git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH v3] t3310: avoid hiding failures from rev-parse in command substitutions

From
Francesco Paparatto <francescopaparatto@gmail.com>
Date
Mar 7, 2026, 10:17 UTC
Message-ID
<CAEaT9_9_F_kcPbTrisX_At6RANJ9MHCEGka6M=WRezTO+-3A-g@mail.gmail.com>
In-Reply-To
<CAPig+cQWCK48GJEnGX7bP6exu847WR8HU3Y8sna525w6NEhmmw@mail.gmail.com>
Eric Sunshine <sunshine@sunshineco.com> writes:
Show 41 quoted lines
> > diff --git a/t/t3310-notes-merge-manual-resolve.sh b/t/t3310-notes-merge-manual-resolve.sh
> > @@ -569,13 +578,15 @@ EOF
> >         test_grep -q "refs/notes/m" output &&
> > -       test_grep -q "$(git rev-parse refs/notes/m)" output &&
> > -       test_grep -q "$(git rev-parse NOTES_MERGE_PARTIAL^1)" output &&
> > +       git rev-parse refs/notes/m >actual &&
> > +       test_grep -q "$(cat actual)" output &&
> > +       git rev-parse NOTES_MERGE_PARTIAL^1 >actual &&
> > +       test_grep -q "$(cat actual)" output &&
>
> Storing the output of git-rev-parse in a file only to read it back out
> of that file a moment later is unnecessarily roundabout. It would
> instead be cleaner to do it this way:
>
>     oid=$(git rev-parse refs/notes/m) &&
>     test_grep -q "$oid" output &&
>     oid=$(git rev-parse NOTES_MERGE_PARTIAL^1) &&
>     test_grep -q "$oid" output &&
>
> Unlike this original in which git-rev-parse's exit code was lost due
> to being embedded in the test_grep invocation, this rewrite is safe
> because the exit code of git-rev-parse becomes the exit code of the
> variable assignment, thus correctly aborts the test (due to the
> &&-chain) if git-rev-parse fails.
>
> > @@ -606,8 +617,9 @@ test_expect_success 'switch cwd before committing notes merge' '
> >         test_must_fail git notes merge refs/notes/other &&
> >         (
> >                 cd .git/NOTES_MERGE_WORKTREE &&
> > -               echo "foo" > $(git rev-parse HEAD) &&
> > -               echo "bar" >> $(git rev-parse HEAD) &&
> > +               oid=$(git rev-parse HEAD) &&
> > +               echo "foo" >"$oid" &&
> > +               echo "bar" >>"$oid" &&
>
> This is purely subjective and you don't have to take the suggestion,
> but although yours is a faithful rewrite (which is good), I probably
> would have simplified this to:
>
>     oid=$(git rev-parse HEAD) &&
>     test_write_lines foo bar >"$oid" &&

Thanks for the review. Both suggestions make sense. I'll use the variable assignment for the rev-parse cases and test_write_lines for the foo/bar case. I will send v4 shortly.

Previous: Eric SunshineNext: Francesco Paparatto
Message 12 of 15 in “t3310: avoid hiding failures from rev-parse in command substitutions”
  1. t3310: avoid hiding failures from rev-parse in command substitutionsFrancesco Paparatto, Mar 4, 2026
  2. Eric SunshineMar 4, 2026
  3. t3310: avoid hiding failures from rev-parse in command substitutionsFrancesco Paparatto, Mar 5, 2026
  4. Junio C HamanoMar 5, 2026
  5. Eric SunshineMar 5, 2026
  6. Francesco PaparattoMar 5, 2026
  7. Junio C HamanoMar 5, 2026
  8. Junio C HamanoMar 5, 2026
  9. Francesco PaparattoMar 5, 2026
  10. t3310: avoid hiding failures from rev-parse in command substitutionsFrancesco Paparatto, Mar 5, 2026
  11. Eric SunshineMar 7, 2026
  12. Francesco PaparattoMar 7, 2026
  13. t3310: avoid hiding failures from rev-parse in command substitutionsFrancesco Paparatto, Mar 7, 2026
  14. Eric SunshineMar 8, 2026
  15. Junio C HamanoMar 8, 2026

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.