Re: PATCH in Java constant propagation

[Martin Kahlert <martin dot kahlert at infineon dot com>]

Thank you for your work!

On Wed, Sep 11, 2002 at 01:51:51PM -0600, tromey@redhat.com wrote:
> >>>>> "Per" == Per Bothner <per@bothner.com> writes:
> 
> Per> I believe this fixes the bug that martin.kahlert@infineon.com
> Per> reported (subject: Is this a bug?).  I *think* it doesn't break
> Per> anything else, but I did get a lot of testsuite errors, so I'm
> Per> not sure.  I looks like I will be inaccessible for a the rest of
> Per> the week, so I'd appreciate it if someone could test it, and if
> Per> it works, check it in.
> 
> I applied and tested this patch.  I rebuilt libgcj from scratch with
> it, and I ran the test suite including Mauve.  I get just the ordinary
> expected problems, and no regressions (against a compiler without the
> patch -- some of the currently-expected failures seem to represent
> regressions against earlier compilers).
> 
> I'm checking this in.

I applied the patch to a clean gcc version 3.2.1 20020902 (prerelease)
snapshot. It applied cleanly. The resulting compiler seems to
work correctly (only the hex stuff in the output is ugly).

Could this patch be applied in the 3.2 branch, too?
If it improves things there and does not cause extra work,
why not include it into the 3.2.1 release?
As i understand it, 3.2.1 is a bug fix release for 3.2 which
would have become a 3.1.2 without the API changed).
3.3 will last quite some time from now on.

Thanks a lot for your help
Martin.

-- 
The early bird catches the worm. If you want something else for       
breakfast, get up later.