From: Phillip Wood Date: Tue, 06 Oct 2026 09:52:02 GMT Subject: Re: [PATCH v6 0/2] ci: link failure and leak annotations to the test script Message-ID: <2554cf1e-9612-4dbf-958f-f9ff7ea71dc7@gmail.com> In-Reply-To: Hi Harald On 06/10/2026 07:56, Harald Nordgren via GitGitGadget wrote: > Link failure and leak annotations in CI to the test script, so both can be > found from the job summary. > > V7 CI job where failures and leaks are reported: > > * https://github.com/git/git/actions/runs/37205086805?pr=2426 > > Changes in v7: > > * Revert the linking in annotation message, direct direct line fails when > error pointed to an unchanged file in that PR, which regresses the > experiences in many cases. Instead now show the file and line as plain > text in the annotation. > * Don't name the the line number as 1 for leaks, where it's never > available, just omit the line number. This has been sent as v6, but the range-diff looks good and matches the changes listed here. Lets get this merged! Thanks Phillip > Changes in v6: > > * Update commit message. > > Changes in v5: > > * Removed % escaping entirely, verified on CI that it isn't needed. Every > existing test description that uses % renders correctly unescaped. > * Rewrote the file/line commit message with a concrete example (failed: > t1060.17 partial clone of corrupted repository). > > Changes in v4: > > * Clarify commit messages and simplify escaping logic. > > Changes in v3: > > * Fixed bug in the --immediate exit ordering: the --immediate && > --invert-exit-code path called exit 0 before the test's annotation was > written, now a single unconditional call covers both exit paths. > * github_escape_message_ no longer relies on \r being a portable sed escape > sequence (not POSIX-guaranteed and BSD sed implementations can differ), > it splices in the literal carriage-return byte via printf instead. > * Reverted unrelated test-tool line back to its original form. > > Changes in v2: > > * Split into two commits, each explaining its own reasoning. > * Leak output is no longer capped or embedded in the message, it's now an > uncapped fold, so multiple leaks in the same test both show in full. A > second leak in a different test still won't show in the same run, > --immediate stops the script at the first failure, but it no longer gets > buried under every later test falsely reporting "not ok" either. > * Drops the giant unfolded message that annotations used to carry, which is > what probably caused the scrolling behavior. > > Harald Nordgren (2): > ci: annotate leaks and stop a leak-sanitizer script at its first > failure > ci: point test failures and fixed known breakages at their file and > line > > ci/lib.sh | 1 + > t/test-lib-github-workflow-markup.sh | 39 +++++++++++++++++++++++----- > t/test-lib.sh | 6 ++++- > 3 files changed, 39 insertions(+), 7 deletions(-) > > > base-commit: 8103b446517e0c44e67561b9d0ccce56efa60a71 > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-2419%2FHaraldNordgren%2Fci-annotation-file-line-v6 > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-2419/HaraldNordgren/ci-annotation-file-line-v6 > Pull-Request: https://github.com/git/git/pull/2419 > > Range-diff vs v5: > > 1: 851efeec8b ! 1: 917f373f91 ci: annotate leaks and stop a leak-sanitizer script at its first failure > @@ Commit message > > Process completed with exit code 1. > > - Give a leak its own annotation. Point it at the test script, the exact > - line isn't known, only which script the leak turned up in, and put the > - sanitizer report in a log group next to it, so it stays visible. > + Give a leak its own annotation, naming the script it turned up in, the > + exact line isn't known, only which script: > + > + memory leak logged in t1060 (t1060-object-corruption.sh) > + > + Put the sanitizer report in a log group next to it, so it stays > + visible. > > Once a script has one leak, it keeps running: the sanitizer log > directory is never cleared between tests, so every later test in the > @@ t/test-lib-github-workflow-markup.sh: start_test_output () { > >$github_markup_output > GIT_TEST_TEE_OFFSET=0 > + github_markup_script_name=${0##*/} > -+} > -+ > -+github_annotation_ () { > -+ echo >>$github_markup_output "::$1 file=$2,line=$3::$4" > } > > # No need to override start_test_case_output > @@ t/test-lib-github-workflow-markup.sh: finalize_test_case_output () { > } > > +finalize_test_leak_output () { > -+ # The exact line the leak turned up on isn't known, only the script, > -+ # so point at line 1. > -+ github_annotation_ error "t/$github_markup_script_name" 1 \ > -+ "memory leak logged in $this_test" > ++ echo >>$github_markup_output \ > ++ "::error::memory leak logged in $this_test ($github_markup_script_name)" > + > + echo >>$github_markup_output "::group::leak: $this_test.$test_count" > + cat "$TEST_RESULTS_SAN_FILE".* >>$github_markup_output > 2: 46f93a9e16 ! 2: acf1fbd250 ci: point test failures and fixed known breakages at their file and line > @@ Metadata > ## Commit message ## > ci: point test failures and fixed known breakages at their file and line > > - When a test fails, GitHub shows an annotation naming it, for example: > + A failing test gets an annotation in the Annotations list on its job's > + summary page, naming it, for example: > > failed: t1060.17 partial clone of corrupted repository > > - but the location GitHub attaches to that annotation is the CI > - workflow file itself, not the test script, so there is nothing > - pointing at where the test actually lives. > + with no indication of where that test lives. > > Find the line a test is defined on by searching its script for the > test's own description as a fixed string, using the first match, and > - attach that file and line to the annotation instead. Fall back to > - line 1 when the description is not found verbatim, which happens when > - a test builds its description at runtime instead of writing it out > - literally. > + add the file and line to the annotation's own message text: > + > + failed: t1060.17 partial clone of corrupted repository (t1060-object-corruption.sh:141) > + > + Fall back to naming just the script, with no line, when the > + description is not found verbatim, which happens when a test builds > + its description at runtime instead of writing it out literally. > > Signed-off-by: Harald Nordgren > > @@ t/test-lib-github-workflow-markup.sh: start_test_output () { > + head -n 1 | cut -d: -f1 > +} > + > - github_annotation_ () { > - echo >>$github_markup_output "::$1 file=$2,line=$3::$4" > - } > -@@ t/test-lib-github-workflow-markup.sh: github_annotation_ () { > + # No need to override start_test_case_output > + > finalize_test_case_output () { > test_case_result=$1 > shift > @@ t/test-lib-github-workflow-markup.sh: github_annotation_ () { > + esac > + > + test_case_line=$(find_test_case_line_ "$1") > ++ test_case_where="$github_markup_script_name${test_case_line:+:$test_case_line}" > + > case "$test_case_result" in > failure) > - echo >>$github_markup_output "::error::failed: $this_test.$test_count $1" > -+ github_annotation_ error "t/$github_markup_script_name" "${test_case_line:-1}" \ > -+ "failed: $this_test.$test_count $1" > ++ echo >>$github_markup_output "::error::failed: $this_test.$test_count $1 ($test_case_where)" > ;; > fixed) > - echo >>$github_markup_output "::notice::fixed: $this_test.$test_count $1" > @@ t/test-lib-github-workflow-markup.sh: github_annotation_ () { > - ok|broken) > - # Exit without printing the "ok" or ""broken" tests > - return > -+ github_annotation_ notice "t/$github_markup_script_name" "${test_case_line:-1}" \ > -+ "fixed: $this_test.$test_count $1" > ++ echo >>$github_markup_output "::notice::fixed: $this_test.$test_count $1 ($test_case_where)" > ;; > esac > + >