This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tuples] Convert loop header copying
- From: Zdenek Dvorak <rakdver at kam dot mff dot cuni dot cz>
- To: Diego Novillo <dnovillo at google dot com>
- Cc: gcc-patches at gcc dot gnu dot org, William Maddox <maddox at google dot com>, Aldy Hernandez <aldyh at redhat dot com>
- Date: Tue, 12 Feb 2008 16:28:52 +0100
- Subject: Re: [tuples] Convert loop header copying
- References: <20080212143647.GA10259@kam.mff.cuni.cz> <47B1BA0C.6030500@google.com>
Hi,
> >this patch converts loop header copying, and related cfg manipulation
> >and code size estimation functions. Also, it makes us avoid using
> >labels in GIMPLE_COND statements once cfg is build as in mainline
> >(restores the parts of the code in cfg building and pretty-printing
> >that got lost -- cfg manipulation and rtl expansion, and probably
> >all the other places already expect that this is the case).
>
> Thanks. The patch is fine, however, there is some coordination we need
> to do:
>
> 1- Bill has a patch to estimate_num_insns that I had already approved
> before he had a chance to post it.
>
> 2- Aldy is now doing a merge which is giving him some grief because of
> some OMP changes in mainline.
>
> So, I will ask folks to refrain from committing stuff to the branch
> until Aldy is done with the merge.
OK
> Also, Zdenek could you commit your
> patch after Bill's change? Your patch is a superset of Bill's, so once
> his stuff is in, your final patch should be smaller.
Actually, if that is the case, it perhaps would make more sense to
commit just my patch (I would avoid having to resolve the conflicts, and
estimate_num_insns would be rewritten anyway)? But I'm just guessing,
let Bill post his patch and we may discuss it then,
Zdenek