Re: [PATCH] Cleanup patches to build libgcj on x86/linux

[Mark Mitchell <mark at codesourcery dot com>]

>>>>> "David" == David Edelsohn <dje@watson.ibm.com> writes:

    David> 	My build now works with the patch where it failed
    David> yesterday.

Very well, I will check it in.

    David> 	I think that you need to test bootstraps on a wider
    David> variety of systems.

I'm not sure who "you" is, but I assume you mean "everyone"?  Or did
you mean Nathan, who checked in the patch, or myself personally?

Many people don't have access to more than one system.  This is
especially true of contributors that are not commercially affiiliated.
We've discussed this in the past, and I think that we agreed that one
system was a reasonable balance between carefulness and time spent.

It would be great if there were a multi-platform GCC build server that
we could submit patches to in order to verify a bootstrap on multiple
platforms before check in.  In fact, one could even automate this;
check ins would go on a branch, and be comitted to the trunk only if
the multi-platform build succeeded.  Failure would result in mail
being sent back to the contributor.

Without an automated system, though, I don't think we should impose a
multi-system bootstrap requirement.

In this particular case, I agree that it would have been a good idea
to test this patch on another system.  I take full responsibility for
approving the patch; I should have realized that the #ifdef's implied
a need for testing on an alternate platform.

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