Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[cdt-debug-dev] [cdt-patch] Proposed CDT 1.2 patches 1/3 (fwd)

The PR number is Bug 45530. Description `CDT does not display back trace if gdb/MI encounters an unwind error.'

From what I've seen, the second command succeeds when gdb can successfully unwind at least a few frames. I believe this happens because gdb reports the unwind error only once as long as you haven't switched to a different thread.

For the scenario that I implemented the fix, the second info-depth would always give the right answer. Hopefully, this will be the case with Solaris as well.


Ashish.



To: cdt-debug-dev@xxxxxxxxxxx
Date: Sat, 22 Nov 2003 19:17:06 -0500 (EST)
From: "Alain Magloire" <alain@xxxxxxx>


I see this kind of behaviour on solaris, stack-info-depth fails when debugging a threaded program,
particularly when the thread is not map to a LWP.  In this case, the exception is thrown
and the CDT/Debug/UI just print the thread with a message that the backtrace is not available.

 threa-1 <stackframe not available>

or something similar.

How often that issuing a second command will succeed when the first one fails ?
Did you make a PR for this ?








Back to the top