Re: PR 13722 candidate fix

["Zack Weinberg" <zack at codesourcery dot com>]

Jim Wilson <wilson@specifixinc.com> writes:

> My latest bootstrap succeeded.  I didn't run the testsuite, and I still
> had some debug code in there just to be safe, but I don't think either
> of those issues should matter.  With the following patch on top of your
> patches, the IA-64 gcc port bootstrap again including Ada.  This patch
> adds the REG_INC notes for auto-inc addresses in the IN operands.  There
> is also the potential problem I pointed out earlier where we use INTVAL
> without checking to see if we have a CONST_INT.  That isn't in this
> patch, but I think it should go in also.

Okay.  I have incorporated your changes into my patch, added the
CONST_INT check, and am running a bootstrap on ia64-hpux; if it
succeeds I'll repost the combined patch.

[...]
> So what I am asking here is that if you check in a patch that
> breaks the IA-64 compiler, and you are not able to solve the problem
> yourself, that you revert the patch until I am able to help you.  I am
> willing to help you with these kinds of things, but it isn't fair to
> others to let you monopolize my time, and it isn't fair to others to
> leave the IA-64 compiler broken for long periods of time.

I am willing to do this for most breaks.

However, it is my understanding that breaks that affect only Ada are
*not* considered critical, and in fact that there is no obligation for
anyone to attempt to build the Ada front end.  I normally do not
bother building it even on platforms that can handle it (ia64-hpux
cannot, as the runtime has not been ported).  This is why I did not
respond very fast to this bug report.  I am willing to change my
habits if the policy has changed, but not otherwise.

> If you want to be really helpful, you could try looking at some of the
> IA-64 gcc bug reports in bugzilla.  PR 7198 is on my list, and is one
> that you could handle.

Sure, I'll look at that.

zw