Re: [Fwd: RFA: [infrun.c] Fix to "nexti".]

[Fernando Nasser <fnasser at cygnus dot com>]

Michael Snyder wrote:
> 
> Christopher Faylor wrote:
> >
> > On Mon, Jan 15, 2001 at 10:35:27AM -0500, Fernando Nasser wrote:
> > >If there is a maintainer for this code please speak up.
> > >
> > >Otherwise I will check this in under the "obvious fix" rule.
> >
> > Do we really have an "obvious fix" rule?  It seems that there is
> > some confusion on this issue.
> 
> We do (I think), but when you're discussing infrun.c, I'm not
> sure that any change can be regarded as "obviously correct".
> At least not in the wait_for_inferior/handle_event area.
> 

Well, someone must be able to evaluate it and make a decision.

This thing has no specific maintainer (I could not find in the 
MAINTAINERS file) and people keep avoiding it.  But it is broken 
and has to be fixed.  What are we going to to about it?

Or will we add in the Users manual that a "nexti" command should
not be used inside a function prologue or it will work as a "continue"?


> By eyeball, this change looks correct to me, or at least
> "not obviously incorrect".  I would like to see it tested,
> and perhaps the best way to do that is to apply it and then
> notice if there's a sudden uptick in testsuite failures.
> 

Well, of course I run the testsuite on it on a couple of targets and
with a couple of versions of gcc.  This means little with this code
as it is used by all targets but what else can I do?

When the code that was originally there was removed years ago and
replaced with a very simplified test, all the knowledge that was
accumulated over the years about different situations that can arise
was lost.  All that we can do know is to start accumulating that 
knowledge again by adding the appropriate tests for the cases where
it fails.

But if the bug fixes are just ignored and remain as a lost message in the 
list archives we will never evolve from where we are now.

We could have a page with known bugs and patches that fix them, with some
mechanism for feedback.   But at the moment, all we can do is to install
the fix and see if there is some hidden condition that was not accounted for
and either revert it or refine it further.


-- 
Fernando Nasser
Red Hat - Toronto                       E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9