Vector indexing patch

[Richard Guenther]

On Fri, Jun 4, 2010 at 6:38 PM, Joseph S. Myers <joseph@codesourcery.com> wrote:
> On Tue, 1 Jun 2010, Artem Shinkarov wrote:
>
>> This is a reworked patch of Andrew Pinski "Subscripting on vector
>> types" in terms of GSoC 2010 [Artjoms Sinkarovs].
>
> We can't consider it without a copyright assignment.
>
>> The documentation was not changed. May be there could be some more
>> ideas to implement.
>
> The original patch had documentation.  You should not make patches worse
> in the course of reworking or updating them; documentation is required for
> all patches with user-visible features.  The original also had ChangeLog
> entries following the usual conventions; please make sure to follow the
> style of existing entries.
>
> Since I previously approved the C changes in
> <http://gcc.gnu.org/ml/gcc-patches/2009-10/msg00562.html>, and do not wish

The mark_addressable_vector change was at least semi-ugly.  I think
it is reasonable to require a constant index for register qualified vectors.
And it is reasonable to emit middle-end friendly BIT_FIELD_REFs for
constant indices as well.

Artem is one of our Goole Summer of Code students btw., and I
requested these changes.

Your comment about the ChangeLog entry is valid - Artem, please
look at existing entries in gcc/ChangeLog and follow that style.
Also please keep the existing documentation changes and/or adjust
them accordingly.

Thanks,
Richard.

> to duplicate review effort, I am not inclined to review any reworked patch
> without adequate explanation of what was changed from the approved version
> and why it needed to be changed.  The initial basis should be to make only
> those changes to the approved patch that are strictly necessary to update
> it to newer sources or to fix bugs, so that only those clearly identified
> changes, and any parts not previously reviewed, need to be reviewed this
> time.
>
> --
> Joseph S. Myers
> joseph@codesourcery.com
>