[PUSHED:3-5] fix STL assertion in vcl graphite code

[Petr Mladek]

Martin Hosken píše v Po 19. 03. 2012 v 15:22 +0100:
> Dear Petr,
> 
> Thank you for the reminder of this issue. My apologies for not getting to it earlier.
> 
> > > 
> > > I do not understand the code either. I just wonder if it would make
> > > sense to check for:
> > > 
> > > 	while (++nCharIndex - mnMinCharPos <
> > >                     static_cast<int>(mvChar2BaseGlyph.size()))

> I would go with the first of these. It makes no sense to compare a
>  charIndex against a glyph->char mapping array, whose size is the
>  number of glyphs, not characters. I think it is simply a bug.

Sounds reasonable. I have updated the code in master, see
http://cgit.freedesktop.org/libreoffice/core/commit/?id=3a878d3dbfdb11cf2f0cce9dbf28a408c130d556

And pushed it into 3-5 branch, see
http://cgit.freedesktop.org/libreoffice/core/commit/?id=3a878d3dbfdb11cf2f0cce9dbf28a408c130d556


> Notice that I am not claiming authoritative understanding of this code
>  since it was written by Keith, who unfortunately is no longer with us
>  and so cannot comment. I am hoping that when I eventually get time to
>  work on this code, and particularly the mac port, I can refactor the
>  code and gain the necessary authoritative confidence.

It would be great. Thanks a lot for your opinion.

Best Regards,
Petr