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

Re: [PATCH 4/6] test-lib: simplify leak-log checking

From
Patrick Steinhardt <ps@pks.im>
Date
Jan 6, 2025, 07:56 UTC
Message-ID
<Z3uMyQ-YfQFI8qmH@pks.im>
In-Reply-To
<20250103202410.GC3212696@coredump.intra.peff.net>
On Fri, Jan 03, 2025 at 03:24:10PM -0500, Jeff King wrote:
Show 29 quoted lines
> On Fri, Jan 03, 2025 at 01:05:45PM +0100, Patrick Steinhardt wrote:
> 
> > On Wed, Jan 01, 2025 at 03:17:21PM -0500, Jeff King wrote:
> > > @@ -1181,8 +1170,14 @@ test_atexit_handler () {
> > >  }
> > >  
> > >  check_test_results_san_file_empty_ () {
> > > -	test -z "$TEST_RESULTS_SAN_FILE" ||
> > > -	test "$(nr_san_dir_leaks_)" = 0
> > > +	test -z "$TEST_RESULTS_SAN_FILE" && return 0
> > > +
> > > +	# stderr piped to /dev/null because the directory may have
> > > +	# been "rmdir"'d already.
> > > +	! find "$TEST_RESULTS_SAN_DIR" \
> > > +		-type f \
> > > +		-name "$TEST_RESULTS_SAN_FILE_PFX.*" 2>/dev/null |
> > > +	xargs grep -qv "Unable to get registers from thread"
> > 
> > Can't we use `-exec grep -qv "Unable to get registers from thread" {}
> > \+` instead of using xargs? Or is that unportable? Might make it a bit
> > easier to reason about the `!` in the presence of a pipe.
> 
> I don't think that saves us from negating, though. The "grep" will tell
> us if it matched any "real" lines, but we want to report that we found
> no real lines.
> 
> Plus I don't think "find" propagates the exit code from -exec anyway. I
> think you can check the exit status with more find logic, so you'd then
> use a conditional -print for each file like:
It should. Quoting find(1):
    If any invocation with the `+' form returns a non-zero value as exit
    status, then find returns a non-zero exit status.
Show 6 quoted lines
>   find ... \
>     -exec grep -qv "Unable to get registers from thread" \{} \; \
>     -print
> 
> and you have to check whether the output is empty. The easiest way to do
> that is with another grep! Which also needs negated. ;)

Yup, I didn't mean to say that we can drop the negation, but that it makes it easier to reason about what the negation applies to (the whole pipe or just the find(1) command)).

Show 17 quoted lines
> I think if we really want to drop the negation, we'd be best to flip the
> function's return, like:
> 
>   have_leaks() {
> 	# not leak-checking
> 	test -z "$TEST_RESULTS_SAN_FILE" && return 1
> 
> 	find "$TEST_RESULTS_SAN_DIR" \
> 		-type f \
> 		-name "$TEST_RESULTS_SAN_FILE_PFX.*" 2>/dev/null |
> 	xargs grep ^DEDUP_TOKEN |
> 	grep -qv sanitizer::GetThreadStackTopAndBottom
>   }
> 
> And then you could switch the initial "grep" to -exec if you want, but
> there's no negation to get rid of, so it is only a preference of -exec
> versus xargs.
Yup.
Patrick
Previous: Jeff KingNext: Jeff King
Message 14 of 37 in “What's cooking in git.git (Dec 2024, #11; Mon, 30)”
  1. Junio C HamanoDec 30, 2024
  2. René ScharfeDec 31, 2024
  3. Patrick SteinhardtJan 3, 2025
  4. a less-invasive racy-leak fix, was Re: What's cooking in git.git (Dec 2024, #11; Mon, 30)Jeff King, Jan 1, 2025
  5. 0/6 a less-invasive racy-leak fixJeff King, Jan 1, 2025
  6. 1/6 test-lib: use individual lsan dir for --stress runsJeff King, Jan 1, 2025
  7. 2/6 Revert "index-pack: spawn threads atomically"Jeff King, Jan 1, 2025
  8. 3/6 test-lib: rely on logs to detect leaksJeff King, Jan 1, 2025
  9. Patrick SteinhardtJan 3, 2025
  10. Jeff KingJan 3, 2025
  11. 4/6 test-lib: simplify leak-log checkingJeff King, Jan 1, 2025
  12. Patrick SteinhardtJan 3, 2025
  13. Jeff KingJan 3, 2025
  14. Patrick SteinhardtJan 6, 2025
  15. Jeff KingJan 7, 2025
  16. 5/6 test-lib: check leak logs for presence of DEDUP_TOKENJeff King, Jan 1, 2025
  17. 6/6 test-lib: ignore leaks in the sanitizer's thread codeJeff King, Jan 1, 2025
  18. Patrick SteinhardtJan 3, 2025
  19. Jeff KingJan 3, 2025
  20. Patrick SteinhardtJan 6, 2025
  21. 0/3 lsan test-lib readabilityJeff King, Jan 7, 2025
  22. 1/3 test-lib: invert return value of check_test_results_san_file_emptyJeff King, Jan 7, 2025
  23. 2/3 test-lib: simplify lsan results checkJeff King, Jan 7, 2025
  24. Patrick SteinhardtJan 7, 2025
  25. Jeff KingJan 9, 2025
  26. Patrick SteinhardtJan 9, 2025
  27. Junio C HamanoJan 7, 2025
  28. Jeff KingJan 9, 2025
  29. 3/3 test-lib: add a few comments to LSan log checkingJeff King, Jan 7, 2025
  30. Patrick SteinhardtJan 7, 2025
  31. Junio C HamanoJan 2, 2025
  32. Jeff KingJan 2, 2025
  33. Chris TorekJan 2, 2025
  34. Junio C HamanoJan 2, 2025
  35. Jeff KingJan 2, 2025
  36. Junio C HamanoJan 2, 2025
  37. Jeff KingJan 2, 2025

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.