refactor(src): execution_report を廃止し write_result_to_comment（stdout → PR コメント）に置き換え#82
Merged
HappyOnigiri merged 9 commits intomainfrom Mar 13, 2026
Conversation
Contributor
|
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)
📝 WalkthroughWalkthrough
Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested labels
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches
🧪 Generate unit tests (beta)
📝 Coding Plan
Comment |
HappyOnigiri
added
the
refix: running
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 3
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/auto_fixer.py (1)
343-373:⚠️ Potential issue | 🟠 Major失敗した Claude 実行の出力が result log に残りません。
src/claude_runner.py:83-205は失敗時もstdout/stderrを例外に保持していますが、ここでは成功時にしかresult_blocksに追加していません。最初のフェーズで失敗するとresult_blocksが空のままなので、state comment に何も残らず、今回追加したwrite_result_to_commentの主目的を満たせません。少なくともClaudeCommandFailedError/ClaudeUsageLimitErrorのstdout/stderrからブロックを作ってからupsert_state_comment()する必要があります。Also applies to: 456-488, 573-595, 734-766
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/auto_fixer.py` around lines 343 - 373, The CI-fix exception handler currently doesn't persist failed Claude outputs into result_blocks, so when run_claude_prompt (used in the ci-fix phase) raises (e.g., ClaudeCommandFailedError / ClaudeUsageLimitError) the stdout/stderr carried by the exception are lost; update the except block around run_claude_prompt to detect if the caught exception exposes stdout/stderr, create a phase result block via build_phase_result_entry("ci-fix", stdout_or_stderr, ctx.state_comment_timezone) and append it to result_blocks (merging with load_state_comment as done now) before calling upsert_state_comment, and apply the same change to the equivalent handlers for the other phases (the similar try/except regions referenced at the other locations) so failed-run outputs are always saved to the state comment.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@src/auto_fixer.py`:
- Around line 1140-1156: The upsert_state_comment failure is currently being
treated as a success path because state_saved isn't reliably updated on error;
initialize and keep state_saved = False before the write attempt, set
state_saved = True only inside the successful upsert_state_comment try block,
and in the except block explicitly leave it False (or set to False) so
downstream cacheability/CI-only flow (the code that decides skip/complete for
CI-only PRs) reads the correct failure state; update any subsequent cacheability
check that references state_saved to use this flag so failed writes won't be
treated as saved.
In `@src/result_report.py`:
- Around line 30-38: The current static triple-backtick fence in
result_report.py breaks the report when stdout_text contains ```; update the
stdout rendering to compute a fence string that does not appear in the output
(e.g., start with "`" and append backticks while the fence is found in
stdout_text/stripped_stdout), then use that dynamic fence variable instead of
the fixed "```" when building the lines list (refer to variables/fields fence,
stdout_text, stripped_stdout and the list construction around lines.extend that
adds the "<summary>stdout</summary>" block).
In `@src/state_manager.py`:
- Around line 211-232: The truncation currently slices raw characters in
_truncate_result_log_body_to_fit which can leave unclosed headings, <details>
blocks, or code fences created by the generator in result_report.py and breaks
strip_result_log_section()/extract_result_log_body(); change the logic to parse
result_log_body into its logical "phase" sections using the same
delimiters/structure that result_report.py emits (e.g., the "####" phase
headings plus their following <details> and fenced code blocks), then
iteratively drop entire trailing phases until the remaining joined phases plus
the state comment scaffold (computed via _build_state_comment_body) fit within
max_length, and finally append truncation_notice; ensure the parser guarantees
balanced tags/fences so returned text never contains partial/unclosed tags or
fences and handle the case where no full phase can be kept by returning "".
---
Outside diff comments:
In `@src/auto_fixer.py`:
- Around line 343-373: The CI-fix exception handler currently doesn't persist
failed Claude outputs into result_blocks, so when run_claude_prompt (used in the
ci-fix phase) raises (e.g., ClaudeCommandFailedError / ClaudeUsageLimitError)
the stdout/stderr carried by the exception are lost; update the except block
around run_claude_prompt to detect if the caught exception exposes
stdout/stderr, create a phase result block via
build_phase_result_entry("ci-fix", stdout_or_stderr, ctx.state_comment_timezone)
and append it to result_blocks (merging with load_state_comment as done now)
before calling upsert_state_comment, and apply the same change to the equivalent
handlers for the other phases (the similar try/except regions referenced at the
other locations) so failed-run outputs are always saved to the state comment.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 855552f3-3791-48c2-b768-441d75aec516
📒 Files selected for processing (15)
.refix.yaml.sampleREADME.ja.mdREADME.mdsrc/auto_fixer.pysrc/claude_runner.pysrc/config.pysrc/report.pysrc/result_report.pysrc/state_manager.pytests/test_auto_fixer.pytests/test_claude_runner.pytests/test_config.pytests/test_report.pytests/test_result_report.pytests/test_state_manager.py
💤 Files with no reviewable changes (2)
- tests/test_report.py
- src/report.py
Owner
Author
🤖 Refix Status対応済みレビュー一覧
実行レポートCI 修正2026-03-13 11:00:00 UTC src/auto_fixer.py CI make ci diff の修正開始CIの失敗原因: 2026-03-13 11:05:00 UTC src/auto_fixer.py CI修正完了4箇所の行長超過（88文字超）をruffフォーマッタが期待する形式に整形してcommit & push完了。
|
Contributor
There was a problem hiding this comment.
♻️ Duplicate comments (1)
src/auto_fixer.py (1)
1167-1187:⚠️ Potential issue | 🟠 Major
_cacheable判定にstate_savedを含めてください。Line 1168-1180 で保存失敗時に
state_saved=Falseになるよう修正されていますが、Line 1229-1234 の_cacheableがstate_savedを見ていないため、結果ログ保存失敗でも cacheable 扱いになる可能性が残っています。修正案
- _cacheable = ( - not dry_run - and not bool(active_rate_limit) - and not bool(active_review_failed) - and not _ci_grace - ) + _cacheable = ( + not dry_run + and state_saved + and not bool(active_rate_limit) + and not bool(active_review_failed) + and not _ci_grace + )Also applies to: 1206-1214, 1229-1234
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/auto_fixer.py` around lines 1167 - 1187, The _cacheable determination must consider whether the state result log was actually saved; update the code paths that compute the _cacheable flag (the block that currently sets _cacheable around the later section and the similar blocks near where state_saved is set) to include state_saved (i.e., make _cacheable false when state_saved is False). Locate the state save logic that sets state_saved (the try/except using load_state_comment and upsert_state_comment) and the subsequent _cacheable computation, and change the boolean expression for _cacheable to AND with state_saved (or explicitly set _cacheable = False when state_saved is False) so failed result-log saves never produce a cacheable result; apply the same change to the other similar occurrence mentioned (the blocks around lines 1206-1214 and 1229-1234).
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Duplicate comments:
In `@src/auto_fixer.py`:
- Around line 1167-1187: The _cacheable determination must consider whether the
state result log was actually saved; update the code paths that compute the
_cacheable flag (the block that currently sets _cacheable around the later
section and the similar blocks near where state_saved is set) to include
state_saved (i.e., make _cacheable false when state_saved is False). Locate the
state save logic that sets state_saved (the try/except using load_state_comment
and upsert_state_comment) and the subsequent _cacheable computation, and change
the boolean expression for _cacheable to AND with state_saved (or explicitly set
_cacheable = False when state_saved is False) so failed result-log saves never
produce a cacheable result; apply the same change to the other similar
occurrence mentioned (the blocks around lines 1206-1214 and 1229-1234).
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: ac6c1902-feff-4914-996f-940c89ec4f13
📒 Files selected for processing (1)
src/auto_fixer.py
…lse ブロック） 結果ログ保存失敗時（state_saved=False）でも cacheable 扱いになる問題を修正。
HappyOnigiri
added
refix: done
HappyOnigiri
deleted the
refactor/replace-execution-report-with-write-result-to-comment
branch
HappyOnigiri
added
refix: auto-merge-requested
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.
概要 / Summary
execution_report（Claude によるファイル書き込みレポート）を廃止し、Claude の stdout を直接 PR ステータスコメントに書き出すwrite_result_to_comment機能に置き換える。変更内容 / Changes
src/report.pyを削除し、src/result_report.pyを新規作成（format_phase_result_block,merge_result_log_body,build_phase_result_entry）src/config.py:execution_report: false（デフォルト）→write_result_to_comment: true（デフォルト）に差し替えsrc/claude_runner.py:report_path/report_enabledパラメータを削除、戻り値をstr→tuple[str, str]（commits, stdout）に変更src/state_manager.py:report_body/REPORT_SECTION_*→result_log_body/RESULT_LOG_SECTION_*に置き換え、折りたたみ見出しを「実行レポート」→「実行ログ」に変更src/auto_fixer.py:PRContextからexecution_report_enabled/reports_dirを削除しwrite_result_to_commentを追加、各フェーズで stdout をresult_blocksに蓄積してupsert_state_comment(result_log_body=...)で保存、レビュー修正フェーズのみ対象コメント URL を付与.refix.yaml.sample/README.md/README.ja.md: 設定キーの説明を更新テスト / Testing
make ci通過済み（lint + 全 206 テスト）関連 issues / Related issues
なし
Summary by CodeRabbit
新機能
ドキュメント