patch rereview requested for PRs 6860, 10467 and 11741

[Mark Mitchell]

> I'd like to express agreement with this specific point -- particularly
> now, when we are closing in on the end of stage 2, and we may not
> *want* the 'good fix' because it would be too invasive.  People should
> start thinking in terms of what would be a good patch for a release
> branch.

Without commenting on Roger's original patch, etc., I'd like to second this
point.  The users don't care *how* the bug is fixed; only *whether* it is
fixed.  It's not like GCC is a picture of perfection at this point; we can
live with adding an occasional wart if it fixes an important bug.

-- Mark