This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Proposed targets to deprecate for 3.4


On Wed, 1 Oct 2003, Zack Weinberg wrote:
> Hans-Peter Nilsson <hp@bitrange.com> writes:
> >> There has got to be a better way to do that.
> > Perhaps a (new) TARGET_ASM_FUNCTION_BEGIN_PROLOGUE?  If you're
> > opposed to TARGET_ASM_FUNCTION_END_PROLOGUE, look at alpha.
> No, I'm not opposed to ..._END_PROLOGUE; it has a well-defined task
> that doesn't overlap with another mechanism.  Could you use
> ASM_FUNCTION_START for what you need, maybe?

What's that?  A new macro? :-)  (Sorry, I see no useful macro
matching *FUNCTION*.)  Still, it'd just be trading one for
another.

> But I was thinking more in terms of marking prologue insns specially
> so that the "avoid renumbering" effect doesn't depend on magic state
> outside the RTL.  Say, with a port-specific REG_NOTE.  (I don't know
> if this is possible.)

It sounds possible.  I'll investigate.  Thanks.

> It's a nuisance when I'm doing a sweep through all the files in the
> config/ directory: imagine typing C-x C-f mm TAB mm TAB .c RET
> C-s SOME_MACRO without stopping for breath only to notice that
> everything after the RET got eaten by y-or-n-p...  and then I usually
> reflexively hit C-g which aborts the entire find-file and I have to
> start over. :-(

Oh, *that*. :-)

> I suppose you have your default style set to something else?  I do so
> little work on anything other than GNU-indented code that I just leave
> that the default (it's not my preferred style, although it has become
> habitual).

Hmm, the same goes for me these days.  It used to be just those files...

brgds, H-P


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]