Re: MATE Desktop Environment in Debian
I agree with Josselin here. There is no point in forking gconf and other libraries. Even worse, this will increase the incompatibility between different desktops. For example, an app making use of gconftool-2 will not work when there's mateconftool-2 in the system instead. If you want to fix a bug in GNOME libraries, you should propose a patch to their bugzilla, they will be happy to accept it. Also, we try to remove existing packages that use deprecated libraries (like libgnome* and libbonobo*), and I think it's not good at all to add new packages relying on those libraries. -- Dmitry Shachnev (Debian Maintainer) -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1343127220.7568.5.camel@eeepc
Re: MATE Desktop Environment in Debian
I'd like to support the proposal to package MATE. +1 While I agree there is code duplication, that's obviously something that will take some time to remove, however I also believe it's important not to wait until it's cleaned. ++1 I strongly agree with Dave, we should not wait to solve this issues. There is enough potential for Mate to live on a non competing space to GNOME3 which will not run on modest or old hardware (e.g. arm based computers, computer with no hardware acceleration). There is also enough murmur and frustration about the lack of feedback acceptance from from GNOME developers. Personally, I feel the applications in GNOME are dumbed versions of their counterparts in GNOME2 only re-written in GTK3. This is not usable to me. And I am not alone out there. There are enough people who devote the time making whole Debian distros including MATE: http://distrowatch.com/table.php?distribution=descentos http://distrowatch.com/table.php?distribution=leeenux http://distrowatch.com/table.php?distribution=snowlinux http://linuxmint.com/ Not debian based: http://distrowatch.com/table.php?distribution=salix This works could be directly invested in Debian. Project like Skolelinux, or Debian Edu, which already used GNOME2 based environment will suffer badly if they will be forced to migrade to GNOME3 on thin clients. MATE will be an excellent continuation for their project. And will also save a lot of time re-designing XCFE to feet into thin clients. Please, do not block MATE in Debian, and let it live in peace side by side to GNOME3. I rolled up my sleeves and learned how to package mate to debian, and I willing to contribute to the efforts of cleaning duplicate codes and making MATE a viable alternative in Debian. Best Regards, Oz -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cadkmuwnbgvu6pglhooyeyp5enwnfyz9h-sogtagjwzcdlop...@mail.gmail.com
Re: MATE Desktop Environment in Debian
On Sun, Jul 22, 2012 at 11:08:32AM +0200, Oz Nahum Tiram wrote: There is also enough murmur and frustration about the lack of feedback acceptance from from GNOME developers. Personally, I feel the applications in GNOME are dumbed versions of their counterparts in GNOME2 only re-written in GTK3. This is not usable to me. That may only be true when considering complete rewrites. Porting an application from Gtk2 to Gtk3 does not at all require that many changes that you'd need to do a rewrite. (Basically they took that opportunity to drop old deprecated APIs[1].) For instance I don't think gedit lost features, but was instead just ported to Gtk3. Kind regards Philipp Kern [1] And then they went and added new deprecations in Gtk3 along the way. Meh. signature.asc Description: Digital signature
Re: MATE Desktop Environment in Debian
On Sun, 22 Jul 2012 15:34:01 +0200 Philipp Kern pk...@debian.org wrote: On Sun, Jul 22, 2012 at 11:08:32AM +0200, Oz Nahum Tiram wrote: There is also enough murmur and frustration about the lack of feedback acceptance from from GNOME developers. Personally, I feel the applications in GNOME are dumbed versions of their counterparts in GNOME2 only re-written in GTK3. This is not usable to me. That may only be true when considering complete rewrites. Porting an application from Gtk2 to Gtk3 does not at all require that many changes that you'd need to do a rewrite. (Basically they took that opportunity to drop old deprecated APIs[1].) I disagree. I have a number of upstream projects which started with early versions of GTK2 which I ported to more recent versions and in which I then implemented support for the DISABLE_DEPRECATED macros of glib2.0 and gtk+2.0 such that the current versions of these packages in unstable build fine against the current gtk+-2.0 with DISABLE_DEPRECATED set - yet none of these will build against gtk+3.0. Some of my upstreams would have to be complete rewrites to port to gtk3, at which point I start considering if it would be easier to rewrite for something other than gtk or abandon the upstream. [1] And then they went and added new deprecations in Gtk3 along the way. Meh. Not so meh. When gtk+2.0 finally moves into oldlibs, my upstream projects may have to be removed as the amount of work required to move to gtk+3.0 from gtk+2.0 with DISABLE_DEPRECATED is larger than the amount of work to set DISABLE_DEPRECATED on gtk+2.0 in the first place. That wasn't in the plan. It took months to make the most recent transition. I wouldn't mind so much if porting to gtk3 added some functionality to my apps, it doesn't, it just makes it harder to let my apps behave as they should. Gtk2 was a step forward from gtk1. Gtk3 is a failure IMHO. Unless third party apps buy into the entire shell monoculture (via a rewrite), gtk3 makes it all but impossible to migrate. Gtk3 is an aggressive move to strip out apps which don't fit the GNOME3 model, a model which I find abhorrent and completely unsuited to how I want my projects to behave. What I can't reimplement in Qt4 (not KDE, Qt) will simply be abandoned, thanks to Gtk3. It augers badly for XFCE and LXDE too as gtk3 makes life difficult for them too, including adding more bloat for no increased functionality when a lack of bloat is the one thing which sets XFCE and LXDE apart from GNOME. http://mail.xfce.org/pipermail/xfce4-dev/2012-May/029843.html -- Neil Williams = http://www.linux.codehelp.co.uk/ pgpa742o3jdqQ.pgp Description: PGP signature
Re: MATE Desktop Environment in Debian
Le dimanche 22 juillet 2012 à 16:09 +0100, Neil Williams a écrit : I disagree. I have a number of upstream projects which started with early versions of GTK2 which I ported to more recent versions and in which I then implemented support for the DISABLE_DEPRECATED macros of glib2.0 and gtk+2.0 such that the current versions of these packages in unstable build fine against the current gtk+-2.0 with DISABLE_DEPRECATED set - yet none of these will build against gtk+3.0. Some of my upstreams would have to be complete rewrites to port to gtk3, at which point I start considering if it would be easier to rewrite for something other than gtk or abandon the upstream. Just because an application doesn’t build doesn’t mean you have to completely rewrite it. http://developer.gnome.org/gtk3/stable/ch24s02.html Gtk2 was a step forward from gtk1. Gtk3 is a failure IMHO. Unless third party apps buy into the entire shell monoculture (via a rewrite), gtk3 makes it all but impossible to migrate. Gtk3 is an aggressive move to strip out apps which don't fit the GNOME3 model, a model which I find abhorrent and completely unsuited to how I want my projects to behave. I have no idea where you got this bullshit from. The changes between GTK 2.x to 3.x are minor and imply in no way a change in a “model” you haven’t defined anyway (object model? widget model? event model? none of them have changed). -- .''`. Josselin Mouette : :' : `. `' `- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1342990116.24016.6.camel@tomoyo
Re: MATE Desktop Environment in Debian
On Sun, 22 Jul 2012 22:48:36 +0200 Josselin Mouette j...@debian.org wrote: Le dimanche 22 juillet 2012 à 16:09 +0100, Neil Williams a écrit : I disagree. I have a number of upstream projects which started with early versions of GTK2 which I ported to more recent versions and in which I then implemented support for the DISABLE_DEPRECATED macros of glib2.0 and gtk+2.0 such that the current versions of these packages in unstable build fine against the current gtk+-2.0 with DISABLE_DEPRECATED set - yet none of these will build against gtk+3.0. Some of my upstreams would have to be complete rewrites to port to gtk3, at which point I start considering if it would be easier to rewrite for something other than gtk or abandon the upstream. Just because an application doesn’t build doesn’t mean you have to completely rewrite it. (I wouldn't have put it that way if it was a small fix.) http://developer.gnome.org/gtk3/stable/ch24s02.html OK, well I'll have another look at the issues after Wheezy. Gtk2 was a step forward from gtk1. Gtk3 is a failure IMHO. Unless third party apps buy into the entire shell monoculture (via a rewrite), gtk3 makes it all but impossible to migrate. Gtk3 is an aggressive move to strip out apps which don't fit the GNOME3 model, a model which I find abhorrent and completely unsuited to how I want my projects to behave. I have no idea where you got this bullshit from. Bitter experience. My main problem is that I don't want to follow the direction set out by GNOME3 with the shell and lack of usable panels, I just want to use GTK without GNOME. In the latter days of GTK2, this seemed to be the way that things were going but I'm no longer convinced that GTK developers care about upstreams who are not signed-up to GNOME. The one upstream which does compile, doesn't run - despite deprecated functions being removed. The changes between GTK 2.x to 3.x are minor and imply in no way a change in a “model” you haven’t defined anyway (object model? widget model? event model? none of them have changed). The model is the redesign of the GNOME3 shell / UI. It's not so much about a source code model, it's how the code gets translated into a UI which looks nothing like what I wanted to create and behaves like a GNOME app even when used in XFCE. However, I will revisit all this after Wheezy - there's no point discussing this ad infinitum at this stage. I just wanted to note that the port is not necessarily as simple as originally outlined. Let's leave it at that until after Wheezy. -- Neil Williams = http://www.linux.codehelp.co.uk/ pgpmRb3rnvMW6.pgp Description: PGP signature
Re: MATE Desktop Environment in Debian
On Jul 22, Neil Williams codeh...@debian.org wrote: Bitter experience. My main problem is that I don't want to follow the direction set out by GNOME3 with the shell and lack of usable panels, You don't have to. I use gnome applications and the panel with fvwm. -- ciao, Marco signature.asc Description: Digital signature
Re: MATE Desktop Environment in Debian
I agree with Josselin here. There is no point in forking gconf and other libraries. Even worse, this will increase the incompatibility between different desktops. For example, an app making use of gconftool-2 will not work when there's mateconftool-2 in the system instead. If you want to fix a bug in GNOME libraries, you should propose a patch to their bugzilla, they will be happy to accept it. Also, we try to remove existing packages that use deprecated libraries (like libgnome* and libbonobo*), and I think it's not good at all to add new packages relying on those libraries. -- Dmitry Shachnev (Debian Maintainer) -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1342780899.5199.17.camel@eeepc
Re: MATE Desktop Environment in Debian
Le mercredi 08 février 2012 à 00:53 +0100, Stefano Karapetsas a écrit : Many users are using it well. Now that this is enough stable, I begun the process for ask the inclusion in Debian. The first package is mate-common. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=658783 (ITP) http://lists.debian.org/debian-mentors/2012/02/msg00115.html (RFS) With this mail I wish to have opinions and suggestions about our work from Debian Developers. MATE introduces a lot of code duplication, which is considered bad in Debian, and is based on obsolete technologies - not just GTK2, which will of course remain for a long time, but also things like Bonobo which very few people really understand, and which are the cause of a lot of not-well-understood bugs. For these reasons I object to having MATE in Debian. OTOH I invite you to contribute to GNOME 3 packaging to make it look great and fix remaining regressions. I am of course not the one to decide whether your packages can be accepted; the FTP masters will. Cheers, -- .''`. Josselin Mouette : :' : `. `' `- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1328691338.3202.364.camel@pi0307572
Re: MATE Desktop Environment in Debian
On 08/02/12 09:55, Josselin Mouette wrote: Le mercredi 08 février 2012 à 00:53 +0100, Stefano Karapetsas a écrit : Many users are using it well. Now that this is enough stable, I begun the process for ask the inclusion in Debian. The first package is mate-common. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=658783 (ITP) http://lists.debian.org/debian-mentors/2012/02/msg00115.html (RFS) With this mail I wish to have opinions and suggestions about our work from Debian Developers. MATE introduces a lot of code duplication, which is considered bad in Debian, and is based on obsolete technologies - not just GTK2, which will of course remain for a long time, but also things like Bonobo which very few people really understand, and which are the cause of a lot of not-well-understood bugs. Maybe this could benefit to both parties if MATE team tries to reduce usage of obsolete libraries to a bare minimum, and avoid using bug monsters (like Bonobo and others). I guess this means a lot of work, but that's the price to pay to ease its maintainability on the long term and having it packaged within Debian. Did MATE team consider such a plan? If yes, what was the outcome of the discussion? Regards, -- Mehdi -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4f325217.8050...@dogguy.org
Re: MATE Desktop Environment in Debian
Il 2012-02-08 11:44 Mehdi Dogguy ha scritto: On 08/02/12 09:55, Josselin Mouette wrote: MATE introduces a lot of code duplication, which is considered bad in Debian, and is based on obsolete technologies - not just GTK2, which will of course remain for a long time, but also things like Bonobo which very few people really understand, and which are the cause of a lot of not-well-understood bugs. Maybe this could benefit to both parties if MATE team tries to reduce usage of obsolete libraries to a bare minimum, and avoid using bug monsters (like Bonobo and others). I guess this means a lot of work, but that's the price to pay to ease its maintainability on the long term and having it packaged within Debian. Did MATE team consider such a plan? If yes, what was the outcome of the discussion? Yeah. In our roadmap there is the dismissal of the obsolete libraries, like the replacement of MateConf (the fork of GConf) with GSettings, and so on. We are studying how replace Bonobo and ORBIT, too. About this point I talked some time ago with Karen Sandler to find a meeting point with GNOME3 developers to share more libraries possible between the two desktop environments, to reduce the duplicated work. I dont think work only on fallback session is the solution, because GNOME3 is based on a different idea. I saw some gnome design team mockups of all applications, and I find its far from GNOME2. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1c760a5fa4a51c4b2868ef9f9f288...@karapetsas.com
Re: MATE Desktop Environment in Debian
On 08/02/12 14:05, Stefano Karapetsas wrote: I saw some gnome design team mockups of all applications, and I find its far from GNOME2. Then, why don't you help them? (It is easier than re-packaging and maintaining Gnome2). Regards, -- Mehdi -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4f327766.4010...@dogguy.org
Re: MATE Desktop Environment in Debian
Le mercredi 08 février 2012 à 14:05 +0100, Stefano Karapetsas a écrit : Yeah. In our roadmap there is the dismissal of the obsolete libraries, like the replacement of MateConf (the fork of GConf) with GSettings, and so on. Sorry but what is the point of *forking* GConf? What does it bring, apart from more work for you and more processes for your users? -- .''`. Josselin Mouette : :' : `. `' `- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1328708469.3202.417.camel@pi0307572
Re: Bug#658783: MATE Desktop Environment in Debian
Il 2012-02-08 14:23 Mehdi Dogguy ha scritto: I saw some gnome design team mockups of all applications, and I find its far from GNOME2. Then, why don't you help them? (It is easier than re-packaging and maintaining Gnome2). Regards, I just answered before. GNOME3 is a completely different desktop, and I dont think it can have two ideas in same environment. It should follow it main goal to have an attractive desktop. We started to work on MATE and we understand that it's possible and human to work and maintain it. Best regards, Stefano -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1d1cccb7365b7c593fdac90c0fbb8...@karapetsas.com
Re: Bug#658783: MATE Desktop Environment in Debian
Il 2012-02-08 14:41 Josselin Mouette ha scritto: Le mercredi 08 février 2012 à 14:05 +0100, Stefano Karapetsas a écrit : Yeah. In our roadmap there is the dismissal of the obsolete libraries, like the replacement of MateConf (the fork of GConf) with GSettings, and so on. Sorry but what is the point of *forking* GConf? What does it bring, apart from more work for you and more processes for your users? GConf is deprecated. MateConf is born to have a temporary solution until we choose the replacement for it. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/9e93a1a64a5422eb07cefad74af92...@karapetsas.com
Re: Bug#658783: MATE Desktop Environment in Debian
Le mercredi 08 février 2012 à 14:52 +0100, Stefano Karapetsas a écrit : GConf is deprecated. MateConf is born to have a temporary solution until we choose the replacement for it. GConf is deprecated, but it is still maintained. It is still used e.g. by evolution. I don’t see the point of renaming it, starting another daemon, and whatnot, if you provide the same functionality. If you want to keep maintaining GConf for longer than they plan, I don’t think upstream developers will prevent you from doing it. -- .''`. Josselin Mouette : :' : `. `' `- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1328710995.3202.429.camel@pi0307572
Re: Bug#658783: MATE Desktop Environment in Debian
Il 2012-02-08 15:23 Josselin Mouette ha scritto: GConf is deprecated, but it is still maintained. It is still used e.g. by evolution. I don’t see the point of renaming it, starting another daemon, and whatnot, if you provide the same functionality. If you want to keep maintaining GConf for longer than they plan, I don’t think upstream developers will prevent you from doing it. This is a good news. I didnt know that GConf is still maintained. This is a good point where we can start to share our forces! -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/85e5d3caf3db02ce87e346f3b460f...@karapetsas.com
MATE Desktop Environment in Debian
Hi, I am a developer of MATE Desktop Environment (http://www.mate-desktop.org), a fork of GNOME2 started in June 2011. Currently MATE has a little team of developers and contributors, and many users thanks to ready packages for some distributions (Debian, Ubuntu, Arch, Fedora) and thanks to the inclusion in the latest Linux Mint edition. A lot of work is done from the initial renaming script. We solved many bugs and fixed all things with the latest libraries since GNOME2 has gone, and we fixed conflicts with GNOME3. We added some new features, too. I think to give to users the chance to use the desktop that they used for many years is a great thing for a distribution, if this is supported and has a team who work on it. Someone can object that MATE is based on a dead technology, but GTK3 support is on our roadmap and we already started to work on it. Someone else can object Why dont you help GNOME3 fallback session or XFCE?. Well, our goal is to keep things like GNOME2. In GNOME3 there are a lot of changes in all software, e.g. in Nautilus and in many others. XFCE is very different from GNOME2. We want to be focused on this purpose to support a desktop like GNOME2. The nice peculiarity of the free software is anyone can develop applications at will. Our work doesnt damage other software. Completed the brief introduction about MATE, I'm going to write about the goal of this mail. I am a Debian user for many years and I know well all his components, like dpkg, the debian policy and the process to have new packages. I started to contribute on MATE as packager for Debian. We have already a repository for Wheezy (http://wiki.mate-desktop.org/download#debian). Many users are using it well. Now that this is enough stable, I begun the process for ask the inclusion in Debian. The first package is mate-common. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=658783 (ITP) http://lists.debian.org/debian-mentors/2012/02/msg00115.html (RFS) With this mail I wish to have opinions and suggestions about our work from Debian Developers. Thank you! Stefano Karapetsas -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e944551ee6cb118622d33c2a01ebe...@karapetsas.com