Re: RFA: implement all missing macro expansion features

[Joel Brobecker <brobecker at adacore dot com>]

> Joel> Thanks for adding the comment describing the function. Can you add an
> Joel> extra empty line before the function declaration?
> 
> Sure.  I was not sure which style is the preferred one -- though I see
> now that this is in gdbint.texinfo.  That said, both styles are common
> in gdb, and in macroexp.c in particular, the blank line only appears
> in one case.

Yes, that's true. I am not too fussy about this as it doesn't make
a whole lot of difference to me. But I try to correct them when I see
them.

> >> +	  /* Just ignore a stray token splicing operator.  Really this
> >> +	     is an error, but there's no reason to penalize the
> >> +	     user.  */
> 
> Joel> I don't see how this would be penalizing the user.  when could
> Joel> this error actually happen?
> 
> Some versions of gcc will pass a macro like this through.
> Trunk rejects code that tries to do this.  By 'this' I mean error
> cases like having '##' at the start or end of the macro definition.
> 
> So, it can only happen in very weird circumstances.
> 
> I can make it an error.  I doubt anybody will notice.
> I'll resubmit the patch a bit later.

I will let you decide what makes the most sense. If you think that
it makes better sense to accept it, then I'll go with that. GNU tools
in general try to be very open in what they accept so your approach
is reasonable. Perhaps a warning explaining that the splicing operator
has been ignored? (just some thoughts)

-- 
Joel