Interdependent, failing commits

[Noel Grandin]

On Tue, 7 Jan 2025 at 18:12, Stephan Bergmann <
stephan.bergmann at allotropia.de> wrote:

> On 1/7/25 15:08, Noel Grandin wrote:
> > Should be fixed once https://gerrit.libreoffice.org/c/core/+/179890
> > <https://gerrit.libreoffice.org/c/core/+/179890> goes in
>
> What exactly should be fixed by that?


That fixes a failure that looks like
sw/inc/calbck.hxx:85:55: runtime error: member access within address
0x6cbf82939ea0 which does not point to an object of type 'const
sw::LegacyModifyHint'
0x6cbf82939ea0: note: object is of type 'sw::AttrSetChangeHint'
 00 00 00 00  d8 aa a7 43 bf 6c 00 00  b2 00 00 00 bf 6c 00 00  20 9e 93 82
bf 6c 00 00  60 9e 93 82
              ^~~~~~~~~~~~~~~~~~~~~~~
              vptr for 'sw::AttrSetChangeHint'
    #0 0x6cbf38ee73be in sw::LegacyModifyHint::GetWhich() const
sw/inc/calbck.hxx:85:55
    #1 0x6cbf3aa4a21c in SwSectionFormat::SwClientNotify(SwModify const&,
SfxHint const&) sw/source/core/docnode/section.cxx:803:34
    #2 0x6cbf38e99fc3 in SwModify::CallSwClientNotify(SfxHint const&) const
sw/source/core/attr/calbck.cxx:231:18

in CppunitTest_sw_rtfexport4.




> With the above change
> cherry-picked on top of recent master
> 61c12dd1c80e4b7227a614227bb690b00feb78ad, I still get both the two
> different failures described at
>

Yeah, I can't help you there, my UBSAN build is happy with both those unit
tests.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice/attachments/20250107/f657ed9d/attachment.htm>