On 12/11/2010 16:06, Joern Rennecke wrote:
> Quoting Dave Korn:
>
>> Is that documented anywhere? I couldn't find it in the constraints
>> chapters of the internals manual.
>
> It's in the reload source code...
Had a good look through and still couldn't find it, would you mind giving me
a pointer to where should I start reading?
>>> And there are subtle register priority implications of having merged or
>>> separate alternatives.
>>
>> Sorry, is this separate from the thing about appending register
>> constraints to memory ones or part of it?
>
> It's separate.
>
>> Again, no documentation is apparent that
>> suggests or mentions such a concept as "merged" alternatives.
>
> Well, these things are only really needed to generate good code, and to
> generate good code, the way things are now, a port maintainer should
> really understand how reload works.
There are different levels of familiarity with "how reload works", ranging
from a vague top-level overview to intimate familiarity with every single line
of code. I'm somewhere in the middle of that, I'd say, but these features
seem to be more toward the obscure/esoteric end of reload knowledge.
> Of course, you can call this a shortcoming of reload and/or the
> documentation and/or call the continued need for reload a shortcoming, and
> propose patches for any of these issues.
I'm probably not going to write a patch documenting some feature of reload
that I don't know about!
cheers,
DaveK
