Re: Loop unroll fixes

["Joseph S. Myers" <jsm28@cam.ac.uk>]

On Mon, 17 Sep 2001, Richard Henderson wrote:

> Suppose the patch is wrong in some way that causes it to fail for
> some obscure target under some conditions.  Suppose this is not
> discovered for a year.  If the patch is well documented, one can
> quickly recall what the original problem was, as opposed to either
> (1) spending lots of time re-examining the original bug, or
> (2) guessing the that the old patch was wrong and reverting it.
> 
> Both (1) and (2) have happened many times in gcc history.

There should have been a testcase added for the original problem.

Two consequences of this:

* We should get Ada into the tree as soon as possible (in particular,
before the cut-off on October 15 for major new features for 3.1), since
sometimes these sorts of problems have occurred with patches written to
fix bugs showing up with Ada testcases.  What caused the delay from the
date of June 1 given in
<URL: http://gcc.gnu.org/ml/gcc/2001-05/msg01148.html >?

* Where <URL: http://gcc.gnu.org/ml/gcc/2001-05/msg01273.html > says (of
creating synthetic testcases for bugs shown up by confidential code)
"sometimes specific tests for fixed bugs are definitely cases of closing
the doors after the cows have fled", this is misleading since such
testcases may be necessary in future to debug problems with the fix; so
creating such testcases would not be as described there "work that
competes with everything else going on" but something at which a
reasonable effort *must* be made before a fix is installed in FSF GCC.

-- 
Joseph S. Myers
jsm28@cam.ac.uk