2.4.1

[Jürgen Spitzmüller]

Am Donnerstag, dem 04.07.2024 um 17:12 +0200 schrieb Jean-Marc
Lasgouttes:
> > I suppose caching this vector and only recreating it when necessary
> > will make the case in the bug report usable again (the long list
> > causes
> > a noticeable delay, but it's bearable if done once IMHO), and it
> > will
> > probably generally speed up document loading with many citations.
> 
>  From what I understand, it exhausts the memory of the computer by 
> allocating memory somewhere in the regex machinery (see the backtrace
> in the report). So the problem is not just about caching.

Probably, but I see here that expandFormat() is repeatedly called,
which calls getAuthorOrEditorList() which calls getAuthorList() which
recreates the vector which takes time with this huge author list.

expandFormat() also has this at its entrance:

	// incorrect use of macros could put us in an infinite loop
	static int const max_passes = 5000;

this is probably what happens here. 5000 passes at that speed are a
long time.

So I wonder which macro we do not use correctly here (in my example
case, attached, it is %!abbrvciteauthor% (but I do not see yet why).

-- 
Jürgen
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bibbug.bib
Type: text/x-bibtex
Size: 31791 bytes
Desc: not available
URL: <http://lists.lyx.org/pipermail/lyx-devel/attachments/20240704/6cb25406/attachment-0001.bib>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bibbug.lyx
Type: application/x-lyx
Size: 2286 bytes
Desc: not available
URL: <http://lists.lyx.org/pipermail/lyx-devel/attachments/20240704/6cb25406/attachment-0001.lyx>