GCC-3.2.2 pre-release (third iteration)

[Joe Buck]

Volker Reichelt <reichelt@igpm.rwth-aachen.de> writes:
> | I know it's a little late now, but I just stumbled over two PR's
> | (PR 8555 and PR 8766) that are regressions *on* the 3.2 branch and have
> | not been fixed yet.

On Tue, Feb 04, 2003 at 09:24:46PM +0100, Gabriel Dos Reis wrote:
> According to the audit trails, PR 8555 was introduced in 3.1 by a
> patch of Jakub's; and PR 8766 already failed with 3.2.1 so, they are not
> showstoppers.  I believe we can look into them for 3.2.3, if that happens
> to exist.
> 
> | Does that affect the final release of gcc 3.2.2?
> 
> No, I'm afraid it is too late now.

Agreed; after three pre-release tarballs the window should be shut tight.
We've given Gaby the job of making the decisions, but at this point, I
would not even consider any bug fix, unless the bug exists in gcc
3.2.2-pre but not in gcc 3.2.1, and is significant.  Time to wrap it up.
We'll never get a release that fixes all bugs, and you approach a point
of diminishing returns (where the risk of a fix causing new bugs starts
to get larger than you would like).

Our goal is to make sure that

ForAll n, Quality(GCC-x.y.(n+1)) > Quality(GCC-x.y.n)

and one of the main measures of that is regressions.  But we can't hope
to fix every single regression from any previous GCC release.