GD bug in 1 pixel line antialiasing

[dmorissette]

This has been discussed before on the -dev list but I didn't find a bug
documenting the issue, so here we go:

Steve Lime wrote:

>> Hi folks: Sean and I found a bug in the antialiasing code (single
>> pixel lines) within GD that causes segfaults under certain
>> circumstances. It's a problem with the way they implement clipping
>> for aa lines. I contacted Thomas Boutell late last night and actually
>> heard back this morning. The issue has been fixed in the upcoming GD
>> 2.0.34 release (no date on that release).
>> 
>> 

thomas bonfort wrote:
> Is that the bug I talked to you about a couple of months ago (I also
> sent you a mapfile and shapefile that consistently segfaulted for me)?
> If this isn't the case could you also see with Thomas Boutell to get
> the one line patch in 2.0.34 (I posted the patch to the mailing list
> here:
>
http://www.nabble.com/Re%3A-UMN_MAPSERVER-USERS-antialiasing-line-features..-p1303400.html
> cheers,
> tb

[dmorissette]

From Thomas' previous email to the list:

the patch is to add this near line 3035 of gd.c at the beginning of
static void gdImageSetAAPixelColor(gdImagePtr im, int x, int y, int
color, int t)

if(x<0||y<0||x>=im->sx||y>=im->sy)
{
       return;
} 

[dmorissette]

Steve Lime wrote:
> Hi Thomas- Same bug, same fix roughly. Your patched version is ok with
MapServer 4.8.
> 
> Steve
> 


Closing as WONTFIX. There is nothing we can do except wait for GD 2.0.34 or
patching our local copy of the lib.