[patch] for PR 18687

[Giovanni Bajo]

Zdenek Dvorak <rakdver@atrey.karlin.mff.cuni.cz> wrote:

>> Did you expect your change to have other side-effects?
>
> no, this should have been purely compile time improvement.  I should
> have time to check what's going on during this week.

This is also a consequence of ivopts having *no* tests in the testsuite.
We're letting large patches going in without any way to test they don't
cause performance regressions (stopping to do some optimization, causing
pessimization), and I'm sure it's hard for Zdenek to do e.g. visual
inspection of all SPEC to check that nothing pessimized.

I think it should be a high priority for 4.1 to setup an infrastructure for
regression-test ivopts (assuming that writing enough details in the dumps
and scanning for them like we do everywhere else is not enough).
-- 
Giovanni Bajo