[PATCH] Improve i386 prologue analyzer

[Andrew Cagney]

>>>> > And I still don't understand what is the rush to release the MIPS
>>>> > patch without waiting for another week or two and then releasing the
>>>> > i386 patch as well.
>>
>>> 
>>> If I were a MIPS user (hmm, I'm even the maintainer), I'd be pretty 
>>> cheesed off that a fix to get `break main; run' working was being held 
>>> back due the inistance that it be bundled with an unrelated i386 fix.
> 
> 
> Sorry, Andrew, this is just reiterating what I already said I didn't
> understand: if we care so much for the MIPS, why did you refuse to
> wait for it to be fixed in 6.2?

I think I answered this:
> The MIPS breakage wasn't important enough to deny our mainline users a new release of GDB.  Especially when it fixed so many bugs, and especially when the delay could easily be a month..

I've since spoken to some i386 users that benefited from the fixes, they 
are most pleased - threading debugging is way better - dwarf2 back 
traces are much better.  They clearly benefited from pushing forward 
with the 6.2 release.  Only thing is, they are now asking if we can fix 
.... :-)

> And if 6.2 could hit the street with
> MIPS broken, why cannot it stay broken for a week or two more, instead
> of letting the i386 remain broken longer?

And I also think I've covered this:
> However, now that we've got the MIPS patched up, I see no good reason for holding it back.

It's no significant work on my part, and provides us with the 
satisfaction of delivering a better GDB (for MIPS users) sooner.  I'm 
sure the MIPS users would have appreciated the kind gesture.

Andrew