On Mon, 2014-11-03 at 15:29 -0500, Fred Smith wrote:
I installed TC-3 in a VirtualBox VM (latest vbox release, 64-bit)
t'other
day. Mostly it went quite well with one really oddball exception. I don't
know if it's a general problem (haven't seen other mentions of it)
or not, but it's weird enough I thought to mention it here.
When creating a user in Anaconda, I entered the password twice, as prompted,
and it said they didn't match. so I did it again. and again. and again. It
insisted they didn't match.
tried the same thing for the root password and got the same result.
So, I entered a one-character password, in both fields, and it did not
say they didn't match. so I tried two chars. ok so far. by working my way
up to the full password, one char at a time in both fields, it allowed me
to get the whole thing in successfully. This worked for both my username
and for root.
Never seen any such thing before!
If anyone wants me to, I can go try more installs and see if I can
reproduce. If I CAN, what sorts of extra info would be helpful when i
report it?
That's, um, a rather strange one. It's hard to know what kind of info is
best apart from the obvious 'how can you reproduce it' - AFAIK those are
pretty standard GTK+ text entry boxes and all anaconda does is compare
their contents.
So basically the bug is this:
1. type 'correcthorsebatterystaple' in box 1
2. type 'correcthorsebatterystaple' in box 2
NO MATCH
1. type 'c' in box 1
2. type 'c' in box 2
3. type 'o' in box 1
4. type 'o' in box 2
5. type 'r' in box 1
6. type 'r' in box 2
(SOME TIME LATER)
49. type 'e' in box 1
50. type 'e' in box 2
MATCH
right? I guess the other info is whether it depends on the actual
password used, and if so, what's a password that triggers it.
I definitely haven't seen that in my F21 testing. I tend to use the
password '111111', but I do use 'correcthorse@"' sometimes for
keyboard
layout tests.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net