Re: [PATCH v2 5/5] Stash frame id of current frame before stashing frame id for previous frame

[Kevin Buettner <kevin at buettner dot to>]

On Wed, 9 Nov 2016 14:48:46 +0000
Pedro Alves <palves@redhat.com> wrote:

> On 11/02/2016 10:26 PM, Kevin Buettner wrote:
> 
> > I will first note that the frame id for frame has not been computed yet.  (This
> > was verified by placing a breakpoint on compute_frame_id().)
> > 
> > The call to get_prev_frame() causes the the frame id to (eventually) be
> > computed for the previous frame.  Here's a backtrace showing how we
> > get there:
> > 
> >     at gdb/frame.c:496
> >     at gdb/frame.c:1871
> >     at gdb/frame.c:2045
> >     at gdb/frame.c:2061
> >     at gdb/frame.c:2303
> >     at gdb/python/py-frame.c:381  
> 
> Function names would make that backtrace soooo much easier to read.  :-)

Thanks for catching this.  It had me puzzled for a while.

Since some of the backtrace lines were long, I pasted the backtrace
into my commit message without any leading whitespace.  As a consequence,
the # for the backtrace ended up in column 0 and was interpreted by git
as a comment.

I added the backtraces back into the commit message with leading whitespace.

After reading the documentation, I see that "git commit --cleanup=verbatim"
could be used to prevent git from removing lines starting with #.  Though
that means that the committer needs to manually remove other comment lines
added by git.

> 
> > gdb/ChangeLog:
> >     
> >     	* frame.c (get_prev_frame): Stash frame id for current frame
> >     	prior to computing frame id for previous frame.  
> 
> 
> I'm fine with this solution.  LGTM.

Pushed.

Kevin