View previous topic :: View next topic |
Author |
Message |
keinfarbton
Joined: 03 Dec 2005 Posts: 224 Location: Stuttgart - Germany
|
Posted: Tue Aug 15, 2006 7:35 am Post subject: Problem with backtrace |
|
|
Hi
GDB is one of the keys to the development. Thanks for your patches, I could not work with D without them.
The problem I have: In many cases the backtrace is not working correctly. I often get something like "corrupted stack" when running programs with threads.
If an exception is thrown the backtrace often show only the frames until the code position, where the exception is thrown. But it would be very important for me, to also see the callers from this code.
E.g. if writeln throws a formating exception, it could be hundreds of places where the format string is not correct. Or array bounds is another example.
Frank |
|
Back to top |
|
|
teqdruid
Joined: 11 May 2004 Posts: 390 Location: UMD
|
Posted: Tue Aug 15, 2006 12:52 pm Post subject: |
|
|
Yes, I've seen this issue as well. I don't think it's an issue with the patches so much as it is an omission. Unfortunately, my expertise with that section of GDB is very limited, and my time is more limited so I haven't had time to look into it.
It might be that there's something Walter can change with DMD that fixes this issue, but I'm not certain. You might trying running your program with a vanilla CVS build of GDB and checking that the issue exists there, too. The vanilla CVS won't have symbol demangling, but it shouldn't crash on a dmd -g compiled program anymore. If the CVS version fails, send your D program to the GDB guys and ask for a fix. They'd be able to fix it quicker than myself.
~John |
|
Back to top |
|
|
keinfarbton
Joined: 03 Dec 2005 Posts: 224 Location: Stuttgart - Germany
|
Posted: Wed Aug 16, 2006 3:55 am Post subject: |
|
|
I added a bug report to the gdb bug database with number 2157 |
|
Back to top |
|
|
teqdruid
Joined: 11 May 2004 Posts: 390 Location: UMD
|
Posted: Thu Aug 17, 2006 7:45 pm Post subject: |
|
|
keinfarbton wrote: | I added a bug report to the gdb bug database with number 2157 |
Thanks for following up on this. Please let us know what the GDB guys say.
~John |
|
Back to top |
|
|
|