[Testing] Announcing the first 8.2.1 release candidate.

[Carl Klitscher]

Got a bit of time testing some of this at the Wellington test session.

Access point is a wrt54g v1.1 running dd-wrt v23. tested using WEP and WPA 
only. XO is running candidate-800 as a soft upgrade from 8.2.1-25

WEP - Minor user error on session setup but once the connection was made then 
it was pretty solid. Consistently connected immediately after reboot. Forgot 
to test suspend/resume.

WPA - Had no luck at all getting TKIP to connect, just a constant recycle of 
the key request pop-up and no actual connection however given subsequent 
results this may not be a reflection on TKIP.

WPA with AES - Had switched from TKIP to AES to see if that worked and 
initially got nowhere either. No amount of entering key would initiate 
connection. AP symbol stayed as a lock. Cancelled all dialogs and went and 
got a coffee. Tried one more time to set up a connection (hover, click 
connect and enter AP key) and it all worked perfectly...
On reboot the AP symbol had changed from a lock to a star so we know the key 
but initial connect (hover/connect) failed asking for key again. Rebooted but 
no change... cancelled dialogs, waited a minute, reconnected perfectly. 
Rebooted, cancelled initial connection dialog, initiated connection manually 
and reconnected perfectly. All subsequent reboot/cancel/connect sequences 
worked. 
Suspend/resume. Lost all connections on s/r and the ap was never part of the 
resume sequence, only the mesh. First attempt at manual ap connection after 
resume failed but the second after the cancel/reconnect sequence worked.

Didn't get a chance to go back and test the TKIP connection with the 
cancel/reconnect sequence unfortunately but I suspect that given the 
encrypted ap connection is solid once it's up this is something about the 
initial dialog sequence that is the issue. Some field not being populated 
correctly right at the start??

Carl   

On Wednesday 11 February 2009 12:53, Chris Ball wrote:
> Some particular areas we'd like to see testing of:
>
> * Connecting to encrypted wireless access points.  We believe there are
>   regressions here that we're trying to understand.  The test procedure
>   is:
>   - connect to a WEP/WPA AP
>   - suspend/resume using the power button, see what happens
>   - reboot, see what happens
>
>   Some hypotheses we're trying to investigate are:
>   - WEP always works
>   - WPA works with some percentage chance of success