Hi, Le 5 févr. 08 à 22:13, Matthew Allum a écrit :
Does attached patch further improve at all (totally untested btw, only Intel here) ?
Yes, it does but I am not sure if this is the correct fix. I mean, I had a look at pango SVN sources (pangoft2-render.c), the Clutter code is very similar to it, but pango does not have any special case for this situation. So, either (i) pango has the same bug and it's not exhausting on my system because gtk is using another renderer (Xft?), or (ii) they somewhat handled the situation differently and earlier?
A colateral question is: can bitmap.buffer be NULL even if FT_Render_Glyph() returns successfully?
BTW, on second thought, I also had a look at cairo (cairo-ft-font.c), and it does have a special case for zero width and height.
-- Gwenolé.-- To unsubscribe send a mail to [EMAIL PROTECTED]
