When fixing an "out of buffer crash" few monthes ago, pdf14_put_image was improved with accounting the 'rect'. Now we noticed that pdf14_cmykspot_put_image and pdf14_custom_put_image still have same problem. I do know how to fix that, but I have no test suits for debugging it. So giving to Ray : please give us a test case, which Dan should have had somewhere when coded these functions. If no test suits available, please pass it back to me with a comment how to make them.
Setting P2 because it's a part of the current project, and it's a crash when somebody will call these functions.
pdf14_put_image was improved with revision 7064. pdf14_cmykspot_put_image and pdf14_custom_put_image were added by revision 7795 as a merge from the custom color callback branch. That branch missed the improvement 7064 due to branching before it.
Patch to HEAD : http://ghostscript.com/pipermail/gs-cvs/2007-October/007858.html The patch was not tested at all. We don't close the bug now because it still requires an appropriate testing. Please provide test cases which involves pdf14_cmykspot_put_image and pdf14_custom_put_image. Likely the test file of the bug 688631 can help.
Thanks for the patch. I have test files for this (from customer 190) and am working on other problems with the transparency code going to cmykspot devices.
Reassigning back to Leo. Ray please add test files (ref comment #4) and provide build instructions for custom color callback code. Thanks.
Assigning back to Ray. Ray please follow comment #5.
Changing the bug title for a better reflection of the problem. The old one is "transparency: The buffer backdroping code can write outside bufers.".
As with other requests for extended testing of some compiled in modes, I am assigning to Marcos, at the appropriate priority. Note that this isP5 since Michael's new ICC color workflow will deprecate the custom color callback as it exists today.