Re: Help on Stack trace
Re: Help on Stack trace
- Subject: Re: Help on Stack trace
- From: Scott Ribe <email@hidden>
- Date: Sat, 07 Feb 2009 10:52:47 -0700
- Thread-topic: Help on Stack trace
> Please let me know the approach mentioned above does it make sense considering
> the reusabilty of the existing framework.
Well, to be clear, I don't think the existing framework is reusable. What
you can do, is let it do its normal thing, then capture its output. There
are 2 basic approaches to this:
- <http://smartcrashreports.com/> This approach is somewhat controversial,
because it depends on subverting a system-wide facility (input managers)
that loads code into *every* application launched. Most of the controversy
was over applications that did this without giving the user a choice--and I
agree that's not a good idea.
- Have your app launch a background process which looks for a crash log when
your app terminates. This is the approach I use. (Mine is not fit for
general release, it only runs on a custom application that I control, is a
Ruby script for speed of development, and so doesn't make use of appropriate
OS facilities: uses hard-coded paths, polls for file changes rather than
using kqueue, is subject to a race condition if an app is restarted too
quickly, and also has lots of code for monitoring my app's log during
execution and reporting non-fatal errors. In other words, no, I won't
disclose it, because it would just be more embarrassing than helpful!) The
one thing to watch out for, 10.4 has a single crash log file per app that
gets appended to per crash, while 10.5 creates a separate log file per
crash. One other hint: use vfork/exec to launch the background process ;-)
--
Scott Ribe
email@hidden
http://www.killerbytes.com/
(303) 722-0567 voice
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden