On 02/28/2011 08:25 AM, Adam Williamson wrote:
On Sun, 2011-02-27 at 23:06 -0600, Jason D. Clinton wrote:
> , running 'startx'
> from a vt just gives me a blank background. I'm posting this
> from Xfce.
>
> Don't know the root cause but using the fallback force method posted
> by mclasen will get you in to fallback mode until it's fixed.
The root cause turned out to be at-spi2 breakage. Again.
I don't want this to come across the wrong way, but this:
[...]
really doesn't make me happy. See caillon's comment starting
"Note that
in order to get things working, one needs to disable a11y" - which makes
it clear that he knows, unless you manually disable a feature by using
an arcane command or by removing a package, GNOME will fail to run at
all...yet he gives the update +1 karma? This really isn't how the update
approval system is supposed to work. If you (corporate 'you', Fedora
desktop team) know about such a critical bug, you should fix it and edit
the update. Unless this gets fixed soon, we are in a state where anyone
who installs the Alpha and does an update will find their system
apparently utterly broken (it boots to a crashing gdm and if you boot to
'runlevel 3' and do startx, GNOME fails to run). I know we're in early
pre-release state and we don't guarantee anything, but it doesn't really
excuse submitting and then pushing to 'stable' code you know to be badly
broken.
I upgraded my home box to F15-pre-alpha over the weekend, first with yum
and when the result was a utterly broken heap thought oh well it's the
unsupported way, then reinstalled from scratch and got the very same
broken heap... and couldn't help thinking "we're about to release
THIS??"
Anaconda seems to default to enabling updates-testing, so unless the
testing-repo is in a reasonable shape at the time of alpha release
there's going to be a lot of grumpy testers.
- Panu -