Comment 10 for bug 221712

Revision history for this message
Rygle (rygle) wrote : Re: [Bug 221712] Re: the random Disappearing crash

Scribblette wrote:
> If there's any way I can provide more information on this, let me know.
> I don't know if DirectX reports etc are relevant... thought perhaps not.
Not really relevant, as they aren't specific to this code.

> I don't know what the error report means, really
Me either, just trying to help you diagnose it well, so someone else who
does know can step up.

> Not one for technical jargon - all the compiling and programming jargon
> scares the bejeebus outta me.
It's really quite hard to damage anything. I've tried to make all the
steps here on a "for dummies" level, firstly because I'm only a few
steps ahead of you, and second because I want to give people an easy leg
up. You're on a steep learning curve, but it does get easier.

If you can do the Inkcl thing, then the gdb thing (the second method) is
only a little harder. If you can get some results there, it will
generally be much better. My hunch is that you won't get results because
of the nature of the crash, but it just may help, and you'll have gained
valuable experience. I was exactly where you are less than two months
ago and just had to push through a few confidence barriers.

As I said, it's pretty hard to damage anything with Inkscape crashes and
gdb. All the Inkscape stuff is quite independent of the Windows core
operating system files. The only file I've *ever* seen corrupted after
probably about 100 crashes is the Inkscape settings file, which again
doesn't affect anything but Inkscape, and even then it just goes back to
defaults, and you delete it and it remakes it.

Give it a go, ask questions and take one step at a time. You've already
started, so keep going.

I will try to improve the debugging page on the wiki today. Had it
pretty good last night, but hadn't saved and then my beta version of
Firefox crashed and I lost it, but it's mostly still here in this bug
report.

> Have been using SVG files at time of crashes, one file and its backup
> are all I've been using. I'll try using another, brand new file today,
> for other work, and see if it happens.

Probably good to try a new file to see if it crashes then, but it could
be related to the contents of the files you've mentioned.

Instead of trying to minimise the chances of it crashing, try and
maximise the chances and figure out how to most easily make it crash. A
pattern, a series of steps, a file where you reduce the number of
objects/elements bit by bit until it stops crashing, and then you go
back one step to see what was removed and you might have your cause.