> I think you could add a SPICE_CHANNEL_ERROR_AUTH_USER_AND_PASS (and a
> property "username" on the session)
This does not work because auth is expected to signal different error, for
example
SPICE_CHANNEL_ERROR_LINK.
___
Spice-devel mailing list
Spic
On Tue, Oct 22, 2013 at 08:36:55AM +, Dietmar Maurer wrote:
> > I don't think mandating that the username we use for SASL is the unix user
> > name
> > the client runs as makes a lot of sense. Imo it would be better if we
> > always asked
> > for username/password when SASL asks for it, and p
> I don't think mandating that the username we use for SASL is the unix user
> name
> the client runs as makes a lot of sense. Imo it would be better if we always
> asked
> for username/password when SASL asks for it, and prefill the username field
> with the local unix user name.
Exactly. I can
On Sat, Oct 19, 2013 at 11:53:55AM -0400, Marc-André Lureau wrote:
> > Why? Using SASL without username is a bit limited - or do I miss something?
>
> Current SASL implementation in Spice doesn't support username.
>
> > And you can still use tickets if you check for strlen(username) == 0 ?
>
> T
> Current SASL implementation in Spice doesn't support username.
>
> > And you can still use tickets if you check for strlen(username) == 0 ?
>
> There is noneed to ask the user a username for the current methods. He
> wouldn't do what to put there.
Sure. That is why I asked if you would accept
- Original Message -
> > > Would you accept a patch for that, or do you think above code is all we
> > > need?
> >
> > We shouldn't ask username for the current auth methods.
>
> Why? Using SASL without username is a bit limited - or do I miss something?
Current SASL implementation in
> > Would you accept a patch for that, or do you think above code is all we
> > need?
>
> We shouldn't ask username for the current auth methods.
Why? Using SASL without username is a bit limited - or do I miss something?
And you can still use tickets if you check for strlen(username) == 0 ?
_
Hi
- Original Message -
>
>
> Current remote viewer only ask for a password when connecting to a spice
> session
>
> which requires auth:
>
>
>
> from virt-viewer-session-spice.c:
>
>
>
> case SPICE_CHANNEL_ERROR_AUTH:
>
> DEBUG_LOG("main channel: auth failure (wrong password?)")
Current remote viewer only ask for a password when connecting to a spice session
which requires auth:
from virt-viewer-session-spice.c:
case SPICE_CHANNEL_ERROR_AUTH:
DEBUG_LOG("main channel: auth failure (wrong password?)");
int ret = virt_viewer_auth_collect_credentials(self
