fix: Added caching of the file view provider's one file. (#935)

SCWells72 · web-flow · commit 91b7b9a1cb8d · 2025-03-28T00:09:44.000+01:00
* Added caching of the view provider's one file. * Okay, let's try this approach as recommended by @FalsePattern. * Fixed a rather subtle but insidious issue where semantic tokens-based elements were being indexed for their end offsets inclusive when they should have been exclusive. Put differently, for a file that uses the semantic tokens file view provider, file.getElementAt(offset) should return an element for [startOffset, endOffset) where those are the start/end offsets from element.getTextRange(). Without this change, it was [startOffset, endOffset], and that was causing an infinite loop in XDebuggerUtilImpl which was using endOffset as the new offset when searching for the next element. I've also removed the file cache that was added to try to resolve this issue before it was understood.
diff --git a/src/main/java/com/redhat/devtools/lsp4ij/features/semanticTokens/viewProvider/LSPSemanticTokensFileViewProviderHelper.java b/src/main/java/com/redhat/devtools/lsp4ij/features/semanticTokens/viewProvider/LSPSemanticTokensFileViewProviderHelper.java
@@ -169,8 +169,8 @@ public void addSemanticToken(@NotNull TextRange textRange,
         if (semanticTokensByOffset != null) {
             LSPSemanticToken semanticToken = new LSPSemanticToken(file, textRange, tokenType, tokenModifiers);
 
-            // Index the token for its text range
-            for (int offset = textRange.getStartOffset(); offset <= textRange.getEndOffset(); offset++) {
+            // Index the token for its text range up to but not including the end offset
+            for (int offset = textRange.getStartOffset(); offset < textRange.getEndOffset(); offset++) {
                 semanticTokensByOffset.put(offset, semanticToken);
             }
         }

Original file line number	Diff line number	Diff line change
`@@ -169,8 +169,8 @@ public void addSemanticToken(@NotNull TextRange textRange,`
`169`	`169`	`if (semanticTokensByOffset != null) {`
`170`	`170`	`LSPSemanticToken semanticToken = new LSPSemanticToken(file, textRange, tokenType, tokenModifiers);`
`171`	`171`
`172`		`- // Index the token for its text range`
`173`		`- for (int offset = textRange.getStartOffset(); offset <= textRange.getEndOffset(); offset++) {`
	`172`	`+ // Index the token for its text range up to but not including the end offset`
	`173`	`+ for (int offset = textRange.getStartOffset(); offset < textRange.getEndOffset(); offset++) {`
`174`	`174`	`semanticTokensByOffset.put(offset, semanticToken);`
`175`	`175`	`}`
`176`	`176`	`}`