Re: [PATCH 1/3] leak tests: mark a handful of tests as leak-free

From: Jeff King <peff@peff.net>
Date: Aug 24, 2023, 21:02 UTC
Message-ID: <20230824210238.GA940724@coredump.intra.peff.net>
In-Reply-To: <b1711c4c817f95011bc477a9485c115b4926c7da.1692902414.git.me@ttaylorr.com>

On Thu, Aug 24, 2023 at 02:40:38PM -0400, Taylor Blau wrote:

Show 18 quoted lines

> In the topic merged via 5a4f8381b6 (Merge branch
> 'ab/mark-leak-free-tests', 2021-10-25), a handful of tests in the suite
> were marked as leak-free.
> 
> As far as I can tell, each patch from that series ran tests from a
> handful of subject areas, such as "some ls-files tests", or "all trace2
> tests". This left some gaps in which tests had and hadn't been audited
> to be leak-free.
> 
> This patch closes those gaps by exporting TEST_PASSES_SANITIZE_LEAK=true
> before sourcing t/test-lib.sh on most remaining leak-free tests. This
> list was compiled by doing:
> 
>     $ make SANITIZE=leak
>     $ make \
>         GIT_TEST_PASSING_SANITIZE_LEAK=check \
>         GIT_TEST_SANITIZE_LEAK_LOG=true \
>         GIT_TEST_OPTS=-vi test

So having resolved my "oops, lsan logs are racy" problem, my system now agrees with yours on which tests are now leak-free. And we definitely _were_ leak free less than year ago when I posted that other patch. So I'm not sure I buy the "these were missed in 5a4f8381b6" logic.

The one in t5571, I mentioned earlier that I bisected to 861c56f6f9 (branch: fix a leak in setup_tracking, 2023-06-11).

The one in t7516 seems to have been fixed by 866b43e644 (do_read_index(): always mark index as initialized unless erroring out, 2023-06-29).

I found both by bisecting between v2.39.0 (which shows the leak) and v2.42.0 (which doesn't).

-Peff