[Portland] RE: RELEASED: xdg-utils 1.0 beta2

[Bastian, Waldo]

>Hi Waldo,
>
>Bastian, Waldo wrote:
>> Yippie, Xdg-utils 1.0 beta2 is released!
>>
>> Head over to
>> http://portland.freedesktop.org/download/xdg-utils-1.0beta2.tgz to
play
>> with this puppie.
>
>I'm happy to see that some of the problems I had with the
>xdg-screensaver script [1] have been addressed.  However, some have
not.
>
>This iteration of the script has the following:
>
>        suspend)
>         # Don't know how to suspend properly
>         gnome-screensaver-command --exit > /dev/null 2> /dev/null
>         result=0
>         ;;
>
>It is unacceptable to cause the screensaver daemon process to exit.
I'm
>surprised that you would release a script that does this.
>
>If you are set on using this script based approach,  I think that you
>can do something with the gnome-screensaver-command --inhibit option,
if
>it exists, or else a loop using the --poke option.

I will consider that.

>To reiterate some of my previous objections:
>
>* The terminology the script uses is faulty.  Inhibit is clearly a
>better verb to use for temporarily attempting to restrain screensaver
>activation.  Nothing is really "suspended".

I'm afraid the nuance is lost on me, but then again, I'm not a native
english speaker. I will consider it.

> It is critical that even
>when the screensaver activation is inhibited that the screen be able to
>lock.

That's a good point.

>* You must make a distinction between activation and explicit locking.
>For example, you should provide both --lock and --activate options.

I will consider it.

>* The window-id approach is broken and reveals implementation details.

I disagree. I think it's a platform detail but I think it is appropriate
to be platform specific in this case. I just notice that my reply to
http://lists.freedesktop.org/archives/xdg/2006-July/008404.html never
made it to the list. I think that the suggestion to use
   gnome-screensaver-command --inhibit \
   --application-name="FooPlayer" \
   --reason="Playing in fullscreen mode" &
   [do something]
   kill %1
instead fails to address the issue that if the [do something] part
crashes, the gnome-screensaver-command will be left running. It will
work when called from a shell script but not if the
gnome-screensaver-command is called from an executable.

>* It is difficult to promote two standard approaches for screensaver
>control.  The reasoning that has been given for why this isn't an issue
>is that the scripts are for immediate deployment and the D-Bus API is
>for future deployment.  This script won't work at all on any existing
>deployment that uses GNOME < 2.14 since it doesn't support
xscreensaver.

Good point.

Cheers,
Waldo