Re: Adding -file-list-exec-source-file command to GDB/MI

[Jason Molenda <jason-swarelist at molenda dot com>]

On Mon, Mar 31, 2003 at 09:31:36AM -0500, Andrew Cagney wrote:
> 
> > -stack-list-frames
> > ^done,stack=[frame={level="0",addr="0x00001dc8",fp="0xbffffca0",func="foo",file="f/a.c",line="3",dir="/tmp/e/"}, [...]
> 
> [...]
> 
> > The other difference is the addition of FP here -- our UI uses the
> > PC and FP to uniquely identify stack frames
> 
> That doesn't work.  The MI needs to be updated so that it uses frame 
> ID's when identifying frames.

Uh, we have a bit of an existance proof that it does--in the form
of a PB+GDB that have been shipping like this for over a year.

I don't mean to indicate that this is the only thing the UI does,
but after an exec-next I think the sequence is something like -
retrieve frame #0 and #1, get the depth of the stack, send a
stack-list-frames-lite command (which returns just the PC and FP
for the frames - filling out the stack frame cache is really
expensive for a big application with a deep stack.)  I don't remember
the exact sequence offhand, but that's basically it.

Our last source merge was early in your frame cache reworkings,
and we won't have another merge until we branch off for our next
release, so we can't look at using it for this release.

(the stack-list-frames-lite is recognized as a hack that is not
relevant in the face of all your frame code reworkings; I expect
we'll be switching over first thing after this release branches.)

Jason