I think that the mail daemon should get some attention. Right now if
yo uhave more that one IMAP adresses it will confuse itself and it
will mix up both IMAP accounts.
+1
I also 'fixed' the intel extreme driver in a way that breaks it for some
other people. I hope I'll find time and motivation to fix this before
r1a3 is out.
On a more general view about the release process, I still think we can't
keep going this way forever, planning alpha releases as we see fit, or
we are going to stay in alpha mode for 10 more years. Sure, it's a way
to try a different release process for each release, but we still have
no clear roadmap for going to R1. So a release pops out, then there are
6 month of void, and people start talking about the next one. The 6
month boundary seems to work fine for alpha stage, so why not settling
for it formally ? A release in november, after GSoC, with new features,
and one in may, before GSoC, that should be more stable :). Note I'm
usually more biased towards setting release on features and not dates,
but that doesn't apply for alphas.
The target should be R1, with a well-defined set of features. We can't
prevent getting more (Stack and tile comes to mind), but at least we
should have an estimate of what's left to do. Interpolating the features
in the two previous releases, we should then get a rough estimate of the
number of alphas needed to get all the features in. The goal of alphas
is to expose new features. For alpha 3, we'd get :
* Usual bugfixes (that doesn't count)
* POSIX Locale + some improvements on native Locale API (still WIP)
* API-Stable Layout system (not sure it is ABI-Stable yet)
* BNotification / infopopper (WIP)
* ext3 (experimental)
* stack and tile (no API yet)
* some network improvements I don't know enough about to talk
