Re: [gdm-list] GDM D-Bus on Solaris

["William Jon McCann" <mccann jhu edu>]

On 8/28/07, Brian Cameron <Brian Cameron sun com> wrote:
...
> Thanks for updating the gdm-gobject branch with my latest patch.  Now aside from
> the utmp code, it all builds and installs fine.  What are the plans for utmp?

I guess I'll merge the changes from trunk to make it work...

> Could you review the following bugs?  If the next bug should be fixed in 2.18,
> then it would be great if you could provide a patch.
>
>    http://bugzilla.gnome.org/show_bug.cgi?id=461056

Patch from reporter looks reasonsable to me if he has tested it and
can confirm it works.

> The next bug was introduced by some of the cleanup work you did in 2.19.  The
> problem was that you rewrote the way this code worked so it tied each
> server-foo section to a display.  I fixed it so it works as it did before.
>
> However, the code still is only reading in [server-foo] sections that
> are referenced in the [servers] section.  Really it should read in all
> [server-foo] sections so it knows what they are.  Currently gdmsetup is
> broken if you try to edit the [server-foo] section because of this.  I
> think you understand the configuration section better than I, so perhaps
> you could explain how this could be fixed?  I'd like to get this fixed
> in 2.18 if possible.
>
>    http://bugzilla.gnome.org/show_bug.cgi?id=462613
>
> You can refer to the SVN commit I did on 2007-08-09 to see the changes
> that I made to fix the bug partially.

Yeah, that is unfortunate.  However, my personal feeling is to just
leave it broken at this point.  I think we're kinda lucky that this
was one of the very few bugs that fell out of all that reworking.  I
think that the risk of fixing this might outweigh the benefit.  This
is all gone in the gobject branch anyway...

> We should probably also start looking at the GDM head daemon bugs that
> have been fixed since the fork.  I'd recommend looking over the patches
> in these bug reports.

It is helpful to include links to the bugs.

>     #470728
http://bugzilla.gnome.org/show_bug.cgi?id=470728

n/a  - we don't use this verify code

>     #457998
http://bugzilla.gnome.org/show_bug.cgi?id=457998

Yeah, I haven't done anything with the modules in the branch.  We
aren't yet setting the module list for the greeter.

>     #460407
http://bugzilla.gnome.org/show_bug.cgi?id=460407

n/a  - we don't have gdmsetup yet.

>     #443557
http://bugzilla.gnome.org/show_bug.cgi?id=443557

Hmm, this seems dubious at best.  I think the reporter was being too
generous by claiming there was some kind of agreement about this
variable.  I've read the entire thread and I can't find any kind of
spec for this nor any reasonable amount of buy in.  I don't think we
should support this as it is.

>     #349835
http://bugzilla.gnome.org/show_bug.cgi?id=349835

Yes.

>     #331059
http://bugzilla.gnome.org/show_bug.cgi?id=331059

n/a - no more (explicit) socket files

>     #434813
http://bugzilla.gnome.org/show_bug.cgi?id=434813

Well we don't yet have any of those tools in the branch...

>     #435552
http://bugzilla.gnome.org/show_bug.cgi?id=435552

n/a  - but we haven't added migration support to the branch yet

>     #436808
http://bugzilla.gnome.org/show_bug.cgi?id=436808

n/a

>     #436803
http://bugzilla.gnome.org/show_bug.cgi?id=436803

Yeah, I guess.

>     #436812
http://bugzilla.gnome.org/show_bug.cgi?id=436812

Yeah, I suppose.  But at the moment we don't handle PAM errors separately...

>     #436797
http://bugzilla.gnome.org/show_bug.cgi?id=436797

OK, but we don't have support for timed logins yet.

>     #462613 (the partial fix discussed above)
>     - The SVN commit I did on 2007-07-11 to fix CVE-2007-3381.
http://bugzilla.gnome.org/show_bug.cgi?id=462613

n/a - these are gone


Jon