We should've known it was too good to last!
Martin Banas fortunately caught a major bug in RC2 overnight:
so we are currently testing RC3 which includes a fix for that. RC3 has
not technically been released yet, but the main images - DVD, netinst,
and release blocking live images - for both primary arches are already
built and in place, so we can get the important testing done.
The RC2->RC3 delta is very small; just a change in anaconda's text mode
software selection spoke, and an incidental change in comps which we
were forced to include as it's hard to back out of comps changes. So
most RC1 and RC2 testing should remain valid.
I have transferred the RC2 results to the RC3 matrices, leaving
strategic gaps for tests we really ought to re-run for safety. In
particular we should at least test that _every release blocking image_ -
those mentioned above - is bootable, under size, works with checksum
verification and isn't broken in some weird way. We should verify the
blocker fix (I've done that already) and run a few other text installs
just to test there's no other major borkage in text mode.
We're aiming to get essential testing done on RC3 today and, assuming it
passes and no other blockers are unearthed, go ahead and sign off on
release later today: the Go/No-Go meeting is being held open to allow
this. So if people could help out with the essential RC3 tests that'd be
really great. The ones left 'open' on the matrix are the key ones we
should run, plus any 'previous RC1 run' or 'previous RC2 runs' that we
can replace with an RC3 test is great. It's probably not necessary to
bother with Base or Desktop testing as they should be 100% unchanged
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora