This is the mail archive of the
insight@sources.redhat.com
mailing list for the Insight project.
Re: Note on stack window
- To: Tom Tromey <tromey at redhat dot com>
- Subject: Re: Note on stack window
- From: Keith Seitz <keiths at cygnus dot com>
- Date: Sat, 13 Oct 2001 11:30:44 -0700 (PDT)
- cc: Insight List <insight at sourceware dot cygnus dot com>
On 13 Oct 2001, Tom Tromey wrote:
> Today I noticed something I didn't see before, though -- whenever the
> failure occurred, this was printed in the console window:
>
> No symbol "gdb_prompt" in current context.
>
> On a hunch I move the mouse to the source window and held it over a
> function name until I got the variable tooltip to appear. Once I did
> that, the stack resynchronized.
This sounds reminiscent of a uncaught longjmp, probably in varobj. I
hacked at varobj all day Friday, and I did notice a few uncaught gdb
functions. (The way varobj is written, the code should NEVER longjmp.)
The other possibility: does this happen when you re-run an executable?
This is bad because all the data that varobj uses is actually in gdb
obstacks. That's what makes varobj pretty fast: it just moves pointers
around.
My guess is that it is the longjmp issue, and when I do a little more
testing of my varobj change (which, given its size, will take many days to
check in), we can see if this is the problem.
Keith