Garrett Rooney wrote:
I've been holding off on rolling new release tarballs because wrowe
said he had some stuff he wanted to get in, but at this point I'm
caring less and less. What we've got is certainly an improvement over
previous versions, so I'm not sure there's any justification for not
releasing now.
That's a good point. I guess I aughta appologize, I've been spending too
much time overseeing the stdcxx podling, where bugs get fixed before folks
suggest to release. Different strokes...
Neither test is a hang/fail issue on APR 0.9 - the unix port didn't grow
an apr_os_uuid_get implementation until APR 1, and the testsockets issue
is specific to (and more complex in) 1.2 and trunk.
Only Solaris failure is Socket Creation sendto_receivefrom, errno 126.
No hang. OS/X is building now (better than good old 10.1, but yet and
still OS/X is good for a coffee/smoke break.)
Only Windows failures are Poll pollset_remove expected 5, saw 4, (will
glance at this - but with the major refactoring I'm not planning to dwell
in 0.9-land), the test_fail_read_flush failure Brian also observed and already
patched on apr trunk for OS2 - will port to Win32 tonight, and finally the
test_stat_eq_finfo which I need to reexamine as it's the source of some svn
issues.
So I reaffirm I'm -1 on 1.2.x branch and trunk at this moment. As for 0.9.x
branch, well, whatever. <shrug>
Bill
