Re: [gdm-list] My feedback about lately GDM development
- From: Brian Cameron <brian cameron oracle com>
- To: Thamawij Pirajnaraporn <jokekidnan gmail com>
- Cc: gdm-list gnome org
- Subject: Re: [gdm-list] My feedback about lately GDM development
- Date: Mon, 10 May 2010 12:52:03 -0500
Thamawij:
I am just a plain user with not so much knowledge and I know that you
guys working for us without a pay so if I wrote something mistaken or
harsh by no means, please excuse.
Thank you for your feedback. It is always good to hear from users and
fans.
I am a fan of Gnome for quite a time, so I am experience GDM most of the
time, using linux. But lately, I am quite disappointed for the release
and this is the 2nd time so I decide to send this feedback.
* The first time I really mess up is when GDM upgrade to version 2.
Note that GDM was rewritten for version 2.22. I assume you mean that
you updated from a version 2.20 or older to version 2.22 or later.
I think you already had many feedback about the theme system that
can no longer compatible with the previous themes. For the users,
this is just buggy, but for those theme developers, this is just
like throwing their devoted good works into trash and this made
many developers turn their back. I am not the developer so I did
not feel so sad. But as you guys are developers, I think you guys
feel the same if the same occurred to you.
Yes, removing support for gdmgreeter style themes was a hard decision.
However, the GNOME community is dropping support for libgnomecanvas
which these gdmgreeter themes depend upon. It probably isn't too hard
of a job to port the old gdmgreeter-style theme code to work with the
new GDM. What is hard is deciding which canvas should be used now that
libgnomecanvas is being dropped. Or does it make more sense to use
something like clutter to support GDM themes? A clutter-based solution
could also work with the same XML file format that the old gdmgreeter
style themes used.
Note that one big problem with gdmgreeter themes is that they never
supported accessibility very well. So, while the new GDM may not be as
nice to theme, it at least works better for users with accessibility
needs. So, hopefully, if someone fixes the new GDM to support themes
again, the accessibility issue will be better addressed.
Note that clutter also does not support accessibility well. Note that
GNOME Shell also uses clutter, and I think it will be easier to decide
if clutter is a good way to theme GDM once the accessibility issues
are solved in clutter for GNOME Shell.
* I am not sure if this is because Ubuntu developers or the GDMs, I
found that in Lucid (Ubuntu and Xubuntu) GDM can no longer enable
root allowed GUI. I understand that this is for security reason.
But GUI development is mostly aims for desktop user (if I am not
mistaken), I think "easy to use" is more weight. Either for
individual or enterprise users, provide choices is more efficient
than forcing users against their will. In many cases, making the
proper one default and let user override the settings on their own
with a warning message is the best solution. I think the most
efficient and security system is nothing compare to the necessary
and easy to use.
I assume you are referring to the login GUI configuration program. The
lack of a gdmsetup/gdmconfig program to configure GDM is a big
regression since the old GDM. There is currently work underway to
better supporting configuring GDM via a GUI configuration, and hopefully
this will work better in an upcoming release of GNOME.
Note that GDM is still configurable. Some configuration options can
be specified in the /etc/gdm/custom.conf file, and GUI-specific keys
are stored in the "gdm" user's GConf settings. The GDM manual explains
what configuration options are supported and how to change them:
http://library.gnome.org/admin/gdm/2.30/configuration.html.en
I am a fresh graduate whom working for just 1 - 2 years so I may be only
a little one in IT world, compare to you guys, but I hope my feedback
and my view could be a bit of help. From my point of view, I think linux
should improve the users management and . Security, performance,
complexity and convenient are not always trade off as there are always a
better one in the next era.
It is a trade off. The new GDM does have some serious regressions
compared to the old GDM, as you highlight above. However, a rewrite of
GDM was long overdue. The old GDM code had not been significantly
reworked for over 10 years, and the code was not taking advantage of
newer GNOME technologies. As we approach GNOME 3.0 there was a real
push to move away from deprecated technologies such as libgnomecanvas
as I explained above.
Now GDM uses much more standard interfaces such as D-Bus for IPC and
better uses GObject style programming for better ongoing
maintainability. We have made a lot of progress fixing the various
regressions caused by the update to the new GDM rewrite, but there is
more work to do.
PS. If there any help you might need, please let me know. I am certainly
respond if I could.
If you know how to program, we are always looking for help fixing bugs,
adding enhancements, and fixing regressions. You can refer to
http://bugzilla.gnome.org and look in the "gdm" category for open bugs
if you want to review what work needs to be done.
Brian
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]