SIGSEGV on master regarding math completion

[Scott Kostyshak]

On Tue, Jan 25, 2022 at 05:12:06PM +0100, Jean-Marc Lasgouttes wrote:
> Le 25/01/2022 à 16:07, Scott Kostyshak a écrit :
> > On testing a different commit I received the following terminal message:
> > 
> >    Undo.cpp (319): There is no group open (creating one)
> 
> If you can reproduce this, I am interested.

It is reproducible but I think my message was not clear. By "a different
commit" I mean "an older commit", and wanted to give the idea that I'm
not sure the bisect would be linear (because of several differences in
behavior). I can reproduce that terminal message with the commit
cc0ef52a (2019-01-14). I cannot reproduce that message on master with
the original recipe.

I have seen that terminal message on master in other contexts (I think
it had something to do with undefined layout insets). I guessed that it
was not a problem since it did not assert. Should I report an MWE
whenever I see that message?

Scott


> > JMarc, the backtrace suggests that painting code is involved. If a
> > bisect would be helpful, which change of behavior should I bisect?
> 
> The problem seems to be that the inset is created without setting properly
> its buffer. And the drawing code uses the value of this buffer.
> 
> JMarc
> 
> -- 
> lyx-devel mailing list
> lyx-devel at lists.lyx.org
> http://lists.lyx.org/mailman/listinfo/lyx-devel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.lyx.org/pipermail/lyx-devel/attachments/20220125/1384446f/attachment-0001.asc>