Re: [rfc] Tighten meaning of gdbarch_convert_register_p

[Joel Brobecker <brobecker at adacore dot com>]

Note to the other maintainers:

I feel I need to say this once, and then I'll stop saying it. I'm
trying to help more in the design decisions, but I still have a long
way to go. My goal right now is to look at as many design discussions
as time allows me, and try to think about it. For a little while, if it
touches areas I've never really worked on, chances are I won't have much
to say about it. Nonetheless, I think that it is still helpful to say
that I have looked at it and saw nothing wrong with it. I'll probably
get burnt a number of times, but I'll learn, and hopefully you'll
appreciate the fact that someone looked at your work, and thought it
looked good. Just take my comments with a grain of salt for now ;-).

> This patch tightens the definition of gdbarch_convert_register_p so
> that a non-zero result means that a conversion is necessary for the
> supplied TYPE, not just that one might be necessary for some type.
> This let me add an assertion that gdbarch_register_convert_p was
> zero for any unwound register values in the new value-based unwinding
> that I wrote yesterday.
[...]
> 
> What do you think?

This looks like a good idea to me - I can see how this simplifies
a bit your other patch (using values in unwind). It complicates a bit
gdbarch_register_convert_p, but really not all that much, and it should
simplify the conversion methods in return.

-- 
Joel