I see this a few times a week while fuzzing. I would be willing to try to capture more info if needed. Here are some recent .dmp files that cannot be processed by minidump_stackwalk.

:gsvelto I think we spoke about this issue briefly in Berlin. I have also seen .dmp files that trigger assertions in minidump_stackwalk but they are less common. Is there away to tell if the minidump_stackwalk failure is interesting (or not) by looking at the stderror from minidump_stackwalk? If it is not interesting and/or nothing can be done I'd prefer to just ignore those issues in automation.

After bug 1596210 minidump_stackwalk shouldn't print anything to stderr at all. It's possible that it does when it assert()s but I'm not 100% sure and it might only be happening in debug builds. I'll have to check, leaving the NI? for now.

(In reply to Gabriele Svelto [:gsvelto] from comment #4)

After bug 1596210 minidump_stackwalk shouldn't print anything to stderr at all.

Oh I mean when I run minidump_stackwalk on the dmp files with our tools.

(In reply to Tyson Smith [:tsmith] from comment #5)

Oh I mean when I run minidump_stackwalk on the dmp files with our tools.

That should spit out a lot of stuff. If the exit code is zero then I'd say it's not interesting but if it's non-zero there might be something useful in there.

(In reply to Gabriele Svelto [:gsvelto] from comment #6)

If the exit code is zero then I'd say it's not interesting but if it's non-zero there might be something useful in there.

OK good, that is how things currently work[1]. Problem is that it happens more often than I'd expect.

[1] https://github.com/MozillaSecurity/ffpuppet/blob/master/ffpuppet/minidump_parser.py#L49