I committed a fix last night so the identity manager will use the supplied
nai and password and not show any ui in this case, but convincing
it to not use the keyring is trickier; we're still working on a solution.
On 10/22/2013 11:16 AM, [log in to unmask] wrote:
> Let me know what you want to do... am happy to install debug versions, run any other bits... This is a 64-bit CentOS system.
>
> Stefan
>
>> -----Original Message-----
>> From: Sam Hartman [mailto:[log in to unmask]]
>> Sent: 22 October 2013 16:02
>> To: Paetow, Stefan (DLSLtd,RAL,LSCI)
>> Cc: [log in to unmask]; [log in to unmask]
>> Subject: Re: Moonshot-webp + default keyring
>>
>> Kevin, it looks like pam_gss is doing the right thing.
>> Namely, it's passing a username and password into
>> gss_acquire_cred_with_password, which means an identity and password
>> should already be given as an input to the ui.
>>
>> In that case, I'd expect us to look up a trust anchor if present, but
>> use the supplied identity.
>> Can you take a look?
>>
>> Also, the segfault is interesting.
>
|