On 03/02/12 01:04, Timo Sirainen wrote:
> The difference between your previously working system and currently
> working system is the GSSAPI/Kerberos libraries.
Just to close this thread off, seems that the bug was in the cyrus-sasl
libraries that kmail uses. Reverting from the 2.1.25 version t
On Fri, 2012-03-02 at 00:57 +1300, Mark Davies wrote:
>
> On 03/02/12 00:52, Timo Sirainen wrote:
> >> so what bit of the code should I be looking at to see what happens
> >> between the "security context state completed" and the "client out"?
> >
> > All of the code is in mech-gssapi.c
>
> Yes,
On 03/02/12 00:52, Timo Sirainen wrote:
>> so what bit of the code should I be looking at to see what happens
>> between the "security context state completed" and the "client out"?
>
> All of the code is in mech-gssapi.c
Yes, I'm just trying to work out the flow of the calls in and out of ther
On Fri, 2012-03-02 at 00:49 +1300, Mark Davies wrote:
> so what bit of the code should I be looking at to see what happens
> between the "security context state completed" and the "client out"?
All of the code is in mech-gssapi.c
Are these working vs. non-working Dovecots in same or different se
On 02/29/12 23:41, Timo Sirainen wrote:
>> Oh, yes, this is a bug in Dovecot..
>
> Hmm. Or it looked like a bug, since it replied only with "+", so I thought
> all auth mechanisms would have such a bug, but no.. So I'm not really sure
> why it's not sending more data. I don't have a Kerberos s
On 29.2.2012, at 12.34, Timo Sirainen wrote:
> On 29.2.2012, at 12.15, Mark Davies wrote:
>
>> 1 AUTHENTICATE GSSAPI YIICWgYJKoZIhvcSAQICAQBugg[...]wfuKg4VUptzPwb\r\n
>>
>> and receives
>>
>> + \r\n
>>
>> from dovecot, which it doesn't like and reports
>> clientAuthenticate: sasl_client_step f
On 29.2.2012, at 12.15, Mark Davies wrote:
> 1 AUTHENTICATE GSSAPI YIICWgYJKoZIhvcSAQICAQBugg[...]wfuKg4VUptzPwb\r\n
>
> and receives
>
> + \r\n
>
> from dovecot, which it doesn't like and reports
> clientAuthenticate: sasl_client_step failed with: -1
Oh, yes, this is a bug in Dovecot..
On 02/28/12 00:11, Timo Sirainen wrote:
>>> Looks like kmail is sending some kind of garbage to Dovecot. Set
>>> auth_debug_passwords=yes to make Dovecot log the auth traffic.
>>
>> Yeah, I did a network trace and it seems kmail is not sending the
>> full authentication request before trying to c
On 27.2.2012, at 12.50, Mark Davies wrote:
> On 02/27/12 23:32, Timo Sirainen wrote:
>> Looks like kmail is sending some kind of garbage to Dovecot. Set
>> auth_debug_passwords=yes to make Dovecot log the auth traffic.
>
> Yeah, I did a network trace and it seems kmail is not sending the
> full a
On 02/27/12 23:32, Timo Sirainen wrote:
> Looks like kmail is sending some kind of garbage to Dovecot. Set
> auth_debug_passwords=yes to make Dovecot log the auth traffic.
Yeah, I did a network trace and it seems kmail is not sending the
full authentication request before trying to carry on.
>
On Mon, 2012-02-27 at 14:17 +1300, Mark Davies wrote:
> Feb 27 13:42:18 bats dovecot: auth:
> GSSAPI(m...@ecs.vuw.ac.nz,130.195.5.88): Invalid base64 data in
> continued response
Looks like kmail is sending some kind of garbage to Dovecot. Set
auth_debug_passwords=yes to make Dovecot log the auth
Hi,
have dovecot 2.0.13 running fine on a NetBSD box, with lots of clients
happily authenticating to it with GSSAPI.
I've just updated an ArchLinux box to latest new version which has meant
its switched from using heimdal to using mit krb5 and its kmail is now
also slightly newer.
Thunderbird s
12 matches
Mail list logo
