Created attachment 83607 [details] GDB backtrace of the segfault Printing certain pages from Evince results in segfaults in libcairo. However, when I ran evince through Valgrind, no segfault occured. Version used: evince: 3.8.3 poppler: 0.22.5 cairo: 1.12.14 I have a GDB backtrace attached to this bug report. Until now, only one document showed this behaviour for me. The problem is, that this document is my bachelor thesis and I am not allowed by my department to share it on the Internet. If you need the file, I would send it to you as a private e-mail, but I do not want to post it publicly available. If you need further information, don't hesitate to ask.
Created attachment 83608 [details] valgrind report
I'll need the file to be able to fix this. You can send it to me at ajohnson@redneon.com.
I am unable to reproduce the bug. What distro are you using? 32 or 64 bit? Does the bug occur if you run: pdftocairo -pdf evince-segfault.pdf out.pdf
Hi, i am running Arch Linux x86_64. The bug does not occur when using pdfcairo.
Are you able to build cairo with debugging enabled? The attached gdb and valgrind reports are not very useful without the debug information. If you have poppler-glib-demo installed you could see if that reproduces the bug. Does printing to a pdf file cause the bug or do you have to print to a real printer?
Hi, > Are you able to build cairo with debugging enabled? The attached gdb and > valgrind reports are not very useful without the debug information. I am currently trying to build cairo with debugging enabled and redo the tests. > If you have poppler-glib-demo installed you could see if that reproduces the > bug. The bug does not occur in poppler-glib-demo. If tried printing to file and to a real printer and it both works. Curiously, if I now try to print the "printed" PDF from poppler-glib-demo in evince it passes the test. Is there a way to do diffs on PDF files? > Does printing to a pdf file cause the bug or do you have to print to a real > printer? Does not matter. Occurs in both types.
Created attachment 83856 [details] gdb output with debugging enabled I just built cairo with CFLAGS set to "-g". Now the stack trace contains some more output, but I am not sure if I have done everything right. There are still a lot of "No symbol table info available".
Try building with "-g3 -O0" and see if that gets more information out of the gdb. If that works you could also do another run with valgrind. > Curiously, if I now try to print the > "printed" PDF from poppler-glib-demo in evince it passes the test. Is there > a way to do diffs on PDF files? Not easily, and it probably won't be useful. I really need to either reproduce it myself or get enough debug info from you to see where the bug is. I will try installing Arch Linux in a VM sometime but I won't have time to do that for a while.
Another idea is to run evince with cairo-trace while printing. But I'm not sure how to separate the trace of the pdf printing from the screen drawing. Maybe Chris can advise the best way to do this.
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/cairo/cairo/issues/255.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.