From: Matthew W. O'P. <mwe...@gm...> - 2005-08-29 15:28:13
|
I decided to try out the 0install Injector a few weeks ago, and have been m= uch impressed -- I feel it's a vast improvement to the lazyfs/0install combo in= the past in terms of user-friendliness and Doing the Right Thing. HOWEVER... this morning my injector updated itself, and the new version (0.= 12?) now overrides LIBDIRPATH -- which means my ROX-CLib apps now no longer know where to look for ROX-CLib. Is there a way to prevent the injector from doing this? It should prepend i= tself to the LIBDIRPATH instead of overriding it, in my opinion; a user may have = a number of settings specific to the work they do, and overriding it is the w= rong thing to do here. --=20 Matthew Weier O'Phinney mwe...@gm... http://weierophinney.net/matthew/ |
From: Thomas L. <ta...@ec...> - 2005-08-29 16:59:20
|
On Mon, Aug 29, 2005 at 11:28:05AM -0400, Matthew Weier O'Phinney wrote: > I decided to try out the 0install Injector a few weeks ago, and have > been much impressed -- I feel it's a vast improvement to the > lazyfs/0install combo in the past in terms of user-friendliness and > Doing the Right Thing. > > HOWEVER... this morning my injector updated itself, and the new > version (0.12?) now overrides LIBDIRPATH -- which means my ROX-CLib > apps now no longer know where to look for ROX-CLib. > > Is there a way to prevent the injector from doing this? It should > prepend itself to the LIBDIRPATH instead of overriding it, in my > opinion; a user may have a number of settings specific to the work > they do, and overriding it is the wrong thing to do here. First, I haven't changed anything in the injector, so it's probably the application you are running that changed (ROX-Session would be a likely candidate). Most likely, the problem is that LIBDIRPATH isn't set at all when you run the injector, so there's nothing for it to prepend to. The work-around is to set LIBDIRPATH to its default value before running the injector. Eg, in your .xsession, do: export LIBDIRPATH=$HOME/lib:/usr/local/lib:/usr/lib In the next version of the injector, you'll be able to set 'default=' in the interface for environment variables. However, setting LIBDIRPATH was itself only a work-around for ROX-Session 0.1.25 having an old findrox. I've just updated it not to do this for 0.26, so try upgrading ROX-Session and logging in again... Let me know if that helps... -- Dr Thomas Leonard http://rox.sourceforge.net GPG: 9242 9807 C985 3C07 44A6 8B9A AE07 8280 59A5 3CC1 |
From: Matthew W. O'P. <mwe...@gm...> - 2005-08-29 18:26:13
|
On 8/29/05, Thomas Leonard <ta...@ec...> wrote: > On Mon, Aug 29, 2005 at 11:28:05AM -0400, Matthew Weier O'Phinney wrote: > > HOWEVER... this morning my injector updated itself, and the new > > version (0.12?) now overrides LIBDIRPATH -- which means my ROX-CLib > > apps now no longer know where to look for ROX-CLib. > > > > Is there a way to prevent the injector from doing this? It should > > prepend itself to the LIBDIRPATH instead of overriding it, in my > > opinion; a user may have a number of settings specific to the work > > they do, and overriding it is the wrong thing to do here. >=20 > First, I haven't changed anything in the injector, so it's probably the > application you are running that changed (ROX-Session would be a likely > candidate). Yes, it appears to be tied to ROX-Session. > Most likely, the problem is that LIBDIRPATH isn't set at all when you > run the injector, so there's nothing for it to prepend to. The > work-around is to set LIBDIRPATH to its default value before running the > injector. >=20 > Eg, in your .xsession, do: >=20 > export LIBDIRPATH=3D$HOME/lib:/usr/local/lib:/usr/lib I had actually already tried this, and it didn't work. What I had to do was= to symlink the ROX-CLib directory into the directory reported by LIBDIRPATH (i= n one of the .cache/0install.net/implementations directories). > In the next version of the injector, you'll be able to set 'default=3D' i= n > the interface for environment variables. That would be nice. > However, setting LIBDIRPATH was itself only a work-around for > ROX-Session 0.1.25 having an old findrox. I've just updated it not to do > this for 0.26, so try upgrading ROX-Session and logging in again... I upgraded ROX-Session to 0.26 using the injector, and this corrected the problem, but introduced a new one: numlock was on by default, and the only = way to turn this off was to manually change it in the Settings.xml file. This s= hould be off by default, I feel -- I use a laptop, and having it switched on mean= t that I started typing a lot of stuff like the f63365ng n6nsense beca4se the n40-ad 5s -art 6f the 2eyb6ard (translation: "following nonsense because th= e numpad is part of the keyboard"). Will there be a GUI option for this in the ROX-Session settings dialog? Thi= s would alleviate my concern as well. --=20 Matthew Weier O'Phinney mwe...@gm... http://weierophinney.net/matthew/ |
From: Guido S. <__g...@we...> - 2005-08-29 18:32:39
|
On Mon, 29 Aug 2005 14:26:05 -0400 Matthew Weier O'Phinney <mwe...@gm...> wrote: > Will there be a GUI option for this in the ROX-Session settings dialog? This > would alleviate my concern as well. You can control this with the "Keyboard" config tool. |
From: Matthew W. O'P. <mwe...@gm...> - 2005-08-29 18:47:55
|
On 8/29/05, Guido Schimmels <__g...@we...> wrote: > On Mon, 29 Aug 2005 14:26:05 -0400 > Matthew Weier O'Phinney <mwe...@gm...> wrote: > > Will there be a GUI option for this in the ROX-Session settings dialog?= This > > would alleviate my concern as well. >=20 > You can control this with the "Keyboard" config tool. Where? I'm not seeing it in the one available with the 0.9.7.9 release of OroboROX. If this isn't the proper config tool, where *is* the proper one t= o utilize? --=20 Matthew Weier O'Phinney mwe...@gm... http://weierophinney.net/matthew/ |
From: Jonatan L. <th...@ho...> - 2005-08-29 21:28:54
|
On Mon, 29 Aug 2005 14:47:48 -0400 Matthew Weier O'Phinney <mwe...@gm...> wrote: > On 8/29/05, Guido Schimmels <__g...@we...> wrote: > > On Mon, 29 Aug 2005 14:26:05 -0400 > > Matthew Weier O'Phinney <mwe...@gm...> wrote: > > > Will there be a GUI option for this in the ROX-Session settings > > > dialog? This would alleviate my concern as well. > > > > You can control this with the "Keyboard" config tool. > > Where? I'm not seeing it in the one available with the 0.9.7.9 > release of OroboROX. If this isn't the proper config tool, where *is* > the proper one to utilize? It's somewhere on Guido's site... I really think we should have a "Configlets" section on the core components rox wiki page, together with rox-filer, rox-session, rox-lib2, oroborox. actually, i think it's a little strange to supply the configlets separated from rox-session, as the configlets anyhow actually just tells rox-session (via dbus) to make the change! I think they should be part of rox-session, and put under ROX-Session/Config or something like that. And put an appmenu entry "Options" on rox-session. /Jonatan -=( http://kymatica.com )=- |
From: Thomas L. <ta...@ec...> - 2005-09-01 19:41:26
|
On Mon, Aug 29, 2005 at 10:35:06PM +0200, Jonatan Liljedahl wrote: > On Mon, 29 Aug 2005 14:47:48 -0400 > Matthew Weier O'Phinney <mwe...@gm...> wrote: > > > On 8/29/05, Guido Schimmels <__g...@we...> wrote: > > > On Mon, 29 Aug 2005 14:26:05 -0400 > > > Matthew Weier O'Phinney <mwe...@gm...> wrote: > > > > Will there be a GUI option for this in the ROX-Session settings > > > > dialog? This would alleviate my concern as well. > > > > > > You can control this with the "Keyboard" config tool. > > > > Where? I'm not seeing it in the one available with the 0.9.7.9 > > release of OroboROX. If this isn't the proper config tool, where *is* > > the proper one to utilize? > > It's somewhere on Guido's site... I really think we should have a > "Configlets" section on the core components rox wiki page, together with > rox-filer, rox-session, rox-lib2, oroborox. Good idea. I've added Mouse and Keyboard to the ROX-All package (0.2): http://sourceforge.net/project/showfiles.php?group_id=7023&package_id=161987&release_id=353534 Guido: it currently downloads them from roxos.sunsite.dk. Will old versions always be available there, or should I mirror them on sf.net? -- Dr Thomas Leonard http://rox.sourceforge.net GPG: 9242 9807 C985 3C07 44A6 8B9A AE07 8280 59A5 3CC1 |
From: Guido S. <__g...@we...> - 2005-09-10 16:49:52
|
On Thu, 1 Sep 2005 20:39:56 +0100 Thomas Leonard <ta...@ec...> wrote: > Guido: it currently downloads them from roxos.sunsite.dk. Will old > versions always be available there, or should I mirror them on sf.net? Due to resource limits, I delete older versions. |
From: Thomas L. <ta...@ec...> - 2005-09-15 17:12:51
|
On Sat, Sep 10, 2005 at 06:48:44PM +0200, Guido Schimmels wrote: > On Thu, 1 Sep 2005 20:39:56 +0100 > Thomas Leonard <ta...@ec...> wrote: > > > Guido: it currently downloads them from roxos.sunsite.dk. Will old > > versions always be available there, or should I mirror them on sf.net? > > Due to resource limits, I delete older versions. OK, I've put up versions of sf.net. I also replaced findrox with a newer version in both cases, as it was causing problems with 0launch. -- Dr Thomas Leonard http://rox.sourceforge.net GPG: 9242 9807 C985 3C07 44A6 8B9A AE07 8280 59A5 3CC1 |
From: Thomas L. <ta...@ec...> - 2005-08-29 19:05:38
|
On Mon, Aug 29, 2005 at 02:26:05PM -0400, Matthew Weier O'Phinney wrote: [...] > I upgraded ROX-Session to 0.26 using the injector, and this corrected the > problem, but introduced a new one: numlock was on by default, and the > only way to turn this off was to manually change it in the > Settings.xml file. This should be off by default, I feel -- I use a > laptop, and having it switched on meant that I started typing a lot of > stuff like the f63365ng n6nsense beca4se the OK, that sounds like a pretty convincing reason to have it off by default! (my laptop doesn't do that, so I hadn't noticed the problem) -- Dr Thomas Leonard http://rox.sourceforge.net GPG: 9242 9807 C985 3C07 44A6 8B9A AE07 8280 59A5 3CC1 |
From: Guido S. <__g...@we...> - 2005-08-29 22:52:01
|
On Mon, 29 Aug 2005 20:04:06 +0100 Thomas Leonard <ta...@ec...> wrote: > On Mon, Aug 29, 2005 at 02:26:05PM -0400, Matthew Weier O'Phinney wrote: > [...] > > I upgraded ROX-Session to 0.26 using the injector, and this corrected the > > problem, but introduced a new one: numlock was on by default, and the > > only way to turn this off was to manually change it in the > > Settings.xml file. This should be off by default, I feel -- I use a > > laptop, and having it switched on meant that I started typing a lot of > > stuff like the f63365ng n6nsense beca4se the > > OK, that sounds like a pretty convincing reason to have it off by > default! (my laptop doesn't do that, so I hadn't noticed the problem) We could check for keyboard type and do the right thing. Don't know if that would be any hard. |
From: Kurt B C. <kur...@te...> - 2005-08-30 00:52:36
|
On Tue, 2005-08-30 at 00:50 +0200, Guido Schimmels wrote: > On Mon, 29 Aug 2005 20:04:06 +0100 > Thomas Leonard <ta...@ec...> wrote: > > > On Mon, Aug 29, 2005 at 02:26:05PM -0400, Matthew Weier O'Phinney wrote: > > [...] > > > I upgraded ROX-Session to 0.26 using the injector, and this corrected the > > > problem, but introduced a new one: numlock was on by default, and the > > > only way to turn this off was to manually change it in the > > > Settings.xml file. This should be off by default, I feel -- I use a > > > laptop, and having it switched on meant that I started typing a lot of > > > stuff like the f63365ng n6nsense beca4se the > > > > OK, that sounds like a pretty convincing reason to have it off by > > default! (my laptop doesn't do that, so I hadn't noticed the problem) > > We could check for keyboard type and do the right thing. > Don't know if that would be any hard. > > > ------------------------------------------------------- > SF.Net email is Sponsored by the Better Software Conference & EXPO > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf Why not just have rox-session remember whether numlock was on or off on the last logout. |
From: Jonatan L. <th...@ho...> - 2005-08-30 20:43:57
|
On Mon, 29 Aug 2005 18:50:19 -0600 Kurt B Cox <kur...@te...> wrote: > > > OK, that sounds like a pretty convincing reason to have it off by > > > default! (my laptop doesn't do that, so I hadn't noticed the > > > problem) > > > > We could check for keyboard type and do the right thing. > > Don't know if that would be any hard. > > Why not just have rox-session remember whether numlock was on or off > on the last logout. I vote for that. /Jonatan -=( http://kymatica.com )=- |
From: <pi...@wg...> - 2005-09-02 14:27:47
|
On Tue, Aug 30, 2005 at 09:45:24PM +0200, Jonatan Liljedahl wrote: > On Mon, 29 Aug 2005 18:50:19 -0600 > Kurt B Cox <kur...@te...> wrote: > > > > > OK, that sounds like a pretty convincing reason to have it off by > > > > default! (my laptop doesn't do that, so I hadn't noticed the > > > > problem) > > > > > > We could check for keyboard type and do the right thing. > > > Don't know if that would be any hard. > > > > Why not just have rox-session remember whether numlock was on or off > > on the last logout. > > I vote for that. Me too. The option in the config tool should then be replaced with a message that tells user that NumLock state is saved on exit. Or, it could change into tri state "NumLock at login"->(on, off, last) peter. |