Nabble - Gdm-list

Re: Request for feedback: Proposed enhancements to GDM greeter interface

2009-12-04T17:53:59Z

Robert:

> I am proposing to do the following work for GDM 2.29 on the greeter code:
> 1. Produce documentation on the interface between the greeter and the
> server

This would be very useful. It might be useful to add a chapter to the
GDM docs in gdm/docs/C/gdm.xml to explain this. I believe the daemon
and the greeter primarily communicate via D-Bus calls.

However, there is probably some interactions that are assumed that
greeters will do. For example, how the greeter reacts when GConf
keys are modified, or when files like /etc/passwd are modified.

> 2. Move code / produce a greeter library to minimise the amount of code
> in the simple-greeter

This sounds like a good idea. Some real thought will need to go into
item #2 above. For example, should this library only include interfaces
that are GUI-agnostic, such as how D-Bus works in the greeter? Or
should the library contain some GUI elements such as the panel or the
face browser that might be shared across multiple greeters?

> 3. Produce bindings to the new interface to allow flexible greeter
> development (e.g. Python/Javascript).

Neat.

> These changes will allow the following improvements:
> - New greeters will be able to be developed without limitations - thus
> making GDM completely themeable

There are some "limitations" that are imposed by accessibility. So, we
should think about how providing accessibility might affect the
interfaces provided to write a greeter. We should make it easy for
users to write an accessible greeter.

The old GDM 2.20 gdmgreeter technique of using libgnomecanvas was never
very accessible. Its GUI elements did not support ATK, and since it ran
without a window manager it did not work well with struts based programs
like GOK. Hopefully any themed greeters that we develop in the future
will have better accessibility support than gdmgreeter had.

Also, there are some "limitations" that are imposed by PAM. For
example, some people have suggested it would be nice to display both the
username and password prompts together on the same dialog, but this
would not be possible without significant enhancements to how PAM works.

> - A greeter can be developed which is easy for artists to them (e.g.
> by embedding webkit and using using CSS for theming)

I encourage people to experiment and try new interfaces for the greeter.

However, remember that the login greeter is a program that people with
access to the machine can use without authentication, and the greeter is
used to manage security sensitive information. So, care should be taken
to ensure that the greeter does not inadvertently give the user any
unintended privileges or access to the machine.

I am all for making GDM easier for artists to use, but we need to be
careful to do this in a way that is also easy to keep maintainable and
secure.

> - Greeters could be developed for varying hardware to make use of
> graphics performance (i.e. flashy 3d effects)

Yes, a clutter-based greeter would be interesting. Though I'd like
to see how gnome-shell deals with a11y before deciding the best way
to go about a clutter-based login GUI.

> - Other applications will be able to access GDM information (e.g.
> application switchers)

Again, care needs to be taken to ensure that other applications do not
gain access to information that is inappropriate. Since the main GDM
daemon runs as root, there are certain functionalities that greeters
are allowed to do that normal user programs should not be able to do.
Such as how Xauth keys are managed, for example.

I believe GDM is already well integrated into the User Switch applet.
ConsoleKit already provides programs like the User Switch applet far
more useful information than the old GDM 2.20 supported via its
private gdmflexiserver sockets interface.

I would be interested to hear what specific interactions between the
GDM daemon and login greeter you think would be useful to provide
to other applications. Do you have any use cases you are thinking
about here?

> - GDM will be usable by other desktops (KDE, XFCE)

I am not sure I follow what you mean here. Could you elaborate?

Brian
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Request for feedback: Proposed enhancements to GDM greeter interface

2009-12-03T20:53:14Z

Hi all,

I am proposing to do the following work for GDM 2.29 on the greeter code:
1. Produce documentation on the interface between the greeter and the server
2. Move code / produce a greeter library to minimise the amount of code
in the simple-greeter
3. Produce bindings to the new interface to allow flexible greeter
development (e.g. Python/Javascript).

These changes will allow the following improvements:
- New greeters will be able to be developed without limitations - thus
making GDM completely themeable
- A greeter can be developed which is easy for artists to them (e.g.
by embedding webkit and using using CSS for theming)
- Greeters could be developed for varying hardware to make use of
graphics performance (i.e. flashy 3d effects)
- Other applications will be able to access GDM information (e.g.
application switchers)
- GDM will be usable by other desktops (KDE, XFCE)

Please provide some feedback!

For more details see the Ubuntu blueprint:
https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-gdm-custom-greeter-support

--Robert Ancell

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: How to remove a specific user from the greeter list

2009-11-30T10:39:36Z

The user docs are probably easier to read:

http://library.gnome.org/admin/gdm/2.29/configuration.html.en#daemonconfig

Note the new IncludeAll, Include, and Exclude configuration options
described in section 5.4.4 of the GDM User's Manual.

This feature is only available in the latest unstable 2.29.0 release, so
don't expect this configuration option to work with older versions. You
can check which version you are using by running with one of these
commands:

gdmflexiserver --command="VERSION"
gdmfelxiserver --version

Brian

Ray Strode wrote:

> Hi,
>
>> I want to remove a specific user from the greeter list in gdm.
>> But I do not want to delete the user, nor to disable the greeter list
>> completely.
>>
>> Is this possible?
>
> Yes, in recent versions, see:
>
> http://git.gnome.org/cgit/gdm/commit/?id=14f686a863dacffd1a192a309b3ab0f4cbc18b22
>
> --Ray
> _______________________________________________
> gdm-list mailing list
> gdm-list@...
> http://mail.gnome.org/mailman/listinfo/gdm-list

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: How to remove a specific user from the greeter list

2009-11-29T13:33:09Z

Hi,

> I want to remove a specific user from the greeter list in gdm.
> But I do not want to delete the user, nor to disable the greeter list
> completely.
>
> Is this possible?

Yes, in recent versions, see:

http://git.gnome.org/cgit/gdm/commit/?id=14f686a863dacffd1a192a309b3ab0f4cbc18b22

--Ray
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

How to remove a specific user from the greeter list

2009-11-29T03:01:01Z

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear list,

I want to remove a specific user from the greeter list in gdm.
But I do not want to delete the user, nor to disable the greeter list
completely.

Is this possible?

Thanks,
Jan.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAksSVG0ACgkQvuFuLCp9giA1bwCeOeY0WdTI36MG4n6UoIFyD8bx
RQQAn1dQ8hhuRiloGDm2twChr+bh8Rck
=J/GT
-----END PGP SIGNATURE-----

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: User feedback on multi-seat gdm and ConsoleKit

2009-11-24T07:47:43Z

Hello,

First, thanks to everyone for the responses. My followup:

On Mon, 23 Nov 2009, Halton Huo wrote:

> On Sat, 2009-11-21 at 16:15 -0800, Wayne Whitney wrote:
>
>> 1) VT switching breaks things.

OK, as others have pointed out, the VT system should just apply to the
primary console display. That still leaves me with a few issues:

1) If I switch VTs, sometimes GDM spawns a third X process (:2), using the
primary .display file. It is as if X respawns prematurely. One time I
had three Xs going on the primary display, on VTs 7, 8 and 9.

2) If switch VT to a text console and back to X, for some reason X on the
primary display starts accepting input from both keyboards and mice. The
only "Device reopened" messages in Xorg.log refer to the desired primary
input devices, so I have no idea why this is happening.

3) I need to figure out how to tell the kernel to ignore input from the
secondary keyboard and mouse, for the case that I have a text console on
the primary display.

4) I still have the problem that logging out dumps me at VT1 instead of
the respawned X. I'm not actually clear on the timing of the X
respawning, it seems like it respawns if I switch VTs (e.g. if I switch to
VT2 from VT1, it respawns and switches VTs for me).

>> 3) My .display files have $vt as a command line option, but when I use ps
>> to look at the Xorg commands run, there is no vtX option passed to Xorg.
>
> This is strange, could you move $vt at the end of Exec line and try
> again? (I'm not sure whether Xorg accept "/usr/bin/Xorg $display $vt
> -layout Layout0 -br -verbose -auth $auth -nolisten tcp")
>
> To be simple, you could try the single session with your
> 00-primary.seat. To disable your second seat, you can switch the key
> Hidden to true and restart your console-kit-daemon and gdm.

Neither of these had any effect. I still don't get any vtX option on the
command line as reported by ps. I seem to recall the Case2 testing files
did work as far as the vt command line option. I also realized that I
don't actually want my secondary display to be on a different VT from the
primary; if it were set up that way, only one of them would be running at
once, depending on which VT is active.

>> 5) I don't know how to configure the greeters I end up with

Is there an alternative to gdm-simple-greeter I could use?

> I do not think disable sould for greeter is supported, should be an
> enhancement for GDM.

Well, I tried the following, not sure if it works yet:

gconftool-2 --config-source=xml:readwrite:/var/lib/gdm/.gconf --set /desktop/gnome/sound/event_sounds --type bool false

Also, I notice that gdm is spawning an instance of pulseaudio, I'm not
sure where that happens and if I can disable it.

> Are you saying you can not use tab to switch the "Login" button? If so,
> should be an A11y bug for GDM.

No, I can use tab, that is an OK workaround. I would just prefer a
greeter that starts off with a "Username:" prompt.

Thanks, Wayne Whitney
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Can I disable progress bar when shutting down?

2009-11-23T18:44:32Z

Can I disable progress bar when shutting down?
My graphics card shows up a mess when shutting down. It's really really ugly.
Thank you.

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: User feedback on multi-seat gdm and ConsoleKit

2009-11-23T09:59:40Z

Wayne:

> This works, in that I get two separate X processes, two greeters, and
> two people can login. There are a few rough edges, though:
>
> 1) VT switching breaks things. I'm not really clear how this should
> work; I would think the VTs should be per seat, although this would
> require multiconsole in the kernel. Anyway VT switching appears to just
> affect the first seat. Ctrl-Alt-Fn to switch from either keyboard
> causes the first seat to switch VTs; if I go to a text console, then
> that console get keystrokes from both keyboards. Worst is that if I
> then switch back to VT7, I get the X session back on the primary seat,
> but it is now also getting keystrokes from both keyboards.
>
> 4) Audio device ownership isn't getting set at all from either seat.

My understanding is that VT switching only works on the console display,
and not on multiple attached displays. Did VT switching on multiple
displays ever work for you with the old GDM, or in any other
configuration?

Regarding your issue with a session accepting keystrokes from multiple
keyboards and audio device permissions. Note that different distros
manage device allocation and permissions differently, and the Xserver
configuration also plays a role in how this works. So, these sorts of
problems may not be closely related to GDM.

> 5) I don't know how to configure the greeters I end up with; I like to
> disable sound and avoid having to click "Login" before entering my
> username.

The new GDM always shows the "Login" button. There is currently no way
to configure it to not show that button. Also, the new GDM does not
currently provide any configuration options to disable sound.

The GDM Init script gets run before displaying the GUI, so if Fedora
provides CLI to turn off sound, you could run it from there.

Brian
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: User feedback on multi-seat gdm and ConsoleKit

2009-11-23T00:45:45Z

On Mon, 23 Nov 2009, Halton Huo wrote:

> On Sat, 2009-11-21 at 16:15 -0800, Wayne Whitney wrote:

>> 4) Audio device ownership isn't getting set at all from either seat.
> Current ConsoleKit does not finish implementation to assign the devices
> for given seat. Sorry that I can not help you since I have bare
> knowledge about the hardware thing.

I do not know what's the current exact plan to proceed with these h/w
access issues, but in princible the place where these are to be handled is
related to udev. I personally deal the soundcard issue with udev rules
that assign the specific pcm devices to specific users because they are
always the same. In the future, udev would assign certain devices to
certain users based on the seat information, one way or other.

--
Jukka
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: User feedback on multi-seat gdm and ConsoleKit

2009-11-22T23:00:17Z

On Sat, 2009-11-21 at 16:15 -0800, Wayne Whitney wrote:

> Hello,
>
> I'm a long time multi-seat home user, since Fedora 7 at least. I recently
> updated to Fedora 12, which broke the gdm-2.20 based solution I had been
> using. So I decided to try the current multi-seat branches of gdm and
> ConsoleKit, and I'm pleased to report partial success. I'm sure I'm doing
> some things wrong, though, since I'm used to the previous solution; any
> guidance would be very welcome.
>
> First let me say that I have an x86_64 machine with two ATI video cards,
> two keyboards, and two mice. I have configured xorg with two different
> layouts, each one using one keyboard, one mouse, and one video card (via
> IsolateDevice). So all I really need gdm to do is run two instances of X
> simultaneously and provide each with a greeter.
>
> After figuring out the rudiments of git, I grabbed the appropriate
> branches of gdm and ConsoleKit, configured them to match the Fedora RPM
> configurations, and then installed them over the existing RPMS. I was
> able to complete the first two testings cases listed on the page
> <http://wiki.genunix.org/wiki/index.php/testing_for_newgdm_consolekit_multiseat_multidisplay>.

Good to know you find the right place.

>
> Then I decided to set up a simple configuration with two seat files in
> seats.d, each with their own session file in sessions.d, and their own
> display file in displays.d. I've attached them at the end, stripped of
> comments. The main difference is that the display file for the secondary
> seat runs Xorg with a different -layout option and with -sharevts and
> -noswitch.
Your configurations seem good to me.

>
> This works, in that I get two separate X processes, two greeters, and two
> people can login. There are a few rough edges, though:
>
> 1) VT switching breaks things. I'm not really clear how this should work;
> I would think the VTs should be per seat, although this would require
> multiconsole in the kernel. Anyway VT switching appears to just affect
> the first seat. Ctrl-Alt-Fn to switch from either keyboard causes the
> first seat to switch VTs; if I go to a text console, then that console get
> keystrokes from both keyboards. Worst is that if I then switch back to
> VT7, I get the X session back on the primary seat, but it is now also
> getting keystrokes from both keyboards.

Could be possible with your #3 issue.

>
> 2) I can logout of the secondary seat and log back in fine. However, for
> some reason when I first log out of the primary seat, X gets respawned,
> but I end up on VT1 instead of VT7. So I have to manually switch back to
> VT7. Unfortunately this has the effect of tripping the first problem.
X respawned is desired behavior. But should be on the same VT and
display you have configuraed in your .session. Should be related with
your #3 issue also.

>
> 3) My .display files have $vt as a command line option, but when I use ps
> to look at the Xorg commands run, there is no vtX option passed to Xorg.
This is strange, could you move $vt at the end of Exec line and try
again? (I'm not sure whether Xorg accept "/usr/bin/Xorg $display $vt
-layout Layout0 -br -verbose -auth $auth -nolisten tcp")

To be simple, you could try the single session with your
00-primary.seat. To disable your second seat, you can switch the key
Hidden to true and restart your console-kit-daemon and gdm.
>
> 4) Audio device ownership isn't getting set at all from either seat.
Current ConsoleKit does not finish implementation to assign the devices
for given seat. Sorry that I can not help you since I have bare
knowledge about the hardware thing.
>
> 5) I don't know how to configure the greeters I end up with; I like to
> disable sound
I do not think disable sould for greeter is supported, should be an
enhancement for GDM.

> and avoid having to click "Login" before entering my
> username.
Are you saying you can not use tab to switch the "Login" button? If so,
should be an A11y bug for GDM.

>
> Yours, Wayne Whitney
>
>
> %%% 00-primary.seat %%%
> [Seat Entry]
> Version=1.0
> Name=Primary seat
> ID=StaticSeat0
> Description=Wayne's Desk
> Hidden=false
> Devices=
> Sessions=Local0;
>
> %%% 01-secondary.seat %%%
> [Seat Entry]
> Version=1.0
> Name=Secondary seat
> ID=StaticSeat1
> Description=Diane's desk
> Hidden=false
> Devices=
> Sessions=Local1;
>
> %%% Local0.session %%%
> [Session Entry]
> Name=Local0
> Type=LoginWindow
> Description=Local Login Screen
> DisplayTemplate=Local0
> [Local]
> display=:0
> vt=vt7
>
> %%% Local1.session %%%
> [Session Entry]
> Name=Local1
> Type=LoginWindow
> Description=Local Login Screen
> DisplayTemplate=Local1
> [Local]
> display=:1
> vt=vt8
>
> %%% Local0.display %%%
> [Display]
> Type=X11
> [X11]
> Exec=/usr/bin/Xorg $display $vt -layout Layout0 -br -verbose -auth $auth -nolisten tcp
>
> %%% Local1.display %%%
> [Display]
> Type=X11
> [X11]
> Exec=/usr/bin/Xorg $display $vt -layout Layout1 -sharevts -novtswitch -br -verbose -auth $auth -nolisten tcp
>
>
>
>
> _______________________________________________
> gdm-list mailing list
> gdm-list@...
> http://mail.gnome.org/mailman/listinfo/gdm-list

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

gdm Autologin GUI mode

2009-11-22T12:39:18Z

Hi All,

I have filed for an enhancement of AutoLogin mode in gdm with the
Keyboard/Language explicitly popped up for the user. This is different
than the TimedLogin mode in that it aims to provide the user with
explicitly popped up keyboard/language menu items before Auto log in her
to gnome desktop.

Pl. take a look at the issue filed at
https://bugzilla.gnome.org/show_bug.cgi?id=602663
regarding the need for a AutoLogin GUI mode for gdm and add your
comments and suggestions in there.

Regards,
Suresh
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

User feedback on multi-seat gdm and ConsoleKit

2009-11-21T16:15:18Z

Hello,

I'm a long time multi-seat home user, since Fedora 7 at least. I recently
updated to Fedora 12, which broke the gdm-2.20 based solution I had been
using. So I decided to try the current multi-seat branches of gdm and
ConsoleKit, and I'm pleased to report partial success. I'm sure I'm doing
some things wrong, though, since I'm used to the previous solution; any
guidance would be very welcome.

First let me say that I have an x86_64 machine with two ATI video cards,
two keyboards, and two mice. I have configured xorg with two different
layouts, each one using one keyboard, one mouse, and one video card (via
IsolateDevice). So all I really need gdm to do is run two instances of X
simultaneously and provide each with a greeter.

After figuring out the rudiments of git, I grabbed the appropriate
branches of gdm and ConsoleKit, configured them to match the Fedora RPM
configurations, and then installed them over the existing RPMS. I was
able to complete the first two testings cases listed on the page
<http://wiki.genunix.org/wiki/index.php/testing_for_newgdm_consolekit_multiseat_multidisplay>.

Then I decided to set up a simple configuration with two seat files in
seats.d, each with their own session file in sessions.d, and their own
display file in displays.d. I've attached them at the end, stripped of
comments. The main difference is that the display file for the secondary
seat runs Xorg with a different -layout option and with -sharevts and
-noswitch.

This works, in that I get two separate X processes, two greeters, and two
people can login. There are a few rough edges, though:

1) VT switching breaks things. I'm not really clear how this should work;
I would think the VTs should be per seat, although this would require
multiconsole in the kernel. Anyway VT switching appears to just affect
the first seat. Ctrl-Alt-Fn to switch from either keyboard causes the
first seat to switch VTs; if I go to a text console, then that console get
keystrokes from both keyboards. Worst is that if I then switch back to
VT7, I get the X session back on the primary seat, but it is now also
getting keystrokes from both keyboards.

2) I can logout of the secondary seat and log back in fine. However, for
some reason when I first log out of the primary seat, X gets respawned,
but I end up on VT1 instead of VT7. So I have to manually switch back to
VT7. Unfortunately this has the effect of tripping the first problem.

3) My .display files have $vt as a command line option, but when I use ps
to look at the Xorg commands run, there is no vtX option passed to Xorg.

4) Audio device ownership isn't getting set at all from either seat.

5) I don't know how to configure the greeters I end up with; I like to
disable sound and avoid having to click "Login" before entering my
username.

Yours, Wayne Whitney

%%% 00-primary.seat %%%
[Seat Entry]
Version=1.0
Name=Primary seat
ID=StaticSeat0
Description=Wayne's Desk
Hidden=false
Devices=
Sessions=Local0;

%%% 01-secondary.seat %%%
[Seat Entry]
Version=1.0
Name=Secondary seat
ID=StaticSeat1
Description=Diane's desk
Hidden=false
Devices=
Sessions=Local1;

%%% Local0.session %%%
[Session Entry]
Name=Local0
Type=LoginWindow
Description=Local Login Screen
DisplayTemplate=Local0
[Local]
display=:0
vt=vt7

%%% Local1.session %%%
[Session Entry]
Name=Local1
Type=LoginWindow
Description=Local Login Screen
DisplayTemplate=Local1
[Local]
display=:1
vt=vt8

%%% Local0.display %%%
[Display]
Type=X11
[X11]
Exec=/usr/bin/Xorg $display $vt -layout Layout0 -br -verbose -auth $auth -nolisten tcp

%%% Local1.display %%%
[Display]
Type=X11
[X11]
Exec=/usr/bin/Xorg $display $vt -layout Layout1 -sharevts -novtswitch -br -verbose -auth $auth -nolisten tcp

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Multi-seat issues with GDM 2.16

2009-11-16T08:50:23Z

Alexander:

gdmlogin and gdmgreeter pretty much do the same thing, so it is unusual
for one to crash and the other to work. When I have seen problems like
this, it is caused because:

- There is some library that gdmgreeter needs that is not available on
the system. If you type "ldd /usr/lib/gdmgreeter", and if any of the
libraries report problems (like they cannot be found), then this is
obviously a problem.

- If gdmgreeter is crashing for some reason.

Try this. Using gdmlogin, log into a normal session and set this
environment variable: "DOING_GDM_DEVELOPMENT=1". Then run gdmgreeter
from a terminal window within the normal user session.

This lets you run a debug version of gdmgreeter. Some of the features
are turned off, so it is not an exact test of what code is executed when
it is launched via the main GDM daemon. However, it is a good test to
run. If gdmgreeter crashes on startup when DOING_GDM_DEVELOPMENT is
set, then it is often easier to track down why.

If this doesn't highlight the problem (and if turning on GDM debug does
not help), then it would likely be necessary to build a debug version of
GDM, add a sleep() statement to gdmgreeter when it starts up, and attach
a debugger to the process to see why things are going awry.

Brian

> I'm experimenting with multi-seat configuration using DisplayLink devices as per
> instructions on their wiki [1].
>
> The displaylink kernel and Xorg drivers compiled and appear to work, the udev
> scripts on the wiki needed slight change to make it work for me and only the
> last part doesn't seem to work.
>
> When the system boots in runlevel 3 I start /usr/sbin/gdm manually and execute
> ./usbseat.sh 2 to start X and gdm login screen on the terminal with ID 2.
>
> The script uses the following commands:
>
> /usr/bin/gdmdynamic -v -t 2 -s 1 -a "$1=/usr/bin/X :$1 -br -audit 0 -nolisten
> tcp -novtswitch -sharevts -config $TMPFILE"
> /usr/bin/gdmdynamic -v -r $1
>
> Where the TMPFILE is a Xorg config file with proper grouping for display, mouse
> and keyboard.
>
> Xorg starts (I've tested it manually too). The mouse pointer is displayed using
> the default theme for my distro (RHEL5.4), there's a blue circle around the
> pointer for a while indicating that stuff is loading.
>
> Then all of this shows only a black screen and the mouse pointer. It doesn't
> show the login screen as one expects.
>
> The pstree output looks ok:
> ├─gdm-binary───gdm-binary─┬─X
> │ └─gdmgreeter
>
> IIRC gdmgreeter is the application that shows the login screen.
>
> The associated Xorg and gdm logs don't indicate any warnings or errors.
>
> Does anyone have an idea how to debug this or why it's not working? I'm using
> 2.16 which is relatively old and multi-seat support is reported to work on older
> versions.
>
> [1] - http://libdlo.freedesktop.org/wiki/MultiSeatTerminal
>
> Thanks,
> Alexander.
> _______________________________________________
> gdm-list mailing list
> gdm-list@...
> http://mail.gnome.org/mailman/listinfo/gdm-list

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Multi-seat issues with GDM 2.16

2009-11-16T05:37:33Z

Alexander Todorov wrote:
>
> Xorg starts (I've tested it manually too). The mouse pointer is displayed using
> the default theme for my distro (RHEL5.4), there's a blue circle around the
> pointer for a while indicating that stuff is loading.
Adding
Greeter=/usr/libexec/gdmlogin

to gdm.conf fixed it for me. I'm seeing a login prompt and I'm able to login.

--
Alexander.
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Multi-seat issues with GDM 2.16

2009-11-16T03:39:36Z

Hello,
I'm experimenting with multi-seat configuration using DisplayLink devices as per
instructions on their wiki [1].

The displaylink kernel and Xorg drivers compiled and appear to work, the udev
scripts on the wiki needed slight change to make it work for me and only the
last part doesn't seem to work.

When the system boots in runlevel 3 I start /usr/sbin/gdm manually and execute
./usbseat.sh 2 to start X and gdm login screen on the terminal with ID 2.

The script uses the following commands:

/usr/bin/gdmdynamic -v -t 2 -s 1 -a "$1=/usr/bin/X :$1 -br -audit 0 -nolisten
tcp -novtswitch -sharevts -config $TMPFILE"
/usr/bin/gdmdynamic -v -r $1

Where the TMPFILE is a Xorg config file with proper grouping for display, mouse
and keyboard.

Xorg starts (I've tested it manually too). The mouse pointer is displayed using
the default theme for my distro (RHEL5.4), there's a blue circle around the
pointer for a while indicating that stuff is loading.

Then all of this shows only a black screen and the mouse pointer. It doesn't
show the login screen as one expects.

The pstree output looks ok:
├─gdm-binary───gdm-binary─┬─X
│ └─gdmgreeter

IIRC gdmgreeter is the application that shows the login screen.

The associated Xorg and gdm logs don't indicate any warnings or errors.

Does anyone have an idea how to debug this or why it's not working? I'm using
2.16 which is relatively old and multi-seat support is reported to work on older
versions.

[1] - http://libdlo.freedesktop.org/wiki/MultiSeatTerminal

Thanks,
Alexander.
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: [ConsoleKit] Multi-seat update

2009-11-15T18:10:23Z

Ray,

Thanks for your response. To save your time, you could review the
changes later than your last commit.

ConsoleKit(multi-seat) : 0b17fb32027beffa658c8ac866c2acbcb3bee603
GDM(display-configuration): a9cf528e725f7553dcd8fd9c07c54a87fd383869

Cheers,
Halton

On Fri, 2009-11-13 at 10:43 -0500, Ray Strode wrote:

> Hi,
>
> On Thu, Nov 12, 2009 at 9:53 PM, Halton Huo <Halton.Huo@...> wrote:
> > We have sees several emails to the consolekit and gdm-list mailing lists
> > where people have asked for multi-seat and multi-display support.
> > Obviously this feature is very important. Currently the patches are very
> > long and difficult to maintain for applying on master and released
> > tarballs.
> Right, we should get these integrated. There was a bit of a hiatus
> because we knew they weren't going to make 2.28 so priorities shifted.
>
> > Ray and Jon, could you please review these changes as soon as you can?
> > If so, it would be very much appreciated.
> >
> > Note the schedule:
> >
> > http://live.gnome.org/TwoPointTwentyseven/
> >
> > - 2.29.2 is due on November 16th
> > - 2.29.3 on November 30th.
> > - 2.29.4 on December 21st
> > - 2.29.5 on January 11th and this is string freeze.
> >
> > We really, at the very latest should get this into 2.29.5 if multi-seat
> > support is going to be delivered with GNOME 2.30. Considering that this
> > feature should be soak tested for as much time as possible, it would be
> > best to integrate well before then. We should ask Jon and Ray to try
> > and commit to a particular release. Is it possible it can go into the
> > 2.29.2 release? If so, then we need to move quickly.
> I don't think 2.29.2 is realistic. You're right, though, we need to
> land these for 2.30.
>
> I'll take another look soon,
> Ray
> _______________________________________________
> ConsoleKit mailing list
> ConsoleKit@...
> http://lists.freedesktop.org/mailman/listinfo/consolekit

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Multi-seat update

2009-11-13T07:43:08Z

Hi,

On Thu, Nov 12, 2009 at 9:53 PM, Halton Huo <Halton.Huo@...> wrote:
> We have sees several emails to the consolekit and gdm-list mailing lists
> where people have asked for multi-seat and multi-display support.
> Obviously this feature is very important. Currently the patches are very
> long and difficult to maintain for applying on master and released
> tarballs.
Right, we should get these integrated. There was a bit of a hiatus
because we knew they weren't going to make 2.28 so priorities shifted.

> Ray and Jon, could you please review these changes as soon as you can?
> If so, it would be very much appreciated.
>
> Note the schedule:
>
> http://live.gnome.org/TwoPointTwentyseven/
>
> - 2.29.2 is due on November 16th
> - 2.29.3 on November 30th.
> - 2.29.4 on December 21st
> - 2.29.5 on January 11th and this is string freeze.
>
> We really, at the very latest should get this into 2.29.5 if multi-seat
> support is going to be delivered with GNOME 2.30. Considering that this
> feature should be soak tested for as much time as possible, it would be
> best to integrate well before then. We should ask Jon and Ray to try
> and commit to a particular release. Is it possible it can go into the
> 2.29.2 release? If so, then we need to move quickly.

I don't think 2.29.2 is realistic. You're right, though, we need to
land these for 2.30.

I'll take another look soon,
Ray
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Multi-seat update

2009-11-13T02:23:26Z

Halton Huo schreef:
> Hi,
>
> I'd like to give an update on multi-seat and multi-display feature.
>
>>From this past May, Ray Strode and myself worked together to commit
> multi-seat support into the ConsoleKit multi-seat branch[3] and GDM
> display-configuration branch[4].

Really great ;-)

With regards,
Paul van der Vlis.

--
http://www.vandervlis.nl/

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Multi-seat update

2009-11-12T18:53:58Z

Hi,

I'd like to give an update on multi-seat and multi-display feature.

>From this past May, Ray Strode and myself worked together to commit
multi-seat support into the ConsoleKit multi-seat branch[3] and GDM
display-configuration branch[4].

In total, we committed 78 times in these branches and fixed a large
number of bugs. Now I'm glad to say this patch is ready for review and
good enough to commit back to master. You can refer to the general
testing result page[2].

We have sees several emails to the consolekit and gdm-list mailing lists
where people have asked for multi-seat and multi-display support.
Obviously this feature is very important. Currently the patches are very
long and difficult to maintain for applying on master and released
tarballs.

Ray and Jon, could you please review these changes as soon as you can?
If so, it would be very much appreciated.

Note the schedule:

http://live.gnome.org/TwoPointTwentyseven/

- 2.29.2 is due on November 16th
- 2.29.3 on November 30th.
- 2.29.4 on December 21st
- 2.29.5 on January 11th and this is string freeze.

We really, at the very latest should get this into 2.29.5 if multi-seat
support is going to be delivered with GNOME 2.30. Considering that this
feature should be soak tested for as much time as possible, it would be
best to integrate well before then. We should ask Jon and Ray to try
and commit to a particular release. Is it possible it can go into the
2.29.2 release? If so, then we need to move quickly.

The latest patches that can be applied on git-master can be found here:
ConsoleKit patch: https://bugs.freedesktop.org/attachment.cgi?id=31124
GDM patch: https://bugzilla.gnome.org/attachment.cgi?id=147528

References
-----------
[1] Design document:
http://wiki.genunix.org/wiki/index.php/design_for_newgdm_consolekit_multiseat_multidisplay

[2] Test report:
http://wiki.genunix.org/wiki/index.php/testing_for_newgdm_consolekit_multiseat_multidisplay

[3] ConsoleKit multi-seat branch:
http://cgit.freedesktop.org/ConsoleKit/tree/?h=multi-seat

[4] GDM display-configuration branch:
http://git.gnome.org/cgit/gdm/?h=display-configuration

[5] ConsoleKit bug:
https://bugs.freedesktop.org/show_bug.cgi?id=19333

[6] GDM bug:
https://bugzilla.gnome.org/show_bug.cgi?id=536355

Thanks,
Halton.

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

double login warning box

2009-11-10T07:47:21Z

Is the double login warning box configurable?
I want to hide two buttons and keep just one: the right one; break off the
login, so it will return to the login box of the gdmgreeter

regards,
guy
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Your gdm changes and SunRay...

2009-11-05T03:01:52Z

Hi Halton,

Am 05.11.09 11:26, schrieb Halton Huo:

> With my knowledge so far, you need do following things to let SRSS work
> with ConsoleKit and new GDM:
>
> 1. Apply the ConsoleKit latest patch
> http://src.opensolaris.org/source/xref/jds/spec-files/trunk/patches/ConsoleKit-01-ck-dynamic.diff
>
> 2. Apply the Sunray specific patch
> http://src.opensolaris.org/source/xref/jds/spec-files/trunk/patches/ConsoleKit-02-add-sunray-type.diff
>
> 3. Apply the GDM patch
> http://src.opensolaris.org/source/xref/jds/spec-files/trunk/patches/gdm-04-dynamic-display.diff
>
> 4. Get gdmdynamic wrapper
> http://src.opensolaris.org/source/xref/jds/spec-files/trunk/ext-sources/gdmdynamic
>
> Let me know if you have any issue.

Oh sorry. I think I didn't make it clear enough. Of coure I applied the
above mentioned changes. In fact, I already used the latest code from
the repective git repositories. However, the problems I am facing are
still present. So please read through my last email again and take
special attention on my proposed fixes for the two problems that are
still present with your latest changes (crash of console-kit-daemon and
broken behaviour of gdm_dynamic_display_finish function).

best regards,
jens
--
Jens Langner Ph: +49-351-2602757
Forschungszentrum Dresden-Rossendorf e.V.
Institute of Radiopharmacy - PET Center J.Langner@...
Germany http://www.fzd.de/
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Your gdm changes and SunRay...

2009-11-05T02:26:03Z

Hi Jens,

With my knowledge so far, you need do following things to let SRSS work
with ConsoleKit and new GDM:

1. Apply the ConsoleKit latest patch
http://src.opensolaris.org/source/xref/jds/spec-files/trunk/patches/ConsoleKit-01-ck-dynamic.diff

2. Apply the Sunray specific patch
http://src.opensolaris.org/source/xref/jds/spec-files/trunk/patches/ConsoleKit-02-add-sunray-type.diff

3. Apply the GDM patch
http://src.opensolaris.org/source/xref/jds/spec-files/trunk/patches/gdm-04-dynamic-display.diff

4. Get gdmdynamic wrapper
http://src.opensolaris.org/source/xref/jds/spec-files/trunk/ext-sources/gdmdynamic

Let me know if you have any issue.

Thanks,
Halton.

On Wed, 2009-11-04 at 15:36 +0100, Jens Langner wrote:

> Hi,
>
> as Brian Cameron asked me to continue the discussion on the gdm mailing
> list I am sending my replay about the latest MultiSeat changes in gdm
> here now.
>
> As an introduction: We are currently trying to get the latest SunRay
> Server (SRS) 4.2 version running with Ubuntu 9.10 which in turn uses gdm
> 2.28. However, as SRS is heavily relying on the previous "gdmdynamic"
> feature the SRS software is not working out-of-the-box anymore with
> systems using gdm 2.28+. After searching the internet we found the
> submitted patches from Brian Cameron and Halton Huo for gdm and
> consolekit and tried to apply those to the sources used in Ubuntu 9.10.
> However, as the work from Brian and Halton is still in progress we found
> some still existing problems and thus are reporting them here now.
>
> Am 03.11.09 20:11, schrieb Brian Cameron:
>
> >> Can you please send me a link to the git repository so that I may
> >> eventually create own patch files while you are not able to update them
> >> at the above locations. Thanks.
> >
> > The latest code is here:
> >
> > http://cgit.freedesktop.org/ConsoleKit/log/?h=multi-seat
> > http://cgit.freedesktop.org/ConsoleKit/tree/?h=multi-seat
> >
> > I believe your problems will go away if you use the latest code in that
> > branch. Let me know if you have any further problems, or if this fixes
> > the issues you have noticed.
> >
> > The patches at opensolaris.org should be updated to include the latest
> > fixes from the above git branch in the next day or two, by the end of
> > the week at the latest. Halton or I will make sure that the bugs in
> > bugzilla are updated with the latest patches also so that people do not
> > use the old patches by accident, if they don't know about git.
>
> Unfortunately the latest patches available in the ConsoleKit git as well
> as in the gdm git did not fully solve our problems here.
>
> >>>> -- cut here --
> >>>> if (kind == CK_SEAT_KIND_DYNAMIC) {
> >>>> g_hash_table_steal (manager->priv->sessions, orig_ssid);
> >>>> g_hash_table_remove (manager->priv->sessions, orig_ssid);
> >>>> }
> >>>> -- cut here --
> >>>>
> >>>> now it doesn't crash anymore here. After that change we have been
> >>>> finally able to log in to our SunRay Server. However, after a short
> >>>> test we came across a second problem which I want to outline below:
> >>>
> >>> Halton, have we fixed this problem in our latest patches?
> >>
> >> FYI: It seems this bug is still present, at least in the patch file
> >> provided at the above location. I also verified that it still crashes by
> >> using your patches. However, after inserting the above
> >> g_hash_table_steal() call the segmentation fault vanishes.
> >
> > Do you still see this problem using the latest code in the git branch?
> > I believe that this is one of the issues that Halton recently fixed,
> > though I am not sure.
>
> Yes, the problem is still present in the latest code in the git branch.
> Here I can perfectly reproduce a crash of console-kit-daemon as soon as
> I log into the SRS. However, by adding the above "g_hash_table_steal()"
> line right before the remove() finally solves that problem and
> console-kit-daemon does not crash anymore.
>
> >>>> 2) With your gdm patches applied it seems that gdm itself does not
> >>>> restart the corresponding X server anymore if either the X server
> >>>> process is killed via (kill PID) or if the session is killed by
> >>>> Ctrl-Bspc-Bspc. This seems to not only be true for the dynamic SunRay
> >>>> sessions, but also for any local session (e.g. on display :0). Please
> >>>> find attached the debug output where you can hopefully spot the problem
> >>>> that gdm tries to start another gdm-simple-slave, however it seems to
> >>>> fail and thus it ends up being aborted. Unfortunately we haven't
> >>>> found a workaround or fix for this problem yet as we are not that
> >>>> trained with gdm<>consolekit interactions. However, I believe that
> >>>> this could be caused either by the reason that the display isn't
> >>>> freed that quickly so that the new gdm slave can not start a new X
> >>>> server that fast or it might be because there is still something
> >>>> missing in the gdm<>consolekit interaction which causes the
> >>>> X server restart to fail.
> >>>
> >>> I believe that this problem is also fixed? Halton, can you confirm?
> >>
> >> I just grabed the patches from the above directory now and recompiled
> >> consolekit 0.3.1 and gdm 2.28.1 for Ubuntu 9.10 here. However, the net
> >> result is the same. I am able to connect to the running SunRay service,
> >> but as soon as I kill a session with Ctrl+Bspc+Bspc gdm never returns on
> >> that display. The same happens with the local display :0, btw. However,
> >> the debug output of gdm is now slightly different to the first one I
> >> sent to you today. Now gdm is not even trying to restart
> >> gdm-simple-slave as it seems. Perhaps you guys have an idea what might
> >> cause that.
> >
> > Do you still see this problem with the latest code in git? Again, I
> > think this is an issue that Halton recently fixed, though I am again not
> > sure.
>
> Yes, unfortunately this problem is also still present with the latest
> changes from the "display-configuration" branch in the gdm git. I can
> perfectly verify that by simply manually killing the Xorg process. As
> soon as I kill it here a new Xorg will never be automatically started on
> display :0 like it is normally done. The same happens on the SRS
> displays. Only the initial session gets an X-server process but as soon
> as Xnewt/Xorg is killed it never returns on that display.
>
> However, after a longer investigation my colleagues and I tracked the
> problem down to the newly added gdm_dynamic_display_finish() function in
> daemon/gdm-dynamic-display.c. We found out that there the following
> statements are used:
>
> -- cut here --
> gdm_dynamic_display_finish (GdmDisplay *display)
> {
> g_return_val_if_fail (GDM_IS_DISPLAY (display), FALSE);
>
> GDM_DISPLAY_CLASS (gdm_dynamic_display_parent_class)->finish (display);
>
> gdm_display_unmanage (display);
>
> return TRUE;
> }
> -- cut here --
>
> However, this seems to cause the display to be finished but never
> restarted like it is done with a normal local display or with the
> previous unpatched gdm 2.28. Even the local display :0 seems to be
> treated as a "dynamic" display now. By replacing the function by the
> same one used in daemon/gdm-static-display.c it seems to work:
>
> -- cut here --
> gdm_dynamic_display_finish (GdmDisplay *display)
> {
> int status;
>
> g_return_val_if_fail (GDM_IS_DISPLAY (display), FALSE);
>
> /* Don't call parent's finish since we don't ever
> want to be put in the FINISHED state */
>
> /* restart dynamic displays */
> gdm_display_unmanage (display);
>
> status = gdm_display_get_status (display);
> if (status != GDM_DISPLAY_FAILED) {
> gdm_display_manage (display);
> }
>
> return TRUE;
> }
> -- cut here --
>
> After that change the local :0 display gets a new Xorg process as soon
> as the old one is killed. In addition, with that change the Xserver of
> the Sunray Server (Xnewt) also gets respawned correctly.
>
> Finally the above differences allowed us to get SRSS 4.2 running with
> gdm 2.28 under Ubuntu 9.10.
>
> It would, however, really be nice if you guys could review my proposed
> changes and perhaps integrate them into your latest patchsets and/or
> change them in case my changes are not appropriate for the general case.
>
> No matter what, I would like to thank Halton and Brian for getting this
> done in gdm 2.28+. And I would like to request that their changes should
> be integrated into the latest gdm trunk ASAP so that future Ubuntu
> versions (e.g. comming LTS 10.04) ship already with a Sunray compatible
> gdm version.
>
> best regards,
> jens

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Your gdm changes and SunRay...

2009-11-04T06:36:46Z

Hi,

as Brian Cameron asked me to continue the discussion on the gdm mailing
list I am sending my replay about the latest MultiSeat changes in gdm
here now.

As an introduction: We are currently trying to get the latest SunRay
Server (SRS) 4.2 version running with Ubuntu 9.10 which in turn uses gdm
2.28. However, as SRS is heavily relying on the previous "gdmdynamic"
feature the SRS software is not working out-of-the-box anymore with
systems using gdm 2.28+. After searching the internet we found the
submitted patches from Brian Cameron and Halton Huo for gdm and
consolekit and tried to apply those to the sources used in Ubuntu 9.10.
However, as the work from Brian and Halton is still in progress we found
some still existing problems and thus are reporting them here now.

Am 03.11.09 20:11, schrieb Brian Cameron:

>> Can you please send me a link to the git repository so that I may
>> eventually create own patch files while you are not able to update them
>> at the above locations. Thanks.
>
> The latest code is here:
>
> http://cgit.freedesktop.org/ConsoleKit/log/?h=multi-seat
> http://cgit.freedesktop.org/ConsoleKit/tree/?h=multi-seat
>
> I believe your problems will go away if you use the latest code in that
> branch. Let me know if you have any further problems, or if this fixes
> the issues you have noticed.
>
> The patches at opensolaris.org should be updated to include the latest
> fixes from the above git branch in the next day or two, by the end of
> the week at the latest. Halton or I will make sure that the bugs in
> bugzilla are updated with the latest patches also so that people do not
> use the old patches by accident, if they don't know about git.

Unfortunately the latest patches available in the ConsoleKit git as well
as in the gdm git did not fully solve our problems here.

>>>> -- cut here --
>>>> if (kind == CK_SEAT_KIND_DYNAMIC) {
>>>> g_hash_table_steal (manager->priv->sessions, orig_ssid);
>>>> g_hash_table_remove (manager->priv->sessions, orig_ssid);
>>>> }
>>>> -- cut here --
>>>>
>>>> now it doesn't crash anymore here. After that change we have been
>>>> finally able to log in to our SunRay Server. However, after a short
>>>> test we came across a second problem which I want to outline below:

>>>

>>> Halton, have we fixed this problem in our latest patches?
>>
>> FYI: It seems this bug is still present, at least in the patch file
>> provided at the above location. I also verified that it still crashes by
>> using your patches. However, after inserting the above
>> g_hash_table_steal() call the segmentation fault vanishes.
>
> Do you still see this problem using the latest code in the git branch?
> I believe that this is one of the issues that Halton recently fixed,
> though I am not sure.

Yes, the problem is still present in the latest code in the git branch.
Here I can perfectly reproduce a crash of console-kit-daemon as soon as
I log into the SRS. However, by adding the above "g_hash_table_steal()"
line right before the remove() finally solves that problem and
console-kit-daemon does not crash anymore.

>>>> 2) With your gdm patches applied it seems that gdm itself does not
>>>> restart the corresponding X server anymore if either the X server
>>>> process is killed via (kill PID) or if the session is killed by
>>>> Ctrl-Bspc-Bspc. This seems to not only be true for the dynamic SunRay
>>>> sessions, but also for any local session (e.g. on display :0). Please
>>>> find attached the debug output where you can hopefully spot the problem
>>>> that gdm tries to start another gdm-simple-slave, however it seems to
>>>> fail and thus it ends up being aborted. Unfortunately we haven't
>>>> found a workaround or fix for this problem yet as we are not that
>>>> trained with gdm<>consolekit interactions. However, I believe that
>>>> this could be caused either by the reason that the display isn't
>>>> freed that quickly so that the new gdm slave can not start a new X
>>>> server that fast or it might be because there is still something
>>>> missing in the gdm<>consolekit interaction which causes the
>>>> X server restart to fail.
>>>

>>> I believe that this problem is also fixed? Halton, can you confirm?
>>
>> I just grabed the patches from the above directory now and recompiled
>> consolekit 0.3.1 and gdm 2.28.1 for Ubuntu 9.10 here. However, the net
>> result is the same. I am able to connect to the running SunRay service,
>> but as soon as I kill a session with Ctrl+Bspc+Bspc gdm never returns on
>> that display. The same happens with the local display :0, btw. However,
>> the debug output of gdm is now slightly different to the first one I
>> sent to you today. Now gdm is not even trying to restart
>> gdm-simple-slave as it seems. Perhaps you guys have an idea what might
>> cause that.
>
> Do you still see this problem with the latest code in git? Again, I
> think this is an issue that Halton recently fixed, though I am again not
> sure.

Yes, unfortunately this problem is also still present with the latest
changes from the "display-configuration" branch in the gdm git. I can
perfectly verify that by simply manually killing the Xorg process. As
soon as I kill it here a new Xorg will never be automatically started on
display :0 like it is normally done. The same happens on the SRS
displays. Only the initial session gets an X-server process but as soon
as Xnewt/Xorg is killed it never returns on that display.

However, after a longer investigation my colleagues and I tracked the
problem down to the newly added gdm_dynamic_display_finish() function in
daemon/gdm-dynamic-display.c. We found out that there the following
statements are used:

-- cut here --
gdm_dynamic_display_finish (GdmDisplay *display)
{
g_return_val_if_fail (GDM_IS_DISPLAY (display), FALSE);

GDM_DISPLAY_CLASS (gdm_dynamic_display_parent_class)->finish (display);

gdm_display_unmanage (display);

return TRUE;
}
-- cut here --

However, this seems to cause the display to be finished but never
restarted like it is done with a normal local display or with the
previous unpatched gdm 2.28. Even the local display :0 seems to be
treated as a "dynamic" display now. By replacing the function by the
same one used in daemon/gdm-static-display.c it seems to work:

-- cut here --
gdm_dynamic_display_finish (GdmDisplay *display)
{
int status;

g_return_val_if_fail (GDM_IS_DISPLAY (display), FALSE);

/* Don't call parent's finish since we don't ever
want to be put in the FINISHED state */

/* restart dynamic displays */
gdm_display_unmanage (display);

status = gdm_display_get_status (display);
if (status != GDM_DISPLAY_FAILED) {
gdm_display_manage (display);
}

return TRUE;
}
-- cut here --

After that change the local :0 display gets a new Xorg process as soon
as the old one is killed. In addition, with that change the Xserver of
the Sunray Server (Xnewt) also gets respawned correctly.

Finally the above differences allowed us to get SRSS 4.2 running with
gdm 2.28 under Ubuntu 9.10.

It would, however, really be nice if you guys could review my proposed
changes and perhaps integrate them into your latest patchsets and/or
change them in case my changes are not appropriate for the general case.

No matter what, I would like to thank Halton and Brian for getting this
done in gdm 2.28+. And I would like to request that their changes should
be integrated into the latest gdm trunk ASAP so that future Ubuntu
versions (e.g. comming LTS 10.04) ship already with a Sunray compatible
gdm version.

best regards,
jens
--
Jens Langner Ph: +49-351-2602757
Forschungszentrum Dresden-Rossendorf e.V.
Institute of Radiopharmacy - PET Center J.Langner@...
Germany http://www.fzd.de/
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: l10n question: Is the word "display" meant in the X11 sense?

2009-11-04T04:09:24Z

On Sat, 2009-10-31 at 20:18 +0100, Åsmund Skjæveland wrote:

> Is the word “display” as it appears in various strings in GDM meant to
> be in the strict X11 sense, where “display”, “screen”, and “monitor” are
> different things, or is it used in a more colloquial sense? I ask
> because if it's an X11 term it can't be translated, but if not it must
> be translated.

I'm not sure a conscious decision has been made in all instances, but
having had a look at the strings, I think it'd be safest to assume that
it's being used in the strict X11 sense.

If you encounter a string that sounds like it's referring to a physical
display device (e.g. "What you see on your display..."), that is a good
indicator that it's being used in the colloquial sense, which could
warrant an exception.

--
Hans Petter

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

GDM Development Issue - Please Advise

2009-11-01T14:33:28Z

Hello,

First off thanks for the new GDM.. it looks great! Seems to be a
lot more streamlined than prior versions which is nice however there is
one very useful feature that seems to have been cut out. The ability to
do an XDMCP login was GREAT to have. It enabled me to log remotely into
my main box (home server) from my laptop or the MythTV box in my living
room which was incredibly handy. Not to mention it seemed to handle
logging into a single monitor machine from a dual monitor machine and
vise versa very well which is a feature most remote login clients sorely
lack. If nothing else can someone please email me back so I know what
is going on with the issue.. if need be I'll just go back to legacy GDM
but I love how the new GDM looks/fuctions, other than the lack of XDMCP.

Thanks again for all your work,
Jesse Klaus
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

l10n question: Is the word "display" meant in the X11 sense?

2009-10-31T12:18:57Z

Is the word “display” as it appears in various strings in GDM meant to
be in the strict X11 sense, where “display”, “screen”, and “monitor” are
different things, or is it used in a more colloquial sense? I ask
because if it's an X11 term it can't be translated, but if not it must
be translated.

--
Åsmund Skjæveland, Norwegian Nynorsk translator/reviewer/committer

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

signature.asc (269 bytes) Download Attachment

GDM modification question

2009-10-31T08:20:43Z

Hallo, I am working on a project that needs modified GDM. I managed to
do what I needed in the 2.20.x and it worked nicely but we have to
upgrade to the newer one. So the problem appeared I just can not find
a way to pass 2 character strings from
/daemon/gdm-xdmcp-display-factory.c to /daemon/gdm-session-worker.c. I
have found two diagrams on the developer page, but they describe only
login procedure and XDMCP is not mentioned anywhere. Would you please
give me a clue how can it be done?

regards
Rumen

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Don't invoke cpp with xrdb [22590328fab9aacf17d4ce32e1c0678c330b26c0]

2009-10-28T15:05:14Z

Hi,

On Wed, Oct 28, 2009 at 22:28, Martin Pitt <martin.pitt@...> wrote:

> Bert Wesarg [2009-10-28 22:11 +0100]:
>> I would like to raise my objection for your reasons to disable the cpp
>> while loading Xresource files. If the number of users who have a
>> .Xresource file is very low, than I think the likelihood that they use
>> cpp directives is very height. And because you pay the price for the
>> high startup time, only if you have this file in the first place there
>> is no reason to use the -nocpp option.
>
> Well, it was a patch that someone applied to the Debian/Ubuntu
> packages ages ago. It makes sense to me, but I'm not attached to it a
> lot, so if upstream wants to revert his patch, I won't object.

I have just upgrade to Ubuntu 9.10, and haven't the problem before
9.10. Maybe my .Xresource was loaded from an other script (I have
something under /etc/X11/Xsession.d/ in mind).

>
> This should perhaps be discussed at
> https://bugzilla.gnome.org/show_bug.cgi?id=588849 and the bug
> reopened?
Back from reading old threads. Now that I'm aware of this problem, I
will simply put this on my checklist after upgrading to the next
release.

Anyway, thanks for the pointer.

Bert

>
> Thanks,
>
> Martin
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Don't invoke cpp with xrdb [22590328fab9aacf17d4ce32e1c0678c330b26c0]

2009-10-28T14:28:50Z

Hello Bert,

Bert Wesarg [2009-10-28 22:11 +0100]:
> I would like to raise my objection for your reasons to disable the cpp
> while loading Xresource files. If the number of users who have a
> .Xresource file is very low, than I think the likelihood that they use
> cpp directives is very height. And because you pay the price for the
> high startup time, only if you have this file in the first place there
> is no reason to use the -nocpp option.

Well, it was a patch that someone applied to the Debian/Ubuntu
packages ages ago. It makes sense to me, but I'm not attached to it a
lot, so if upstream wants to revert his patch, I won't object.

This should perhaps be discussed at
https://bugzilla.gnome.org/show_bug.cgi?id=588849 and the bug
reopened?

Thanks,

Martin

--
Martin Pitt | http://www.piware.de
Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org)
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: Don't invoke cpp with xrdb [22590328fab9aacf17d4ce32e1c0678c330b26c0]

2009-10-28T14:11:29Z

Hi Martin,

I would like to raise my objection for your reasons to disable the cpp
while loading Xresource files. If the number of users who have a
.Xresource file is very low, than I think the likelihood that they use
cpp directives is very height. And because you pay the price for the
high startup time, only if you have this file in the first place there
is no reason to use the -nocpp option.

Please reconsider your decision.

Kind regards,
Bert Wesarg
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: double login

2009-10-27T12:31:44Z

On Tuesday 27 October 2009 20:15:46 you wrote:

> > Maybe some additional info:
> > if Vtallocation is set to false, it's not working whatsoever.
> > If vtallocation is set to true or commented out,it's working.
> > here is the syslog debug with gdmgreeter in gdm.conf but crashing:
> > http://www.multiseatcomputer.be/syslog
> >
> > and here with gdmlogin in gdm.conf working:
> > http://www.multiseatcomputer.be/syslogok
>
> It looks like GDM is treating your displays as type FLEXI, but the
> displays you are starting are obviously not flexi. Setting Vtallocation
> to false, just disables the VT switching support, and thus avoids the
> problem.
>
> In your GDM configuration file, if you have "flexible=true" in the
> [server-Foo] section associated with the server you are trying to start,
> that would cause this problem if the display doesn't really support VT.
>
> Brian

No, all servers are flexible=false

all i am getting is a box with :the greeter is stopped unexpected,try to use
another one.Than i get the old style debian login box.But keyboard is not
working at that moment.

Guy

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: double login

2009-10-27T12:17:21Z

On Tuesday 27 October 2009 19:54:16 duportail wrote:

> On Tuesday 27 October 2009 19:07:40 you wrote:
> > > Ok, I got multiseat working with ubuntu 9.10 with gdm-2.20 .
> > > It cost me some time to figure out the gdmgreeter was crashing.(but not
> > > in gdm 2.18)Changing this setting to gdmlogin works ok(but it's the old
> > > style debian greeter).I believe that gdm is too much attached to
> > > gnome,because with dual screens, not only the greeter is crashing also
> > > gnome-panel.
> > > Maybe something to do with xrandr.
> >
> > Glad to hear you got things working. If you could provide a stack trace
> > of gdmgreeter, or any further information, I'd be happy to help debug.
> > But glad to hear that gdmlogin is a workaround.
> >
> > Sorry to hear about your problem with gnome-panel. I'd recommend filing
> > a bug at http://bugzilla.gnome.org/ under gnome-panel with details about
> > the crash. It's probably not a related issue.
> >
> > Brian
>
> Maybe some additional info:
> if Vtallocation is set to false, it's not working whatsoever.
> If vtallocation is set to true or commented out,it's working.
> here is the syslog debug with gdmgreeter in gdm.conf but crashing:
> http://www.multiseatcomputer.be/syslog
>
> and here with gdmlogin in gdm.conf working:
> http://www.multiseatcomputer.be/syslogok
>
> maybe you see something.(do not pay attention to kdm-kde3,i've installed
> kde3 on the machine,but same errors without kde3)
>
>
> guy
>
> _______________________________________________
> gdm-list mailing list
> gdm-list@...
> http://mail.gnome.org/mailman/listinfo/gdm-list

And this is a working setup with gdm 2.18 and de gdmgreeter:
http://www.multiseatcomputer.be/syslog218
but gdm changed the inlogscreen language(not in dutch anymore)
guy
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: double login

2009-10-27T12:15:46Z

> Maybe some additional info:
> if Vtallocation is set to false, it's not working whatsoever.
> If vtallocation is set to true or commented out,it's working.
> here is the syslog debug with gdmgreeter in gdm.conf but crashing:
> http://www.multiseatcomputer.be/syslog
>
> and here with gdmlogin in gdm.conf working:
> http://www.multiseatcomputer.be/syslogok

It looks like GDM is treating your displays as type FLEXI, but the
displays you are starting are obviously not flexi. Setting Vtallocation
to false, just disables the VT switching support, and thus avoids the
problem.

In your GDM configuration file, if you have "flexible=true" in the
[server-Foo] section associated with the server you are trying to start,
that would cause this problem if the display doesn't really support VT.

Brian
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: double login

2009-10-27T11:53:39Z

On Tuesday 27 October 2009 19:07:40 you wrote:

> > Ok, I got multiseat working with ubuntu 9.10 with gdm-2.20 .
> > It cost me some time to figure out the gdmgreeter was crashing.(but not
> > in gdm 2.18)Changing this setting to gdmlogin works ok(but it's the old
> > style debian greeter).I believe that gdm is too much attached to
> > gnome,because with dual screens, not only the greeter is crashing also
> > gnome-panel.
> > Maybe something to do with xrandr.
>
> Glad to hear you got things working. If you could provide a stack trace
> of gdmgreeter, or any further information, I'd be happy to help debug.
> But glad to hear that gdmlogin is a workaround.
>
> Sorry to hear about your problem with gnome-panel. I'd recommend filing
> a bug at http://bugzilla.gnome.org/ under gnome-panel with details about
> the crash. It's probably not a related issue.
>
> Brian

Maybe some additional info:
if Vtallocation is set to false, it's not working whatsoever.
If vtallocation is set to true or commented out,it's working.
here is the syslog debug with gdmgreeter in gdm.conf but crashing:
http://www.multiseatcomputer.be/syslog

and here with gdmlogin in gdm.conf working:
http://www.multiseatcomputer.be/syslogok

maybe you see something.(do not pay attention to kdm-kde3,i've installed kde3
on the machine,but same errors without kde3)

guy

_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list

Re: double login

2009-10-27T11:07:40Z

> Ok, I got multiseat working with ubuntu 9.10 with gdm-2.20 .
> It cost me some time to figure out the gdmgreeter was crashing.(but not in gdm
> 2.18)Changing this setting to gdmlogin works ok(but it's the old style debian
> greeter).I believe that gdm is too much attached to gnome,because with dual
> screens, not only the greeter is crashing also gnome-panel.
> Maybe something to do with xrandr.

Glad to hear you got things working. If you could provide a stack trace
of gdmgreeter, or any further information, I'd be happy to help debug.
But glad to hear that gdmlogin is a workaround.

Sorry to hear about your problem with gnome-panel. I'd recommend filing
a bug at http://bugzilla.gnome.org/ under gnome-panel with details about
the crash. It's probably not a related issue.

Brian
_______________________________________________
gdm-list mailing list
gdm-list@...
http://mail.gnome.org/mailman/listinfo/gdm-list