Release criteria - code quality...

[Mark Mitchell]

>>>>> "Jan" == Jan Hubicka <hubicka@atrey.karlin.mff.cuni.cz> writes:

    Jan> Perhaps we can work together on making this suite usable for
    Jan> release criteria.  It contains number of small tests so it is
    Jan> quite easy to figure out whats wrong and yet I believe it is
    Jan> superrior in testing to the gzip/stephanov benchmark (at
    Jan> least it include stephanov benchmark and many pieces of code
    Jan> similar to gzip) It is also able to do comparisons in code
    Jan> length etc.

I think it's great that you're working on the benchmark suite.
However, there are already lots of experimental aspects to this
release -- and I don't want to add more.  I think the existing
benchmark criteria are good enough for minimum release criteria.

However, if the benchmark shows up serious regressions, then I would
certainly consider holding up the release long enough to understand
them.

Thanks,

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com