Re: [PATCH 1/3] Revert "EXA: Accumulate arbitrary number of glyphs without flushing."

2009-11-30 Thread Adam Jackson
On Mon, 2009-11-30 at 10:58 -0800, Jeremy Huddleston wrote: > On Nov 30, 2009, at 08:41, Keith Packard wrote: > > As a reminder to others -- alloca should never be used in the X server; > > it cannot be relied upon to return NULL on failure, doing various things > > From crashing the application to

Re: [PATCH 1/3] Revert "EXA: Accumulate arbitrary number of glyphs without flushing."

2009-11-30 Thread Jeremy Huddleston
On Nov 30, 2009, at 08:41, Keith Packard wrote: > On Mon, 30 Nov 2009 13:17:51 +0100, Michel Dänzer wrote: > >> Not sure what I was thinking, turns out alloca() of a size derived from >> client >> input is a bad idea. > > That was the sole remaining use of alloca outside of the quartz > back-

Re: [PATCH 1/3] Revert "EXA: Accumulate arbitrary number of glyphs without flushing."

2009-11-30 Thread Keith Packard
On Mon, 30 Nov 2009 13:17:51 +0100, Michel Dänzer wrote: > Not sure what I was thinking, turns out alloca() of a size derived from client > input is a bad idea. That was the sole remaining use of alloca outside of the quartz back-end. As a reminder to others -- alloca should never be used in th

[PATCH 1/3] Revert "EXA: Accumulate arbitrary number of glyphs without flushing."

2009-11-30 Thread Michel Dänzer
From: Michel Dänzer This reverts commit c11678cc189551f2a01eaa7a63969c16950739b4. Not sure what I was thinking, turns out alloca() of a size derived from client input is a bad idea. Signed-off-by: Michel Dänzer --- exa/exa_glyphs.c | 47 +-- 1 fil