Re: [patch/rfc] Add a sentinel frame

[Daniel Jacobowitz <drow@mvista.com>]

On Wed, Feb 19, 2003 at 04:25:52PM -0500, Andrew Cagney wrote:
> 
> >Oh, but you're misunderstanding.  There's more than one frame in there. 
> >The call stack in glibc looks like:
> >  _start
> >  calls __libc_start_main
> >  calls main
> 
> Nope, that assumes glibc.  Remember, i debugged this using the d10v.

No, it doesn't assume glibc, it uses glibc as an example.  That check
would prevent backtraces entering _start; on the d10v, that would
override backtrace-beyond-main, but on a glibc system, it wouldn't.

> >_start is written in assembly; it generally doesn't have a frame worth
> >talking about.  Even if we want to show __libc_start_main, we can't
> >safely backtrace into _start.  That's what the inside_entry_file
> >(frame_pc_unwind (fi)) is for.
> 
> Why not?  If someone wants to do that, why should we stand in their way :-)

... uhm, I guess, I don't really find that convincing.  The check was
there because it won't work.  But now that it's not the default I don't
care if you want to break it.

> >The missing test I mentioned above inside_entry_func, not
> >inside_entry_file.  Where'd that go?
> 
> Left until something that does need it surfaces.

If you're going to yank code that way please add a comment saying so,
before you yank frame_chain_valid as dead and someone else discovers
all the inside_entry_func support is now dead code and purges it.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer