| Re: [AD] Electricfence positive with 4.0.3-rc2 stretch_blit |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On 2003-03-24, Eric Botcazou <ebotcazou@xxxxxxxxxx> wrote:
> The stretcher uses self-modifying code, that is we build code
> within scratch memory. Could this possibly harm ElectricFence?
I doubt it, because exscale.c works ok, and now I've discovered
the corner case. This was the line I posted which crashed:
stretch_blit(src, bmp, 0, 0, src->w, src->h, 0, 0, bmp->w, bmp->h);
Ok, put src->w - x or src->w - x, being x any value in the range 0
< limit and it works. IOW, the stretch fails with efence when the
source bitmap is the whole bitmap. Modifications to the destination
don't seem to trigger anything. Could this be related to the
recent fix in the scaling routine which wasn't scaling correctly
the bottom-right corner of the bitmap? What is the file I should
revert to verify this hipotesis?