sqlitebrowser</code> for a graphical interface or use the data base as the foundation of your very own tooling around the analysis. <h2>How does it work?</h2> Litestats reads the dump of the profiler and creates a normalized data base with tree tables: <ul> <li><code>functions</code>: contains each function (callers and callees) with filename, line number and function name</li> <li><code>stats</code> contains the statistics (primitive/total calls, total/cumulative time) for all functions</li> <li><code>calls</code> a caller-callee mapping</li> </ul> While this provides an exact representation of the dump, those tables would be cumbersome to use. So litestats additionally creates three views emulating <code>pstats</code> <code>print_stats()</code>, <code>print_callers()</code> and <code>print_callees()</code> functionality: <ul> <li><code>pstats</code></li> <li><code>callers</code></li> <li><code>callees</code></li> </ul> <h2>Installation</h2> Litestats has no requirements other than Python itself and is available on <a href="https://pypi.org/project/litestats/">PyPI</a>: <div class="codehilite"><pre><code>$ pip install litestats </code></pre></div> <h2>Usage</h2> <div class="codehilite"><pre><code>$ # run the profiler and dump the output $ python3 -m cProfile -o example.prof example.py $ # convert dump to sqlite3 db $ litestats example.prof $ # example.prof.sqlite created </code></pre></div> You can now use the sqlite3 data base to investigate the profiler dump: <div class="codehilite"><pre><code>sqlite> select * ...> from pstats ...> order by cumtime desc ...> limit 20; ncalls tottime ttpercall cumtime ctpercall filename:lineno(function) ---------- ---------- -------------------- ---------- ---------- ------------------------------------ 18/1 0.000161 8.94444444444444e-06 0.067797 0.067797 ~:0(<built-in method builtins.exec>) 1 1.0e-06 1.0e-06 0.067755 0.067755 <string>:1(<module>) 1 4.0e-06 4.0e-06 0.067754 0.067754 /usr/lib/python3.7/runpy.py:195(run_ 1 6.0e-06 6.0e-06 0.066135 0.066135 /usr/lib/python3.7/runpy.py:62(_run_ 1 1.1e-05 1.1e-05 0.066113 0.066113 /home/venthur/Documents/projects/lit 1 6.6e-05 6.6e-05 0.055152 0.055152 /home/venthur/Documents/projects/lit 1 4.1e-05 4.1e-05 0.0549 0.0549 /home/venthur/Documents/projects/lit 1 0.050196 0.050196 0.050196 0.050196 ~:0(<method 'executescript' of 'sqli 20/3 8.9e-05 4.45e-06 0.011064 0.003688 <frozen importlib._bootstrap>:978(_f 20/3 4.8e-05 2.4e-06 0.011005 0.00366833 <frozen importlib._bootstrap>:948(_f 20/3 7.5e-05 3.75e-06 0.01083 0.00361 <frozen importlib._bootstrap>:663(_l 15/3 3.5e-05 2.33333333333333e-06 0.01073 0.00357666 <frozen importlib._bootstrap_externa 29/5 2.5e-05 8.62068965517241e-07 0.010215 0.002043 <frozen importlib._bootstrap>:211(_c 3 6.0e-06 2.0e-06 0.010087 0.00336233 ~:0(<built-in method builtins.__impo 28/6 9.0e-06 3.21428571428571e-07 0.008977 0.00149616 <frozen importlib._bootstrap>:1009(_ 1 9.0e-06 9.0e-06 0.00841 0.00841 /home/venthur/Documents/projects/lit 16 0.000138 8.625e-06 0.004802 0.00030012 <frozen importlib._bootstrap_externa 1 4.5e-05 4.5e-05 0.004143 0.004143 /usr/lib/python3.7/logging/__init__. 1 0.004038 0.004038 0.004038 0.004038 ~:0(<method 'commit' of 'sqlite3.Con 13 3.3e-05 2.53846153846154e-06 0.002368 0.00018215 <frozen importlib._bootstrap_externa </code></pre></div>Syncing Exchange Server with Thunderbird 60 (again)2018-09-02T00:00:00+02:002018-09-02T00:00:00+02:00Bastian Venthurtag:venthur.de,2018-09-02:/2018-09-02-thunderbird-ews.html

Syncing Exchange Server with Thunderbird 60 (again)

A couple weeks ago Mozilla Thunderbird <a href="https://www.thunderbird.net/en-US/thunderbird/60.0/releasenotes/">released version 60</a> and a lot of extensions became incompatible. Amongst them <a href="https://github.com/ExchangeCalendar/exchangecalendar">Exchange Calendar</a>, an extension that allows to sync calendars, tasks and contacts with a Microsoft Exchange Server. That extension has served me well over the last years in my job where we have to use the Exchange Server. Unfortunately it seems that the extension will not be continued anymore, and I had to search for something else. After trying out some alternatives, I found the extension <a href="https://addons.thunderbird.net/en-US/thunderbird/addon/tbsync/">TbSync</a>. It it uses Exchange Active Sync instead of EWS, so there are currently some limitations with accessing shared calendars and contacts but for everything else it works quite nice. Also the author is actively maintaining it and working on <a href="https://github.com/jobisoft/EWS-4-TbSync">EWS support</a> as well.Thanks!2015-10-13T14:41:00+02:002015-10-13T14:41:00+02:00Bastian Venthurtag:venthur.de,2015-10-13:/2015-10-13-thanks.html

Thanks!

Two months ago I <a href="2015-08-18-please-help-to-port-python-debianbts-to-python3.html">asked</a> for help porting <a href="https://github.com/venthur/python-debianbts">python-debianbts</a>, to Python3. Python-debianbts is a tiny little library that allows for querying Debian’s bug tracking system. Since Debian’s reportbug depends on it, the library is installed on ca 80% of the Debian installations. The main blocker back then was that python-debianbts depended on SOAPpy which was not available for Python3. So before we could port python-debianbts to Python3 we had to migrate from SOAPpy to pysimplesoap – a daunting task given the beast that SOAP is. Fortunately, <a href="https://github.com/gaetano-guerriero">Gaetano Guerriero</a> heard my call and helped a lot. First he migrated python-debianbts to pysimplesoap. Then, he ported python-debianbts to Python3 and now he is still busy fixing bugs and providing me with pull requests. I’m very satisfied with the outcome. His pull requests are very high-quality and come usually with the appropriate unit tests included. While he is doing the major grunt work, I merely do some occasional nitpicking and uploading to Debian/unstable. Thank you very much Gaetano! If you ever come to Berlin, please drop me a note so I can invite you to a beer or two.Please Help to Port python-debianbts to Python32015-08-18T13:07:00+02:002015-08-18T13:07:00+02:00Bastian Venthurtag:venthur.de,2015-08-18:/2015-08-18-please-help-to-port-python-debianbts-to-python3.html

Please Help to Port python-debianbts to Python3

Dear Lazyweb, I’m currently trying to find a way to port <a href="https://tracker.debian.org/pkg/python-debianbts">python-debianbts</a> to Python3. Debian’s standard bugreport tool <a href="https://tracker.debian.org/pkg/reportbug">reportbug</a> depends on python-debianbts and can thus not convert to Python3 if python-debianbts does not as well. Unfortunately python-debianbts depends on <a href="https://tracker.debian.org/pkg/python-soappy">SoapPy</a> for parsing the Debian bugtracker’s responses, and that library is not ported to Python3 yet, and probably never will. I’m planning to replace SoapPy with <a href="https://tracker.debian.org/pkg/python-pysimplesoap">pysimplesoap</a> which is available for Python2 and Python3. Unfortunately <a href="https://www.debian.org/Bugs/">debbugs</a> does not support <a href="https://en.wikipedia.org/wiki/Web_Services_Description_Language">WSDL</a> which makes parsing of the replies extremely painful and error-prone. I wonder if there is a SOAP/Python expert out there who’d be willing to give some assistance in porting python-reportbug to pysimplesoap? python-reportbug’s <a href="https://github.com/venthur/python-debianbts">repository</a> is on GitHub and patches are very welcome. Since <a href="https://en.wikipedia.org/wiki/SOAP">SOAP</a> is quite a beast and debbugs uses it for read-only purposes only, another attractive solution would be to replace/augment debbugs’ API with something much more simple, like <a href="https://en.wikipedia.org/wiki/JSON">JSON</a>. That would make parsing extremely easy as many programming languages including Python support JSON without any external libraries. In theory this could be quite easy but requires some serious Perl skills.General Resolution is not required2014-11-19T09:21:00+01:002014-11-19T09:21:00+01:00Bastian Venthurtag:venthur.de,2014-11-19:/2014-11-19-general-resolution-is-not-required.html

General Resolution is not required

The result for the <a href="https://www.debian.org/vote/2014/vote_003">General Resolution</a> about the init system coupling is out and the result is, not quite surprisingly, “General Resolution is not required”. When skimming over -devel or -private from time to time, one easily gets the impression that we are all a bunch of zealots, all too eager for fighting. People argue in the worst possible ways. People make bold statements about the future of Debian if solution X is preferred over Y. People call each other names. People leave the project. At some point you realize, we’re not all a bunch of zealots, it is usually only the same small subset of people always involved in those discussions. It’s reassuring that we still seem to have a silent majority in Debian that, without much fuss, just do what they can to make Debian better. In this sense: A General Resolution is not required.What are the most popular .vimrc options?2013-09-11T10:21:00+02:002013-09-11T10:21:00+02:00Bastian Venthurtag:venthur.de,2013-09-11:/2013-09-11-what-are-the-most-popular-vimrc-options.html

What are the most popular .vimrc options?

Hi always wondered what the most popular options are, you usually find in <code>.vimrc</code> files. So I downloaded 155 <code>.vimrc</code> files from the net (mostly from <a href="https://dotfiles.org">dotfiles.org</a> and <a href="https://github.com">github.com</a>), and wrote a little <a href="https://gist.github.com/venthur/6521228">script</a> which counts the number of times an option has been set. Since most options come in normal- and shortcut form, I mapped the shortcuts to the long version whenever I recognized them. So without further ado, here are the most popular <code>.vimrc</code> options (without values!). The number specifies the number of times this option has been set. The most popular option is on the bottom: <blockquote> 10 tselect 10 dictionary 10 runtimepath 11 mousehide 11 t_vb 11 foldlevel 11 foldopen 12 suffixes 12 matchtime 12 fileencoding 13 modelines 13 wrap 14 sidescrolloff 14 clipboard 14 lines 14 novisualbell 15 linebreak 15 cursorline 15 fileformats 15 columns 15 cindent 16 undolevels 16 shiftround 16 lazyredraw 16 completeopt 18 modeline 18 whichwrap 18 comments 18 wildignore 19 list 19 autowrite 19 foldcolumn 19 grepprg 19 titlestring 20 autoread 20 title 21 foldenable 21 cmdheight 22 pastetoggle 23 formatoptions 23 fileencodings 24 tags 24 directory 25 ttyfast 26 termencoding 26 complete 27 nohlsearch 27 noerrorbells 27 visualbell 28 shortmess 30 showmode 31 wildmode 32 t_Co 32 listchars 32 backupdir 34 hidden 34 backup 35 smarttab 35 foldmethod 36 viminfo 36 textwidth 37 scrolloff 37 nobackup 41 nowrap 44 encoding 47 guifont 51 guioptions 53 smartcase 54 wildmenu 57 smartindent 60 mouse 63 background 64 softtabstop 66 history 70 showmatch 72 ignorecase 74 showcmd 74 laststatus 79 number 82 hlsearch 91 statusline 94 expandtab 94 ruler 96 autoindent 96 backspace 99 tabstop 109 incsearch 114 shiftwidth 124 nocompatible Out of 155 .vimrcs </blockquote> Fun fact: <code>nocompatible</code> is the most popular, but also most useless one. The fact that you have an <code>.vimrc</code> automatically implies the <code>nocompatible</code> mode in vim.How to get the most precise time, comparable between processes in Python?2013-05-15T11:49:00+02:002013-05-15T11:49:00+02:00Bastian Venthurtag:venthur.de,2013-05-15:/2013-05-15-how-to-get-the-most-precise-time-comparable-between-processes-in-python.html

How to get the most precise time, comparable between processes in Python?

Let’s consider the following scenario: I have two Python processes receiving the same events and I have to measure the delay between when process A received the event and when process B received it, as precisely as possible (i.e. less than 1ms). Using Python 2.7 and a Unix system you can use the <code>time.time</code> method which provides the time in seconds since Epoch and has a typical resolution of a fraction of a ms on Unix. You can use it on different processes and still compare the results, since both processes receive the time since Epoch, a defined and fixed time in the past. On Windows time.time also provides the time since Epoch, but the resolution is in the range of 10ms, which is not suitable for my application. There is also <code>time.clock</code> which is super precise on Windows, and much less precise on Unix. The mayor drawback is that it returns the time since the process started or since the first call of <code>time.clock</code> within that processes. This means you cannot compare the results of <code>time.clock</code> between two processes as they are not calibrated to a common t-zero. I had high hopes for Python 3.3 where the <code>time</code> module was revamped and I was reading about <code>time.monotonic</code> and <code>time.perf_counter</code>. Especially <code>time.perf_counter</code> looked like it would suit my needs as the <a href="https://www.python.org/dev/peps/pep-0418/">documentation</a> said it provides the highest available resolution for the system and was “system-wide”, in contrast to for example the new <code>time.process_time</code> which was “process_wide”. Unfortunately it turned out that <code>time.perf_counter</code> acts similar to <code>time.clock</code> on Python 2.7 as it provides you with the time since the process started or the first time the method was called within the process. The results of <code>time.monotonic</code> are comparable between processes, but again not precise enough on Windows. Here is a small script which demonstrates how the times provided by <code>time.clock</code> and <code>time.perf_counter</code> are not comparable between processes. It starts two processes and lets both of them print out the output of the timer to stdout. In the output the times should be monotonically increasing. Since I let process 2 sleep for one second before calling the timer method for the first time, the output of this process is usually one second smaller when using <code>time.clock</code> or <code>time.perf_counter</code>. <div class="codehilite"><pre><code>#!/usr/bin/env python from multiprocessing import Process import time timers = ['clock', 'time', 'monotonic', 'perf_counter'] def proc(timer): timer = getattr(time, timer) time.sleep(1) for i in range(3): print('P2 {time}'.format(time=timer())) time.sleep(1) if __name__ == '__main__': for t in timers: print("Using {timer}".format(timer=t)) p = Process(target=proc, args=(t,)) timer = getattr(time, t) p.start() for i in range(3): print('P1 {time}'.format(time=timer())) time.sleep(1) p.join() </code></pre></div> The result when running on Windows with Python 3.3: <div class="codehilite"><pre><code>$ python timertest.py Using clock P1 6.146032526480321e-06 P1 0.9926582847820045 P2 2.9612702173041547e-05 P1 1.9941743992602412 P2 1.0008579302676737 P2 2.0022709590185346 Using time P1 1368614235.509732 P1 1368614236.511172 P2 1368614236.601301 P1 1368614237.512612 P2 1368614237.602741 P2 1368614238.604181 Using monotonic P1 484.636 P1 485.63800000000003 P2 485.738 P1 486.639 P2 486.73900000000003 P2 487.741 Using perf_counter P1 12.390910576623565 P1 13.39050745276285 P2 7.542858100680394e-06 P1 14.39190763071843 P2 1.0014012954160376 P2 2.0041399116368144 </code></pre></div> So as far as I see it, there is no way of getting comparable times between two processes on Windows with more precision than 10ms. Is that correct or am I missing something?Wee! Wheezy is out (better late than never)2013-05-11T10:57:00+02:002013-05-11T10:57:00+02:00Bastian Venthurtag:venthur.de,2013-05-11:/2013-05-11-wee-wheezy-is-out-better-late-than-never.html

Wee! Wheezy is out (better late than never)

Last week we <a href="https://lists.debian.org/debian-announce/2013/msg00002.html">released Wheezy</a>, roughly two years after our last release Squeeze. I’d like to thank all the contributors in- and outside of Debian for your fine work! Every single contribution – no matter how big or small – summed up to the wonderful release we finished last week. Without you this release would not have been possible. Keep up the good work guys and make Jessie rock even harder! PS: It is very nice to see once again fresh packages rolling into unstable and spending some time fixing broken dependencies :)Synchronizing Google Mail Contacts with Thunderbird2013-01-30T12:37:00+01:002013-01-30T12:37:00+01:00Bastian Venthurtag:venthur.de,2013-01-30:/2013-01-30-synchronizing-google-mail-contacts-with-thunderbird.html

Synchronizing Google Mail Contacts with Thunderbird

Dear Lazyweb, can anyone recommend a good Thunderbird extension which allows for synchronizing the address book with Google mail? So far I tried <a href="https://addons.mozilla.org/en-us/thunderbird/addon/google-contacts/">Google Contacts</a>, but something went wrong with the syncing and some contacts where deleted on both sides. To avoid this problem, one can use Google Contacts in read-only mode (it will only fetch contacts from Google, but never write to it) but then you have to import new Thunderbird contacts to Google mail manually. Google introduced <a href="https://googleblog.blogspot.de/2012/12/winter-cleaning.html">CardDav</a> in December 2012, which allows for syncing of contacts, but since Thunderbird’s development is apparently on hold this is probably not gonna be supported out-of-the-box. There are some other extensions for Thunderbird, but since synchronization is hard and a lot more complicated than: “replace newer version with older one” I’m looking for something mature and well tested. Before someone suggests it, I know Evolution has this feature built-in. I gave it a try last week and found so many other grave bugs with the calendar and newsgroups that Evolution is simply unfit for my needs. I really like Thunderbird and want to stick with it for a few more years until I have to look for something else. Yours truly, BastiShiny new iPod Nano 6G... fffffffuuuuuuuuuuuu2013-01-08T11:59:00+01:002013-01-08T11:59:00+01:00Bastian Venthurtag:venthur.de,2013-01-08:/2013-01-08-shiny-new-ipod-nano-6g-fffffffuuuuuuuuuuuu.html

Shiny new iPod Nano 6G... fffffffuuuuuuuuuuuu

So I got an <a href="https://en.wikipedia.org/wiki/IPod_Nano">iPod Nano (6th generation)</a> for Christmas this year, just in time since my trusty old iPod Mini started beg for retirement after almost 8 years of usage. Since my old iPod was working like a charm all those years I expected a smooth sailing when I plugged in my new iPod Nano. Gnome recognized it correctly and mounted the device. The iPod showed up in <a href="https://projects.gnome.org/rhythmbox/">Rhythmbox</a> as I was used to and I started to fill it with some music. Everything worked as expected: Rhythmbox copied the music to the iPod without complaining and after unmounting the iPod and starting it – it was emtpy. Whait, what? Why is it empty? Didn’t I just… So I tried again, and again with the same result. Half an hour later I found out that <a href="https://www.gtkpod.org/wiki/Libgpod">libgpod</a> (the iPod “driver” for Linux) supports all iPods except the iPod Nano 6G. Bummer. Apparently <a href="https://apple.com">Apple</a> changed the algorithm to calculate the checksums of the files on the device and since that algorithm is unknown you cannot successfully write on it with free software. That means technically this device is unusable for Linux users since iTunes doesn’t run on Linux. However, there seems to be a way to use the iPod Nano with libgpod if you are trusting <a href="https://franck78.ath.cx/">this guy</a> and willing to use his binary (only!) file with libgpod (which I am not). And somehow the guys over at <a href="https://www.spotify.com/">Spotify</a> managed to get the iPod Nano running in their <a href="https://www.spotify.com/de/download/previews/">Linux client</a> but they don’t provide the code either. Being already more than two years old, I don’t have much hope that the iPod Nano will work on Linux with libgpod in the foreseeable future. On the bright side I can say the device is not totally useless as it comes with FM radio…Introducing The Art of Asking2012-11-02T09:29:00+01:002012-11-02T09:29:00+01:00Bastian Venthurtag:venthur.de,2012-11-02:/2012-11-02-introducing-the-art-of-asking.html

Introducing The Art of Asking

Since October 2011 my flatmate and I where quite busy realizing a little pet project of ours called <a href="https://theartofasking.com">The Art of Asking</a>. The ultimate goal is to visualize the world’s opinion in an intuitive fashion and make it easy for everyone to play around with the data. The idea behind The Art of Asking is that users submit interesting questions which are answered by users around the world. But instead of showing only the boring result, we want to provide interesting insights and statistics about the answers given. For now the users can see the results of the question visualized by geographical regions. For example on the page for the question <a href="https://theartofasking.com/question/wjucdg6k">‘How are you today?’</a> you can see the interactive map with the pie chart. The map shows the average/dominating answer for each continent encoded by color, and the pie chart the distribution of the different answers for that region. If you move the mouse over a continent the pie chart updates and shows the distribution of answers for that continent. You can also click to zoom into the map to see the same for countries and regions. This allows you to investigate how the answers are distributed around the world. <img alt="" src="/images/taoa1.png" title="The Art of Asking Screenshot"> This is already quite nice and fancy to play around with, but of course we want much more. Right now we’re working on a feature which will users allow to combine two arbitrary questions and see how the answers are related. This doesn’t sound like much, but it is very addictive to crawl through the list of questions and find interesting correlations. Here is a little plot how it could look like (the data is from the actual data base of answers). <img alt="" src="/images/gay_marrigage_vs_fashion.png" title="Support of gay marriage vs Importance of fashion"> For each possible combination of answers it shows the percentage of people who answered in that combination. Of course those plots only make sense when enough users answered for both questions you want to compare which is right now not very often the case, so I guess we’ll roll out that feature sometime later when we have more data. But we have lots of ideas and are working on further ways to investigate the answers. Obviously one low hanging fruit for example would be the distribution of the answers over time. We’ve been working for over a year now on this project in our spare time and we built it more or less from scratch. I wrote a small <a href="https://en.wikipedia.org/wiki/Web_Server_Gateway_Interface">WSGI</a> framework in <a href="https://python.org">Python</a> and on top of that the WSGI application which runs the site. We use <a href="https://www.mongodb.org/">MongoDB</a> for the storage of the data, <a href="https://projects.unbit.it/uwsgi/">uWSGI</a> and <a href="https://nginx.org/">nginx</a> for the server, <a href="https://jinja.pocoo.org/docs/">Jinja2</a> and <a href="https://twitter.github.com/bootstrap/">Bootstrap</a> for the HTML and <a href="https://d3js.org/">D3.js</a> for the visualization of the data, where <a href="https://www.user.tu-berlin.de/fazli/">Maci</a> did a wonderful job realizing the interactive map and charts. We’re running this site since July 2012 now and are already quite satisfied with the number of users, and the quality of the questions. But of course we could always use more (especially more answers). So if you want to try it out, go to <a href="https://theartofasking.com">theartofasking.com</a> and fill out the blank spots on the map! We’re happy about every answer and question we can get and are eager to hear your suggestions.Give Camp Berlin looking for volunteers2012-10-26T15:05:00+02:002012-10-26T15:05:00+02:00Bastian Venthurtag:venthur.de,2012-10-26:/2012-10-26-give-camp-berlin-looking-for-volunteers.html

Give Camp Berlin looking for volunteers

My friend Martijn from <a href="https://talentspender.org">Talentspender</a> is co-organizing a Give Camp in Berlin. They are looking for IT professionals and designers who want to spend one weekend of their time to support non-profit organizations to solve a specific problem at the Give Camp. It is for a good cause and there is no further commitment after the Camp. Plus you will be provided with free food and drinks. So if you are interested and happen to be in Berlin between 30. November and 02. December 2012 have a look at their <a href="https://berlin.givecamp.de/en">website</a> and <a href="https://berlin.givecamp.de/tech-registration/">register</a> for the GiveCamp. Quoting from their flyer: <blockquote> A GiveCamp is a weekend-long event where technology professionals donate their time to provide custom solutions for non-profit organizations. Voluntarily, without further commitment and for a good cause: the long-term strengthening of the organizations. How does it work? <ul> <li>Teamwork of experts during one weekend, with regular input from the NPOs</li> <li>Clearly defined projects, to be completed at the GiveCamp</li> <li>Young professionals mentored by experienced experts</li> <li>Meals and drinks are provided</li> </ul> Who are we looking for? <ol> <li>Software developers, database administrators, designers and entrepreneurs</li> <li>From students to senior experts</li> <li>Team players with enthusiasm for interdisciplinary projects</li> </ol> </blockquote>Awesome Firefox Campaign2012-08-30T20:07:00+02:002012-08-30T20:07:00+02:00Bastian Venthurtag:venthur.de,2012-08-30:/2012-08-30-awesome-firefox-campaign.html

Awesome Firefox Campaign

Gil writes about the awesome <a href="https://gil.badall.net/2012/08/29/red-pandas-are-taking-over-alexanderplatz/">Firefox ad campaign</a> running currently in Berlin Alexanderplatz. Good stuff!Happy Birthday Debian!2012-08-16T08:27:00+02:002012-08-16T08:27:00+02:00Bastian Venthurtag:venthur.de,2012-08-16:/2012-08-16-happy-birthday-debian-3.html

Happy Birthday Debian!

19 Years and still rocking! Huge thanks to all contributors for making this happen and a happy birthday Debian!Usability of git-http-backend with self signed SSL certificates2012-08-03T09:29:00+02:002012-08-03T09:29:00+02:00Bastian Venthurtag:venthur.de,2012-08-03:/2012-08-03-usability-of-git-http-backend-with-self-signed-ssl-certificates.html

Usability of git-http-backend with self signed SSL certificates

Why makes it git so hard to use a self signed SSL certificate in conjunction with the https protocol? At work we have a server for shared git repositories. For some reasons we can’t use the ssh protocoll to acces the repositories so we looked into the <a href="https://www.kernel.org/pub/software/scm/git/docs/git-http-backend.html">git-http-backend</a>. So far so good but we want it encrypted, of course. So we used SSL with our self signed certificate: <div class="codehilite"><pre><code>git clone https://git.example.com/public Cloning into 'public'... error: server certificate verification failed. CAfile: /etc/ssl/certs/ca-certificates.crt CRLfile: none while accessing https://git.example.com/public/info/refs fatal: HTTP request failed` </code></pre></div> Huh? Looking deeper into the problem it turns out that git uses <a href="https://curl.haxx.se/">curl</a> for the http(s) transport and curl refuses to work with SSL certificates it cannot verify. Ok, that’s not a bad thing. To circumvent that you can either set an environment variable (<code>GIT_SSL_NO_VERIFY=1</code>) to make curl ignore the verification or install the certificate on your machine. The first option is not very attractive on the long term as you’d have to do it on every operation with the remote server, the second one is not very attractive when dealing with multiple developers working on different operating systems. You’ll have to explain to them how to install the certificate on their machine, and that has to be done every time a new developer joins the team, yada, yada. There is also an option in git (<code>http.sslverify</code>) you can set where you can tell git to ignore the verification of the SSL certificate for that repository. The thing is you still have to set the environment variable on the first clone and then you have to tell git to permanently ignore this issue for that repository with the configuration option – a lot of stuff to remember. Heck, looking on the interwebs I see may of the people with that problem suggesting to shut of SSL cert verification permanently by setting it globally. I really wonder why git cannot simply tell the user that the SSL certificate cannot be verified and if you want to accept it permanently, temporarily or not. Every browser does that. Right now it just quits with an error and leaves the user with a cryptic error. On the other side, when using git with a server providing the repositories via ssh. Git simply asks if you want to accept the key when you access the server for the first time and never bothers you again.Suggestions for a Graphics Tablet?2012-07-26T10:02:00+02:002012-07-26T10:02:00+02:00Bastian Venthurtag:venthur.de,2012-07-26:/2012-07-26-suggestions-for-a-graphics-tablet.html

Suggestions for a Graphics Tablet?

Dear Lazyweb, I’m looking for a good graphics tablet. I’ll mainly use it for post processing DSLR photos with <a href="https://rawtherapee.com/">Raw Therapee</a>, Bibble, and <a href="https://www.gimp.org/">Gimp</a>. It should be cheaper than 200 EUR and must work with Debian/Sid (of course). I’d rather not have to install any fancy drivers, it should work out of the box. Any suggestions?.inputrc2012-06-26T14:46:00+02:002012-06-26T14:46:00+02:00Bastian Venthurtag:venthur.de,2012-06-26:/2012-06-26-inputrc.html

.inputrc

A few days ago I found <a href="https://blog.sanctum.geek.nz/lazier-tab-completion/">this blog</a> explaining how to improve the tab completion by tinkering with your .inputrc. The magic lines are: <div class="codehilite"><pre><code>set show-all-if-ambiguous on set completion-ignore-case on set completion-map-case on </code></pre></div> Last two lines make tab-completion ignore case, hyphens, underscores, the first one spares you one tab when more than one match was found. Very neat! I would have never found out, since there wasn’t even an .inputrc in my home directory. Do you know any other cool .inputrc-tricks?Can aptitude show older versions of packages if available in /var/cache/apt/archives?2012-02-09T11:10:00+01:002012-02-09T11:10:00+01:00Bastian Venthurtag:venthur.de,2012-02-09:/2012-02-09-can-aptitude-show-older-versions-of-packages-if-available-in-varcacheaptarchives.html

Can aptitude show older versions of packages if available in /var/cache/apt/archives?

Dear Lazyweb, is it possible to tell aptitude to show older versions of a package next to the currently available one if it is still present in <code>/var/cache/apt/archives</code>? Like it does when you use unstable and experimental side by side? I know that aptitude does not really support downgrades of packages, but showing those packages directly in aptitude if available in the cache is a lot easier than searching them in the file system and installing them manually, especially if you don’t know where they hide.Helping awstats to correctly interpret lighttpd's log format2012-01-26T11:53:00+01:002012-01-26T11:53:00+01:00Bastian Venthurtag:venthur.de,2012-01-26:/2012-01-26-helping-awstats-to-correctly-interpret-lighttpds-log-format.html

Helping awstats to correctly interpret lighttpd's log format

Note to self: when configuring awstats using lighttpd logfiles, you have to use the following log format: <div class="codehilite"><pre><code>LogFormat="%host %virtualname %logname %time1 %methodurl %code %bytesd %refererquot %uaquot" </code></pre></div> neither 1 nor 4 will work correctly.How to get a list of manually installed packages (and remove the other ones)2011-12-13T23:31:00+01:002011-12-13T23:31:00+01:00Bastian Venthurtag:venthur.de,2011-12-13:/2011-12-13-how-to-get-a-list-of-manually-installed-packages.html

How to get a list of manually installed packages (and remove the other ones)

<h2>The Good</h2> Every other year or so I feel the need to clean up my Debian system and remove the installed packages I’m not interested in anymore. I remember there was a nice aptitude pattern to search for packages which I have manually installed (i.e. which were not installed to satisfy a dependency). Ideally I would then go through the (presumably short) list of packages and remove the ones I don’t need any more. Since I always forget the aptitude pattern to search for those packages, I google for something like “list of manually installed packages” and find a solution like: <code>aptitude search ‘~i !~M’</code>. Although this solution is not wrong, it is not quite what I was looking for. Sure, it will find you all packages which are installed and not installed to satisfy a dependency, but it also contains packages of priorities: required, important and standard which you usually don’t want to deinstall. So the actual solution should also exclude those packages. <div class="codehilite"><pre><code>aptitude search '~i !~M !~pstandard !~pimportant !~prequired' </code></pre></div> The above line will do the trick and present you a list of packages which are: installed and not dragged in to satisfy a dependency and not of priority standard, important or required. If you where only interested in the solution of the quest of getting a list of all manually installed packages you can stop reading here. <h2>The Bad</h2> If you’re running your system for longer than, say, a few weeks you will probably notice that the list is very long and contains much more packages than you actually installed manually. You will probably notice a lot of libs and other packages in this list and wonder why they aren’t marked <code>auto</code> as they should since you didn’t install them on purpose. Unfortunately I don’t have an answer for that question. I’m using aptitude exclusively to install packages and rely on the fact that aptitude does install the dependencies automatically and more importantly: also removes them if not needed any more. But it seems like something is broken in aptitude (or somewhere else) and packages simply lose their <code>auto</code> status instead of being removed and thus stay on your system forever if you don’t manually clean your system. <h2>The Ugly</h2> To solve that issue I started aptitude and filtered the package tree with the above pattern (press <code>l</code> in aptitude and enter <code>~i !~M !~pstandard !~pimportant !~prequired</code>). Now the package view shows only packages which are safe to remove since they are not of of priority standard or higher and manually installed by you. Now I navigate with the cursor to the list of installed packages an mark all of them as <code>auto</code> by pressing <code>M</code>. For each package in this list, aptitude will try to find an installed package which depends on that package and mark that package <code>auto</code>. The remaining packages which no other packages depend on, aptitude will mark for deletion. If you now press <code>g</code> one time, aptitude will show you which packages will be removed. I then go through that list and mark every single of the remaining packages I really want on my system with <code>+</code> as manually installed. After I’m done I proceed by pressing <code>g</code> again which will remove all remaining packages. In my case nearly 600 packages (1GB) where removed. <h2>Bugreport!</h2> I assume there is a bug in aptitude or some related package which sometimes(!) removes a packages <code>auto</code> status instead of removing it. Unfortunately I did not find a way to reproduce that (except that it obviously does happen). If someone could find a way to reproduce that problem, one could fill a bugreport and help to get that bug fixed. Right now, a Debian system suffers (like many other operating systems) from a slow but steadily growing bloat of installed software. In my case 600 packages with 1GB accumulated over one or two years.No sound with Gnome 3 and pulseaudio?2011-11-16T08:43:00+01:002011-11-16T08:43:00+01:00Bastian Venthurtag:venthur.de,2011-11-16:/2011-11-16-no-sound-with-gnome-3-and-pulseaudio.html

No sound with Gnome 3 and pulseaudio?

Dear Lazyweb, I recently switched from KDE to Gnome 3 and I am so far very happy with the decision. One thing that bothers me since then is that the sound does not work when restarting the computer or waking it up from suspend. After a while I figured that a simple <code>sudo killall pulseaudio</code> fixes the problem – but only until the next reboot. Has anyone a hint what the problem could be/ I found quite a few similar problem descriptions on the internet but with no satisfying solutions so far.Speech Recognition Software2011-10-05T12:10:00+02:002011-10-05T12:10:00+02:00Bastian Venthurtag:venthur.de,2011-10-05:/2011-10-05-speech-recognition-software.html

Speech Recognition Software

Dear Lazyweb, Having broke my elbow last week, I’m experiencing serious trouble typing with only one hand on the computer. I know that there is speech to text software, but I was never really interested in that – until now of course. Is there any good software available on Linux you can recommend?KDE4's annoying taskbar2011-08-25T09:12:00+02:002011-08-25T09:12:00+02:00Bastian Venthurtag:venthur.de,2011-08-25:/2011-08-25-kde4s-annoying-taskbar.html

KDE4's annoying taskbar

Dear Lazyweb, since KDE4 is in unstable, the taskbar behaves very annoyingly when working with two monitors with different resolutions. I have a laptop. When I’m at work, I put it into a docking station which is connected to a huge monitor. I’m not working with dualhead or something, I’m just using the huge monitor at work and the laptop’s screen at home. I always want the taskbar to be maximized. The problem is, that KDE’s taskbar fails to adjust the size of the taskbar correctly: Starting KDE at work, the taskbar is always too small (as wide as the laptop’s monitor) and my first task is to maximize the taskbar. Starting KDE when I’m back home, the taskbar is too wide (as wide as the external monitor’s size) for the smaller monitor and I have to (ironically) maximize it again to fit on the now smaller monitor. Note that I’m not using suspend or something. I’m booting the laptop with the big screen connected or not. I also had that problem with KDE3, but that was fixed upstream long time ago. Has anyone a solution how to avoid this problem?Happy Birthday Debian!2011-08-17T11:25:00+02:002011-08-17T11:25:00+02:00Bastian Venthurtag:venthur.de,2011-08-17:/2011-08-17-happy-birthday-debian-2.html

Happy Birthday Debian!

<a href="https://www.debian.org/News/2011/20110816">18 Years</a> and still rocking!Not coming to DebConf2011-06-30T10:22:00+02:002011-06-30T10:22:00+02:00Bastian Venthurtag:venthur.de,2011-06-30:/2011-06-30-not-coming-to-debconf.html

Not coming to DebConf

Unfortunately Debian cannot sponsor my trip to Banja Luka this year and it’s too expensive to pay it from my own budget. Thus I just cancelled my attendance for the conference. I’m very sad, since my first DebConf last year in NY was a lot of fun and I was really looking forward to see you guys again this year. I wish you a lot of fun anyways. Have a productive and fun conference!I'm going to DebConf112011-05-30T09:27:00+02:002011-05-30T09:27:00+02:00Bastian Venthurtag:venthur.de,2011-05-30:/2011-05-30-im-going-to-debconf11.html

I'm going to DebConf11

Weee! After the excellent experience last year in New York, I couldn’t resist :) See you all in Banja Luka!Installing Dropbox in Debian/Sid2011-04-28T20:07:00+02:002011-04-28T20:07:00+02:00Bastian Venthurtag:venthur.de,2011-04-28:/2011-04-28-installing-dropbox-in-debiansid.html

Installing Dropbox in Debian/Sid

A few months ago dropbox was a Debian packge in non-free. Recently I noticed that there are several new versions available and tried to notifiy the maintainer and ask him to update the package. But unfortunately it looks like the package is not available any more (<a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=610300">Here’s why</a>). However getting Dropbox is quite easy if you know what to do. The obvious choice would be to go to <a href="https://dropbox.com">dropbox.com</a> and download either the source package and compile it yourself or try to install the precompiled Ubuntu packages. The first option is not really an option for most users and the second one does currently not work on Debian/Sid (aka unstable). However, there is an easy solution: on the download page search for the CLI script (currently it is named dropbox.py) Download that and run it with <code>python dropbox.py start -i</code>. That will download the propritary binaries into <code>~/.dropbox-dist</code>. Now go into that folder and start <code>dropboxd</code>. There you go, you have a fully working systray app for dropbox. Depending on your desktop environment you might want to put <code>dropboxd</code> in your “autostart”. So you don’t have to start the daemon every time you login. And that’s it. It is not really ideal, since a Debian package would be much better for updates, security, yada yada, but it’s still better than nothing.Introducing python-ardrone2011-04-20T14:47:00+02:002011-04-20T14:47:00+02:00Bastian Venthurtag:venthur.de,2011-04-20:/2011-04-20-introducing-python-ardrone.html

Introducing python-ardrone

<img alt="Flying AR.Drone" src="/images/2011-04-15_10-41-55-small.jpg" title="AR.Drone"> The last weeks I spend quite a lot of time hacking on a Python library for the <a href="https://ardrone.parrot.com/">AR.Drone</a>. The AR.Drone is a nice toy for nerds. You connect to it via WIFI and soon you’ll realize that it has 4 ports open. Reading the specs you’ll find, that on one port it listens for AT-Commands with which you can remote control the drone, on the other two ports it waits for an incoming package which will trigger the drone to send the navdata (speed, angles, battery status, etc) and the video stream. Heck, you can even telnet into the drone… Unfortunately it comes without a proper software to control the drone, only an iPhone app (w/o iPhone of course). But given the documentation, it should be easy to write your own. While getting the beast to fly was relatively easy, decoding the “almost”-jpg-video-stream was not. Almost-jpg, since the images the drone sends are more or less jpg with a small difference which makes it impossible to decode them using standard multi-media libraries. Anyways, the format is documented and implementing a decoder was not that hard. The tricky part was to get the framerates from unacceptable 0.5 FPS to 12-22 FPS – the whole decoder is written in Python. I’m cheating a bit by using <a href="https://psyco.sourceforge.net/">psyco</a>, but the code in arvideo.py is heavily optimized to minimize calculations and to please psyco. In the code is also a small demo app which uses <a href="https://pygame.org/">Pygame</a> to display the video stream and allows to control the AR.Drone with the keyboard. It should be ready-to-use as soon as you are connected to the drone via WIFI. The git repository is <a href="https://github.com/venthur/python-ardrone">here</a>, the license is MIT. Suggestions and patches are welcome. <a href="https://youtu.be/EmQOEX9a3tM">Here</a> is a video of the drone flying through the office.Dear Lazyweb2011-04-06T10:31:00+02:002011-04-06T10:31:00+02:00Bastian Venthurtag:venthur.de,2011-04-06:/2011-04-06-dear-lazyweb-2.html

Dear Lazyweb

After todays or yesterdays daily package update (Sid) several kernel modules, namely <code>thinkpad_acpi</code>, <code>snd-hda-intel</code> and probably others are not loaded automatically anymore. So my <a href="2008-11-25-lenovo-t500-shiny-new-laptop.html">Thinkpad T500</a> had no support for audio and several power management functions anymore. Once I found out that the missing kernel modules where the problem, the fix was easy: just add the corresponding module names to <code>/etc/modules</code> to enforce loading of the modules. But I wonder what caused the sudden change? Why are the modules not loaded automatically anymore? I usually update my Sid packages every morning so some package update from yesterday or today might have caused this. Looking at the packages which where updated, I found nothing suspicious. Has anyone an idea at which package I should look? Update: <a href="https://bugs.debian.org/620995">Looks</a> like the problem was related to the recent update of <code>base-files</code> to version 6.2 which introduced the new <code>/run</code> directory in combination with udev. Downgrading <code>base-files</code> to version 6.1 (via <a href="https://snapshot.debian.org/package/base-files/6.1/#base-files_6.1">snapshot.debian.org</a>) fixed the problem.Summer School: Advanced Scientific Programming in Python2011-03-01T16:03:00+01:002011-03-01T16:03:00+01:00Bastian Venthurtag:venthur.de,2011-03-01:/2011-03-01-summer-school-advanced-scientific-programming-in-python.html

Summer School: Advanced Scientific Programming in Python

Once again, there will be another round of the Summer School “Advanced Scientific Programming in Python”. This year in St. Andrews, UK. Quoting from the <a href="https://python.g-node.org/wiki/">official announcement</a>: <blockquote> Scientists spend more and more time writing, maintaining, and debugging software. While techniques for doing this efficiently have evolved, only few scientists actually use them. As a result, instead of doing their research, they spend far too much time writing deficient code and reinventing the wheel. In this course we will present a selection of advanced programming techniques, incorporating theoretical lectures and practical exercises tailored to the needs of a programming scientist. New skills will be tested in a real programming project: we will team up to develop an entertaining scientific computer game. We use the Python programming language for the entire course. Python works as a simple programming language for beginners, but more importantly, it also works great in scientific simulations and data analysis. We show how clean language design, ease of extensibility, and the great wealth of open source libraries for scientific computing and data visualization are driving Python to become a standard tool for the programming scientist. This school is targeted at PhD students and Post-docs from all areas of science. Competence in Python or in another language such as Java, C/C++, MATLAB, or Mathematica is absolutely required. Basic knowledge of Python is assumed. Participants without any prior experience with Python should work through the proposed introductory materials before the course. </blockquote> You can apply on-line at <a href="https://python.g-node.org">https://python.g-node.org</a> Applications must be submitted before May 29, 2011. Notifications of acceptance will be sent by June 19, 2011. No fee is charged but participants should take care of travel, living, and accommodation expenses. Candidates will be selected on the basis of their profile. Places are limited: acceptance rate in past editions was around 30%. Prerequisites: You are supposed to know the basics of Python to participate in the lectures. Please consult the website for a list of introductory material. If you’re a scientist and interested in Python, I cannot recommend this summer school highly enough. The Summer School is always fun, the faculty members are very nice and always willing to help and answer questions. The balance between lectures and exercises is in my opinion very good, and you’ll learn a lot about Python during the week. The highlight of every Python School is always the Pac Man tournament, where groups of students will implement their own Pac Man agent (in Python) and compete against the other agents in a tournament ‘till the bitter end! The students of the previous editions were usually very happy during and after the School. Speaking of happiness, since no fee is charged for the Summer School, you’ll have more money left over to invest into the excellent Scottish pubs!Squeeze2011-02-06T11:31:00+01:002011-02-06T11:31:00+01:00Bastian Venthurtag:venthur.de,2011-02-06:/2011-02-06-squeeze.html

Squeeze

Tonight Squeeze was released! Along with the release came a redesign of our <a href="https://debian.org">website</a> and <a href="https://planet.debian.org">planet</a>. Nice! I’m particularly happy that we’ll see lots of fresh software and updates in unstable again, now that it is not frozen anymore. Thanks to all contributors. Keep up the good work!Canon IXUS 1302010-12-30T01:30:00+01:002010-12-30T01:30:00+01:00Bastian Venthurtag:venthur.de,2010-12-30:/2010-12-30-canon-ixus-130.html

Canon IXUS 130

Dear Lazyweb, Today I bought me a Canon IXUS 130 and was quite disappointed to see that it couldn’t properly connect to my Debian/Sid machine via USB. Apparenty the camera is not recognized as a USB Mass Storage Device but utilizes <a href="https://en.wikipedia.org/wiki/Picture_Transfer_Protocol">PTP</a>. I upgraded to libgphoto2 (2.4.10.1-3) from experimental and added a proper udev rule with vendor- and product-id but it did not help. Is there anything more I can try? Update: Accessing the camera actually works, even with ligbphoto2 (2.4.6) from unstable. I can access the camera with gphoto2 and download pictures (but not with KDE’s filemanager). My problem was that I had <a href="https://projects.gnome.org/rhythmbox/">rhythmbox</a> running and the camera was automatically mounted by it as a media device. When KDE’s filemanager or gphoto2 tried to access the camera I got an error like: (‘Could not lock the device’): Camera is already in use., which is probably also a bug. Closing rhythmbox solved the issue and I was able to download images with gphoto2.Google Summer of Code 2010 Mentor Summit Report2010-10-25T22:09:00+02:002010-10-25T22:09:00+02:00Bastian Venthurtag:venthur.de,2010-10-25:/2010-10-25-google-summer-of-code-2010-mentor-summit-report.html

Google Summer of Code 2010 Mentor Summit Report

As requested by our beloved leader, here’s my report on this year’s Google Summer of Code 2010 Mentor Summit. The summit took place in Google’s HQ in Mountain View, California and was what they call an “unconference”. In the beginning of the unconference the attendees had 10 minutes to write down proposals for a session and put them on a huge white board. After that, every attendee voted for the sessions in which he was interested by putting a small sticker on the proposal. This way we could evaluate the interest on a specific topic and schedule the sessions to the differently sized rooms according to their interest. Neat trick! The the rest of the summit was basically like a conference: we had 11 tracks filled with 5-6 sessions per day, which ran from 10:00 to 17:30. Google provided shuttle buses from our wonderful hotel to the googleplex in the morning and back to the hotel in the evening. The hotel was very nice, it was equipped a swimming pool and a hot tub (both outside), which was extensively used in the morning and of course during the night. Speaking of the night, Google sponsored Thai food in the first night and ye olde Pizza and Beer in the second. Some of the attendees brought their keyboard, guitar, bass and electric clarinet, and where playing some sweet ass Jazz and Pop (on special request…). The sessions itself where not quite what I expected. I expected some proper talks about specific topics, but it was more like and open discussion round where everyone was invited to participate. For some topics it worked pretty well, for others not. It was still totally worth coming, since you meet a lot of like minded people and there is plenty stuff to talk about. Of course I was very exited to see the googleplex. And the rumors are true: the googleplex seems to be a very nice place to work. They provide free breakfast, lunch and dinner for their employees. They have micro kitchens everywhere where you can grab free snacks and beverages. There are lot’s of toys lying around, they have bicycles everywhere on the campus which you can just grab to move from one building to another. There is a dinosaur on the yard, a spaceship over the stairwell and lots of other things (testing on the toilet, anyone?), and we have probably only seen a small fraction, since we were not allowed to leave the restricted areas for the unconference. Despite the horrible yet lag (leaving Germany on Friday morning and leaving San Francisco on Sunday night), I’m glad I was able to attend to the summit and can only recommend aspiring mentors and students to participate to the next summer of code!Introducing python-popcon2010-10-16T17:10:00+02:002010-10-16T17:10:00+02:00Bastian Venthurtag:venthur.de,2010-10-16:/2010-10-16-introducing-python-popcon.html

Introducing python-popcon

<a href="https://github.com/venthur/python-popcon">Python-popcon</a> is a small Python library which allows to query Debian’s popcon database. The usage is very trivial: <div class="codehilite"><pre><code>>>> import popcon >>> popcon.package('python-popcon') {'python-popcon': 2} >>> popcon.package('icedove', 'iceweasel') {'icedove': 12140, 'iceweasel': 45666} >>> popcon.package('foobarbaz') {} </code></pre></div> You can call the <code>package</code> method with an arbitrary number of package names and it will return a dictionary with <code>package name : popcon value</code> mappings for all packages found. If a package is not found in the popcon database it will not be in the resulting dictionary or if it was the only package in the method call, the method will return an empty dictionary. The popcon value is the number of installations according to Debian’s Popularity Contest service. There is also a second method which gives some more information <div class="codehilite"><pre><code>>>> popcon.package_raw('icedove', 'iceweasel') {'icedove': [8065, 2195, 1879, 1], 'iceweasel': [27857, 10681, 7120, 8]} </code></pre></div> It returns a list of <code>[vote, old, recent, no-files]</code> for the given package. How does it work? Upon a query, popcon downloads https://popcon.debian.org/all-popcon-results.txt.gz, extracts and parses the file and returns the desired information. Since downloading and extraction this file is expensive, it saves the file and tries to re-use it is not older than 7 days. So while the first call of one of the above methods can take a few seconds, the result will appear almost instantaneously for the next week until the cache file expired. That’s it, no rocket science but a convenient way to get popcon information from within python. The package python-popcon is available in Debian/Sid for a few days.Weird Konsole Split2010-09-21T22:13:00+02:002010-09-21T22:13:00+02:00Bastian Venthurtag:venthur.de,2010-09-21:/2010-09-21-weird-konsole-split.html

Weird Konsole Split

Can someone explain me the purpose of (KDE’s) konsole’s ability to split? <img alt="" src="/images/konsole-split.png" title="konsole-split"> You can split horizontally or(!) vertically but the resulting terminals are just mirroring the activity of the first one. What is the usecase for that? <a href="https://extragear.kde.org/apps/yakuake/">Yakuake</a> does splitting right. You can split several times horizontally and vertically and the resulting terminals are independent. You can also switch the focus of the terminals easily via keystrokes which makes yakuake an invaluable tool when working on the terminal under KDE. <img alt="" src="/images/konsole-split2.png" title="konsole-split2"> So what’s the point of konsole’s variant of splitting?Happy Birthday Debian!2010-08-16T18:25:00+02:002010-08-16T18:25:00+02:00Bastian Venthurtag:venthur.de,2010-08-16:/2010-08-16-happy-birthday-debian.html

Happy Birthday Debian!

<img alt="" src="/images/birthdaylolcat.jpg" title="Happy Birthday Debian!"> 17 years… I guess the puberty is almost a thing of the past and you’re getting all grown up and stuff. You certainly changed for the better in the last years. Ok, a few rough edges here and there but overall a quite impressive development. You’re already forking regularly and even your offspring is quite handsome. No wonders with those genes if you ask me. Thanks to <a href="https://en.wikipedia.org/wiki/Ian_Murdock">Ian</a> for starting probably the best Linux distribution out there. I hope you are proud of what you’ve started (I certainly was when I met you <a href="2008-05-31-linuxtag-2008.html">two years ago</a>) and someone is killing a few beers with you right now. Also big thanks to all the Debian Developers, Maintainers and contributors for making Debian what it is today. Keep up the good work!Goodbye DebConf102010-08-09T18:28:00+02:002010-08-09T18:28:00+02:00Bastian Venthurtag:venthur.de,2010-08-09:/2010-08-09-goodbye-debconf10.html

Goodbye DebConf10

DebConf was really an awesome conference. I’m happy that I have finally met so many of the fellow Debian people I only knew from the mailing lists or planet. There where lot’s of interesting talks, the hacklabs where always busy, and the overall atmosphere was very good. Now that I’m back in Berlin, I’m already missing it – I’m by the way also missing my luggage which is apparently still at the JFK airport. This was my first DebConf, but definitively not my last. I encourage people who never made it to DebConf: Get your butt up and register for the next conference! You don’t know what you’re missing. If money is an issue: there is a sponsoring which you can apply for and chances are good that you’ll get at least partially sponsored. Big thanks to all the volunteers who helped to organize and run this conference and to SPI for sponsoring my trip, food and accommodation! I’m looking forward seeing some of you again in three weeks at Steve’s place in Cambridge for a nice BBQ and the rest of you next year in Bosnia and Herzegovina!Debconf: Day 22010-08-02T04:28:00+02:002010-08-02T04:28:00+02:00Bastian Venthurtag:venthur.de,2010-08-02:/2010-08-02-debconf-day-2.html

Debconf: Day 2

<img alt="" src="/images/interesting-cat.jpg" title="interesting-cat"> Second day in the big city and after the obligatory pancakes, scrambled eggs and bacon breakfast, I spend almost the entire day in the hacklab fixing the <a href="https://wiki.debian.org/DebbugsSoapInterface">documentation</a> for the debbugs SOAP interface. Thanks to Don I think I finally have the <code>get_status</code> part at an accurate state. Accordingly I was very busy making changes in <a href="https://github.com/venthur/python-debianbts">python-debianbts</a> which uses the SOAP interface to query the BTS. Some Bugreport attributes disappeared, others got their data type fixed, Unittests where added and docstrings updated. Finally I uploaded the new version to unstable. Between that mess I met a lot of nice people, and heard a talk whose slides consisted almost entirely of <a href="https://icanhascheezburger.com/">lolcat images</a> – which was of course awesome! I definitively have to try that in one of my next scientific talks. Hopefully tomorrow I’ll find some time to actually prepare my talk.linen and towels2010-07-30T10:36:00+02:002010-07-30T10:36:00+02:00Bastian Venthurtag:venthur.de,2010-07-30:/2010-07-30-linen-and-towels.html

linen and towels

Dear lazyweb or someone who is already in NY, do I have to bring my own bed linen and towels or are they provided by Furnald? I couldn’t find the information on the website. I guess I have to bring that stuff since it is not really a hotel, right?I'm going to DebConf102010-07-13T21:43:00+02:002010-07-13T21:43:00+02:00Bastian Venthurtag:venthur.de,2010-07-13:/2010-07-13-im-going-to-debconf10.html

I'm going to DebConf10

<img alt="" src="/images/im_going_to_debconf10.png" title="im_going_to_debconf10"> Awesome! Thanks to all the sponsors and all the people helping to organize this year’s DebConf for making it happen! See you soon in New York City!I may go to Debconf 102010-05-30T17:38:00+02:002010-05-30T17:38:00+02:00Bastian Venthurtag:venthur.de,2010-05-30:/2010-05-30-i-may-go-to-debconf-10.html

I may go to Debconf 10

Today I received a mail from the DebConf Travel Sponsorship Team and they informed me that I may get partially (or fully) sponsored. Ok that’s a bit vague for now, but that also means that I may come to Debconf 10 – yeah! I also applied for this Debconf Newbies thingy, as this will be the first DebConf for me, but unfortunately there where too many applicants and not enough money and they had to decline my request.Github announces SVN support2010-05-05T09:09:00+02:002010-05-05T09:09:00+02:00Bastian Venthurtag:venthur.de,2010-05-05:/2010-05-05-github-announces-svn-support.html

Github announces SVN support

This is freakin’ awesome! It is now possible to <a href="https://github.com/blog/626-announcing-svn-support">checkout from-</a> and <a href="https://github.com/blog/644-subversion-write-support">commit to</a> git repositories on github with SVN. This just made my day. This could also mean that this SVN support may later be added into official git (like the CVS emulation), allowing the early movers to use git also on the repository side while staying backwards compatible with the rest who is still using SVN on the client side. Currently it’s more the other way ‘round, the wherever SVN is used, git users have to resort to git-svn.Running Gnome Applets in KDE2010-05-05T08:32:00+02:002010-05-05T08:32:00+02:00Bastian Venthurtag:venthur.de,2010-05-05:/2010-05-05-running-gnome-applets-in-kde.html

Running Gnome Applets in KDE

Dear Lazyweb, does anyone know how to run gnome applets in KDE? I’m talking about <a href="https://packages.debian.org/byzanz">byzanz</a> which is a software to record your desktop in an animated gif, ogg or Flash format. It provides a command line and a gnome applet, which apparently only works under Gnome. Is there a workaround to make gnome applets run under KDE as well?Debbugs RW-SOAP API Project in GSoC 20102010-04-09T10:31:00+02:002010-04-09T10:31:00+02:00Bastian Venthurtag:venthur.de,2010-04-09:/2010-04-09-debbugs-rw-soap-api-project-in-gsoc-2010.html

Debbugs RW-SOAP API Project in GSoC 2010

<img alt="Google Summer of Code 2010" src="/images/gsoc2010_300x267px.jpg" title="gsoc2010"> This year again, I’m <a href="https://wiki.debian.org/SummerOfCode2010/DebbugsAPI">proposing a project</a> for this year’s Summer of Code. Very much like my last year’s proposal, it is about adding submit- and manipulation capabilities to <a href="https://wiki.debian.org/DebbugsSoapInterface">debbugs’ SOAP interface</a>. The idea for this project is to add another way to communicate with our bug tracking system than email. Since we already have a read-only SOAP interface, it seems natural to add write-capabilities. Libraries like <a href="https://packages.qa.debian.org/p/python-debianbts.html">python-debianbts</a> would adopt those features and end user applications like <a href="https://packages.qa.debian.org/r/reportbug.html">reportbug</a> or <a href="https://packages.qa.debian.org/r/reportbug-ng.html">reportbug-ng</a> could use it to enhance usability. One student has already applied. Others are also very <a href="https://wiki.debian.org/SummerOfCode2010/DebbugsAPI">welcome</a> – but be quick, the deadline for student applications is April 9th at 19:00 UTC.apt-get update slow when LANG != C?2010-03-29T09:31:00+02:002010-03-29T09:31:00+02:00Bastian Venthurtag:venthur.de,2010-03-29:/2010-03-29-apt-get-update-slow-when-lang-c.html

apt-get update slow when LANG != C?

For a few weeks now, aptitude is really slow updating the package list – downloading the lists is actually fast as normal but it always waits for a minute or so with a <code>99% [Warten auf Kopfzeilen]</code> (Waiting for headers) message. I tried apt-get update – same problem. Now the funny thing is: <code>LANG=C apt-get update</code> or<code>LANG=C aptitude -u</code> works just fine! I also tested with <code>LANG=de_DE</code>, <code>it_IT</code>, <code>fr_FR</code> and even <code>en_US</code> – always slow, so it looks it always occurs when <code>LANG</code> is not set to C? Anyone else noticed this problem? I skimmed through the bug reports of apt, but didn’t find a similar bug. Update: Looks like google-chrome is the problem! Commenting out the content of <code>/etc/apt/sources.list.d/google-chrome.list</code>, as Jaime suggested in the comments, everything worked fine again.Reportbug-ng can now hide closed bugs2010-02-21T17:30:00+01:002010-02-21T17:30:00+01:00Bastian Venthurtag:venthur.de,2010-02-21:/2010-02-21-reportbug-ng-can-now-hide-closed-bugs.html

Reportbug-ng can now hide closed bugs

Today I finally found the reason why the table in reportbug-ng was not always sorted correctly. The fix was trivial and I’m happy it’s finally corrected. As <a href="2010-02-18-reportbug-ng-now-supports-complex-queries.html">announced</a> last week, reportbug-ng now also can optionally hide closed bugs, which makes reportbug-ng together with the complex queries a great tool for finding easy NMU candidates. To play around with those new features, I also did 5 lazy NMUs today. Most of them where fixes for FTBFS/RC bugs which had already a patch in the BTS.Reportbug-ng now supports complex queries2010-02-18T21:54:00+01:002010-02-18T21:54:00+01:00Bastian Venthurtag:venthur.de,2010-02-18:/2010-02-18-reportbug-ng-now-supports-complex-queries.html

Reportbug-ng now supports complex queries

Until today you could only use reportbug-ng to query the <a href="https://bugs.debian.org/">BTS</a> with simple queries like “packagename”, “bugnumber”, “tag:patch”, etc. But the BTS actually supports composite queries like “severity:grave tag:patch” which returns bugreports with severity grave and a patch. The underlying Python library <a href="https://github.com/venthur/python-debianbts">python-debianbts</a> also supported this right from the start, but reportbug-ng did not make use of it. Last weekend I finally had the time to fix that and the result is on it’s way to unstable. Composite queries provide a very convenient way to find cheap NMU candidates: the query <code>"severity:critical severity:grave severity:serious tag:patch"</code> will return release critical bugs which have a patch. Now you can just go through this list, pick an open bug, test the patch and do what’s necessary to release Squeeze in time. Next item on my list is an option to hide closed bugs, maybe next weekend.Links not opening in Thunderbird 3?2010-02-17T16:42:00+01:002010-02-17T16:42:00+01:00Bastian Venthurtag:venthur.de,2010-02-17:/2010-02-17-links-not-opening-in-thunderbird-3.html

Links not opening in Thunderbird 3?

If you’re using KDE and upgraded to Thunderbird 3 lately you might have the problem that links in <a href="https://bugs.debian.org/566819">emails don’t open in a browser anymore</a>. For my case that happened even without a user visible error message. If that sounds familiar for you, check the error console (Tools/Error-Console). If there is an error like: <div class="codehilite"><pre><code>Error: uncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIExternalProtocolService.loadUrl]" nsresult: "0x80004005 (NS_ERROR_FAILURE)" location: "JS frame :: chrome://communicator/content/contentAreaClick.js :: openLinkExternally :: line 188" data: no] </code></pre></div> Then Thunderbird just (silently) fails to launch the browser. Guido pointed me to gconftool. <code>gconftool -R / | grep url-handlers</code> told me that some components of gnome still had Firefox configured as http-handler (which is strange since it was rebranded to iceweasel ages ago). Resetting them solved the problem for me. The sad part of the story: x-www-browser, sensible-browser and KDE’s http handler where all correctly configured and pointed to iceweasel. All applications behaved correctly only Icedove used a (for me) hidden setting pointing to firefox.Query Google Scholar using Python2010-01-27T17:45:00+01:002010-01-27T17:45:00+01:00Bastian Venthurtag:venthur.de,2010-01-27:/2010-01-27-query-google-scholar-using-python.html

Query Google Scholar using Python

In desperate need to organize my collection of scientific papers, I had a look at various tools which could help me organizing them. Probably one of the best out there is <a href="https://www.mendeley.com/">Mendeley</a>. Mendeley seems to be a very good tool to keep your massive collection of pdfs under control. Unfortunately a very basic function, namely looking up a newly imported paper in <a href="https://scholar.google.com/">Google Scholar</a> to get attributes like: Authors, Year, etc. right, is bundled with a Mendeley account. I guess that’s their way of forcing the user to participate to their community stuff, since without the Google Scholar lookup Mendeley is pretty useless unless you want to fill all the attributes manually. So I decided to write my own tool to make the lookup. Unfortunately Google does not really want to give away that precious data: they don’t provide an API and even block certain User-Agents from accessing the page. Then, there is also the problem of scraping the results page to get the right data. The first problem can be trivially solved by setting a common User-Agent string, the second one can be elegantly circumvented by using the bibtex files provided in the search results. The bibtex entries are however only showed if you enabled them in the settings, which are stored in a cookie. After a few tries, I figured that the <code>CF</code> attribute (citation format?) controls which bibliography format should be offered in the results page and <code>CF=4</code> corresponds to bibtex. Generating a fake cookie is easy, but you have to know what must be included. In this case it looks like a 16 digit hex as ID and the CF attribute is sufficient. The <code>ID</code> is probably supposed to be your id, but a randomly generated one also works like a charm. The resulting cookie looks like this: <code>GSP=ID=762a112b5c765732:CF=4</code> All you have to do now is to query Google Scholar using the user string and the cookie: <div class="codehilite"><pre><code>... # fake google id (looks like it is a 16 elements hex) google_id = hashlib.md5(str(random.random())).hexdigest()[:16] GOOGLE_SCHOLAR_URL = "https://scholar.google.com" HEADERS = {'User-Agent' : 'Mozilla/5.0', 'Cookie' : 'GSP=ID=%s:CF=4' % google_id } def query(searchstr): """Return a list of bibtex items.""" searchstr = '/scholar?q='+urllib2.quote(searchstr) url = GOOGLE_SCHOLAR_URL + searchstr request = urllib2.Request(url, headers=HEADERS) response = urllib2.urlopen(request) html = response.read() # grab the bibtex links ... </code></pre></div> And Google Scholar will offer you links to the bibtex files of the results. Getting those links is easy since they all start with <code>"/scholar.bib"</code>. Just search for those and download the targets. The complete code is available on <a href="https://github.com/venthur/gscholar">github</a>. It can be used as a python library or a standalone application, you just call it like this: <code>gscolar "some author or title"</code> and it will print the first ten results in bibtex to stdout.How to find packages installed/updated yesterday?2010-01-26T10:14:00+01:002010-01-26T10:14:00+01:00Bastian Venthurtag:venthur.de,2010-01-26:/2010-01-26-how-to-find-packages-installedupdated-yesterday.html

How to find packages installed/updated yesterday?

Dear Lazyweb, since yesterday’s daily update, my laptop runs really sluggish. Is there a way to list all packages which where updated or installed yesterday to track down the problem?git bisect, ccache, cowbuilder2010-01-09T16:15:00+01:002010-01-09T16:15:00+01:00Bastian Venthurtag:venthur.de,2010-01-09:/2010-01-09-git-bisect-ccache-cowbuilder.html

git bisect, ccache, cowbuilder

<a href="https://www.kernel.org/pub/software/scm/git/docs/git-bisect.html">Git bisect</a>, <a href="https://ccache.samba.org/">ccache</a> and <a href="https://packages.debian.org/search?keywords=cowbuilder">cowbuilder</a>: a combination made in heaven! Tracking down a commit which introduced an <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=564415">ugly bug</a> with those tools was a breeze. Git bisect is very useful finding a commit which introduced a bug very quickly, ccache massively reduces compiling time. Compiling icedove (thunderbird) on my laptop using cowbuilder takes roughly 30 minutes. Using cowbuilder with ccache, it only takes 10 minutes, where most of the time is spent setting up the build environment.19842010-01-05T11:58:00+01:002010-01-05T11:58:00+01:00Bastian Venthurtag:venthur.de,2010-01-05:/2010-01-05-1984.html

1984

One idiot <a href="https://en.wikipedia.org/wiki/Northwest_Airlines_Flight_253">ignites a bomb</a>, others immediately scream for <a href="https://en.wikipedia.org/wiki/Security_scan">security scanners</a>. Apparently Terrorism works pretty well for countries with increased security needs.Printing2009-12-08T17:26:00+01:002009-12-08T17:26:00+01:00Bastian Venthurtag:venthur.de,2009-12-08:/2009-12-08-printing.html

Printing

Funny coincidence that David <a href="https://www.davidpashley.com/blog/linux/printer-conversation">writes</a> how well Linux and Printers go together for 12 years, while it is apparently <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=558755">impossible</a> to print something with CUPS in unstable since a week ;)The sorry state of Python in Debian2009-12-02T10:45:00+01:002009-12-02T10:45:00+01:00Bastian Venthurtag:venthur.de,2009-12-02:/2009-12-02-the-sorry-state-of-python-in-debian.html

The sorry state of Python in Debian

Looking at the <a href="https://lists.debian.org/debian-devel/2009/12/msg00024.html">sorry</a> state of Python in Debian, makes me wonder if we shouldn’t enforce team maintainance of packages above a certain popularity/importance/whatever threshold. People worked hard in the last months to fix any bugs that would prevent Python2.6 to land in unstable and yet nothing happens. Time passes by and we will eventually end up with Squeeze having a horribly outdated Python version.My eyes! The goggles, they do nothing!2009-11-23T10:26:00+01:002009-11-23T10:26:00+01:00Bastian Venthurtag:venthur.de,2009-11-23:/2009-11-23-my-eyes-the-goggles-they-do-nothing.html

My eyes! The goggles, they do nothing!

Dear Lazyweb, do you know how to theme Tk apps? In the default setting, Tk apps running on my KDE4 desktop, cause spontanous eye cancer: <img alt="Default Tk theme." src="/images/tkeyecancer.png" title="tkeyecancer"> there must be a way to theme them to look like the Qt or at least GTK apps, or not?Bundesverdienstkreuz for Matthias Ettrich2009-11-07T16:11:00+01:002009-11-07T16:11:00+01:00Bastian Venthurtag:venthur.de,2009-11-07:/2009-11-07-bundesverdienstkreuz-for-matthias-ettrich.html

Bundesverdienstkreuz for Matthias Ettrich

<a href="https://en.wikipedia.org/wiki/Matthias_Ettrich">Matthias Ettrich</a>, the founder of KDE <a href="https://www.berlin.de/landespressestelle/archiv/2009/11/04/144882/">earned</a> the Bundesverdienstkreuz (<a href="https://en.wikipedia.org/wiki/Federal_Cross_of_Merit">Federal Cross of Merit</a>) last Friday for his engagement in open source. How cool is that? Congratulations Matthias and thank you (and all the others) for the K Desktop Environment!Python2.6 Blockers2009-11-07T13:42:00+01:002009-11-07T13:42:00+01:00Bastian Venthurtag:venthur.de,2009-11-07:/2009-11-07-python26-blockers.html

Python2.6 Blockers

Today’s work: 5 lazy NMUs (thanks again Kumar). Leaves us with only five open <a href="https://bugs.debian.org/cgi-bin/pkgreport.cgi?users=debian-python@lists.debian.org;tag=python2.6">python2.6 blockers</a> to fix and a whopping 62 of closed ones.reportbug-ng has localization support again2009-10-25T18:58:00+01:002009-10-25T18:58:00+01:00Bastian Venthurtag:venthur.de,2009-10-25:/2009-10-25-reportbug-ng-has-localization-support-again.html

reportbug-ng has localization support again

After having ported reportbug-ng from PyQt3 to PyQt4 <a href="2008-07-13-porting-rng-to-qt4.html">over a year ago</a>, reportbug-ng lost it’s localization, since the gettext based translations where incompatible with Qt4’s translation system. This weekend I finally had the time to have a closer look at this problem. To make a long story short: I have ported the gettext based system to Qt4’s system. All the old .po files where converted to .ts files, but almost all strings are marked as “obsolete” so that they don’t appear in the translated program. But since they are still available in the .ts file, it is easy to get the translations up-to-date. So far only English and German are complete, but eventually other translations will be added. PyQt4 makes it by the way really hard to get non-Qt strings translated.Amarok going downhill2009-10-20T17:13:00+02:002009-10-20T17:13:00+02:00Bastian Venthurtag:venthur.de,2009-10-20:/2009-10-20-amarok-going-downhill.html

Amarok going downhill

Amarok 2.2 is probably the only audio application on the market <a href="https://bugs.debian.org/540203">unable</a> connect properly to the IPod. It is <a href="https://ubuntuforums.org/showthread.php?t=1151706">impossible</a> to copy podcasts within Amarok to the IPod since Amarok 2.x. This functionality simply disappeared during the 1.x to 2.x version jump. You can download and listen to podcasts, you can connect your IPod, but you cannot drag the podcast to the IPod. A few weeks ago you could help yourself if you know where Amarok stores the podcasts (<code>~/.kde/share/apps/amarok/podcasts</code>): just navigate to this directory and drag the audio file to the IPod, but looks this rake was also disabled within the last versions, since Amarok now only segfaults if you dare to do this. Great.Python 2.6 Transition2009-10-17T15:45:00+02:002009-10-17T15:45:00+02:00Bastian Venthurtag:venthur.de,2009-10-17:/2009-10-17-python-26-transition.html

Python 2.6 Transition

Today I NMUed over a dozen of Python packages with <a href="https://bugs.debian.org/cgi-bin/pkgreport.cgi?users=debian-python@lists.debian.org;tag=python2.6">bugs which blocked the Python 2.6 transition</a>. I really want to thank Kumar Appaiah for his work. He provided patches for all the bugs I NMUed today and lots more. I really did not much more than applying, testing and uploading his patches, but Kumar probably invested days of labor to create the patches and test them. Thanks to his effort, the number of 2.6-blockers shrinked considerably so that we now have like ~15 open blockers and ~50 closed ones!python-debianbts 1.0 uploaded to unstable2009-10-10T20:36:00+02:002009-10-10T20:36:00+02:00Bastian Venthurtag:venthur.de,2009-10-10:/2009-10-10-python-debianbts-10-uploaded-to-unstable.html

python-debianbts 1.0 uploaded to unstable

Today I was working all day on <a href="https://github.com/venthur/python-debianbts">python-debianbts</a> 1.0 and uploaded it to unstable a few minutes ago. This version breaks backwards compatibility with previous versions. I removed lots of unneeded old cruft like the HTMLStripper class needed ages ago when I was still using HTML instead of debbugs’ SOAP interface. A new method <code>get_usertag(email, *tags)</code> was introduced. It returns a dict containing usertag-buglist mappings. If tags are given the dict is limited to matching tags, otherwise all available tags of the given user are returned: <div class="codehilite"><pre><code>In [1]: import debianbts as bts In [2]: bts.get_usertag("debian-python@lists.debian.org") Out[2]: {'dist-packages': [547838, 547832, ..., 547858], 'dpmt-todo': [332913], 'policy': [373301, 373302, ..., 377089], 'python-oldnum': [478467, 478442, ..., 478441], 'python2.1': [351108, 351110, ..., 351131], 'python2.2': [351108, 351109, ..., 351161], 'python2.6': [547838, 547832, ... 547858]} In [3]: bts.get_usertag("debian-python@lists.debian.org", "python2.1", "python2.2") Out[3]: {'python2.1': [351108, 351110, ..., 351131], 'python2.2': [351108, 351109, ..., 351161]} </code></pre></div> <code>get_bug_log(nr)</code> now returns a list of dicts with the keys: <code>header</code> (string), <code>body</code> (string), <code>msg_num</code> (int) and <code>attachments</code> (list). Before 1.0 it returned a list of Buglog objects. The Bugreport class now supports every information provided by the SOAP interface. I tried to stay as close as possible to the data SOAP provides, so I renamed existing attributes (like Bugreport.nr which is not supported by SOAP but is now <code>Bugreport.bug_num</code>) and also added the quirky ones like: <code>id</code> and <code>bug_nr</code>, <code>found</code> and <code>found_versions</code>, <code>keywords</code> and <code>tags</code>, <code>fixed</code> and <code>fixed_date</code> which always seem to provide the same data. Instead of the Bugreport.value() method which provided a number representing the openness (in terms of: open, closed and archived) and urgency (like: grave, important, …) to make bugreports sortable by their status, the Bugreport class now has a <code>__cmp__</code> method which makes bugreports comparable. The more open and urgent a bug is, the greater it is. Openness always beats urgency (eg: an open whishlist bug is greater than a closed grave one). While pre 1.0 versions of python-debianbts more or less served the needs of <a href="https://github.com/venthur/reportbug-ng">reportbug-ng</a>, it now tries to stay as close as possible to the data provided by SOAP. As a result many parts of reportbug-ng had to be fixed for the new version. I hope this makes python-debianbts more attractive for other projects dealing with Debian’s bug tracker. As always: python-debianbts is on <a href="https://github.com/venthur/python-debianbts">github</a> and forks, patches or other kinds of collaboration are very welcome. For the curious here a litte quickstart. it shows how to get all important bugs from reportbug-ng and prints out the bugnumber and summary: <div class="codehilite"><pre><code># Get all important bugs of reportbug-ng (returns a list of integers) bugnrlist = bts.get_bugs("package", "reportbug-ng", "severity", "important") bugnrlist [548871, 439203, 542759] # Get the actual bugreports (returns a list of Bugreport-objects) bugs = bts.get_status(bugnrlist) for bug in bugs: print bug.bug_num, bug.subject ....: 542759 [reportbug-ng] Erroneously reports nothing or repeats previous package's report 439203 Doesn't give any explanations of the severities and what they mean 548871 reportbug-ng: does not check for newer versions before reporting a bug </code></pre></div>Please help to complete python-debianbts2009-09-20T15:29:00+02:002009-09-20T15:29:00+02:00Bastian Venthurtag:venthur.de,2009-09-20:/2009-09-20-please-help-to-complete-python-debianbts.html

Please help to complete python-debianbts

I’m currently working on an updated version of <a href="https://packages.debian.org/search?keywords=python-debianbts">python-debianbts</a> a Python interface to Debian’s Bugtracker. The goal is to equip the Bugreport class with all available attributes delivered by the SOAP interface. The problem is, that for some attributes it is not quite clear what data they provide and in which datatype they are wrapped. For example the <code>mergedwith</code> attribute should be a list of bugnumbers, but it seems to be a single Integer when merged with one bug and an empty String when the bug is not merged at all. Some attributes have an ambiguous name and it’s hard to guess what they mean, for example there is an <code>id</code> and a <code>bug_nr</code> and both seem to contain the same information. There is a <a href="https://github.com/venthur/python-debianbts/tree/complete-support">git branch</a> for this task and a <a href="https://wiki.debian.org/DebbugsSoapInterface">wiki page</a> collecting all available information. If you have some experience with SOAP, our BTS and some time to kill, your help would be appreciated.Github2009-09-19T15:57:00+02:002009-09-19T15:57:00+02:00Bastian Venthurtag:venthur.de,2009-09-19:/2009-09-19-github.html

Github

<a href="https://github.com/venthur/reportbug-ng">Reportbug-ng</a> and <a href="https://github.com/venthur/python-debianbts">python-debianbts</a> are now on github. Many thanks to the alioth team who provided this service before.Gridengine and Qlogin Zombies2009-09-12T11:38:00+02:002009-09-12T11:38:00+02:00Bastian Venthurtag:venthur.de,2009-09-12:/2009-09-12-gridengine-and-qlogin-zombies.html

Gridengine and Qlogin Zombies

Dear Lazyweb, we’re using <a href="https://gridengine.sunsource.net/">gridengine</a> 6.2 on our cluster system at work. On this system we have a queue for interactive jobs. The queue has a fixed number of slots. If someone wants to start an interactive session and no free slots are available the interactive job cannot start otherwise it will consume one slot during the session and free it again after the user quit the interactive session. Unfortunately, if the user forgets to quit the interactive session properly (e.g. by entering “quit” in the terminal) and just closes the terminal, the interactive session becomes a zombie. The user cannot use this session anymore, but the process is still running and thus blocking a slot. Since it happens quite often that users fail to quit the interactive session properly, the free slots are used up pretty soon and no one can start new interactive sessions anymore – quite a nice bug in the gridengine software (available in Debian/Lenny by the way…) You can see that those processes are still running using ps and qstat. The question is: how can I effectively search for those processes and kill them? Gridengine does not seem to be aware of those zombies and the user might run other interactive jobs in parallel which must not be killed. So how does one find the zombies and kill them reliably?Package Completion for Rng2009-08-03T19:11:00+02:002009-08-03T19:11:00+02:00Bastian Venthurtag:venthur.de,2009-08-03:/2009-08-03-package-completion-for-rng.html

Package Completion for Rng

Today, I finally found the time to apply a patch contributed by Daniel Schaal to reportbug-ng. Among two minor fixes it brings a huge usability improvement: package completion. Whenever you type something to the input field rng suggests package names matching that prefix: <img alt="" src="/images/rng-package-completion.png" title="rng-package-completion"> That should make bug reporting against packages with cryptic names like libraries a lot easier. Thank you very much Daniel!SO_REUSEADDR2009-07-22T10:39:00+02:002009-07-22T10:39:00+02:00Bastian Venthurtag:venthur.de,2009-07-22:/2009-07-22-so_reuseaddr.html

SO_REUSEADDR

Dear Lazyweb, I have a simple test application where a TCP/IP server listens for incoming connections, reads the data and closes the connection again and a client which opens connections to the server and sends a package and closes the connection as fast as it can: The server looks like this: <div class="codehilite"><pre><code>sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) sock.setblocking(False) sock.bind(("", 12347)) sock.listen(1) slist = [sock] # use select to poll the sockets while 1: l = select.select(slist, [], []) for i in l[0]: conn, addr = i.accept() data = "" while 1: tmp = conn.recv(1024) if not tmp: break data += tmp conn.shutdown(socket.SHUT_RDWR) conn.close() </code></pre></div> The Client: <div class="codehilite"><pre><code># Open a connection, send data and close the connection as fast as possible while 1: sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) sock.connect(("", 12347)) sock.send("foo") sock.shutdown(socket.SHUT_RDWR) sock.close() </code></pre></div> The Problem with this application: After roughly 25.000 Iterations the client quits with a friendly: <blockquote> error: (99, ‘Cannot assign requested address’) </blockquote> Netstat shows the problem: roughly 25.000 of these ones: <div class="codehilite"><pre><code>... tcp 0 0 localhost:56946 localhost:12347 TIME_WAIT tcp 0 0 localhost:47163 localhost:12347 TIME_WAIT tcp 0 0 localhost:42758 localhost:12347 TIME_WAIT ... </code></pre></div> I’m not a TCP/IP expert but I thought <code>SO_REUSEADDR</code> is supposed to address this problem by allowing to reuse those as-good-as-closed connections in <code>TIME_WAIT</code> state, or not? So why does it fail in my test application?Debconf 11 in Berlin?2009-06-03T10:43:00+02:002009-06-03T10:43:00+02:00Bastian Venthurtag:venthur.de,2009-06-03:/2009-06-03-debconf-11-in-berlin.html

Debconf 11 in Berlin?

This is just a quick note for everyone who is interested: Yesterday we had a small meeting discussing the plan to bring Debconf 11 to Berlin, Germany. There will be an official meeting regarding this topic at this year’s LinuxTag. The meeting will probably take place on Friday the 26th of June at a conference room. Everyone is invited to join the discussion. PS: I also have a shiny new key, the fingerprint is <code>8795 AF2E 4A1A 4EFB A068 A449 8E88 9544 D985 000D</code>. The old key will be revoked in a couple of months.LinuxTag 20092009-05-05T08:35:00+02:002009-05-05T08:35:00+02:00Bastian Venthurtag:venthur.de,2009-05-05:/2009-05-05-linuxtag-2009.html

LinuxTag 2009

<a href="https://nion.modprobe.de/blog/">Nion</a> just reminded me of this year’s LinuxTag. Although I certainly won’t have the time to help at the booth every day as in the last years, I hope I can manage to work there at least for one or two days. If you want to help at our booth (you don’t have to be a Debian Developer), <a href="https://wiki.debian.org/DebianEventsDe/2009/LinuxTag">join us</a>! It’s fun and a great way to get in touch with other like minded people and your help is really appreciated.Using rng for bug triaging2009-04-18T12:08:00+02:002009-04-18T12:08:00+02:00Bastian Venthurtag:venthur.de,2009-04-18:/2009-04-18-using-rng-for-bug-triaging.html

Using rng for bug triaging

Reportbug-NG’s ability to filter and sort bugreports in various ways can greatly help triaging bugreports after a certain criterion. In this example I use the bugs of the KDE team, because they have quite a lot of bugs and are currently on a bug triage. Two releases ago I tried to help them cleaning their BTS from obsolete bugeports. My tactic back then was to find bugreports where no action has been taken for a long time. I noticed that there was no easy way to find such bugs since neither reportbug nor the BTS allowed to sort the bugs this way, so finding the candidates I searched for was very cumbersome. So let’s try to find those bugs with rng: To get all bugs belonging to the KDE team start rng with the maintainer as argument: <div class="codehilite"><pre><code>rng debian-qt-kde@lists.debian.org </code></pre></div> After rng started you see a table with all bugreports against the packages of the KDE team. To find potentially solved or unreproducible bugs, click on the last column header to sort the list by the date of the last action. The date of the last action is the date where the bugreport received it’s last mail (including control@b.d.o mails). You will see that there are quite a lot bugs which haven’t been touched for more than two years. Many of them don’t have a unreproducible or moreinfo tag. If you want to help the KDE team, you can now start trying to reproduce the bug. If that’s not possible you can tag them with moreinfo and unreproducible. If there is no reply in roughly one month or so it’s probably safe to close this bug. To quickly find a lot of other candidates to close one can use the filter of rng. Just enter moreinfo in the filter and the table will reduce to all bugreports containing this word or tag. Now you can again sort this list by date of last action and you will see quite a lot bugs bugs where more info was requested but no action took place for more than one year. Those need to be inspected, but probably quite a few of them are safe to close. This example shows how filtering and sorting can greatly help to find certain bugs and support bug triaging. But please: don’t carelessly close very old bugs! Read the reports and try to reproduce them, find matching bugreports in upstream’s BTS and ask the submitter if in doubt.Migrating a project from one SVN repository to another using git-svn2009-04-08T15:13:00+02:002009-04-08T15:13:00+02:00Bastian Venthurtag:venthur.de,2009-04-08:/2009-04-08-migrating-a-project-from-one-svn-repository-to-another-using-git-svn.html

Migrating a project from one SVN repository to another using git-svn

… sounds probably like a stupid idea since you could just copy the repository from the old to the new location, but in this case we don’t have direct access to the repositories, only accounts with commit-access. Ahh, and we want to preserve the complete history, of course. Ok, the plan is to use git-svn to get a copy of the old SVN repository, complete with history and then apply each commit to the new location: <h2>Getting the old repository</h2> <div class="codehilite"><pre><code>git svn clone https://old/repo/ cd repo </code></pre></div> <h2>Removing the traces of the old repository</h2> Now we have a git repository containing the complete history of the old SVN repository. That git repository is connected to the old SVN repository via git-svn. In order to commit it to the new repository we have to remove the traces of git-svn: <div class="codehilite"><pre><code># Remove the remote branch git branch -rD git-svn # Remove the rest rm -rf .git/svn/ </code></pre></div> Edit .git/config and remove the <code>svn-remote</code> stanza <h2>Commiting into the new repository</h2> Now that we can connect our git repository to the new SVN location and commit all the stuff: <div class="codehilite"><pre><code># Create the new repository directory if necessary svn mkdir https://new/repo/ # Init, fetch, rebase and commit: git svn init https://new/repo git svn fetch git rebase git-svn git svn dcommit </code></pre></div> Maybe not the most elegant solution but it works. The new repository contains everything from the old one, only the commit dates are messed up.Reportbug finally has a GUI!2009-03-25T18:22:00+01:002009-03-25T18:22:00+01:00Bastian Venthurtag:venthur.de,2009-03-25:/2009-03-25-reportbug-finally-has-a-gui.html

Reportbug finally has a GUI!

It’s good to see that reportbug 4 finally has a graphical user interface – it’s GTK, but nobody’s perfect :) Anyways, I’m really glad to see reportbug aiming to improve it’s usability. Reportbug finally shows all the bugreports for a package in a list which is easily filterable. So when you’re confronted with a huge list of bugreports but searching for a specific problem, you can just enter a few letters in the filter and the buglist will shrink immediately. This missing feature was basically the most important reason why I developed reportbug-ng: Users are expected to don’t report bugs which are already filed against a package, but the main tool which was used back then to report the bugs didn’t make this task trivial. Now it does and it’s a good thing! <img alt="reportbug and reportbug-ng" src="/images/reportbug-rng.png" title="reportbug-rng"> Compared to reportbug-ng, reportbug shows the bugreports of a package in a tree-view, while rng shows them in a table-view. The difference is that reportbug could theoretically group serveral bugs together, like merged bugs. It currently only uses this feature to group bugs of the same severity and status. Rng does that by coloring the lines. Unfortunately in reportbug all nodes are collapsed by default, which is not very usable since you have to expand every node in order to see all bugreports. This costs the user several clicks in order to see all bugreports. It’s also not possible to sort the bugs via the columns of the table, e.g. sort by date of last action which I think is a huge feature in order to quickly find certain kind of bugs (old bugs, bugs sorted by date of last activity) – but that’s probably easy to fix. Another good news is that Reportbug now brings it’s own editor so the user is not forced anymore to use an console based- and maybe unfamiliar editor to compose his bugreport. This is a huge step forward regarding usability. It also moves the concept of writing a mail to the BTS a bit more away from the user which is a good thing. Maybe the BTS will support an <a href="https://wiki.debian.org/SummerOfCode2009/DebbugsRWSOAP">alternative way</a> to receive bugreports in the future so mail clients may become superfluous for reporting bugs. Under the hood reportbug still parses the HTML from https://bugs.debian.org/ go gain the bugreport instead of the BTS’ SOAP interface. This coupling is of course <a href="2007-06-27-what-does-it-take-to-break-rng.html">dangerous</a> since even the smallest change in the BTS webdesign could break reportbug. But SOAP support is already on the todo list of reportbug’s devs. I can only encourage this step, when reportbug-ng moved away from HTML-parsing to SOAP the actual amount of work it took was much less then I thought, I think it took me just one or two days and as a result there is <a href="https://packages.qa.debian.org/p/python-debianbts.html">python-debianbts</a> which provides a bridge between the SOAP interface and Python. In summary I think this is an important improvement for reportbug. There are still some minor issues like the welcome screen always showing, the nodes collapsed by default and the columns not beeing sortable but those are small issues and easy to fix. I’m very happy to see this improvement, since it will motivate the not so tech-savvy folks to report bugs, and more bugreports will ultimately lead to a better distribution. Good work guys!git-svn branch2009-02-27T20:08:00+01:002009-02-27T20:08:00+01:00Bastian Venthurtag:venthur.de,2009-02-27:/2009-02-27-git-svn-branch.html

git-svn branch

Dear Lazyweb, how do I push a local git branch to a new svn branch? The other way round is quite easy and well documented but I didn’t find a single document describing what I want to do. I’m working with an svn repository with standard layout. I use git-svn to be able to use git locally. Until now I’ve always had my master branch reflect the svn’s trunk. For every new feature, I created a local git branch, merged with the master branch when I was done and dcommit’ed it to the svn’s trunk. But now I need to publish an experimental feature which should not yet go into the trunk so I have to publish my local git branch to an svn branch. I tried <code>git-svn branch foo</code> in the local git branch and it created a svn branch/foo for me but not from the current git branch but from the current master (remote/trunk). That’s of course not what I wanted, I wanted to push the current branch where I’m in and have it linked to the remote svn branch. Neither the manpage of git-svn nor some documentation in the web helped me. I even found a <a href="https://www.amitu.com/blog/2008/march/git-svn-whys-and-hows/">pretty detailed document</a> which describes each and every aspect of git-svn but unfortunately not mine. Is this use case so unusual that no one else uses git-svn this way or is it my weak git-FU?Lenny2009-02-15T16:39:00+01:002009-02-15T16:39:00+01:00Bastian Venthurtag:venthur.de,2009-02-15:/2009-02-15-lenny.html

Lenny

Today Lenny was finally released. According to <a href="https://bugs.debian.org/release-critical/">this page</a> with roughly 120 open release critical bugs. That’s quite a big number according to our standards, but still a good and necessary compromise. Thanks for everyone involved, good work everyone! I’m very happy that the release didn’t take quite as long as I predicted. Three or four more months of Sid being frozen would probably have meant more people playing with semi-supported mixtures of unstable and experimental packages and more migration towards other distributions with more current software packages. So I’m very happy to see over a hundred new packages uploaded to unstable already and roughly 700 new release critical bugs opened :)Valentine's Day!2009-02-02T00:04:00+01:002009-02-02T00:04:00+01:00Bastian Venthurtag:venthur.de,2009-02-02:/2009-02-02-valentines-day.html

Valentine's Day!

According to <a href="https://lists.debian.org/debian-devel-announce/2009/02/msg00000.html">this message</a> on -devel-announce, Lenny might release on this year’s Valentine’s Day which happens to be in two weeks. These are absolutely fantastic news! Can’t wait to comment out the experimental lines in my sources.list and see some <a href="https://kde.org/">fresh</a> (and <a href="https://www.x.org/">badly</a> <a href="https://kernel.org/">needed</a>) packages in unstable again.Nice(r) console fonts2009-01-10T14:05:00+01:002009-01-10T14:05:00+01:00Bastian Venthurtag:venthur.de,2009-01-10:/2009-01-10-nicer-console-fonts.html

Nice(r) console fonts

Ever wanted to get a nicer console font than the default one delivered with Debian? First you need to get a nice font. I chose console-terminus: <div class="codehilite"><pre><code>$ sudo aptitude install console-terminus </code></pre></div> that alone won’t change anything but install a whole bunch of new fonts into <code>/usr/share/consolefonts/</code>. To enable one of them, you have to edit <code>/etc/console-tools/config</code>: <div class="codehilite"><pre><code>set SCREEN_FONT=Uni3-Terminus16 </code></pre></div> Optional: comment out the following lines with <code>SCREEN_FONT_vc*</code>, so every terminal will have the same settings. To test the new font: <div class="codehilite"><pre><code>$ sudo /etc/init.d/console-screen.sh start </code></pre></div> The new font kicks in at the second half of the boot process, which is fixable but sufficient for me. I think the new font is much easier to the eyes than the default one, especially on big monitors with high resolutions.How many bugs have you fixed today?2008-12-16T14:54:00+01:002008-12-16T14:54:00+01:00Bastian Venthurtag:venthur.de,2008-12-16:/2008-12-16-how-many-bugs-have-you-fixed-today.html

How many bugs have you fixed today?

That is the sledgehammer argument people bring up when others dare to criticize the current release process. I think this question is nonsense. While the bug-fix rate was more or less the same since the last two releases, it looks like in this release we simply started the freeze with much more open RC-bugs than before (~3 times more RC-bugs than with Etch). So it was foreseeable that the freeze would take longer this time. We can’t solve the problem by trying to fix bugs faster, we’ve tried it several times with mediocre success. That doesn’t mean we shouldn’t keep up fixing bugs, but we should probably accept that a certain number of people can only fix a certain number of bugs in a given time. In our case roughly 30 RC-bugs per month. So what’s the point of asking how many RC-bugs one has fixed? Does that mean only those are allowed to make suggestions, who fixed an RC bug? That pressure doesn’t help anyone. I believe unstable being frozen for several months is a very important problem which can have drastic consequences regarding the number of our users, and we will not solve it by suggesting to fix our bugs faster.Lenny Release General Resolution2008-12-14T12:34:00+01:002008-12-14T12:34:00+01:00Bastian Venthurtag:venthur.de,2008-12-14:/2008-12-14-lenny-release-general-resolution.html

Lenny Release General Resolution

I agree, this <a href="https://www.debian.org/vote/2008/vote_003">vote</a> is one of the worst formulated I can remember. I think votes should be as easy as possible to understand, so that the voter is absolutely clear about what he’s going to vote. The current vote is the perfect example how not to design such a vote. Reading and understanding the ballot is very exhausting, the wording sounds polemic and given the options, it somehow feels like a trap. Example: The first option is “Reaffirm the social contract”. Sounds nice, of course we want that, but in the footnotes it says <blockquote> […] we will delay the release of Lenny until such point that the work to free the operating system is complete (to the best of our knowledge as of 1 November 2008). </blockquote> So you’re actually voting on delaying the release of Lenny until some technical details are cleared. This is really crap, why not name the option “Delay Lenny until foo” in the first place? I guess to motivate people a bit more to rank this option above all others. I’m very disappointed. Next vote, fever options and clearer wording please.Lenny's release date III2008-12-09T16:07:00+01:002008-12-09T16:07:00+01:00Bastian Venthurtag:venthur.de,2008-12-09:/2008-12-09-lennys-release-date-iii.html

Lenny's release date III

Two months since my <a href="2008-10-07-lennys-release-date.html">last status update</a> of my <a href="2008-08-01-my-guess-for-lennys-release-date.html">initial guess for Lenny’s release date</a>, time for an update: <img alt="My guess for Lenny's release III" src="/images/release-guess-3.png"> I don’t know what the current “official” release target for Lenny is, but I still think we’re heading towards Spring/Summer 2009. The last <a href="https://lists.debian.org/debian-devel-announce/2008/10/msg00000.html">related mail on -devel-announce</a> is from October 2008 from Alexander who hopes that we’ll release Lenny at least by the end of 2008. I’ve read an article with an interview with Steve McIntyre <a href="https://www.theregister.co.uk/2008/10/10/debian_lenny_late/">somewhere on The Register</a> where he said: <blockquote> “Bastian’s being a little bit pessimistic based on the data he’s looking at, but as far as I know, he hasn’t spoken to any of the release team. As far as Lenny goes, we’re ‘almost’ there,” he told The Register. </blockquote> That interview was in the beginning of October. It doesn’t bother me too much, that we release late. What bothers me is that we don’t correct our estimated release dates even when it’s obvious that there is no way we will release in time. It’s really sad that we keep our users in the dark about the fact that we can’t release in time until it’s already too late. I believe many users would appreciate if we’d honestly say: We don’t know when we release, but currently we have x RC bugs open and looking at the last releases we think it will take roughly n months to fix them and release, instead of the good old “We release when it’s ready”. Another important point: Unstable is frozen for almost five months now. If you buy a <a href="2008-11-25-lenovo-t500-shiny-new-laptop.html">current laptop</a>, chances are hight that it might not fully work with Debian “bleeding edge” unstable, since it’s horrible outdated. Some of the newer stuff can be obtained from experimental if you’re brave enough, but if you’re not, unstable is getting more and more unattractive compared to other distribution’s current versions. This is really something we should try to fix after we released Lenny.Lenovo T500 (shiny new laptop)2008-11-25T17:36:00+01:002008-11-25T17:36:00+01:00Bastian Venthurtag:venthur.de,2008-11-25:/2008-11-25-lenovo-t500-shiny-new-laptop.html

Lenovo T500 (shiny new laptop)

Yesterday I exchanged my “old” Thinkpad T60 with a shiny new T500. This time I ordered a Thinkpad with integrated Intel graphics since it’s supposed to be way better supported than the graphic chips from ATI or NVIDIA under Linux. And of course it is supposed to be much more silent. I don’t regret this decision so far, the new T500 is barely hearable even under heavy load. <h2>Debian</h2> I installed Debian unstable (aka. frozen for waaayy to long) without a problem. Sound worked out of the box, graphics worked somewhat. The kernel automatically loaded the correct modules but I had to tell Xorg that I want the “intel” driver instead of Vesa. Unfortunately Xorg’s intel driver in unstable is already too old to allow me to use an external monitor via the docking station’s DVI port, so I had to to upgrade to Xorg from experimental, which is not very cool since this is the machine I depend on at work. Wifi also doesn’t work, since the the driver and firmware available in unstable are outdated. Suspend to RAM works sometimes, but most of the time the laptop refuses to wake up properly so I have to shut it down. I’ve not yet tested if bluetooth and Wimax works. So currently the T500 is usable under unstable but without external monitor and without Wifi. If you are brave enough to use packages from experimental, you get your external monitor to work but still don’t have Wifi. Suspend to RAM does not work reliably. The T500 should probably work fine with Lenny but of course without external Monitor and Wifi support. <h2>Vista</h2> After I’ve installed Debian and moved my /home from the old laptop to my new baby, I noticed that silly “Windows Vista” sticker on my laptop. Of course I tried to remove that thing, which I regretted painfully. After I’ve successfully removed the sticker – which was surprisingly hard – some glue remained on the surface of the laptop. No problem I thought, nothing some nail polish remover can’t handle. Unfortunately it only smeared the glue but did not resolve it as I had expected. So I’ve tried perfume, water, soap, rubber, tape and solvent for several hours, but was not able to remove the glue at all. All I achieved was to smear the compact but thick chunk of glue to a very thin but much more distributed layer of this nasty glue on my laptop. Now it looks like a 5x5cm square of dirt on top of my laptop – not sure if removing the Vista sticker was worth it.rm -rf /2008-11-10T11:14:00+01:002008-11-10T11:14:00+01:00Bastian Venthurtag:venthur.de,2008-11-10:/2008-11-10-rm-rf.html

rm -rf /

… does not work. At least not today on a Debian/Sid system. Looks like there are some things, not even root is allowed to do: <div class="codehilite"><pre><code>root@sid:~# rm -rf / rm: cannot remove root directory ‘/' </code></pre></div> If you’re curious what happens when you wipe your disk on a running system: <div class="codehilite"><pre><code>root@sid:~# cd / root@sid:/# rm -rf *` </code></pre></div> does the trick. Update: That’s with plain rm not safe-rm. Update 2: I looked through the code of <code>rm.c</code> to find that <code>/</code> is protected by default – as a simple <code>rm --help</code> also shows: <div class="codehilite"><pre><code>$ rm --help Usage: rm [OPTION]... FILE... Remove (unlink) the FILE(s). [...] --no-preserve-root do not treat `/' specially --preserve-root do not remove `/' (default) </code></pre></div>Tags2008-10-11T20:29:00+02:002008-10-11T20:29:00+02:00Bastian Venthurtag:venthur.de,2008-10-11:/2008-10-11-tags.html

Tags

Version 1.2 of reportbug-ng now shows the tags of a bug if available. This was long overdue. Now you can for example, quickly search your bugs for the ones with the moreinfo-tag set and sort the resulting list by the date of the last action. This should give you a quick overview of bugs which may need some action (eg. another reminder to the submitter) or can potentially be safely closed. <img alt="rng-tags.png" src="/images/rng-tags.png"> In the next step I’ll improve the filter of rng and add a context menu which will allow to manipulate the tags of one or more bugreports and send all the changes in a single mail to control@bugs.debian.org.Lenny's release date2008-10-07T14:07:00+02:002008-10-07T14:07:00+02:00Bastian Venthurtag:venthur.de,2008-10-07:/2008-10-07-lennys-release-date.html

Lenny's release date

Today, an <a href="https://www.heise.de/newsticker/Debian-5-0-soll-noch-in-diesem-Jahr-fertig-werden--/meldung/117018">article at heise online</a> claimed, that Lenny is supposed to ship this year, quoting <a href="https://blog.schmehl.info//2008/10/07#releasing-lenny">Alexander’s blog entry</a>. While I can’t find a single word in Alexander’s article claiming that we’ll release this year, I’m pretty sorry that just after we failed the first deadline which was obviously too “optimistic”, we get another one which is too short. <a href="2008-08-01-my-guess-for-lennys-release-date.html">Two month’s ago</a> I created a pretty simple estimation regarding Lenny’s release by looking at the rc-bugfix-rate of the last two releases and applying them to the current number of rc-bugs. As a result it looked more likely that we release Lenny in June 2009 than in September 2008. Until now, my guess for Lenny’s release still holds, at least the updated rc-graph lies perfectly on the line I’ve created two months ago. <img alt="rc-graph2.png" src="/images/rc-graph2.png"> While I wish that we’ll release Lenny as soon as possible, I hope that someone will clarify the proposed release date to the media and this time with a more defensive deadline. Some people might actually trust our estimation (I know this sounds silly). And even if everyone knows that Debian notoriously releases too late – we could at least try to propose more realistic dates and update them regularly. Just two weeks ago I still read and heard in the media that Debian will release Lenny in September although it was clear at this point that this is impossible.Gettext and QT4?2008-09-14T14:04:00+02:002008-09-14T14:04:00+02:00Bastian Venthurtag:venthur.de,2008-09-14:/2008-09-14-gettext-and-qt4.html

Gettext and QT4?

Dear Lazyweb, do you know how to translate QT4 applications with gettext? Looks like QT4 now uses it’s own mechanism (lupdate -> .ts -> linguist -> .ts -> lrelease -> .qm) for translating strings, which makes things a bit complicated if the rest of your application (the non GUI part) still uses gettext. The background is – as some of you may have noticed – that rng’s GUI isn’t properly translated anymore since the switch from QT3 to QT4. I don’t know if I just should adopt QT4’s translation mechanism, but for now it would be very nice if I just could use my .po files for the QT4 stuff.Finally, a Debian Live CD2008-08-28T11:05:00+02:002008-08-28T11:05:00+02:00Bastian Venthurtag:venthur.de,2008-08-28:/2008-08-28-finally-a-debian-live-cd.html

Finally, a Debian Live CD

That was just about time. In the last two Linux Tag events people asked us for Live-CDs and we had to explain why we don’t have any, while most other distros have. Now we finally can offer a <a href="https://cdimage.debian.org/cdimage/lenny_live_beta1/">Debian Live-CD</a>. Still beta, but a huge improvement. Thanks Debian Live team!What is group x in /etc/group for and should I be a member?2008-08-03T20:53:00+02:002008-08-03T20:53:00+02:00Bastian Venthurtag:venthur.de,2008-08-03:/2008-08-03-what-is-group-x-in-etcgroup-for-and-should-i-be-a-member.html

What is group x in /etc/group for and should I be a member?

A few days ago I noticed that suspend to RAM didn’t work anymore on my T60 running sid. By chance I stumbled upon <a href="https://bugs.debian.org/484248">#484248</a> which explains that the user now has to be in the <code>powerdev</code> group to be able to suspend the machine. If I hadn’t found this information, I would have had no idea what to change to make me able to suspend my machine again. Shouldn’t we have a central list somewhere in <code>/usr/share/doc</code> explaining all the different groups and what they mean? Most of the times the package a group belongs to already explains what the meaning of this group is, but if you don’t know the package, you’re pretty much screwed. I’m not saying, we already have too many groups, but a list with all the important groups and what they’re for would be a huge usability improvement. We can’t expect the user to find that information for every group alone.My guess for Lenny's release date2008-08-01T11:58:00+02:002008-08-01T11:58:00+02:00Bastian Venthurtag:venthur.de,2008-08-01:/2008-08-01-my-guess-for-lennys-release-date.html

My guess for Lenny's release date

Looking at the <a href="https://bugs.debian.org/release-critical/">current number of release-critical bugs</a>, my guess for the actual release date of Lenny is: around June 2009. <img alt="rc-graph.png" src="/images/rc-graph.png"> The blue lines have exactly the same angle, the greenish line marks the planned release date. I know that’s not exactly scientific, but let’s see how close I am. The blue graph marks by the way the number of release-critical bugs in stable.Rng 1.0 in unstable2008-07-19T15:30:00+02:002008-07-19T15:30:00+02:00Bastian Venthurtag:venthur.de,2008-07-19:/2008-07-19-rng-10-in-unstable.html

Rng 1.0 in unstable

I’ve finished the porting of <a href="https://reportbug-ng.alioth.debian.org/">rng</a> from qt3 to qt4 and uploaded it to unstable. It has some minor regressions compared to the qt3 version, like the localization of the qt4 widgets which isn’t working yet and that the column which sorts the table isn’t saved and restored properly on startup, but it closes many other bugs and it doesn’t crash anymore. The new version doesn’t bring so much new features so far, but since it’s using much of the stuff which came with QT4, like the QTableModel and the QWebView, it runs much faster in many cases, e.g. when sorting or filtering the table. Ah, and last but but not least: thanks to WebKit rng even passes <a href="https://acid1.acidtests.org/">acid1</a> and <a href="https://acid2.acidtests.org/">acid2</a>. Doesn’t yet pass <a href="https://acid3.acidtests.org/">acid3</a> but I guess someone is already working on it :) <img alt="rng passing acid2" src="/images/rng-acid2.png">Porting rng to QT42008-07-13T23:40:00+02:002008-07-13T23:40:00+02:00Bastian Venthurtag:venthur.de,2008-07-13:/2008-07-13-porting-rng-to-qt4.html

Porting rng to QT4

Today I was busy porting reportbug-ng to QT4. One really cool thing I noticed is the new QWebView widget, which is the main widget of the QWebKit module. It works surprisingly easy compared to the good old QTextBrowser of QT3 – now you can just throw an URL at it and it will render the page nicely complete with CSS, images, etc. Here’s the obligatory screenshot: <img alt="reportbug-ng qt4" src="/images/rng-qt4.png"> as you can see, the bugreport looks much better than in the old version of rng. WebKit does a really good job rendering the HTML, I think the fonts look even smoother than under firefox. To ease the transition to from QT3 to QT4 I’ve extracted all the Debian BTS specific stuff to <a href="https://packages.debian.org/sid/python-debianbts">python-debianbts</a> and uploaded it to unstable. Rng uses it to query our Bug Tracking System via it’s SOAP interface and get the data in Python’s native data types. There is still the problem with rng (QT3) crashing when showing bugreports. I’ve investigated the problem a bit and found out that it is not rng’s fault but <a href="https://bugs.debian.org/490340">probably a problem with python-qt3</a> or one of it’s underlying libs. QTextBrowser causes segfaults in nearly 50% of the cases where you try to render html from a different thread: here’s an easy way to reproduce: <div class="codehilite"><pre><code># crash.py # Usage: python crash.py "some text" # python crash.py "some evil html" from qt import * import sys import threading class Form(QMainWindow): def __init__(self,parent = None,name = None,fl = 0): QMainWindow.__init__(self,parent,name,fl) self.browser = QTextBrowser(self, "browser") self.setCentralWidget(self.browser) if __name__ == '__main__': if len(sys.argv) < 2: print 'Usage: %s "some text"' % sys.argv[0] sys.exit() s = str(sys.argv[1]) # setup qt stuff with a browser app = QApplication(sys.argv) app.connect(app, SIGNAL("lastWindowClosed()"), app, SLOT("quit()")) form = Form() app.setMainWidget(form) form.show() # start a different thread and try to set the text in the browser # we wait 1.0 second -- just to make sure the change happens after the # start of app.exec_loop() below t = threading.Timer(1.0, form.browser.setText, (s, )) t.start() app.exec_loop() </code></pre></div> QWebView doesn’t seem to have this problem, the QT4 version of rng didn’t crash a single time. I’m not quite ready with the porting to QT4, and since I’ll be quite busy during the next weeks I suppose it will take a few weeks until the new version will be updated to unstable.Maximize vs Maximize to Preferred Size2008-07-11T18:00:00+02:002008-07-11T18:00:00+02:00Bastian Venthurtag:venthur.de,2008-07-11:/2008-07-11-maximize-vs-maximize-to-preferred-size.html

Maximize vs Maximize to Preferred Size

In times where screen resolutions of 1600x… and above are becoming more and more common, I wonder if the maximize window button which enlarges a window to take all the available space on the desktop is still as useful as it used to be. I noticed that some (if not all, I’m not a Mac user) applications on Mac only maximize their windows depending on the content of the window. For example if you maximize the window of your word processor, the window becomes only as wide as necessary to make the whole page visible. This doesn’t just look better on very large monitors, it is also more usable than a simple maximize since it doesn’t unnecessarily hide other windows. Is there a window manager for Linux which does something similar, or better: is something like this possible with KDE? I know that QT widgets have a “preferred size” – shouldn’t it be possible to have a maximize to preferred size button? It’s surely not useful for every application, but may applications only need a quite small main window and look very ugly when maximized.Segfaults Everywhere2008-07-07T21:09:00+02:002008-07-07T21:09:00+02:00Bastian Venthurtag:venthur.de,2008-07-07:/2008-07-07-segfaults-everywhere.html

Segfaults Everywhere

Dear Lazyweb, A few weeks ago rng started to show some segfaulty behavior. I’ve absolutely no idea what’s going on, since Python code usually doesn’t segfault. I guess the problem lies in the qt3-bindings rng uses – does someone know how to verify this? There is a <a href="https://bugs.debian.org/486212">backtrace</a> which seems to show that the problem lies somewhere in <code>/usr/lib/python2.5/site-packages/qt.so</code> but I’m not entirely sure how to interpret that.Tab Completion in Python's Interactive Mode2008-07-06T22:15:00+02:002008-07-06T22:15:00+02:00Bastian Venthurtag:venthur.de,2008-07-06:/2008-07-06-tab-completion-in-pythons-interactive-mode.html

Tab Completion in Python's Interactive Mode

While browsing the python docs, I accidentally stumbled upon <a href="https://docs.python.org/lib/module-rlcompleter.html">this</a> today. It describes how to get Tab completion in Python’s interactive shell: <ol> <li>Create a <code>~/.pythonrc</code> file with the following content:</li> </ol> <div class="codehilite"><pre><code>try: import readline except ImportError: print "Module readline not available." else: import rlcompleter readline.parse_and_bind("tab: complete") </code></pre></div> <ol> <li>Put this somewhere in your <code>~/.bashrc</code></li> </ol> <div class="codehilite"><pre><code>export PYTHONSTARTUP=~/.pythonrc </code></pre></div> The <code>.pythonrc</code> will be executed every time the Python interpreter is started but not when it executes a script. How does it look like? <div class="codehilite"><pre><code>$ python Python 2.5.2 (r252:60911, Jun 25 2008, 17:58:32) [GCC 4.3.1] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> l = [] >>> l. l.__add__ l.__getslice__ l.__ne__ l.append l.__class__ l.__gt__ l.__new__ l.count l.__contains__ l.__hash__ l.__reduce__ l.extend l.__delattr__ l.__iadd__ l.__reduce_ex__ l.index l.__delitem__ l.__imul__ l.__repr__ l.insert l.__delslice__ l.__init__ l.__reversed__ l.pop l.__doc__ l.__iter__ l.__rmul__ l.remove l.__eq__ l.__le__ l.__setattr__ l.reverse l.__ge__ l.__len__ l.__setitem__ l.sort l.__getattribute__ l.__lt__ l.__setslice__ l.__getitem__ l.__mul__ l.__str__ >>> l. </code></pre></div> Very cool! Why is this hidden so deep in the docs and not enabled by default?Linux Action Show2008-07-03T13:35:00+02:002008-07-03T13:35:00+02:00Bastian Venthurtag:venthur.de,2008-07-03:/2008-07-03-linux-action-show.html

Linux Action Show

<a href="https://mjr.towers.org.uk/writing/reflections/End_of_LugRadio_.html">MJ Ray</a> noticed that LugRadio is going to end soon and asks for alternatives. I’d recommend to give <a href="https://www.jupiterbroadcasting.com/?cat=4">Linux Action Show</a> a try. I’ve always preferred it over LugRadio. It is funny, informative and not quite as vulgar as LugRadio.Lenny won't ship rng2008-06-09T15:43:00+02:002008-06-09T15:43:00+02:00Bastian Venthurtag:venthur.de,2008-06-09:/2008-06-09-lenny-wont-ship-rng.html

Lenny won't ship rng

Philipp <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=422085#198">thinks</a>, the fact that rng is not using the information in /usr/share/bug renders rng “unfit for release” and upgraded the corresponding bugreport from wishlist to serious. Moreover: since I dared to downgrade the report back to wishlist <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=422085#205">he decided</a> to remove rng from testing and block it until the bug is fixed. I don’t want to heat the debate about this bug again, but Philipp’s decision seems arbitrary for me and I wonder if the same standard is applied to every other Debian package. I mean, rng has no release critical defects. It just does not use the aforementioned scripts as additional information in bugreports – does this really render the software “unfit for release”? For example: In Etch we shipped a famous email client with a known bug which lead reproducibly to loss of emails on IMAP accounts, just because removing the program, or one of it’s components was not possible in the remaining time.actual file size does not match size in .changes2008-06-04T20:26:00+02:002008-06-04T20:26:00+02:00Bastian Venthurtag:venthur.de,2008-06-04:/2008-06-04-actual-file-size-does-not-match-size-in-changes.html

actual file size does not match size in .changes

Beh! Looks like it is currently <a href="https://bugs.debian.org/474949">not possible</a> to use debsign from etch to sign packages built in a sid cowbuilder environment. Update: Adam just pointed out that this was already announced on debian-devel-announce. He proposed to get an updated version of devscripts from etch-backports. I wonder if we shouldn’t provide a fixed version of devscripts in etch officially, just in case someone who maintains packages actually *erm* uses it.Linuxtag 20082008-05-31T18:52:00+02:002008-05-31T18:52:00+02:00Bastian Venthurtag:venthur.de,2008-05-31:/2008-05-31-linuxtag-2008.html

Linuxtag 2008

Linuxtag 2008 is over. Like <a href="2007-05-30-linuxtag.html">last year’s Linuxtag</a>, it was a very nice event. We had a lot of visitors, answered a lot more questions and sold many t-shirts, posters and stickers at our eee-pc *cough* Debian booth. It was also very nice to meet some other people from Debian – one of the rare moments in life where I can feel really young again… sorry guys :P Of course there were many interesting talks and booths from other projects at this year’s Linuxtag. After Nico’s and my yearly Linuxtag raid, I had like a dozen different ball pens, a t-shirt, a handful of cookies, and many many gummy bears. Unlike last year, this year we had no Debian Day – but guess who showed up at our booth yesterday: <img alt="Ian Murdock and Bastian Venthur at Linuxtag 2008" src="/images/ian_and_basti.jpg"> that’s right, <a href="https://ianmurdock.com/">Ian Murdock</a> himself! I was so nervous to meet him, that I almost forgot to ask for a picture. Apropos showing up at our booth: although we where fairly good manned at our booth, we felt that a few more helpers wouldn’t have hurt. We (the staff from this year’s Debian booth) just want to say, that you don’t have to be a Debian Developer / Maintainer or whatever to help at our booth. If you’re an enthusiastic Debian user and can spare some time, you’re very welcome. The tickets for the staff are usually free, so don’t hesitate to ask if you can help next year!4 months and 10 days without any new Debian developer...2008-04-15T21:57:00+02:002008-04-15T21:57:00+02:00Bastian Venthurtag:venthur.de,2008-04-15:/2008-04-15-4-months-and-10-days-without-any-new-debian-developer.html

4 months and 10 days without any new Debian developer...

I sadly agree with <a href="https://www.lucas-nussbaum.net/blog/?p=286">Lucas</a>. Waiting several months after you finished all the tests until your account gets finally created is really frustrating and demotivating. In my opinion this is just a lack of manpower and seemingly no one who dares to step in to solve this problem. Why again don’t we just throw a few more people at this problem?Happy Birthday rng!2008-03-27T16:45:00+01:002008-03-27T16:45:00+01:00Bastian Venthurtag:venthur.de,2008-03-27:/2008-03-27-happy-birthday-rng.html

Happy Birthday rng!

Hmm, better late then never, I guess. One year and two weeks ago, <a href="https://reportbug-ng.alioth.debian.org/">reportbug-ng</a> was initially uploaded to Debian unstable. Happy birthday rng! I wish you a lot of <a href="https://reportbug-ng.alioth.debian.org/ponies.html">ponies</a> in the next year!The Eagle has Landed2008-03-13T10:50:00+01:002008-03-13T10:50:00+01:00Bastian Venthurtag:venthur.de,2008-03-13:/2008-03-13-the-eagle-has-landed.html

The Eagle has Landed

Yesterday, I submitted my diploma thesis (one day before the deadline – yay!), so I will hopefully have much more time to work on my Debian packages again.Good Python IDE2008-02-16T16:20:00+01:002008-02-16T16:20:00+01:00Bastian Venthurtag:venthur.de,2008-02-16:/2008-02-16-good-python-ide.html

Good Python IDE

<a href="https://www.philkern.de/weblog/en/debian/python_ide_wanted.html">Philipp</a> asked for a good Python IDE. I’d recommend <a href="https://pydev.sourceforge.net/">Pydev</a>, which is a plugin for Eclipse and works really well. I use it exclusively when coding in Python, and I’m very happy with it. Unfortunately the eclipse version in unstable is not the current one, so you can’t use the latest Pydev version, but even the latest one which is working with eclipse 3.2.x works well enough.Juten Rutsch ins neue Jahr!2007-12-31T11:34:00+01:002007-12-31T11:34:00+01:00Bastian Venthurtag:venthur.de,2007-12-31:/2007-12-31-juten-rutsch-ins-neue-jahr.html

Juten Rutsch ins neue Jahr!

Happy new year everyone! NP: Moby – Jam for the Ladies (Nevins Club Blaster mix)Merry Christmas everyone!2007-12-24T00:52:00+01:002007-12-24T00:52:00+01:00Bastian Venthurtag:venthur.de,2007-12-24:/2007-12-24-merry-christmas-everyone.html

Merry Christmas everyone!

Since it’s already that time of the year again, merry Christmas everyone! I hope you’re able to spend the holidays with the ones you love and enjoy a merry, merry Christmas! Have fun everyone! BTW: Am I the only one who really hates this freakin’ “Last Christmas”-song by WHAM!? Why the heck do the radio stations here in Germany have to have this crap on A-rotation every end of the year? This year I really tried to avoid this song by not hearing radio or watching music television. But I made a major mistake: I went to the Christmas fair… guess what the first song was, I heard when I arrived? Thank god there was enough mulled wine around to ease my pain :) NP: Thomas D - LiebesliedOn #422085 and #4172562007-12-18T01:03:00+01:002007-12-18T01:03:00+01:00Bastian Venthurtag:venthur.de,2007-12-18:/2007-12-18-on-422085-and-417256.html

On #422085 and #417256

Since I received a terrifying amount of insults via mail for not implementing <a href="https://bugs.debian.org/422085">this feature request</a> after my <a href="2007-12-08-want-to-get-money-for-nothing-and-chicks-for-free-too.html">last blog entry</a>, where I asked for help developing rng, I’d like to make my position about this issue clear. Why I was opposed to implement it: <ol> <li>I personally hated that some packages sent a huge amount of unrelated info with every bugreport for this package, even if it’s not meaningful for this bugreport. I made a quick check against my favorite package with a very long output and thought (and still think) that this info is not even relevant for the majority of bugreports of this package. So I thought it was not too much to ask, to write the reporter a friendly mail to post the output of this script, if it’s really needed.</li> <li>I was personally very annoyed by packages with very long presubj text, which I doubt anyone reads anyway. Since I don’t want rng to be annoying to the users, I decided to leave that feature away. An implementation of this feature would mean to pop up a window with some text the user should read before continuing to report the bug. I don’t like popups and don’t want rng to make use of them.</li> <li>I’m definitely opposed to a feature which will pop up a terminal where a user has to do something before he can proceed reporting a bug. Sorry, but this won’t happen in rng. I might consider such a thing if it could be scripted to use QT or even GTK but a terminal popping up in a GUI application is a no-go for me, sorry.</li> <li>I was personally very annoyed by some of the reactions on this bugreport. Since we’re all volunteers and stuff and this feature is maybe a nice-to-have but definitely not a must-have, I decided to put this issue very low on my to do list.</li> </ol> However, I agree that the stuff in /usr/share/bug isn’t completely useless. The opposite is true, it makes the life of maintainers easier and rng should make use of it where possible. So what can we do now? Maybe we can start all over and discuss this issue in a more friendly and constructive way? Here’s my offer: rng will support bugscripts, but it will not feature a terminal popping up asking a user questions. I’m developing a GUI application and a popping up terminal is not very GUI’ish for me. What can we do about this? Is there a way to implement this? Supporting presubj might be okay too, but I don’t want rng to be annoying. Popups are annoying and having to read the same stuff again and again when reporting multiple times on the same package is annoying too. Maybe rng could spawn a pop up with a checkbox “Don’t show this text again when reporting a bug against this package”? Maybe we could have an option in the (not yet implemented) preferences window where we could suppress those messages completely? And please, don’t use abusive language or even insults when contacting me about this issue. My rng-time is currently very limited and my motivation to work especially on this issue is already very low. We’re speaking here about a fully optional feature. Providing the output of some scripts or having to read a presubj is helpful, but not mandatory when reporting a bug. So please, be nice! I really don’t need all that stress right now.Want to get money for nothing and chicks for free, too?2007-12-08T15:49:00+01:002007-12-08T15:49:00+01:00Bastian Venthurtag:venthur.de,2007-12-08:/2007-12-08-want-to-get-money-for-nothing-and-chicks-for-free-too.html

Want to get money for nothing and chicks for free, too?

Is anyone out there interested in helping me developing <a href="https://reportbug-ng.alioth.debian.org/">reportbug-ng</a>? I’m currently very busy with my diploma thesis and can’t spend as much time on my pet bug reporting tool as I’d like to spend. The package itself has some small bugs, is not yet <a href="https://reportbug-ng.alioth.debian.org/dev.html">feature complete</a>, but in a fairly good shape and quite popular too. If you’re interested you should have experience with python, QT and maybe some SOAP-skills. You don’t have to be a Debian Developer or -Maintainer! I haven’t done this before so I don’t know that the best practice for such a situation is. I’m currently thinking of something like: <ol> <li>First you send me patches, which I’ll review, apply and upload</li> <li>After a while you get write access to the BZR repo</li> <li>Finally become co-maintainer of the package if you want</li> </ol> Interested? Then write me a mail to: venthur at debian org. Oh, and please don’t write if you just want to get your <a href="https://bugs.debian.org/422085">pet-bug-which-I-still-refuse-to-fix</a> fixed. PS: If you prefer guys over chicks, you’re welcome too ;)Forcing win32-loader to use i386?2007-12-04T19:19:00+01:002007-12-04T19:19:00+01:00Bastian Venthurtag:venthur.de,2007-12-04:/2007-12-04-forcing-win32-loader-to-use-i368.html

Forcing win32-loader to use i386?

Dear Lazyweb, I’m happily using the win32-loader from <a href="https://goodbye-microsoft.com">goodbye-microsoft.com</a> to install Debian/Etch on several Lenovo X61s subnotebooks which don’t have a DVD drive but a running XP preinstalled. The installer correctly detects a 64bit architecture and therefore installs AMD64. Unfortunately I want to run i386 on those machines. Is there a way to force a specific arch for the win32-loader? Edit: Of course I already tried to install Debian from an USB stick on those machines, but the Debian installer refuses to work at the point where it’s unable to find the DVD/CD drive on the subnotebooks without one. I’ve tried netinst and businesscard, none of them worked from USB for me. Any hints?Time-lapse-video-with-your-webcam mini HOWTO2007-11-27T17:46:00+01:002007-11-27T17:46:00+01:00Bastian Venthurtag:venthur.de,2007-11-27:/2007-11-27-time-lapse-video-with-your-webcam-mini-howto.html

Time-lapse-video-with-your-webcam mini HOWTO

I bought me a cheapo webcam the other day and the first thing (meh… actually the second. The first thing I had to do was obviously getting this little piece of plastic working…) I wanted to do was making a time-lapse video of me working on my diploma thesis or something. With a little help from google and the friendly folks at debian-user-german I finally got it working. <img alt="snapshot.jpg" src="/images/snapshot.jpg"> What you need: <code>camserv</code>, <code>wget</code> and <code>mencoder</code>. Assuming you have your webcam working with v4l, camserv will provide snapshots of your cam via <code>https://localhost:9192</code>, so the following script will make a snapshot every second until you kill it: <div class="codehilite"><pre><code>#!/bin/sh while [ 1 ]; do wget https://localhost:9192 -O `date +%s`.jpg -q sleep 1s done </code></pre></div> Let it run for several hours and then create the movie with the following script: <div class="codehilite"><pre><code>#!/bin/sh mencoder "mf://*.jpg" -mf fps=25 -o output.avi -ovc lavc -lavcopts vcodec=mpeg4:mbd=2:trell=yes:v4mv=yes </code></pre></div> Hint: coding, juggling and even sleeping at your desk looks cool in time-lapse - eating or singing your favorite songs while listening to them does not :P I recorded myself for ca 12 hours and the resulting video was like 130 megs large. The result looks pretty funny, but since there was a lot of the not-so-cool-looking eating and singing stuff in there, I’m afraid I can’t even show it to my closest friends ;) Hint for users having problems getting proper images from camserv: I had to change <code>max_frames</code> value to 1 in the <code>[socket]</code> section of <code>/etc/camserv/camserv.cfg</code> to get proper single images via wget. I find the camserv solution to get the images from the webcam rather suboptimal, has anyone a smarter solution? NP: Gianna Nannini - Bello e impossibile(Almost) Back2007-10-16T11:27:00+02:002007-10-16T11:27:00+02:00Bastian Venthurtag:venthur.de,2007-10-16:/2007-10-16-almost-back.html

(Almost) Back

Due some drastic changes in my personal life, I wasn’t very active in the last two months. Now things are getting normal again, and I’m looking forward working on <a href="https://reportbug-ng.alioth.debian.org/">rng</a> and my <a href="https://qa.debian.org/developer.php?login=venthur">other packages</a> again. Sorry for all the unanswered mails, I’ll need some time to clear the backlog. I’m currently working on my diploma thesis so my Debian related activity level won’t be as high as it was two months ago. But it should be at least high enough to keep the packages in a good shape.Deserializing SOAP replies with ZSI?2007-08-20T18:29:00+02:002007-08-20T18:29:00+02:00Bastian Venthurtag:venthur.de,2007-08-20:/2007-08-20-deserializing-soap-replies-with-zsi.html

Deserializing SOAP replies with ZSI?

Dear Lazyweb, what is the most elegant way to read a SOAP reply containing a complex type into an object using <a href="https://pywebsvcs.sourceforge.net/">ZSI</a>? I’ve read the <a href="https://pywebsvcs.sourceforge.net/zsi_v2_0_0.html">documentation</a> several times and I guess there is an elegant way to solve this, but I still don’t get it. I guess it should work like this: <ul> <li>Define a class Bugreport</li> <li>Define Bugreport’s <a href="https://pywebsvcs.sourceforge.net/zsi.html#SECTION007000000000000000000">typecode</a></li> <li>Parse (how?) the SOAP reply containing bugreport-data into a Bugreport object</li> </ul> What I’m currently doing with soappy in <a href="https://reportbug-ng.alioth.debian.org/">rng</a> is rather crude: I take the SOAP response which is a dictionary holding many interesting facts about a bug report, take those elements I’m interested in and copy them into a bug report object. Not very elegant, but it works. While I’m moving away from soappy to ZSI I’d like to implement the deserializion the ZSI way. Unfortunately I don’t understand their examples. So please Lazyweb, is there any ZSI expert out there who can provide a minimalistic example?Howto setup a print server for Windows (and others) using CUPS and zeroconf2007-08-17T12:03:00+02:002007-08-17T12:03:00+02:00Bastian Venthurtag:venthur.de,2007-08-17:/2007-08-17-howto-setup-a-print-server-for-windows-and-others-using-cups-and-zeroconf.html

Howto setup a print server for Windows (and others) using CUPS and zeroconf

Now that CUPS 1.3 has built in <a href="https://en.wikipedia.org/wiki/Zeroconf">zeroconf</a> support, it is easier than ever to provide print service to Windows machines. You don’t even need <a href="https://samba.org">Samba</a> anymore. Here’s a quick way to setup a CUPS print server and a Windows client: <h2>On the server side</h2> First you need to install CUPS 1.3 or higher on your print server and setup the printer (the web interface makes this task dead easy). Zeroconf isn’t enabled by default so you have to browse the CUPS admin page (https://localhost:631/admin if you’re running it on localhost) and enable the “Share published printers connected to this system” option. “Change Settings” will restart CUPS and your printer should be instantly visible to zeroconf-aware applications in the local network. If you can’t or don’t want to use the web interface (which is the easiest way though), you can edit <code>/etc/cups/cupsd.conf</code> and replace the line <code>Listen localhost:631</code> with <code>Port 631</code> and add the line <code>BrowseAddress @LOCAL</code> Now save the changes and restart CUPS and you’re done. <h2>On the client side</h2> Since most Linux distributions and MacOS support zeroconf out of the box, I’ll just explain how to setup the Windows client. On the Windows side you’ll need to install <a href="https://www.apple.com/support/downloads/bonjourforwindows.html">Bonjour for Windows</a> which is freely available on <a href="https://apple.com">apple.com</a>. Bonjour for Windows will enable Multicast DNS (mDNS) and DNS based Service Discovery (DNS-SD) on your Windows machine. Once it is installed, you can start the printer helper application provided by bonjour which will guide you through the process of installing the network printer. The only problem here was that although it correctly identified my printer, it only let me chose between a generic postscript and a PCL driver. Both were not working correctly so I had to manually change the driver for this printer to HP LaserJet 4 Plus (which is my printer) after the setup via the printer properties, where all other printer drivers where available again. That’s it! Now you should be able to print on every zeroconf aware operating system connected to your network. Oh, one final warning: With this setup, everybody in the LAN will be able to print without prior authentication. This should be what you want in most cases when running a home LAN or a small business LAN anyways, but it’s certainly not so cool when strangers have access to your network. The whole setup process took like 15 minutes today. I remember me a few years ago, not being very experienced with Samba and CUPS wasting a whole weekend to achieve the same result. Thanks apple for making life a bit easier again!My top 5 Firefox/Thunderbird annoyances2007-08-15T12:57:00+02:002007-08-15T12:57:00+02:00Bastian Venthurtag:venthur.de,2007-08-15:/2007-08-15-my-top-5-firefoxthunderbird-annoyances.html

My top 5 Firefox/Thunderbird annoyances

Firefox is constantly gaining marked share (especially here in Europe) and it’s little not-yet-so-popular brother Thunderbird is evolving too. While I’m happy that they’re contributing their part to bring Free Software to the masses, I’m concerned that the quality of their software (especially Firefox under Linux) has decreased in the last months. Here are my top 5 annoyances for Firefox and Thunderbird. Some of them only apply to Debian’s version others are more visible on my fairly old and “slow” (1.5GHz/512MiB) Laptop than on current and therefor faster machines: <h2>Firefox</h2> <ul> <li>Firefox seems to have a <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=362569">massive problem</a> with font kerning. Sometimes text is hardly readable since several letters seem to share the same one-letter space. And marking messed up text like this gives funny results: <img alt="" src="https://people.debian.org/~venthur/stuff/2007-08-firefox-kerningbug.gif"></li> <li>When visiting pages with a fixed background image like <a href="https://linuxactionshow.com/">this</a>, scrolling the page down causes massive CPU load and very slow scrolling even on fairly fast computers. Konqueror seems to handle pages like this much better.</li> <li>When reloading all tabs at once in a session and you have many (read: more than say, 6) tabs open, Firefox takes all the available CPU resources and pretty much becomes unresponsive until all tabs are loaded.</li> <li>Firefox literary freezes while loading very large pages (combine this with the previous annoyance for funny results)</li> <li>I still haven’t figured out why Firefox on my laptop (correctly) wants to launch kpdf when I open a .pdf while Firefox on my desktop seems to prefer xpdf. I’ve checked all the available options on both systems and wasn’t able to find the difference in the configuration.</li> </ul> <h2>Thunderbird</h2> <ul> <li>The most annoying bug in Thunderbird is Debian specific: it doesn’t open links in a browser anymore <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=425790">for months</a>. Since you have to change some non-obvious configuration to fix this, I assume this is the default behavior for many users out there which are now forced to copy-paste links to their browser.</li> <li>Like most other mail clients Thunderbird is also able to manage contacts. But it lacks two important features:</li> <li>No im/export of <a href="https://en.wikipedia.org/wiki/Vcard">vcard</a>: Hmm? Even outlook can do that and vcard is pretty much standard for storing contact information platform independent.</li> <li>No birthdays associated to contacts: Superb! When combining the Mozilla calendar application with Thunderbird you cannot just import all the birthdays from your contacts (which are stored as vcards of course), no you have to separately create those events which then of course are totally unrelated to the persons in your roster.</li> <li>The “search entire message” feature in the search box does absolutely nothing when searching in newsgroups. Why is it enabled then? Or: why does it pretend to do something by hiding all threads when entering something, suggesting no match was found?</li> <li>Wasn’t Thunderbird 1.x able to work with RSS feeds? <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=428421">What happened</a> to this feature in Thunderbird 2.x?</li> <li><a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=425336">What happened</a> to Thunderbird’s icon?</li> </ul> While I’m still happy with Thunderbird (or Iceddove as we call it here) I’m really searching for an alternative to Firefox (Iceweasel). Unfortunately there is currently no alternative available which fits all my needs. Firefox has two extensions no other browser seems to provide and without which I wouldn’t want to work anymore (I wonder if that was a proper English sentence…): <ul> <li><a href="https://addons.mozilla.org/de/firefox/addon/2410">Foxmarks Bookmark Synchronizer</a>: which allows me to store my bookmarks on my (that’s right, not just foxmark’s) server to keep it in sync with all my Firefoxes on various places and machines.</li> <li><a href="https://addons.mozilla.org/de/firefox/addon/10">Adblock</a> + <a href="https://addons.mozilla.org/de/firefox/addon/1136">Filterset.G Updater</a>: which keeps me mostly ad-free and automatically supplies me with the newest filter set rules.</li> </ul> While other browsers have some ad-blocking functionality too, they all seem to lack the filter set updater. The next best alternatives are <a href="https://www.konqueror.org/">Konqueror</a> and <a href="https://www.gnome.org/projects/epiphany/">Epiphany</a> but besides the lack of the above features every one of them has it’s own little quirks which doesn’t really motivate me to make the switch yet. But Konqueror is making progress and I’m really looking forward to KDE4 which will hopefully bring a usable web browser so I can finally get rid of Firefox. Update: Ooops, I nearly forgot my favorite pet annoyance of Firefox under Linux: Why does it have to insult the user’s eyes with this ugly toolkit when drawing check boxes, line edits, buttons etc? Firefox under Windows paints stuff like that according to Windows’ current theme, Konqueror under Linux as well, but what does Firefox for years? <img alt="" src="https://people.debian.org/~venthur/stuff/2007-08-firefox-widgets.png">Rng and others temporary broken2007-08-10T12:02:00+02:002007-08-10T12:02:00+02:00Bastian Venthurtag:venthur.de,2007-08-10:/2007-08-10-rng-and-others-temporary-broken.html

Rng and others temporary broken

Just a short note to let you know that due a change in <a href="https://packages.debian.org/debbugs">debbugs’</a> SOAP interface several applications including <a href="https://packages.debian.org/apt-listbugs">apt-listbugs</a> and <a href="https://reportbug-ng.alioth.debian.org/">rng</a> are temporary broken. The change was not too dramatic and I suppose the problem will be fixed shortly. Meanwhile you can use <a href="https://packages.debian.org/reportbug">reportbug</a> which is not affected, since it still uses the HTML interface. Update: Don fixed the problem and everything should be working as usual again.Implement two features and get the third for free!2007-08-03T13:48:00+02:002007-08-03T13:48:00+02:00Bastian Venthurtag:venthur.de,2007-08-03:/2007-08-03-implement-two-features-and-get-the-third-for-free.html

Implement two features and get the third for free!

Today I’ve implemented a feature which was long overdue: making the table sortable by the available table headers. The second feature I’ve added is that every bug report now has it’s “Last Action” date. While both features are useful for themselves, combining both result in a very handy bug triaging tool. <img alt="screenshot.png" src="/images/screenshot.png"> A few months ago when I was trying to help the KDE team on their bug triage, I’ve missed the feature to sort bugs by their date of the last action in our BTS since I’ve noticed that many of the very old bugs (last action-, not bug number wise) where already fixed or could be closed for other obvious reasons. The problem was just to find all those “stale bugs”. I was able to close dozens of old bug reports but it took much more time to find them, than to decide whether they’re still active or not. And since I hadn’t the tools available to find the “easy targets” on a quick way, I gave up after a few days. Today it’s just as easy as firing up rng and let it search for all bug reports belonging to the qt-kde team: <code>rng debian-qt-kde@lists.debian.org</code> and sorting the resulting list by last action… sweet! The package was just uploaded and should be available tomorrow. Update: Don pointed out that you can get the ordering in the web interface too by appending <code>ordering=age</code> to the URL. Neat.On Debian Maintainers2007-07-26T09:44:00+02:002007-07-26T09:44:00+02:00Bastian Venthurtag:venthur.de,2007-07-26:/2007-07-26-on-debian-maintainers.html

On Debian Maintainers

When voting on the current <a href="https://lists.debian.org/debian-devel-announce/2007/07/msg00009.html">Debian Maintainers General Resolution</a>, I think it’s worth to think about the following questions: What is the difference between a Debian Maintainer and a Debian Developer? Who is interested in becoming a Debian Maintainer and not a Debian Developer and why? Will this Debian Maintainer class really improve the situation for New Maintainers? I suppose most of our New Maintainers will aim to become a Debian Maintainer just to bridge the time until they’re full Debian Developers. And I predict that this is also the major target audience for this new Debian Maintainers class. Some people said on -vote that there are people actually interested in becoming a Debian Maintainer and not a Debian Developer. Personally I find their reasons questionable, but after all I suppose there won’t be many people going this route. On the other side, I don’t see how this new class enhances the situation for the New Maintainers. Now they not only wait for the Application Manager, Front Desk and Debian Account Managers – No, now they will also search for someone who advocates them to become a Debian Maintainer and have to wait for their key being added to the Debian Maintainers keyring. So my question is: If this General Resolution mainly helps New Maintainers, why must it be so complicated? Why another class of contributers, all those complicated rules to set it up and this additional layer of bureaucracy? Why don’t we aim for something simple, like improving our New Maintainer process. <a href="https://lists.debian.org/debian-vote/2007/07/msg00008.html">I would rather see</a> some of those Debian Maintainer rights granted to New Maintainers after a certain point in their New Maintainer career, but not in this complicated and overengineered way the Debian Maintainer class proposes.Small web calendar sought2007-07-22T15:20:00+02:002007-07-22T15:20:00+02:00Bastian Venthurtag:venthur.de,2007-07-22:/2007-07-22-small-web-calendar-sought.html

Small web calendar sought

Dear Lazyweb, does anyone know a small web calendar which has a web frontend and does not rely on a database server? It should save it’s data directly into <a href="https://de.wikipedia.org/wiki/ICalendar">.ical (or .ics)</a> files so I can fetch and alter them with a desktop calendar application too. I don’t need fancy features like multi user support or the ability to publish calendars, I just want do see and edit my calendar/todo stuff when I’m not at home and have the .ical saved on a centralized (my) server to make it available for desktop calendar software. For privacy reasons it has to run on my own server.Improved filtering for rng2007-07-18T15:18:00+02:002007-07-18T15:18:00+02:00Bastian Venthurtag:venthur.de,2007-07-18:/2007-07-18-improved-filtering-for-rng.html

Improved filtering for rng

I’ve just uploaded a new version of <a href="https://reportbug-ng.alioth.debian.org/">reportbug-ng</a> including an Italian and a Hebrew <a href="https://reportbug-ng.alioth.debian.org/ponies.html">pony</a> and a cool enhancement to the buglist filter: Now the filter doesn’t try to match the whole string anymore but splits the input by whitespaces and tries to match all words. Now you can even exclude words by prepending a leading minus to it. The Syntax for the filter is: <code>[-]keyword [[-]keyword2] ...</code> Multiple keywords are separated by whitespaces and can have a leading minus. A leading minus means “exclude” the whitespaces represent the logical AND. Some Examples: <ul> <li><code>foo bar -baz</code> means all bugs containing the words foo, bar and not baz</li> <li><code>outstanding -wishlist</code> means all bugs which are outstanding and not wishlist</li> <li><code>-resolved</code> will show all open bugs</li> </ul> Please note the the filter does look into the actual bugreports (the fulltext), it currently tries to match against the packagename, the bugnumber, the summary, the status and the severity. The best thing about this new filter? It’s not even slower than the old one, the only delay you’ll notice is when filtering packages with a huge amount of bugs (like wnpp) and even there the filter just needs a fraction of a second. The delay of a few secons comes acutally from the QTable when I remove all the rows not matching the filter. <code>QTable.hideRow(int)</code> and <code>-showRow(int)</code> seem to be very expensive even when I temporary disable all graphic updates for this widget during this operation. If someone knows a trick to circumvent this, please tell me (you may get a pony…). This feature isn’t spectacular, but it should make filtering big buglists a lot easier.Reportbug-NG finally in testing!2007-07-08T19:57:00+02:002007-07-08T19:57:00+02:00Bastian Venthurtag:venthur.de,2007-07-08:/2007-07-08-reportbug-ng-finally-in-testing.html

Reportbug-NG finally in testing!

Waiting ten days for rng to slip into testing, I had enough time to stuff the next version with new features and small bugfixes. The most interesting ones are: <ul> <li>Added SOAP support and enabled it by default. Feel free to mess around with the look and feel of our BTS, rng doesn’t depend on it anymore.</li> <li>Support for WNPP bugs. Well, you could use rng to file WNPP bugs before, but now rng actively supports you with a nice template for ITP and RFP reports.</li> <li>Mayor user interface overhaul. <img alt="rng.png" src="/images/rng.png"></li> <li>Use freedesktop.org’s xdg-utils to decide which browser to use to open URLs. This should allow rng to open the browser you set as default in your desktop environment.</li> </ul> In totally unrelated news, I finally managed to master <a href="https://en.wikipedia.org/wiki/Rubenstein%27s_revenge">Rubenstein’s Revenge</a>.Python, os.fork(), webbrowser and debbugs2007-07-01T13:20:00+02:002007-07-01T13:20:00+02:00Bastian Venthurtag:venthur.de,2007-07-01:/2007-07-01-python-osfork-webbrowser-and-debbugs.html

Python, os.fork(), webbrowser and debbugs

<a href="https://grmontesino.blogspot.com/2007/06/python-osfork-and-gsoc.html">Gustavo</a> writes about python, it’s webbrowser module, os.fork and parsing bugreports from our BTS via html. This all sound very familiar, so here are some hints from me: <ul> <li>You don’t have to fork a new process just to call a browser, try threading instead: <code>thread.start_new_thread(webbrowser.open, (url,))</code>. No need to clean up the child process after it has finished, just call <code>start_new_thread</code> with the function and it’s parameters and forget it.</li> <li>Python’s webbrowser module is nice, but it often fails to find the right browser. The problem is to find the users (read: not the system’s) default browser across the different desktop environments. xdg-open from the package xdg-utils (from freedesktop.org) seems to do a very good job. ~~Rng doesn’t use it yet, but I’m definitely planning to replace the webbrowser call with xdg-open.~~ Rng uses it and falls back to python’s webbrowser.open()-call if xdg-open failed.</li> <li>If your project suffers from changes in the html code of our BTS, try SOAP instead. It provides all the infos in machine readable format and also comes with backwards compatibility: the namespace for debbug’s SOAP interface is <code>Debbugs/SOAP</code> which always calls the latest versions of the remote functions. But you can also use <code>Debbugs/SOAP/V1</code> (Version 1 currently matches <code>Debbugs/SOAP</code>). This will ensure that changes in debbugs SOAP interface don’t affect your implementation.</li> </ul>What does it take to break rng?2007-06-27T11:10:00+02:002007-06-27T11:10:00+02:00Bastian Venthurtag:venthur.de,2007-06-27:/2007-06-27-what-does-it-take-to-break-rng.html

What does it take to break rng?

Actually just a small change in the html code of our BTS. The problem was easy to fix and the new version is already uploaded, but this shows again how fragile the connection between BTS’ html output and rng is. I’ve heard other tools where affected too – just one reason more to switch to SOAP as soon as possible. An unfortunate side effect of today’s upload: In the last weeks I tried very hard to finally get rng into testing by not uploading as often as I wanted to. The problem aren’t RC bugs, but archs like SPARC or ARM which seem to have problems keeping up with the build queue. Being of priority extra doesn’t help very much, since packages with higher priorities (like: all other packages but rng) always jump the queue. Unfortunately my request to fix the override disparity I filed two weeks ago wasn’t processed yet and the only one who could fix this is currently on vacation. In fact <a href="https://ftp-master.debian.org/">others</a> could fix this too, but it looks like they’re on vacation too. I guess I’ll just exclude both archs in one of the next uploads until things run smoother again. But not everything is dark: with the current upload, a Dutch, a Frech, a Norwegian and a Polish pony found a <a href="https://reportbug-ng.alioth.debian.org/ponies.html">new home</a>. Update: Looks like the override disparity has been fixed. Whoever fixed this, thank you! Rng should be available on all our archs by tomorrow and in testing in 10 days!Ponies!2007-06-22T21:33:00+02:002007-06-22T21:33:00+02:00Bastian Venthurtag:venthur.de,2007-06-22:/2007-06-22-ponies.html

Ponies!

<img alt="pony.gif" src="/images/pony.gif"> As <a href="2007-06-22-localizing-reportbug-ng-ii.html">promised</a>, everyone who sends in a new translation for reportbug-ng gets a <a href="https://reportbug-ng.alioth.debian.org/ponies.html">pony</a>. Just send a new translation and the picture of your favorite pony. If you don’t send a pony, you’ll get one from me.Localizing reportbug-ng (II)2007-06-22T09:06:00+02:002007-06-22T09:06:00+02:00Bastian Venthurtag:venthur.de,2007-06-22:/2007-06-22-localizing-reportbug-ng-ii.html

Localizing reportbug-ng (II)

A month ago I <a href="2007-05-27-please-help-to-localize-reportbug-ng.html">asked</a> for help to localize <a href="https://reportbug-ng.alioth.debian.org/">reportbug-ng</a>. To my surprise translations came in pretty quick. Junichi Uekawa was the first one, he sent his Japanese translation only a few hours after my request for help. Currently we have rng translated into the following languages: Catalan, Czech, English, German, Japanese, Spanish and Swedish. Thanks David Planella, Miroslav Kure, Junichi Uekawa, Javier Fernández-Sanguino Peña and Daniel Nylander for your translations! But that’s not enough. If your language is missing, please consider translating rng to your language. Just take the source of the current version in unstable, grab the .pot and translate it. There isn’t really that much text to translate. The .pot contains only 42 messages and most of them are very short. When you’re done, please file a l10n+patch bug against rng with the translation attached. I’ll include your translation as soon as possible, thank you in the changelog and will probably name my first pony after you ;) Thanks!1 Kilobyte = 1000 ± 24 bytes2007-06-12T21:28:00+02:002007-06-12T21:28:00+02:00Bastian Venthurtag:venthur.de,2007-06-12:/2007-06-12-1-kilobyte-1000--c2-b1-24-bytes.html

1 Kilobyte = 1000 ± 24 bytes

Could you exactly say how big your hard disk has to be at least when the vendor claims it’s 500 Gigabytes big? Do you know how many bytes a file has when the filemanager says 2 Megabytes? Do you know what your DSL-provider actually means with 2 Megabit per second? Experienced users surely can since they know the meaning of “Kilobyte” depends on the context. Sometimes it’s 1000 bytes, sometimes it’s 1024. Non technical users will tend to assume that a 1 Kilobyte is 1000 bytes (which is correct, since k is the SI prefix for kilo) while users with technical background will tend to assume (depending on the context of course) it’s 1024. There is <a href="https://en.wikipedia.org/wiki/Binary_prefix">a solution</a> for this ambiguity, but unfortunately it doesn’t come for free: First, it sounds really stupid (kibi-what?) and second it will add an extra character (kB -> kiB) and we all know how stuffed our 80 column terminals already are nowadays. If you have better arguments against the binary prefix notation please, join the <a href="https://lists.debian.org/debian-devel/2007/06/msg00500.html">discussion</a>. Personally I’d find it very cool if we could use SI prefixes where we mean n * 10^3 and binary prefixes where we mean n * 2^10 in Debian consistently.Rebuilding the Archive Using Cowbuilder (II)2007-06-06T11:11:00+02:002007-06-06T11:11:00+02:00Bastian Venthurtag:venthur.de,2007-06-06:/2007-06-06-rebuilding-the-archive-using-cowbuilder-ii.html

Rebuilding the Archive Using Cowbuilder (II)

A few months ago I wrote a script to <a href="2007-02-20-rebuilding-the-archive-using-cowbuilder.html">rebuild the archive using cowbuilder</a> This script had a subtle bug: If you where running a sid machine and wanted to rebuild testing, the script would fetch the sources from sid (via apt-get source) and build this package in a testing environment. Thanks to <a href="https://info.comodo.priv.at/">Gregor Herrmann</a> (Currently unassigned <a href="https://nm.debian.org/nmstatus.php?email=gregor%2Bdebian%40comodo.priv.at">NM</a> with ~180 packages in the archive!) this bug is fixed now and it should be possible to rebuild the archive correctly for every host/build-environment combo. To rebuild the archive, just <a href="https://people.debian.org/~venthur/stuff/2007-06-rebuild/">get the scripts</a> and <div class="codehilite"><pre><code>sudo aptitude install cowdancer grep-dctrl sudo cowbuilder --create --distribution lenny --basepath /var/cache/pbuilder/testing-base.cow ./getlist lenny # get the list of packages ./buildall list.lenny.i386 lenny # build all packages in list.lenny.i386 in a lenny environment </code></pre></div> Thanks Gregor for your patches and good luck with your NM process!LinuxTag2007-05-30T21:58:00+02:002007-05-30T21:58:00+02:00Bastian Venthurtag:venthur.de,2007-05-30:/2007-05-30-linuxtag.html

LinuxTag

My first day at LinuxTag is over. I’m a bit tired, but the day was nice. We had many visitors at our booth and some nice conversations. We sold a few t-shirts and gave a way a lot of Debian DVDs. Unfortunately there was also a high demand on Live CDs which we couldn’t fulfill since we still don’t have official Live images. Last but not least, it was very nice to meet some other Debian Developers. Unfortunately I can’t attend at our booth tomorrow, but I’m already looking forward to Friday (Debian Day!).Please help to localize reportbug-ng2007-05-27T16:37:00+02:002007-05-27T16:37:00+02:00Bastian Venthurtag:venthur.de,2007-05-27:/2007-05-27-please-help-to-localize-reportbug-ng.html

Please help to localize reportbug-ng

I’ve just uploaded a new version of <a href="https://reportbug-ng.alioth.debian.org/">reportbug-ng</a>. The biggest change in this release is the new reportbug-ng.pot file which contains some work for our brave localization teams. I’ve already translated rng into German, so rng is now available in German and English. If you want your mother language to appear in this list, please grab the .pot, translate it and file a l10n+patch bug against reportbug-ng with the translation attached. Thanks!Using zeroconf to resolve local hostnames2007-05-25T14:22:00+02:002007-05-25T14:22:00+02:00Bastian Venthurtag:venthur.de,2007-05-25:/2007-05-25-using-zeroconf-to-resolve-local-hostnames.html

Using zeroconf to resolve local hostnames

If you bought one of those cheap WLAN-routers which provides a DHCP server but no DNS (why again?), you probably miss the feature to access your hosts in the LAN by their name, right? If you’re running a fairly recent Linux distribution (yes, Etch works too), just try to ping them with the <a href="https://en.wikipedia.org/wiki/.local">.local DNS suffix</a>, like phoibe.local or hades.local. If that doesn’t work, check if avahi-daemon is installed on the machines you want to resolve. If not (which is unlikely on recent desktop machines), just install it and your machine will be available via hostname.local. I wasn’t aware that I would ever need this “zeroconf crap” until I replaced my rusty old Dell Optiplex GXa router with a less energy consuming and much more silent Linksys WRT54G. Unfortunately I bought Version 7 which only comes with 2MB flash and is therefore not supported by <a href="https://openwrt.org/">openwrt</a>. So without my precious <a href="https://en.wikipedia.org/wiki/Dnsmasq">dnsmasq</a> on my router, I was suddenly blind in my own LAN – and very thankful to find out that <a href="https://en.wikipedia.org/wiki/Zeroconf">zeroconf</a> already took care of my problem.4GB RAM on a T60?2007-05-14T18:50:00+02:002007-05-14T18:50:00+02:00Bastian Venthurtag:venthur.de,2007-05-14:/2007-05-14-4gb-ram-on-a-t60.html

4GB RAM on a T60?

Dear Lazyweb, at work we have bought some of those shiny new ThinkPad T60s with 64Bit. Some of them have 4GB RAM installed, but unfortunately those with 4GB only seem to have 3GB available under Debian/Etch/AMD64. Since those laptops are running on AMD64 kernels, they should work with >= 3GB RAM out of the box, right? Unfortunately they don’t and <a href="https://www.thinkwiki.org/wiki/Category:T60">thinkwiki</a> says: <blockquote> Note: While you can install 4GB of memory, the chipset in the T60 supports a maximum of 3GB usable memory. </blockquote> while the BIOS correctly shows the full amount of installed RAM! So my question is: Has anyone out there a T60 running successfully with 4GB RAM or is it really a chipset constraint?Last chance for LinuxTag 20072007-05-04T08:24:00+02:002007-05-04T08:24:00+02:00Bastian Venthurtag:venthur.de,2007-05-04:/2007-05-04-last-chance-for-linuxtag-2007.html

Last chance for LinuxTag 2007

Alexander Wirt <a href="https://lists.debian.org/debian-events-eu/2007/05/msg00009.html">posted</a> the following request on <a href="https://lists.debian.org/debian-events-eu/">-events-eu</a>. Since I fear it won’t get the attention it deserves, I’m quoting his request here again. <blockquote> Hi, unfortunatly there weren’t enough talks submitted for the Debian Day that I have to cancel it until there won’t be some new submissions in the next few hours. So if you want to have a debian day, please submit a talk via the VCC (https://www.linuxtag.org/vcc/). Otherwise it will be canceled. Alex </blockquote> Update: Mission accomplished. Looks like a few talks were submitted just in time to save the (Debian) day.Making cool panoramic pictures with hugin2007-04-25T18:13:00+02:002007-04-25T18:13:00+02:00Bastian Venthurtag:venthur.de,2007-04-25:/2007-04-25-making-cool-panoramic-pictures-with-hugin.html

Making cool panoramic pictures with hugin

Thanks <a href="https://0pointer.de/blog">Lennart</a> for mentioning <a href="https://hugin.sourceforge.net/">Hugin</a>! This is a really cool piece of software, allowing you to assemble a mosaic of photographs into a complete panorama. <img alt="panorama-640.jpg" src="/images/panorama-640.jpg"> I’ve created this 180 degrees panorama from 16 photos taken out of my window (3 rows: 6 + 6 + 4). The original picture is huge: 5000-something pixels wide and high and approximately 22Megs (jpg) fat. If you want to test it yourself just <code>aptitude install hugin</code> and make sure to get enblend from somewhere else (eg, <a href="https://debian-multimedia.org">debian-multimedia.org</a>). If you don’t install enblend you don’t lose functionality, but the output will be of much lower quality. Thanks Florent Bayle for packaging hugin and Christian Marillat for packaging enblend!New queries for reportbug-ng2007-04-23T17:52:00+02:002007-04-23T17:52:00+02:00Bastian Venthurtag:venthur.de,2007-04-23:/2007-04-23-new-queries-for-reportbug-ng.html

New queries for reportbug-ng

Instead of the boring “package” queries, <a href="https://reportbug-ng.alioth.debian.org/">rng</a> used to support ‘till now, rng now supports all the queries our BTS supports: <ul> <li><code>package</code>: Returns all the bugs belonging to the PACKAGE</li> <li><code>bugnumber</code>: Returns the bug with BUGNUMBER and loads it into the built in browser immediately</li> <li><code>maintainer@foo.bar</code>: Returns all the bugs assigned to MAINTAINER</li> <li><code>src:package</code>: Returns all the bugs belonging to the SOURCEPACKAGE</li> <li><code>from:submitter@foo.bar</code>: Returns all the bugs filed by SUBMITTER</li> <li><code>severity:foo</code>: Returns all the bugs of SEVERITY. Warning this list is probably very long. Recognized are the values: critical, grave, serious, important, normal, minor and wishlist</li> <li><code>tag:bar</code>: Returns all the bugs marked with TAG</li> </ul> Those queries are available on the command line too and behave exacly as if entered in the program: The syntax is: <code>rng [QUERY]</code>, where query is one of the above (and <code>rng</code> a convenient shortcut to <code>reportbug-ng</code>). An example: if you start <code>rng from:venthur@debian.org</code>, you’ll get the full list of all bugreports I’ve ever reported. Now you can filter the list to find all bugreports I’ve filed against, say: qgo… voila. It doesn’t get any easier than that, does it? Those new queries and a few changes under the hood required quite a lot of refactoring in the code and I’m somewhat scared that I’ve introduced some new bugs. So I guess I should stop releasing so often (or for at least 10 days so rng can finally enter testing) and use the time to start writing some unit tests.Reportbug-NG migrated from SVN to BZR2007-04-20T12:48:00+02:002007-04-20T12:48:00+02:00Bastian Venthurtag:venthur.de,2007-04-20:/2007-04-20-reportbug-ng-migrated-from-svn-to-bzr.html

Reportbug-NG migrated from SVN to BZR

The migration was rather easy <div class="codehilite"><pre><code>sudo aptitude install bzr-svn # patch (see below) bzr branch svn://svn.debian.org/reportbug-ng # That's right I branch directly from a SVN repo! # ignore the error bzr push --use-existing-dir sftp://venthur@bzr.debian.org/bzr/reportbug-ng </code></pre></div> To get your own branch: <div class="codehilite"><pre><code>bzr branch https://bzr.debian.org/bzr/reportbug-ng/ </code></pre></div> Bzr-svn really makes it easy to migrate from svn to bzr… well, in theory. In practice we have <a href="https://bugs.debian.org/415721">#415721</a>, but fortunately there is an easy workaround. Just temporarily apply this trivial patch: <div class="codehilite"><pre><code>--- /usr/share/pycentral/bzr-svn/site-packages/bzrlib/plugins/svn/__init__.py 2007-04-20 09:30:00.000000000 +0200 +++ /usr/share/pycentral/bzr-svn/site-packages/bzrlib/plugins/svn/__init__.py.new 2007-04-20 09:30:24.000000000 +0200 @@ -23,7 +23,7 @@ import bzrlib` __version__ = '0.2.0' -required_bzr_version = (0,13) +#required_bzr_version = (0,13) def check_bzrlib_version(desired): """Check that bzrlib is compatible. </code></pre></div> and life is good again. The reason why I switched from a centralized to a distributed RCS was that I really like the option to do (small) local checkins and just do the “real” checkin when the work is done. In terms of BZR it means commits to my local branch for smaller changes and pushes back to the mainline for every major change. Oh, and it’s also nice always having the full repository data locally stored. My requirements for the RCS were: it should be easy to use (no steep learning curve), well tested and somewhat mainstream. BZR fits quite well, especially since Ubuntu heavily relies on it for their launchpad stuff. PS: Thanks to the <a href="https://alioth.debian.org">alioth</a> team for creating the BZR repository so quickly, it just took like 1 minute.What about a contest for the next debian.org theme?2007-04-18T20:53:00+02:002007-04-18T20:53:00+02:00Bastian Venthurtag:venthur.de,2007-04-18:/2007-04-18-what-about-a-contest-for-the-next-debianorg-theme.html

What about a contest for the next debian.org theme?

Currently there is a <a href="https://lists.debian.org/debian-www/2007/04/msg00249.html">discussion</a> about a new theme for <a href="https://debian.org">debian.org</a> going on on <a href="https://lists.debian.org/debian-www/">-www</a>. Some people presented their ideas and build some mockups while others commented them. I strongly suggest to let some professional designers and other (not necessarily Debian related) talented people create some mockups, to avoid moving from one outdated and rusty design to the next. I wonder if we couldn’t open up the process a bit by starting some kind of contest. Skilled webdesigners could create some mockups and we could collect all those and let a vote decide which one to keep.How long does it take to turn a feature into a bug?2007-04-17T20:06:00+02:002007-04-17T20:06:00+02:00Bastian Venthurtag:venthur.de,2007-04-17:/2007-04-17-how-long-does-it-take-to-turn-a-feature-into-a-bug.html

How long does it take to turn a feature into a bug?

Looking at the current spam attack at our <a href="https://bugs.debian.org">BTS</a> I wonder how long it’ll take ‘till the feature that our BTS is controllable via mail becomes a pain in the ass because of the spam. I’m sure we have some spamfilter and some brave people controlling it, since in the last time it was relative quiet and spam free. But the current attack makes me think if we couldn’t raise the barrier for spam somehow. Not sure how to achieve that though.Since More and More People Are Referencing to wiki.debian.org...2007-04-15T19:02:00+02:002007-04-15T19:02:00+02:00Bastian Venthurtag:venthur.de,2007-04-15:/2007-04-15-since-more-and-more-people-are-referencing-to-wikidebianorg.html

Since More and More People Are Referencing to wiki.debian.org...

shouldn’t we finally move away from this ugly MoinMoin theme? It’s not only the most ugly and spartan theme one could imagine (it even beats our main <a href="https://debian.org">web presence</a> in ugliness) – it does not even try to comply with some corporate design. The only things that remind me that I’m actually surfing a Debian related site are the URL and the words “Debian Wiki” on the top left. (There is a chance that the link-colors match the Debian-Red too, but I’m slightly colorblind, so I’m not sure) MoinMoin can do better. Please look at <a href="https://wiki.ubuntu.com/">Ubun…</a> oops I mean <a href="https://bazaar-vcs.org/">bazaar’s homepage</a>, they’re using MoinMoin too and it looks much more appealing than our wiki.Reportbug-NG, Now With Fancy Colors2007-03-28T11:51:00+02:002007-03-28T11:51:00+02:00Bastian Venthurtag:venthur.de,2007-03-28:/2007-03-28-reportbug-ng-now-with-fancy-colors.html

Reportbug-NG, Now With Fancy Colors

After a few hours of hacking I got <a href="https://bugs.debian.org/416132">#416132</a> closed and am proud to say that we finally have fancy colors enabled in <a href="https://reportbug-ng.alioth.debian.org/">Reportbug-NG</a>. <img alt="" src="images/reportbug-ng.gif"> I’m also very proud to say that yesterday – exactly 3 weeks after I wrote the first line of R-NG – R-NG received it’s 100th <a href="https://popcon.debian.org/">popcon</a> submission! Thanks to all of you who tested and reported bugs and wishes. Some of the next goals I’ve currently in mind are <ul> <li>Porting R-NG from qt3 to qt4</li> <li>Include some advanced BTS-features for developers like tagging, reassigning, merging, etc.</li> </ul>Dear SoC enthusiasts,2007-03-22T16:17:00+01:002007-03-22T16:17:00+01:00Bastian Venthurtag:venthur.de,2007-03-22:/2007-03-22-dear-soc-enthusiasts.html

Dear SoC enthusiasts,

please stop writing mails asking me whether I’d like to mentor your yet-another-reportbug-alternative project. It’s not that I’m opposed to the idea of rewriting reportbug(-ng) for gtk or as a webapp – really. I just don’t see why your project needs to be part of SoC. I know working for/with google is kinda sexy nowadays and as a student you can even get some money – but be honest: what was Debian’s benefit of <a href="https://wiki.debian.org/SummerOfCode2006">last year’s SoC projects</a>? My question which of last year’s projects failed and which succeeded is <a href="https://lists.debian.org/debian-project/2007/03/msg00122.html">still unanswered</a>. So no, I’m not yet jumping on this bandwagon everybody seems to be so exited about. Don’t get me wrong: I don’t hate google or the SoC, I’m just not convinced of the success of our last year’s projects, plus I don’t think we need such an event to write useful tools for Debian. NP: The Beatles – Happiness Is a Warm GunThanks Lazyweb,2007-03-20T15:00:00+01:002007-03-20T15:00:00+01:00Bastian Venthurtag:venthur.de,2007-03-20:/2007-03-20-thanks-lazyweb.html

Thanks Lazyweb,

I received quite a few answers on my <a href="2007-03-19-dear-lazyweb.html">request for help</a> yesterday. Thanks to you guys and gals, <a href="https://reportbug-ng.alioth.debian.org/">Reportbug-NG</a> now also supports Opera, KMail, Pine and <a href="https://mail.google.com/">GoogleMail</a> (yes I know, that’s crazy).Dear Lazyweb,2007-03-19T19:43:00+01:002007-03-19T19:43:00+01:00Bastian Venthurtag:venthur.de,2007-03-19:/2007-03-19-dear-lazyweb.html

Dear Lazyweb,

Icedove, Iceape, Evolution, Sylpheed, Sylpheed-Claws and Mutt. Is your favorite mail client missing? Does your mail client accept URLs in mailto-format? If yes, please drop me a note containing the name of your mail client, if it needs a terminal (like mutt) and an example call which starts the client with address, subject an body already filled in. Examples: <ul> <li>Icedove (needs no terminal) <code>icedove "mailto:mail@example.com?subject=foosubject&body=foobody"</code></li> <li>Sylpheed (needs no terminal) <code>sylpheed --compose "mailto:mail@example.com?subject=foosubject&body=foobody"</code></li> </ul> I’ll use this data to include more clients in <a href="https://reportbug-ng.alioth.debian.org/">Reportbug-NG</a>. Please double-check if your mail client really supports mailto-URLs with subject and body. KMail for instance accepts such a parameter but <a href="https://bugs.debian.org/415454">silently drops subject and body</a>.Reportbug-NG in Unstable2007-03-11T15:52:00+01:002007-03-11T15:52:00+01:00Bastian Venthurtag:venthur.de,2007-03-11:/2007-03-11-reportbug-ng-in-unstable.html

Reportbug-NG in Unstable

Just a few minutes ago <a href="https://reportbug-ng.alioth.debian.org/">reportbug-ng</a> was accepted and should be available in unstable tomorrow (For the impatient: you can already get it from <a href="https://incoming.debian.org/">incoming</a>). <img alt="" src="images/reportbug-ng.gif"> If you can spare some time please test it and send me suggestions, bugreports and stuff.Reportbug-NG at Alioth2007-03-08T17:46:00+01:002007-03-08T17:46:00+01:00Bastian Venthurtag:venthur.de,2007-03-08:/2007-03-08-reportbug-ng-at-alioth.html

Reportbug-NG at Alioth

Yesterday <a href="2007-03-07-a-screencast-is-worth-a-thousand-pictures.html">I mentioned</a> a small project of mine called reportbug-ng. The feedback so far was very positive. Looks like quite a few people out there would like to see a desktop version of reportbug. Since I’ve already received a few patches via email, I decided to move the project to alioth so contributers can easily keep track of the changes. So here it is: the <a href="https://reportbug-ng.alioth.debian.org/">official homepage</a> and teh almighty <a href="svn://svn.debian.org/reportbug-ng">SVN repository</a>. Comments and contributions are very welcome! (Hint, hint: The source-tree also contains a TODO list. Once all the points under “urgend” are processed, I’ll consider uploading it to Debian.)A Screencast is Worth a Thousand Pictures2007-03-07T16:14:00+01:002007-03-07T16:14:00+01:00Bastian Venthurtag:venthur.de,2007-03-07:/2007-03-07-a-screencast-is-worth-a-thousand-pictures.html

A Screencast is Worth a Thousand Pictures

<img alt="" src="images/reportbug-ng.gif"> It’s still in a very early stage (I’ve started coding it yesterday), but it’s basic functionality is already there: You can query the BTS, filter the list of bugreports in realtime by typing something and view the full bugreports in HTML. You can even provide additional information for existing bugs or create new bugreports. It currently only supports icedove and mutt as mailclient but adding other clients is very easy. This tool is not designed to replace reportbug, it rather aims for the newbie users feeling uncomfortable typing their mails in vi and would rather have something to point and click. Maybe I can merge the code with reportbug, but after a quick look at it’s code it seems that this will not be as easy as I thought in the beginning. The code is available <a href="https://debian.org/~venthur/stuff/2007-03-reportbug/reportbug-ng.tar.gz">here</a>. You’ll need python-qt3 to run it. Update: The screencast was created with <a href="https://packages.qa.debian.org/b/byzanz.html">byzanz</a> via <div class="codehilite"><pre><code>$ byzanz-record -c -l -d 45 reportbug-ng.gif </code></pre></div> Hint: Best to switch the desktop resolution to something like 800x600 before recording – resizing the gif afterwards is very expensive (cpu- and memory wise).Etch Won't Ship Wordpress2007-03-07T01:09:00+01:002007-03-07T01:09:00+01:00Bastian Venthurtag:venthur.de,2007-03-07:/2007-03-07-etch-wont-ship-wordpress.html

Etch Won't Ship Wordpress

So the security team <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=413269">decided to kick Wordpress from Etch</a> since they assume it won’t be sanely maintainable over the course of 30-36 months. This practice surprises me a bit, especially since upstream plans to support the 2.0.x series until 2010 and the package is currently very well maintained by Kai Hendry. So what’s the problem? Is it manpower? I guess not, since <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=413269#44">Kai even offered</a> to provide security updated packages when they become available, which was rejected with a friendly “<a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=413269#54">EOD for me.</a>”. Anyway, since my blog will become unsupported soon, I need to switch to something else. So dear lazyweb, what is the next best alternative to WP available in Etch? I need good comment spam protection and an option to import my old wordpress entries.Pimp My Boot Process (III)2007-03-02T00:52:00+01:002007-03-02T00:52:00+01:00Bastian Venthurtag:venthur.de,2007-03-02:/2007-03-02-pimp-my-boot-process-iii.html

Pimp My Boot Process (III)

Short version: <div class="codehilite"><pre><code>$ sudo aptitude install desktop-base splashy splashy-themes $ sudo splashy_config --set-theme debian-moreblue $ sudo vim /boot/grub/menu.lst </code></pre></div> <div class="codehilite"><pre><code>[...] timeout 3 hiddenmenu [...] # defoptions=quiet splash [...] </code></pre></div> <div class="codehilite"><pre><code>$ sudo update-grub </code></pre></div> Long version: <a href="2007-02-10-pimp-my-boot-process-ii.html">A few weeks ago</a> I tried to improve the visual appearance of my boot process using usplash and a matching grub splash theme. It took me quite a while to create my first very own usplash theme and it was a rather complicated affair (Thanks to Petter Reinholdtsen for his patch which lead to my first working prototype). Since creating usplash themes is rather tricky and not really suitable for the average user I decided to switch to splashy. Splashy has many advantages above usplash like a very easy way to create themes, support of high resolution images and many themes already available in Debian. So here is my current solution for a nice bootup: First I installed desktop-base in order to get the nice moreblue KDM theme: <div class="codehilite"><pre><code>$ sudo aptitude install desktop-base </code></pre></div> Then I installed splashy and splashy-themes in order to get a theme matching KDM’s moreblue theme: <div class="codehilite"><pre><code>$ sudo aptitude install splashy splashy-themes $ sudo splashy_config --set-theme debian-moreblue </code></pre></div> Since grub has only rudimentary support for theming, I decided to remove the boot menu completely. Most of the time I’ll boot the first kernel in the list anyway and you still have <code>$timeout</code> seconds to press Escape in order to bring the menu back. <div class="codehilite"><pre><code>$ sudo vim /boot/grub/menu.lst </code></pre></div> <div class="codehilite"><pre><code>[...] timeout 3 hiddenmenu [...] # defoptions=quiet splash [...] </code></pre></div> <div class="codehilite"><pre><code>$ sudo update-grub </code></pre></div> The result looks quite cool and gives my boot process a polished an clean look. It’s still not perfect since splashy’s theme does not match exactly KDM’s theme and KDM flicks in an other wallpaper before it starts and after the login, but it’s good enough for now.Decoupling Presentation from Logic2007-03-01T11:43:00+01:002007-03-01T11:43:00+01:00Bastian Venthurtag:venthur.de,2007-03-01:/2007-03-01-decoupling-presentation-from-logic.html

Decoupling Presentation from Logic

Reading <a href="https://www.debian.org/vote/2007/platforms/sho">Sam’s platform</a>, I applaud his plans to sex up our <a href="https://debian.org">web appearance</a> which is IMHO overdue. While redesigning our website shouldn’t be much of a technical problem, redesigning our <a href="https://bugs.debian.org">BTS</a> could become one: I noticed that reportbug and bts are basically querying the BTS via HTTP as a user would, grepping the source of the html-reply for the information they need afterwards. While this works pretty good in most situations it has the major drawback that those tools unnecessarily depend on the presentation of the BTS, making it harder to redesign it. A possible and easy to implement solution would be to allow a special parameter in the URL-request which triggers an answer in a machine readable format (like flatfile or XML) rather than a human readable. This would also have the additional benefit of saving bandwidth and computation power.Rebuilding the Archive Using Cowbuilder2007-02-20T12:23:00+01:002007-02-20T12:23:00+01:00Bastian Venthurtag:venthur.de,2007-02-20:/2007-02-20-rebuilding-the-archive-using-cowbuilder.html

Rebuilding the Archive Using Cowbuilder

Did you ever want to rebuild the archive, but setting up this whole wanna-build/sbuild/buildd stuff somewhat scared you? Then this small script might be for you. All you need is cowbuilder, the rights to build inside a cowbuilder environment and <a href="https://people.debian.org/~venthur/stuff/2007-02-rebuild/rebuild.tar.gz">this script</a>. When you call it with a list of packages as argument it will build every package in a clean cowbuilder environment, saving the buildlog <code>(succeeded-$package or failed-$package)</code> and cleaning up the rest afterwards. To speed up things you can run several instances of this script concurrently – lockfiles will prevent the script from building the same package more than once. It is even safe to kill the script and restart it later. It will automatically skip all packages where a logfile <code>(succeded|failed)-$package</code> already exists. All you have to do after you killed the script is to remove the lockfile in the <code>LOGDIR</code> and the <code>$package-directory</code> in the <code>BUILDDIR</code>. <div class="codehilite"><pre><code>#!/bin/sh # Version: 2007-02-20 LIST=$1 LOGDIR=$(pwd)/"logs" BUILDDIR=$(pwd)/"build" BASEPATH="/var/cache/pbuilder/base-testing.cow" function build { PACKAGE=$1 mkdir -p $BUILDDIR/$PACKAGE cd $BUILDDIR/$PACKAGE apt-get source $PACKAGE &> /dev/null cd $(find . -type d ! -name .) pdebuild --pbuilder cowbuilder --use-pdebuild-internal --buildresult .. -- --basepath $BASEPATH &> $LOGDIR/.$PACKAGE if [ $? -eq 0 ]; then mv $LOGDIR/.$PACKAGE $LOGDIR/succeeded-$PACKAGE else mv $LOGDIR/.$PACKAGE $LOGDIR/failed-$PACKAGE fi cd $BUILDDIR rm -rf $PACKAGE } [ -d $LOGDIR ] || mkdir $LOGDIR if [ ! -e $LIST ]; then echo "No Buildlist given." exit 1 fi while read package; do # Create Lockfile (ln -s is atomic and fails if link already exists) ln -s $package $LOGDIR/.$package.lock 2> /dev/null || continue # Build the package if not already built if [ ! -e $LOGDIR/succeeded-$package -a ! -e $LOGDIR/failed-$package ]; then echo $(date +%F\ %T) Building $package... build $package fi # Remove Lockfile rm -f $LOGDIR/.$package.lock done < $LIST </code></pre></div> I know my bash scripting skills are shitty, so please send patches after you laughed ;) I’m currently running two instances concurrently for 3 days on my Pentium 2,4GHz HT box and have rebuilt ~70% of the testing archive so far. It would be cool to distribute the rebuilt across a few computers via ssh, so that they’re all building on the same list but using their own cowbuilders. It should be sufficient if all computers shared the same <code>LOGDIR</code> across the network, but I’m not sure how to do this. NFS does not the trick since it cannot guarantee the atomicity of the ln-command.DPL 2007, Current Candidates2007-02-14T13:36:00+01:002007-02-14T13:36:00+01:00Bastian Venthurtag:venthur.de,2007-02-14:/2007-02-14-dpl-2007-current-candidates.html

DPL 2007, Current Candidates

For those who don’t follow -vote regularly, it seems like we have currently four official nominations for this year’s DPL vote: <ul> <li><a href="https://lists.debian.org/debian-vote/2007/02/msg00002.html">Gustavo Franco</a></li> <li><a href="https://lists.debian.org/debian-vote/2007/02/msg00003.html">Sven Luther</a></li> <li><a href="https://lists.debian.org/debian-vote/2007/02/msg00004.html">Wouter Verhelst</a></li> <li><a href="https://lists.debian.org/debian-vote/2007/02/msg00037.html">Aigars Mahinovs</a></li> </ul> and two pending ones: <ul> <li>Raphaël Hertzog (undecided)</li> <li>Sam Hocevar (unofficial)</li> </ul> A totally unrelated note: If you want to get really frustrated and angry but don’t feel like participating in Debian’s infamous flamefests – try to learn to juggle a 5 ball cascade. This will keep you busy for several months without getting on everybody’s (well, everybody but your neighbor’s) nerves. Update: Ooops, I forgot Wouter – thanks <a href="https://blog.schmehl.info/Debian/dpl-2007-1">Alexander</a>.Pimp My Boot Process (II)2007-02-10T14:59:00+01:002007-02-10T14:59:00+01:00Bastian Venthurtag:venthur.de,2007-02-10:/2007-02-10-pimp-my-boot-process-ii.html

Pimp My Boot Process (II)

I’m still working on a grub/usplash splashscreen in order to <a href="2007-01-25-pimp-my-boot-process.html">pimp my boot process</a>. Let’s see what we have: <img alt="" src="/images/screenshot_debian_grubsplash.png"> Creating the grub image is rather easy. What’s obviously missing here is the correct position and color of the text stuff. <img alt="" src="/images/screenshot_debian_usplash.png"> Creating a working splash image for usplash was kinda tricky. I tried several HOWTOs which took a long time since I had to reboot in order to test the result. Unfortunately every attempt I made resulted in either a blank screen or the default fallback splash. As a last resort I took the source of debian-edu-artwork which has a working usplash, replaced their image with mine and rebuild the package. I don’t know yet why, but it worked. I plan to create package with combined grub/usplash(/xDM) artwork and several alternative themes, but since our usplash version is <a href="https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=397970">lagging behind</a> upstreams I think I’ll wait until we catched up again. The new version is not backwards compatible with the old one and the new one as quite a lot new cool features like: <ul> <li>scripts which help to create splashscreens</li> <li>support of 256 colors (currently 16)</li> <li>support of different resolutions (currently 640x480 is hardcoded)</li> </ul> which make it a lot easier to create good looking splash screens.Number concerning the next release: 1042007-02-09T19:04:00+01:002007-02-09T19:04:00+01:00Bastian Venthurtag:venthur.de,2007-02-09:/2007-02-09-number-concerning-the-next-release-104.html

Number concerning the next release: 104

Looks like we’ve not made much progress in the last month. Many bugs where closed but unfortunately many new bugs where opened as well. <a href="2007-01-30-number-concerning-the-next-release-116.html">Last time</a> I tried to motivate you to fix a few RC bugs, let’s try a different approach: *Ahem* I don’t have to remind you that… <img alt="" src="/images/debian_god_kills_a_kitten.jpg">First Candidate2007-02-06T14:57:00+01:002007-02-06T14:57:00+01:00Bastian Venthurtag:venthur.de,2007-02-06:/2007-02-06-first-candidate.html

First Candidate

<a href="https://sam.zoy.org/blog/2007-02-06-just-a-little-bit">Looks</a> like we found our first ~~victim~~ candidate for this year’s DPL election.RC Bug Triage2007-02-01T14:54:00+01:002007-02-01T14:54:00+01:00Bastian Venthurtag:venthur.de,2007-02-01:/2007-02-01-rc-bug-triage.html

RC Bug Triage

In the last three days I made 15 NMUs, fixing 11 RC- and 4 l10n-bugs. I noticed that some maintainers were not aware that one of their packages had an open RC bug from the mails I received afterwards. So please double check your QA package for open RC bugs and fix them if you can. If you have some spare time please go to the <a href="https://bts.turmzimmer.net/details.php">unofficial RC-bugs page</a> and check the bugs which are older than 7 days and affect etch and sid. We’re currently in a <a href="https://lists.debian.org/debian-devel-announce/2006/09/msg00020.html">permanent BSP</a> till etch is out, so 0-day NMUs are allowed. If you are an NM or no maintainer at all, you still can help by providing patches, submitting them to the BTS and <a href="https://www.debian.org/Bugs/server-control">tagging the bug</a> with “patch”. Someone will upload ASAP and might even thank you in the corresponding changelog entry.Random Firefox Tweaks2007-01-30T23:32:00+01:002007-01-30T23:32:00+01:00Bastian Venthurtag:venthur.de,2007-01-30:/2007-01-30-random-firefox-tweaks.html

Random Firefox Tweaks

The following options have to be changed or added in about:config. Those tweaks can of course be achieved through extensions, but by setting them directly you can abandon some extensions and hopefully conserve some memory. In order to open a bookmark-folder with mid-click in tabs without overwriting the current tabs change: <div class="codehilite"><pre><code>browser.tabs.loadFolderAndReplace=false </code></pre></div> To remove the close buttons on the tabs (they still can be closed by mid-clicking on them): <div class="codehilite"><pre><code>browser.tabs.closeButtons=2 </code></pre></div> Enable the spellchecker for inputfields and textareas (default is textareas only): <div class="codehilite"><pre><code>layout.spellcheckDefault=2 </code></pre></div> To override the useragent string: <div class="codehilite"><pre><code>general.useragent.override=foobrowser </code></pre></div> And the last one really made my day: Open lastfm://-links directly in amarok: <div class="codehilite"><pre><code>network.protocol-handler.app.lastfm=amarok network.protocol-handler.external.lastfm=true </code></pre></div> BTW: Shouldn’t firef*cough* iceweasel open the composer of your mailclient when clicking on mailto-links?Number concerning the next release: 1162007-01-30T11:55:00+01:002007-01-30T11:55:00+01:00Bastian Venthurtag:venthur.de,2007-01-30:/2007-01-30-number-concerning-the-next-release-116.html

Number concerning the next release: 116

Looks like our RC counter is heading into the wrong direction. Two years ago – when our RC counter was as high as now – it took approximately three months from there to the release. So… <img alt="" src="/images/unclesam_rc_bugs.png">Using My Favorite Mailclient When Reporting Bugs2007-01-27T22:12:00+01:002007-01-27T22:12:00+01:00Bastian Venthurtag:venthur.de,2007-01-27:/2007-01-27-using-my-favorite-mailclient-when-reporting-bugs.html

Using My Favorite Mailclient When Reporting Bugs

Dear Lazyweb, is there an option in reportbug allowing me to use Thunderbird (KMail, foo, …) to write the bugreport? I’m currently using vim, but Thunderbird’s spell checker is much more convenient to use and copy-pasting from KDE into vim is not working very well, making it cumbersome to attach backtraces and stuff.Pimp My Boot Process2007-01-25T12:38:00+01:002007-01-25T12:38:00+01:00Bastian Venthurtag:venthur.de,2007-01-25:/2007-01-25-pimp-my-boot-process.html

Pimp My Boot Process

I’ve tested Ubuntu the other day and I must say I was really impressed how polished the boot process looked. Ubuntu uses the same image as grub splash and as usplash image. This has the nice effect that there is no visible switch between grub and usplash. I don’t remember if they even used the same image for the login as well, but it would surely improve the boot experience even further. Since we have grub and usplash in our repository too, I thought it would be nice to create something similar for Debian. Creating an image and preparing it for grub is not that hard but <a href="https://help.ubuntu.com/community/USplashCustomizationHowto">creating something suitable for usplash</a> actually is a problem. First obstacle: usplash uses certain defined colors from the palette of the splash to draw stuff like text, background and progress bar, so you almost certainly have to re-arrange the position of the colors in the palette of the image manually. Second: installing the image is not exactly user-friendly. You have to create a .c-file from the .png. Compile it to a .so. and regenerate the initramfs in order to activate the theme… While second part is not that hard and can be done by a script, the first problem (re-arranging the color palette entries) seems kinda unsolvable for me and my limited Gimp skills. So does anyone know a tool which allows to create a usplash theme painlessly? I think it would be nice to have some grub/usplash theme packs in our repositories. While it’s certainly uninteresting for servers it surely adds some extra fun for our desktop users.Allow myself to introduce.......... myself!2007-01-23T20:36:00+01:002007-01-23T20:36:00+01:00Bastian Venthurtag:venthur.de,2007-01-23:/2007-01-23-allow-myself-to-introduce-myself.html

Allow myself to introduce.......... myself!

Ok then, this is my first serious blog entry. To break the ice, maybe an introduction would be appropriate: My name is Bastian Venthur I’m currently 26 years old and live in Berlin, Germany. I’m studying computer science at the <a href="https://www.inf.fu-berlin.de">Freie Universität Berlin</a> and I am Debian Developer since October 2006. I maintain a few <a href="https://qa.debian.org/developer.php?login=venthur">packages</a>, mostly KDE related and am quite active in <a href="https://bugs.debian.org/cgi-bin/pkgreport.cgi?submitter=venthur%40debian.org">reporting bugs</a>. One of my current hobbies is to juggle. I started a few months ago and already master some neat tricks like Mills Mess, Over the Head and the famous four ball Async Fountain.Hello World!2007-01-23T20:06:00+01:002007-01-23T20:06:00+01:00Bastian Venthurtag:venthur.de,2007-01-23:/2007-01-23-hello-world-2.html

Hello World!

First post.

Original Source | Taken Source