Re: elementary shutdown kills opened applications
On Tue, Nov 11, 2014 at 3:22 PM, Carl name.is.c...@gmail.com wrote: Hello, I'm trying to investigate a bug [1] we have on elementary freya (which is based on ubuntu 14.04). Do you have an idea of a portion of the code that I should look at? Cheers, Carl. ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Gnome / systemd
(this is just a resend message from devuan list... But i would like to get technical answers and no flamewar) I would like to know how you, as a GNOME Core member, think about systemd - Gnome and Operating Systems like the BSDs. And i already asked this in IRC but what is the result you (The Gnome Team) expect. Is the systemd dependency just because of dbus and logind? The dbus part is the part which i understand. But what is actually the benefit of using logind in comparison to pam. -- Mit freundlichen Grüßen Patrick Erdmann XMPP/Mail: patr...@perdmann.de ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Gnome / systemd
On Fri, Dec 5, 2014 at 11:58 AM, Patrick Erdmann patr...@perdmann.de wrote: (this is just a resend message from devuan list... But i would like to get technical answers and no flamewar) I would like to know how you, as a GNOME Core member, think about systemd - Gnome and Operating Systems like the BSDs. Systemd offers us many useful APIs (in some cases, it directly took over D-Bus APIs that we've had to maintain ourselves before). It makes building a functional desktop on linux much easier and thus is very welcome. And i already asked this in IRC but what is the result you (The Gnome Team) expect. Is the systemd dependency just because of dbus and logind? The dbus part is the part which i understand. But what is actually the benefit of using logind in comparison to pam. The question really doesn't make much sense. First, D-Bus has been around for more than a decade, not just on linux but on bsds as well, and should absolutely not be a controversial dependency. Mentioning it in the same context is 'guilt by association' and demonstrates lack of factual knowledge. Second, pam does not do any of the things that logind does. We are still using pam for what it _does_ offer... ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Gnome / systemd
On Fri, Dec 5, 2014 at 8:58 AM, Patrick Erdmann patr...@perdmann.de wrote: (this is just a resend message from devuan list... But i would like to get technical answers and no flamewar) I would like to know how you, as a GNOME Core member, think about systemd - Gnome and Operating Systems like the BSDs. And i already asked this in IRC but what is the result you (The Gnome Team) expect. Is the systemd dependency just because of dbus and logind? The dbus part is the part which i understand. But what is actually the benefit of using logind in comparison to pam. logind and PAM work well together. In fact, they have some tight integration so logind can do its job correctly: http://cgit.freedesktop.org/systemd/systemd/tree/src/login/pam_systemd.c logind might be a misnomer, since it doesn't actually do anything related to authentication or authorization. What logind does do is track logged in sessions, and help manage that in a central location. For instance, when gdm tries to log a user in, we actually double-check with logind and make sure that if they have any existing sessions, that we switch to that. But it's smart enough to not do that switch if you are logged in via a VT or ssh, etc. When a user tries to shut down, we first make sure that there are no other users logged in, and we query logind for that information. In the new Wayland world, we actually use logind to keep track of which sessions are on which VTs so we know which VT to switch to when the user hits a keybinding, and use it to hand us device nodes for the display and input systems, which we can't directly open due to permissions. It's better to do all of this user session tracking in a central component. This used to be done inside a module called ConsoleKit, but we ran into deep fundamental issues when it wasn't integrated tightly with the init system. That's why it's now part of systemd. Some people haven't liked this approach, and have forked ConsoleKit to make ConsoleKit2. GNOME doesn't officially support the ConsoleKit APIs anymore, but these community members also have patches to add it back in, based on our old ConsoleKit implementation. What PAM handles is the authentication part of logging in. It allows people to make pluggable authentication modules so people can prove who they are who they say they are, through passwords, remote lookup, two-factor authentication, or even a game of rock paper scissors ( https://github.com/nalind/pam_rps ). -- Mit freundlichen Grüßen Patrick Erdmann XMPP/Mail: patr...@perdmann.de ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list -- Jasper ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Gnome / systemd
On Mon, Jan 12, 2015 at 09:50:49AM -0600, Michael Catanzaro wrote: I assume you're contacting us because you're considering GNOME for your default desktop environment. That's something that I would like to encourage. :) I also assume you've already willing to reimplement the various D-Bus interfaces provided by systemd, and that logind is your concern here. I've asked Devuan to post here a while ago. It seems that it was stuck in the moderation queue. Despite various things[1] I've noticed assume people mean well worked quite well. I don't participate in their mailing list anymore as the atmosphere is (was?) toxic. I believe we're sometimes too aggressive in our replies/communication. Regarding Devuan: They've made something which provides a logind API with ConsoleKit2 as a backend. See https://github.com/dimkr/LoginKit. -- Regards, Olav [1] the do our bidding website, huge amount of trolls in their mailing list, simplistic blaming GNOME, etc ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Standardizing the latest dev code Version field value in GNOME Bugzilla?
On Mon, 2015-01-12 at 21:28 +0100, Andre Klapper wrote: 1) Does this variety of names create a real problem? For example, does it make using git-bz harder (remembering the version value, assuming you mostly develop against the latest code)? Do you care / do enough people use git-bz / do people not use git-bz because of this? Other reasons? The occasional times I'm using git-bz, I am either lucky that it works or if not, I usually resort back to creating bug reports manually (never invested enough time to remember the various names used). From this PoV, I think it would be great if git-bz were able to always create an entry against 'latest devel branch', without the occasional contributor having to care for it all. 2) Shall we standardize this? (I volunteer to rename; git master seems to be the most popular option.) Or is that a waste of time? Sounds like a good approach; I don't mind the name that much, as long as git-bz basically 'uses the right default' Cheers, Dominique -- Dimstar / Dominique Leuenberger dims...@opensuse.org ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Standardizing the latest dev code Version field value in GNOME Bugzilla?
Welcome GNOME community to this bikeshed!
Many products in Bugzilla's {Core, Platform, Bindings, Applications}
classifications have entries in the Version field which refer to
latest dev code, not expressed via some version number.
We have 18 different names for describing that in GNOME Bugzilla.
See the list below.
I bring that up on d-d-l because it was brought up ages ago in
https://bugzilla.gnome.org/show_bug.cgi?id=681607 and to some extend
also in https://bugzilla.gnome.org/show_bug.cgi?id=710799
QUESTIONS:
1) Does this variety of names create a real problem?
For example, does it make using git-bz harder (remembering the version
value, assuming you mostly develop against the latest code)?
Do you care / do enough people use git-bz / do people not use git-bz
because of this? Other reasons?
2) Shall we standardize this? (I volunteer to rename; git master seems
to be the most popular option.) Or is that a waste of time?
LIST OF VALUES:
* CVS (head) (2: libxml2, libxslt)
* CVS (1: libxml++)
* CVS HEAD (6: gnome-backgrounds, gnome-menus, gnome-user-share,
atk, anjuta, gconf-editor)
* SVN (2: adwaita-icon-theme, sound-juicer)
* SVN trunk (3: gnome-desktop, gnome-session, libwnck)
* SVN TRUNK (1: anjuta)
* GIT (1: Gnumeric)
* git (2: Gstreamer, cheese)
* git master (56: baobab, dconf, gnome-color-manager,
gnome-dictionary, gnome-font-viewer, gnome-packagekit,
gnome-screenshot, gnome-system-log, gnome-themes-standard,
gsettings-desktop-schemas, gvfs, json-lib, librsvg, mutter,
yelp-tools, yelp-xsl, clutter, cogl, folks, gdk-pixbuf,
gtksourceview, libgdata, libgee, libnotify, libpeas, librest,
NetworkManager, gstreamermm, gstreamermm-plugins-good,
librsvgmm, libvtemm, mm-common, anjuta, caribou, five-or-more,
four-in-a-row, gedit, glade, gnome-chess, gnome-getting-started,
gnome-klotski, gnome-mahjongg, gnome-maps, gnome-mines,
gnome-nettool, gnome-nibbles, gnome-robots, gnome-search-tool,
gnome-sudoku, gnome-tetravex, gnote, iagno, lightsoff,
quadrapassel, swell-foop, transmageddon)
* Git master (1: pygobject)
* Git Master (2: pygtk, pygtksourceview)
* GIT master (1: ekiga)
* git head (2: seed, gitg)
* trunk (5: eog, gnome-control-center, tracker, devhelp,
gnome-games)
* Trunk (2: gnome-keyring, hitori)
* master (10: epiphany, evince, gnome-calculator, gnome-terminal,
gucharmap, vte, brasero, Rygel, vinagre, vino)
* HEAD (4: gnome-common, system-monitor, seahorse, Yelp)
PLEASE NOTE:
This thread is NOT about discussing e.g. Component: General or
Version: unspecified. Also, whether project X should really be in
classification Y, or if project Z is active, is a different topic.
If you want to discuss such topics, start your new thread with a
separate mail subject in a few days. Only one bikeshed at a time!
Cheers,
andre
--
Andre Klapper | ak...@gmx.net
http://blogs.gnome.org/aklapper/
___
desktop-devel-list mailing list
desktop-devel-list@gnome.org
https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Standardizing the latest dev code Version field value in GNOME Bugzilla?
On Mon, 2015-01-12 at 16:43 -0600, Michael Catanzaro wrote: On Mon, Jan 12, 2015 at 2:28 PM, Andre Klapper ak...@gmx.net wrote: 1) Does this variety of names create a real problem? For example, does it make using git-bz harder (remembering the version value, assuming you mostly develop against the latest code)? Do you care / do enough people use git-bz / do people not use git-bz because of this? Other reasons? It's a minor annoyance to have to load Bugzilla in a web browser to figure out the version to use when filing a bug with git-bz. Agreed. I would actually call it a major annoyance, since it entirely breaks the git-bz workflow. I think I'm just being an old man though. Note that as long as there is also no standardized General/general/Miscellaneous/misc component for all products, that would still need to be done anyway to figure out what Component to use when filing bugs with git-bz. Actually, the components in a product can be queried using Bugzilla's XMLRPC API, and if hooked up to tab completion in git-bz, could be used from the command line: https://bugzilla.gnome.org/show_bug.cgi?id=742402 2) Shall we standardize this? (I volunteer to rename; git master seems to be the most popular option.) Or is that a waste of time? I think standardizing on 'git master' would be great. Thanks for volunteering. Agreed. Philip signature.asc Description: This is a digitally signed message part ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Installing DBus interface files for services
Hi all, I was wondering if there's any reason we typically don't install on the system DBus XML interface files for services. On my system, I can see a bunch of definitions in /usr/share/dbus-1/interfaces, but it's by no means a complete list of all the services in the system. Standardizing such a practice would make it easier to write code that uses e.g. gdbus-codegen to automatically generate code for those interfaces; currently a lot of projects need to copy/paste the interface definition in their source tree, which is impractical and can lead to inconsistencies when one version of the interface is updated (in a backwards-compatible way) but not the other side. Thanks, Cosimo ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Gnome / systemd
On Mon, Jan 12, 2015 at 8:53 AM, Olav Vitters o...@vitters.nl wrote: On Mon, Jan 12, 2015 at 08:25:26AM -0800, Sri Ramkrishna wrote: Internet. The greatest way of changing minds and hearts is to be calm and coherent on what we're trying to do. That said, this thread and its replies are a perfect example of how we should approach issues. If we want to gain mindshare, we need to be better than the Internet. :-) Almost forgot: Devuan does have a few people who you can communicate with. Entirely reasonable people who are willing and have put in work. They might (initially) not have a good understanding or maybe think things were done out of forcing, but things changed around quickly with just explaining things. Happy to talk to distributions who are interested in having GNOME front and center. sri ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Gnome / systemd
On Mon, Jan 12, 2015 at 8:02 AM, Olav Vitters o...@vitters.nl wrote: On Mon, Jan 12, 2015 at 09:50:49AM -0600, Michael Catanzaro wrote: I assume you're contacting us because you're considering GNOME for your default desktop environment. That's something that I would like to encourage. :) I also assume you've already willing to reimplement the various D-Bus interfaces provided by systemd, and that logind is your concern here. I've asked Devuan to post here a while ago. It seems that it was stuck in the moderation queue. Despite various things[1] I've noticed assume people mean well worked quite well. I don't participate in their mailing list anymore as the atmosphere is (was?) toxic. I believe we're sometimes too aggressive in our replies/communication. I agree that sometimes we get a little too aggressive. It is a little understandable that we get a little ansy because of what we read on the Internet. The greatest way of changing minds and hearts is to be calm and coherent on what we're trying to do. That said, this thread and its replies are a perfect example of how we should approach issues. If we want to gain mindshare, we need to be better than the Internet. :-) sri ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: Gnome / systemd
On Mon, Jan 12, 2015 at 08:25:26AM -0800, Sri Ramkrishna wrote: Internet. The greatest way of changing minds and hearts is to be calm and coherent on what we're trying to do. That said, this thread and its replies are a perfect example of how we should approach issues. If we want to gain mindshare, we need to be better than the Internet. :-) Almost forgot: Devuan does have a few people who you can communicate with. Entirely reasonable people who are willing and have put in work. They might (initially) not have a good understanding or maybe think things were done out of forcing, but things changed around quickly with just explaining things. -- Regards, Olav ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
Re: elementary shutdown kills opened applications
One idea is that you're killing the X server directly. GTK+ applications take a disconnect from the primary display as an indication that the user's session has closed, and that they should quit (actually, it's just considered an IO error, to which the response of a standard X11 application is fatal by default, but we never changed that behavior for that reason). You need to keep the X server up during that time, and that might require some fairly invasive changes to gnome-session or gdm. On Sat, Nov 22, 2014 at 6:43 PM, Carl name.is.c...@gmail.com wrote: On Tue, Nov 11, 2014 at 3:22 PM, Carl name.is.c...@gmail.com wrote: Hello, I'm trying to investigate a bug [1] we have on elementary freya (which is based on ubuntu 14.04). Do you have an idea of a portion of the code that I should look at? Cheers, Carl. ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list -- Jasper ___ desktop-devel-list mailing list desktop-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/desktop-devel-list
