On Thu, Apr 19, 2012 at 7:59 PM, Adam Williamson <awilliam(a)redhat.com> wrote:
On Thu, 2012-04-12 at 08:02 +0900, Joel Rees wrote:
> On Thu, Apr 12, 2012 at 12:06 AM, Mark Haney <markh(a)abemblem.com> wrote:
> > On 04/11/2012 10:34 AM, Paul W. Frields wrote:
> >
> >>
> >> You can still login at this prompt with your 'root' account and
> >> password. From that point, you can look at /var/log/Xorg.0.log to see
> >> what happened that caused your X Window System to fail. If you want
> >> to post that for people here to look at and offer advice, please
> >> *don't* attach the file to your email. It will probably be too big
> >> and your message won't come through. Instead, post it somewhere like
> >>
http://fpaste.org and send a link to your paste here.
> >>
> >
> >
> > That's true it is a fairly generic question, however, the OP did state
he'd
> > tried to login with root and failed. Sounds to me like X wasn't the only
> > thing that is having issues.
> >
> > Although it could be the password he used. I noticed one time that the
> > password I was using simply wouldn't work on the initial install of Fedora
> > no matter how many times I installed it. It did work however after changing
> > it once I got it installed.
>
> Maybe keyboard definition issues?
>
> My hardware tends to be Japanese, and sometimes the difference in key
> positions has left me with a root password set assuming US English
> layout. Some of the punctuation keys move when the full system boots
> and the keyboard definition is correctly set. If I work out what moved
> where, I can log in.
>
> But sometimes it's easier to just boot single user and set the
> password again. (Don't have all the layouts memorized.)
>
> Lately, I am beginning to doubt the wisdom of always hiding the
> password when you're setting it, especially now that proper passwords
> are generally understood to be long and convoluted. It would sometimes
> be nice to have a "Debug keyboard" or "I've checked, nobody's
looking
> over my shoulder, and I need to see what I'm typing." button.
>
> Setting up a new system is, statistically speaking, sometimes going to
> require some debugging until we can put the WINTEL-pseudo-standard
> infected hardware behind us. (And I don't even see Apple trying to do
> that, now.)
>
> Of course, you can always try the keys that might have moved --
> ()[]{}"'=;:+*-_\| and so forth -- where you'd type a user name. You
> often have to think in reverse, of course, as in, "I thought I was
> typing left-bracket, what would that have been?"
Note that we actually have a test case which is run during validation
testing and is intended to ensure that the same keyboard layout is used
for setting passwords during installation and entering them
post-install, because we had a lot of this kind of trouble back before
we did that. To my knowledge we haven't had a major bug of this type
since F15 or so.
Well, maybe that doesn't get applied to some of the spins?
I'm pretty sure I ended up with keys moving on a password on a live
USB of the F16 security spin. If I notice it again and have time,
maybe I should file a bug?
--
Joel Rees