Re: Last alignment change for MEM tracking

[Geoff Keating <geoffk at geoffk dot org>]

kenner@vlsi1.ultra.nyu.edu (Richard Kenner) writes:

> This large change completes (modulo bugs) the alignment changes that were
> the driving force behind the MEM tracking changes.  It removes the ALIGN
> parameter from many more functions and uses the value from the MEM instead.

Hi Richard,

This patch, and the preceding patches in the sequence, seem to be
making (when taken together) major changes to GCC.  The new
development process requires that major changes be first made on a
branch, where they can be tested on multiple platforms, before being
merged in; and major changes cannot be merged in during certain
periods relative to a release.

I am quite unhappy at the instability these patches have introduced.
Each time one of these aliasing patches has been committed, a
significant number of platforms have broken.  This time, it seems like
_every_ platform broke.  Had the patches been put on a branch and
tested there, these problems would have been caught before disrupting
work on the mainline.

Even if you didn't consider these patches major changes, you would
surely be aware that the last few patches in the sequence caused
problems on a number of platforms, and yet you did not seem to adjust
your testing procedure by testing on more platforms.

Behaviour like this not only makes it difficult for others who are
trying to perform development work on a broken tree, it also sets a
bad example for other developers.  When a maintainer with global write
privileges repeatedly commits patches that immediately cause problems
and do not appear to have been adequately tested, other contributors
feel that thorough testing is unnecessary and resent being asked to do
it.

I therefore ask that you consider voluntarily submitting future
patches for review and approval by other maintainers.  I believe this
would be of benefit to the GCC project.

-- 
- Geoffrey Keating <geoffk@geoffk.org> <geoffk@redhat.com>