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

Re: [PATCH 01/11] test-lib: catch misspelt 'test_expect_successo'

From
Jeff King <peff@peff.net>
Date
Mar 26, 2026, 04:08 UTC
Message-ID
<20260326040828.GA686242@coredump.intra.peff.net>
In-Reply-To
<20260325062114.2067946-2-gitster@pobox.com>
On Tue, Mar 24, 2026 at 11:21:04PM -0700, Junio C Hamano wrote:
>  # Test the binaries we have just built.  The tests are kept in
>  # t/ subdirectory and are run in 'trash directory' subdirectory.
> +
> +set -e
This causes failures in t0005 and t3600 with dash, but not bash.

It looks like the suppression of "-e" on the left-hand-side of an && is different when there is command substitution in play:

  $ dash -c 'OUT=$( ((yes; echo $? 1>&3) | :) 3>&1) && echo out=$OUT'
  out=141
  $ dash -ec 'OUT=$( ((yes; echo $? 1>&3) | :) 3>&1) && echo out=$OUT'
  out=
whereas with bash, both produce 141.

The idea is that $OUT becomes the exit status of "yes" here, and we are expecting to see SIGPIPE. With "-e" in effect, the failing "yes" will terminate before we echo $?.

To demonstrate the effect as we build it up from smaller pieces:
  # produces 141, SIGPIPE from yes
  dash -c '((yes; echo $? 1>&3) | :) 3>&1'
  # produces nothing, "-e" kills subshell after yes fails
  dash -ec '((yes; echo $? 1>&3) | :) 3>&1'
  # produces 141 (and "ok"), as the && suppresses -e
  dash -ec '((yes; echo $? 1>&3) | :) 3>&1 && echo ok'
  # produces "out="; the $() makes us forget that we're on LHS of &&
  dash -ec 'OUT=$( ((yes; echo $? 1>&3) | :) 3>&1) && echo out=$OUT'
The actual failing code in t0005 is:
  OUT=$( ((large_git; echo $? 1>&3) | :) 3>&1 ) &&
  test_match_signal 13 "$OUT"
and the one in t3600 is similar. I guess you could do:
diff --git a/t/t0005-signals.sh b/t/t0005-signals.sh
index afba0fc3fc..0bf1f16750 100755
--- a/t/t0005-signals.sh
+++ b/t/t0005-signals.sh
@@ -42,7 +42,7 @@ test_expect_success 'create blob' '
 '
 
 test_expect_success !MINGW 'a constipated git dies with SIGPIPE' '
-	OUT=$( ((large_git; echo $? 1>&3) | :) 3>&1 ) &&
+	OUT=$( ((large_git || echo $? 1>&3) | :) 3>&1 ) &&
 	test_match_signal 13 "$OUT"
 '
 

That neglects to echo $? when large_git surprisingly succeeds, but that
would mean $OUT is empty, which would cause the test to (correctly)
fail. I kind of hate it, though.

-Peff
Previous: Junio C HamanoNext: Junio C Hamano
Message 3 of 28 in “detect misspelt test_expect_success and friends”
  1. 00/11 detect misspelt test_expect_success and friendsJunio C Hamano, Mar 25, 2026
  2. 01/11 test-lib: catch misspelt 'test_expect_successo'Junio C Hamano, Mar 25, 2026
  3. Jeff KingMar 26, 2026
  4. Junio C HamanoMar 26, 2026
  5. Jeff KingMar 26, 2026
  6. Patrick SteinhardtMar 27, 2026
  7. Junio C HamanoMar 27, 2026
  8. Patrick SteinhardtMar 31, 2026
  9. 02/11 t0008: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  10. 03/11 t6002: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  11. Patrick SteinhardtMar 25, 2026
  12. Junio C HamanoMar 25, 2026
  13. 04/11 t4032: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  14. Patrick SteinhardtMar 25, 2026
  15. 05/11 t7450: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  16. Patrick SteinhardtMar 25, 2026
  17. 06/11 tests: make svn test "set -e" cleanJunio C Hamano, Mar 25, 2026
  18. Patrick SteinhardtMar 25, 2026
  19. 07/11 t7508: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  20. 08/11 t9200: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  21. Patrick SteinhardtMar 25, 2026
  22. 09/11 t940?: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  23. 10/11 t5570: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  24. Patrick SteinhardtMar 25, 2026
  25. Junio C HamanoMar 25, 2026
  26. 11/11 t9902: make test "set -e" cleanJunio C Hamano, Mar 25, 2026
  27. Patrick SteinhardtMar 25, 2026
  28. Junio C HamanoMar 25, 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.