https://bugs.kde.org/show_bug.cgi?id=476108
Mark Wielaard <m...@klomp.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |FIXED
Status|REPORTED |RESOLVED
--- Comment #1 from Mark Wielaard <m...@klomp.org> ---
commit a8b6ee6b5f5efbd759c87fa987e9149800db2899
Author: Mark Wielaard <m...@klomp.org>
Date: Thu Oct 26 12:25:44 2023 +0200
vg_replace_malloc DELETE should not check size
The DELETE replacement functions check the size argument, but this
doesn't actually exist. Only the DELETE_SIZED replacement functions
get a size (and should check it).
On i386 (fedora gnu/linux) this causes the following failures:
memcheck/tests/cxx17_aligned_new (stderr)
memcheck/tests/leak_cpp_interior (stderr)
memcheck/tests/mismatches (stderr)
memcheck/tests/mismatches_xml (stderr)
memcheck/tests/new_aligned_delete_default (stderr)
memcheck/tests/new_nothrow (stderr)
memcheck/tests/realloc_size_zero_mismatch (stderr)
All showing "size" being undefined:
+Conditional jump or move depends on uninitialised value(s)
+ at 0x........: ...operator delete[]... (vg_replace_malloc.c:...)
or
+Mismatched new/delete size value: 4
+ at 0x........: ...operator delete... (vg_replace_malloc.c:...)
Oddly no other architecture seems to show issues. Maybe we just got lucky?
This patch fixes the issues on i386 (and shows no regressions on x86_64)
https://bugs.kde.org/show_bug.cgi?id=476108
--
You are receiving this mail because:
You are watching all bug changes.
