gdb-mi

[Vladimir Prus]

On Thursday 27 April 2006 17:48, you wrote:

[moving to kdevelop-devel]


> > That flag was there before me, and I'm not sure about it's purpose. I plan
> > to remove it soon, but in the meantime, in revision 534560 I've comitted a
> > fix that makes sure this flag is always cleared after successfull 'run'.
> > Can you try with that revision, while I'm looking into removing that flag
> > completely?
> 
> Appears to work consistently now :)

Good.

> Yes, it's really tediously slow, but with dual core + latest valgrind I can
> just put it on display 2 and do something else while I wait...
>
> BTW, the show more frames improvement really rocks... that will be great
> for recursive call crashes (which was unusable previously).

Yes, we have a bug report exactly about that:

  http://bugs.kde.org/show_bug.cgi?id=102319

Now, the feature is implemented in very nasty way, because I've hit 1001-st 
misfeature in gdb:

  http://sources.redhat.com/ml/gdb/2006-04/msg00322.html

The problem is that we still issue one command ('-stack-info-depth 10000') 
that will take quite some time to execute if your program is stuck in 
infinite recursion. As soon as I get reply on gdb mailing list, I'll consider 
how to improve that.

I've seed that Qt4 data models have 'hasMoreData' method that's supposed to 
help with such incremental fetching, but last time I've tried that did not 
work at all. Guess for KDevelop4 I'll try to use that feature to make 
incremental frame display even more smooth.

- Volodya