The new version changed a line to support the ywrap/pan options of the framebuffer. Somehow that segfaults, I've traced around the code to figure out how or why it happens... but I've been unable to find the cause. I doubt I would be able to fix a testcase as this bug seems very environment dependant. I hacked together a small patch that reverts part of the new version. It should not have any effect on performance as I can not see how the code would work with ywrap/pan anyway (I should read up on that). If anything it improves performance. Otherwise the patched version works without problem. Maybe I should submit it upstream or at least alert the maintainers of the issue... The patch might have a bad format, if so -- I'm sorry!
Created attachment 48494 [details, diff] the patch the patch
please add a metadata.xml for your package.
patch format is fine. Submit patch upstream. You haven't described what condition needed to occur for the segv to occur. I'm hesitant about including this patch at this stage.
No response from upstream (one of the e-mails bounced too). I still don't know how the bug is triggered. But I know that the patch would be most stable as it is a partial revert to 0.99 which never segfaulted. Check the #if 0 in the source (might be visible in my patch), the code that has been #if0ed out was what it looked like in the previous version.
added patch to 1.0b.
