0.7.1 (2025-04-07)

• fix: resolves issue where SDK could stall when MaxParallelRequests * MaxBatchSize < body.Checks length
• fix: resolves issue where private singleBatchCheck method in SdkClient prevented mocking

0.7.0 (2025-04-02)

• feat: fix and improve retries and rate limit handling. (#176)
  The SDK now retries on network errors and the default retry handling has been fixed
  for both the calls to the OpenFGA API and the API Token Issuer for those using ClientCredentials
  The SDK now also respects the rate limit headers (Retry-After) returned by the server and will retry the request after the specified time.
  If the header is not sent or on network errors, it will fall back to exponential backoff.
• feat: retry on network errors when calling the token issuer (#182)
• feat: add support for server-side BatchCheck (#187)
• fix: use defaults when transaction options were only partially set (#183)
• chore: log retry attempts when debug mode is enabled (#182)

NOTE:
This release was previously released as v0.6.6, but has been re-released as v0.7.0 due to the breaking changes.

0.6.6 (2025-04-02)

0.6.5 (2025-02-06)

• feat: support assertions with context and contextual tuples (#169)
• fix: populate start_time in ReadChanges request (#166)

0.6.4 (2025-01-29)

• feat: add support for start_time parameter in ReadChanges endpoint (#158)
• fix: correctly set request level storeId in non-transactional write (#162) - thanks @preiter93
• fix: api client should set default telemetry if not specified (#160)
• docs: replace readable names with uuid (#146) - thanks @sccalabr
• fix: support marshaling client.ClientWriteResponse (#145) - thanks @FedotCompot
• fix: update client interface with store and model getter/setter (#155)
• fix: api client should set default telemetry if not specified (#160)

0.6.3 (2024-10-22)

• fix: fix metrics data race issues (#139)

0.6.2 (2024-10-21)

• fix: fix batch check consistency (#131)
• fix: fix data race on TelemetryInstances (#136) - thanks @Kryvchun!

NOTE: TelemetryInstances in telemetry.go has been deprecated, as its usage is intended to be internal. It will be removed in a future release.

0.6.1 (2024-09-23)

• refactor(OpenTelemetry): move configuration API into public package namespace (#122)
• docs(OpenTelemetry): initial documentation and example (#123)

0.6.0 (2024-08-29)

• feat: support OpenTelemetry metrics reporting (#115)
• feat!: support for sending the consistency parameter to the read, check, list users, list objects, and expand endpoints (#117)
• chore(docs): update stale README (#113) - thanks @Code2Life

BREAKING CHANGE:

When the generator converts enums in the open API definition, by default it removes the type prefix. For example, TYPE_NAME_UNSPECIFIED is converted to a const named UNSPECIFIED. This leads to potential collisions with other enums, and as the consistency type is a new enum, we finally got a collision (it was just a matter of time).

The fix for this is to specify "enumClassPrefix": true in the generation config. This will then include the class name on the const name, which resolves collision issues. This means any enum value, such as INT now becomes TYPENAME_INT. The main impact of this is the TypeName const and error codes. The fix is to add the class name prefix, as discussed above.

0.5.0 (2024-06-14)

• fix: correctly set HTTPClient - thanks @wonyx
• chore!: remove excluded users from ListUsers response

⚠️ BREAKING CHANGE ⚠️

This version removes the ExcludedUsers field from the ListUsersResponse and ClientListUsersResponse structs, for more details see the associated API change.