[Gs-code-review] SF bug #458769 - Clipping problem with fill operator

[Dan Coby]

Ray,

This looks okay to me.  I think that the enclosed is a slightly
shorter variant of the same thing.  (It is untested and you are
welcome to ignore it.)

Dan


-----Original Message-----
From: gs-code-review-admin at ghostscript.com
[mailto:gs-code-review-admin at ghostscript.com]On Behalf Of Ray Johnston
Sent: Wednesday, January 30, 2002 8:49 AM
To: GS Code Review List
Subject: [Gs-code-review] SF bug #458769 - Clipping problem with fill
operator


Reviewers,

This was the long standing bug from an important Artifex customer.

The problem is that there was a clip path that was computed by the
function gx_effective_clip_path that was then saved in the gstate
which was not being properly kept track of, resulting in the view
clip path being ignored at times (depending on the number and the
balance of gsave/grestore operations).

The fix involves using the gstate structure element which holds
the 'id' of the effective clip path. In some cases this will be
the same as the id of the clip_path, or the viewclip_path, or it
may be a unique id for the intersection of those paths. When the
effective_clip_path is set, the effective_clip_id is set to the
correct value.

Regards,
Ray
________________________________________________________________
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gspath.c.diff
Type: application/octet-stream
Size: 572 bytes
Desc: not available
Url : http://ghostscript.com/pipermail/gs-code-review/attachments/20020204/222ac4ec/gspath.c.obj