https://bugs.winehq.org/show_bug.cgi?id=54353
Bug ID: 54353 Summary: crypt32:cert - testVerifyRevocation() gets unexpected success in Wine on second run Product: Wine Version: unspecified Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: crypt32 Assignee: wine-bugs@winehq.org Reporter: fgouget@codeweavers.com Distribution: ---
crypt32:cert - testVerifyRevocation() gets unexpected success in some Wine configurations:
cert.c:3910: Test failed: success cert.c:3911: Test failed: got 00000000
See https://test.winehq.org/data/patterns.html#crypt32:cert
In fact the test only succeeds on the first run in a given wineprefix. All later runs fail:
$ rm -rf ~/.wineprefix $ ./wine dlls/crypt32/tests/i386-windows/crypt32_test.exe cert 0700:cert:0.615 608 tests executed (0 marked as todo, 0 as flaky, 0 failures), 0 skipped.
$ ./wine dlls/crypt32/tests/i386-windows/crypt32_test.exe cert cert.c:3910: Test failed: success cert.c:3911: Test failed: got 00000000 078c:cert:0.242 608 tests executed (0 marked as todo, 0 as flaky, 2 failures), 0 skipped. # Running the test again gets the same failures
This is why the test.winehq.org only has failures for my fg_deb64-wow32 test configuration: on my desktop I run the wow64 tests first, and then the wow32 tests in the same wineprefix.
A bisect shows that these failures were introduced by the commit below:
commit 67de946ca9bdc8f39adffca57a336165936e62e2 (refs/bisect/bad) Author: Paul Gofman pgofman@codeweavers.com Date: Wed Jan 11 14:00:48 2023 -0600
cryptnet: Check cached revocation status in verify_cert_revocation().
That is, before this commit, running the test twice in a row was not an issue.