Re: [RFA] New GDB target iq2000

[Daniel Jacobowitz <drow@false.org>]

On Mon, Mar 07, 2005 at 09:17:08PM +0100, Corinna Vinschen wrote:
> On Mar  7 09:03, Daniel Jacobowitz wrote:
> > On Mon, Mar 07, 2005 at 11:08:35AM +0100, Corinna Vinschen wrote:
> > > On Mar  5 15:20, Daniel Jacobowitz wrote:
> > > > On Sat, Mar 05, 2005 at 09:17:45PM +0100, Mark Kettenis wrote:
> > > > > Yup.  Although the lower-limit for first-line breakpoints may cause
> > > > > bogus parameter values to be printed.  I consider that less a problem
> > > > > than my program unexpectedly running to completion though.  The
> > > > > problem is that some people tend to think differently and we never
> > > > > reached consensus about it.
> > > > 
> > > > Well, it makes sense to me.  It's clear that the FRV and submitted
> > > > iq2000 ports have different heuristics for these two cases; it would be
> > > > good to cover both of them in common code.
> > > 
> > > This is rather getting a generic discussion to me.  What about the
> > > iq2000 port itself?  Does it make sense to defer iq2000 until after
> > > the generic problem has been solved?
> > 
> > No, it doesn't.  Did you see the patch I sent you on Friday night?
> 
> I've tried it and it has two FAILs more than the original version of
> iq2000_skip_prologue, the two break.exp FAILs we already know:
> 
>   (gdb) FAIL: gdb.base/break.exp: breakpoint small function, optimized file
>   (gdb) FAIL: gdb.base/break.exp: run until breakpoint set at small function, optimized file

Just a guess but - this is probably because you sent me "break" rather
than "breako2", the actual optimized binary.  May I have another shot
at it?

In the mean time, this is drawing out too long; I'm going to review the
port without attention to this issue this afternoon.

-- 
Daniel Jacobowitz
CodeSourcery, LLC