Re: Require '%' to be at the beginning of a constraint string

[Jeff Law <law@redhat.com>]

On 05/28/14 13:50, Richard Sandiford wrote:
> Thanks for the review.
>
> Jeff Law <law@redhat.com> writes:
>> On 05/26/14 13:21, Richard Sandiford wrote:
>>>> If we're going to change it, then clearly the docs need to change and
>>>> ideally we'd statically check the port's constraints during the build
>>>> process to ensure they meet the tighter definition.
>>>
>>> OK, how does this look?  I built a cc1 for one target per config/
>>> directory to (try to) check that there were no remaining cases.
>>>
>>> This means that we will silently ignore '%'s that are embedded in the
>>> middle of an asm constraint string, but in a way that's already true for
>>> most places that check for commutativity.  An error seems a bit extreme
>>> when '%' is only a hint.  If we want a warning, what should the option
>>> be called?  And should it be under -Wall, -Wextra, or on by default?
>>>
>>> Tested on x86_64-linux-gnu.  OK to install?
>> OK.  My initial thought on adding a warning was to weed out bad
>> constraints.  You've already done that for the in-tree ports.  I'm a lot
>> less inclined to do much more here to help the out-of-tree ports, so
>> upon further review, let's not worry about the warning unless you've
>> already got it ready to go :-)
>
> Well, the new genoutput.c error should catch problems in out-of-tree ports.
Right.

> It's just asms where the non-initial '%'s would be silently accepted
> and have no effect.
Right.  And I believe that it's conservatively correct -- so we're OK 
here as well.

>
> David W suggested off-list that I add "Only input operands can use
> @samp{%}." to the documention as well.  That seemed like it was
> obviously an improvement so I applied the patch with that change (below).
Funny I was thinking about that when looking at the arc changes.  I saw 
the '%' on operand0 and was confused thinking it made no sense to have a 
'%' for an output operand.  Then I realized that 0/1 were inputs and 2 
was the output.

Thanks for pushing this through.

jeff