DUChain crash, reparsing cpp/parser/parser.cpp

[Andreas Pakulat]

On 03.09.08 21:15:50, Hamish Rodda wrote:
> On Wednesday 03 September 2008 03:47:34 Andreas Pakulat wrote:
> > On 02.09.08 17:47:45, David Nolden wrote:
> > > Am Dienstag, 2. September 2008 14:52:10 schrieb Hamish Rodda:
> > > > On Tuesday 02 September 2008 22:24:23 Andreas Pakulat wrote:
> > > > > On 02.09.08 21:20:13, Hamish Rodda wrote:
> > > > > > On Tuesday 02 September 2008 06:10:27 Andreas Pakulat wrote:
> > > > > > > this happened after removing a (manually added) duplicate of the
> > > > > > > hashed-string include at the end of includes. So after removing
> > > > > > > the duplicate include kdevelop started to reparse the file and
> > > > > > > crashed close to the end. I'll leave the core file around in case
> > > > > > > more frames/threads are needed.
> > > > > >
> > > > > > What are the other threads doing?
> > > > >
> > > > > (gdb) thread apply all bt
> > > >
> > > > <snip>
> > > >
> > > > Hmm, doesn't particularly look like a race condition, I wonder what
> > > > david thinks...  is it reproducable? (valgrind log?)
> >
> > Well, I do have occasional crashes when working with kdev4, but I can't
> > just redo the action and get the same crash again. So this one is not
> > reproduceable and thus a valgrind log is a bit of a problem :)
> >
> > > I don't think it has anything to do with the removed include thing.
> > >
> > > Looks like the parent-range was deleted. Maybe a document was closed, and
> > > the editor-integrator didn't notice?
> >
> > No, I only opened parser.cpp, then copied the include, removed it again
> > and then it crashed.
> 
> Hrm, still smells like a race condition.  Did anything else happen at the same 
> time (eg. code completion visible)

No code-completion. I was literally sitting there waiting for it to
finish parsing to check something. All I did was hitting save after
removing the line.

Andreas

-- 
You will win success in whatever calling you adopt.