This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [autovect] Add William Pugh's Omega integer programming solver
On Wed, 2005-04-13 at 23:40 +0000, Joseph S. Myers wrote:
> On Mon, 11 Apr 2005, Sebastian Pop wrote:
>
> > Thanks to Danny for having sorted out the license questions and for
> > having contacted Bill Pugh. I will wait for comments before
> > committing the code.
> >
> > As it is in this patch there was no modification to the original code,
> > except renaming the include name "dd_omega-ip.h" to "omega.h". The
> > code as it is would not pass bootstrap because of some warnings that
> > will be fixed on a later patch.
>
> I hope it is realised that code without proper comments above each
> function (explaining what the function does and the meaning of its
> parameters and return value) is unacceptable for mainline
Of course.
> . (Apart from
> being completely undocumented, this code has many other problems in coding
> style terms; functions and variables should be named in the same style as
> the rest of GCC and macros should be named in uppercase when the GCC
> Coding Conventions say so; macros expanding to statements should be in do
> { ... } while (0); magic numbers should be properly explained and --params
> if meaningfully variable; there's no need to add uses of "register".)
Yes, we know.
Who has said any different?
>
> The tree-ssa developers still haven't added all the required comments,
> despite claims that this was done before the merge to mainline (bug
> 13756).
In files i maintain, there is approximately one function missing a
comment (and the function didn't exist at merge time).
I will commit a patch momentarily to fix that.
> I hope future branch mergers will not try to push through code
> that does not follow the coding standards.
>
Considering the code isn't even hooked up to anything yet, let alone
submitted to mainline, i think you are jumping the gun a bit.