Re: Check-In Requirements

[Mark Mitchell <mark at codesourcery dot com>]

Thanks for elaborating on your proposal.

  This is not a rhetorical question: should we only be sending results
  to gcc-testsuite generated from pristine (w.r.t. the mainline and/or
  tagged releases) source trees?

That's a good question.  It's my opinion that the answer to that
question is yes, i.e., that we should only send in results from
pristine sources.  The reason is that the database of past results is
used to answer questions like "Did I break anything?" and "Does this
bug happen on Solaris too?".  If there are "random" builds in the
database, then it's much harder to use it in that way.

I think the benefits you mention for having people mail this stuff in
are probably real -- but I'm not sure at this point that they're worth
the effort.  I think that if we are going to do this we should have a
separate database for these test results.

Thanks again for your feedback.

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