Repository navigation
refactor(tests): annotation syntax with non-ASCII sigils - #537
Merged
Merged
Conversation
The old grammar reserved $ and @, which collide with LSP snippet
placeholders (${1:), $/cancelRequest, ${workspace} config vars and
Doxygen tags (@PARAM[in] matches @key[...] and asserts on bare @word).
The new grammar reserves no ASCII character at all:
§ nameless point §(name) named point
§⟦...⟧ nameless range §(name)⟦...⟧ named range
Ranges nest via a stack (the bracket-balance heuristic is gone, so
unbalanced brackets inside a range are now expressible), stray sigils
fail loudly, and §(name) requires an identifier so a nameless point
directly before real parentheses stays unambiguous (§()).
Also adds test/snap_region.h: whole-line /// <snap:begin>/<snap:end>
markers plus a containment filter, for snapshot fixtures that only want
a marked region's results.
All ~750 annotations across 40 files migrated mechanically; a semantic
differ verified byte-identical stripped sources and annotation tables
against the old parser (four sites in selection_tests relied on the old
$() quirk and were fixed, one signature-help point was recovered).
Review found the assert-only guards let NDEBUG builds wrap i past npos on an unterminated §( (infinite loop) and pop an empty stack on a stray ⟧ (UB). Malformed annotations now LOG_FATAL unconditionally. Also adds the review-suggested coverage: digit-only names, § at EOF, empty range body, end marker without trailing newline, empty snap region.
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughWalkthroughThe annotation framework now uses ChangesAnnotation grammar and test infrastructure
Estimated code review effort: 4 (Complex) | ~45 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7863e7b1ce
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
tests/unit/feature/document_link_tests.cpp (1)
47-52: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueRemove stray nameless point markers.
The trailing
§markers placed just before the range closing markers (⟧) define unused nameless points. Sincenameless_points()is not utilized in these test cases, they are functionally harmless but likely unintentional leftovers from the marker syntax migration.
tests/unit/feature/document_link_tests.cpp#L47-L52: remove the trailing§from the include arguments (e.g.,⟦"test.h"§⟧->⟦"test.h"⟧).tests/unit/feature/document_link_tests.cpp#L89-L89: remove the trailing§from⟦HEADER§⟧.tests/unit/feature/document_link_tests.cpp#L103-L103: remove the trailing§from⟦"bytes.bin"§⟧.tests/unit/feature/document_link_tests.cpp#L118-L118: remove the trailing§from⟦"data.bin"§⟧.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@tests/unit/feature/document_link_tests.cpp` around lines 47 - 52, Remove the unused trailing nameless-point markers from the marker ranges in tests/unit/feature/document_link_tests.cpp: lines 47-52 include arguments, line 89 HEADER, line 103 "bytes.bin", and line 118 "data.bin"; preserve the surrounding marker syntax and test content.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@tests/unit/test/annotation.cpp`:
- Around line 81-83: Update the point-annotation insertion at
tests/unit/test/annotation.cpp lines 81-83 to assert that insertion succeeds
instead of silently ignoring duplicate keys; update the range-annotation
insertion at lines 91-94 to assert success and explicitly reject a second
nameless range keyed by the empty string, preserving the fail-loudly behavior
for duplicate annotation keys.
---
Nitpick comments:
In `@tests/unit/feature/document_link_tests.cpp`:
- Around line 47-52: Remove the unused trailing nameless-point markers from the
marker ranges in tests/unit/feature/document_link_tests.cpp: lines 47-52 include
arguments, line 89 HEADER, line 103 "bytes.bin", and line 118 "data.bin";
preserve the surrounding marker syntax and test content.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 444f980c-8a0b-413b-ad77-45613a8b3e98
📒 Files selected for processing (44)
tests/data/hover/.clang-formattests/data/hover/attributes.cpptests/data/hover/auto.cpptests/data/hover/basics.cpptests/data/hover/callee_args.cpptests/data/hover/concepts.cpptests/data/hover/decltype.cpptests/data/hover/docs.cpptests/data/hover/expressions.cpptests/data/hover/fields.cpptests/data/hover/functions.cpptests/data/hover/getter_setter.cpptests/data/hover/lambdas.cpptests/data/hover/misc.cpptests/data/hover/no_hover.cpptests/data/hover/no_hover_errors.cpptests/data/hover/pass_types.cpptests/data/hover/spaceship.cpptests/data/hover/tag_decls.cpptests/data/hover/template_params.cpptests/data/hover/templates.cpptests/data/hover/this_expr.cpptests/data/hover/using_decls.cpptests/data/hover/values.cpptests/data/hover/variables.cpptests/unit/compile/directive_tests.cpptests/unit/feature/code_completion_tests.cpptests/unit/feature/document_link_tests.cpptests/unit/feature/folding_range_tests.cpptests/unit/feature/hover_tests.cpptests/unit/feature/inlay_hint_tests.cpptests/unit/feature/semantic_tokens_tests.cpptests/unit/feature/signature_help_tests.cpptests/unit/index/index_query_tests.cpptests/unit/index/merged_index_tests.cpptests/unit/index/preamble_state_tests.cpptests/unit/index/tu_index_tests.cpptests/unit/index/usr_tests.cpptests/unit/semantic/selection_tests.cpptests/unit/server/query_freshness_tests.cpptests/unit/server/query_overlay_tests.cpptests/unit/test/annotation.cpptests/unit/test/annotation_tests.cpptests/unit/test/snap_region.h
try_emplace silently kept the first binding, so a duplicate §(name) or a second nameless range dropped an annotation without a trace.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b46a2f7112
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
The test annotation grammar reserved
$and@, both of which collide with real content that tests need to contain:@param[in]matches the@key[...]annotation shape and destroys the input; a bare Doxygen tag like@briefaborts assertion-enabled builds. Net effect: no test source could contain Doxygen comments — the hover documentation fixtures avoid them entirely today.$collides with LSP snippet placeholders (${1:...}),$/cancelRequest, and${workspace}config variables.[...]range delimiter collides with C++ brackets: the balance heuristic cannot express a range containing unbalanced brackets (lambda captures, partial subscripts).New grammar
The grammar now reserves no ASCII character at all — sigil
§(U+00A7), range delimiters⟦/⟧(U+27E6/E7):§(name)requires an identifier name, keeping a nameless point directly before real parentheses unambiguous (§()).nposand loop forever).Also introduces
test/snap_region.h: whole-line/// <snap:begin [name]>//// <snap:end [name]>markers plus a containment filter, for snapshot fixtures that only want results from a marked region.Migration
~750 annotations across 40 files converted mechanically. A semantic differ replayed the old parser against the old files and the new parser against the new files and verified byte-identical stripped sources and annotation tables. Three findings from that process, all fixed:
$()meant point-plus-literal-parens; they now use the explicit§()()form.$during auditing.Tests