Re: dbus - Was: A thread that shouldn't be mentioned anymore
DBus isn't a problem per se, it just can cause issues, when implemented without thinking about the needs of all users? Right but it's actually much worse than that. Take mozilla firefox even which may or may not have been changed due to me bringing it up on the dev-security list. Without dbus in a chroot it would die, the reason was handling it's text configuration files, which is obviously rediculously pointless and I assume with some confidence, actually quite dumb. -- ___ 'Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface' (Doug McIlroy) ___ -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/556735.15807...@smtp112.mail.ird.yahoo.com
Re: dbus - Was: A thread that shouldn't be mentioned anymore
Le 13/04/2013 19:39, Kevin Chadwick a écrit : DBus isn't a problem per se, it just can cause issues, when implemented without thinking about the needs of all users? Right but it's actually much worse than that. Take mozilla firefox even which may or may not have been changed due to me bringing it up on the dev-security list. Without dbus in a chroot it would die, the reason was handling it's text configuration files, which is obviously rediculously pointless and I assume with some confidence, actually quite dumb. Same thing for emacs-gtk. Compiled with gconf (but is it possible otherwise) thus it sends error messages without dbus, whereas it does not need it. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/51699192.5000...@rail.eu.org
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Sat, Apr 13, 2013 at 10:39 AM, Kevin Chadwick ma1l1i...@yahoo.co.uk wrote: DBus isn't a problem per se, it just can cause issues, when implemented without thinking about the needs of all users? Right but it's actually much worse than that. Take mozilla firefox even which may or may not have been changed due to me bringing it up on the dev-security list. Without dbus in a chroot it would die, the reason was handling it's text configuration files, which is obviously rediculously pointless and I assume with some confidence, actually quite dumb. Are you sure about that? I have never seen anything dbus related in any version of Mozilla or Firefox, aside from one extension that never really when anywhere. Cheers, Kelly Clowers -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAFoWM=-3bnt0ofyokjo50amvjhx8vtgoaukq9dp5epkrnew...@mail.gmail.com
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Sat, 13 Apr 2013 10:34:36 -0700 Kelly Clowers wrote: DBus isn't a problem per se, it just can cause issues, when implemented without thinking about the needs of all users? Right but it's actually much worse than that. Take mozilla firefox even which may or may not have been changed due to me bringing it up on the dev-security list. Without dbus in a chroot it would die, the reason was handling it's text configuration files, which is obviously rediculously pointless and I assume with some confidence, actually quite dumb. Are you sure about that? I have never seen anything dbus related in any version of Mozilla or Firefox, aside from one extension that never really when anywhere. It does as you can see from the output when running it in a chroot and for atleast one release it would die. (firefox:1515): GConf-WARNING **: Client failed to connect to the D-BUS daemon: Failed to connect to socket /tmp/dbus-jEHTNI62oh: Connection refused Fontconfig error: Cannot load default config file -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2572.57457...@smtp112.mail.ir2.yahoo.com
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Sat, Apr 13, 2013 at 5:02 PM, Kevin Chadwick ma1l1i...@yahoo.co.uk wrote: On Sat, 13 Apr 2013 10:34:36 -0700 Kelly Clowers wrote: DBus isn't a problem per se, it just can cause issues, when implemented without thinking about the needs of all users? Right but it's actually much worse than that. Take mozilla firefox even which may or may not have been changed due to me bringing it up on the dev-security list. Without dbus in a chroot it would die, the reason was handling it's text configuration files, which is obviously rediculously pointless and I assume with some confidence, actually quite dumb. Are you sure about that? I have never seen anything dbus related in any version of Mozilla or Firefox, aside from one extension that never really when anywhere. It does as you can see from the output when running it in a chroot and for atleast one release it would die. (firefox:1515): GConf-WARNING **: Client failed to connect to the D-BUS daemon: Failed to connect to socket /tmp/dbus-jEHTNI62oh: Connection refused Fontconfig error: Cannot load default config file Odd. Just looked at Dbus with SeaMonkey and Firefox running, nothing from them. Aurora channel, stock mozilla.org versions. Cheers, Kelly -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAFoWM=_3xkcYta-xvUt_y5+O_W=pgucskxwhyb7nu+wle6f...@mail.gmail.com
dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, 2013-04-09 at 07:21 -1000, Joel Roth wrote: Ralf Mardorf wrote: dbus often is a PITA! Do you have experience with dbus's predecessors, such as CORBA? No. I guess a predecessor won't help, if applications depend on dbus, such as jackd/jackdmp. I'm aware that I can use jackdmp without dbus, since I'm already doing this and I read that others are able to handle this dbus issue even when they run jack with dbus for sessions without X, IIUC. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365528548.2607.255.camel@archlinux
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, 2013-04-09 at 19:29 +0200, Ralf Mardorf wrote: On Tue, 2013-04-09 at 07:21 -1000, Joel Roth wrote: Ralf Mardorf wrote: dbus often is a PITA! Do you have experience with dbus's predecessors, such as CORBA? No. I guess a predecessor won't help, if applications depend on dbus, such as jackd/jackdmp. I'm aware that I can use jackdmp without dbus, since I'm already doing this and I read that others are able to handle this dbus issue even when they run jack with dbus for sessions without X, IIUC. Oops, perhaps you mean that they were less good :D. I confused it with the word successor and noticed it after I sent the mail. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365528739.2607.257.camel@archlinux
Re: dbus - Was: A thread that shouldn't be mentioned anymore
Le 09/04/2013 19:29, Ralf Mardorf a écrit : On Tue, 2013-04-09 at 07:21 -1000, Joel Roth wrote: Ralf Mardorf wrote: dbus often is a PITA! Do you have experience with dbus's predecessors, such as CORBA? No. I guess a predecessor won't help, if applications depend on dbus, such as jackd/jackdmp. I'm aware that I can use jackdmp without dbus, since I'm already doing this and I read that others are able to handle this dbus issue even when they run jack with dbus for sessions without X, IIUC. WHen reporting erreor messages when emacs-gtk from a ssh -X session, I was told that it's normal that it won't work since it uses dbus. Since then I have a bad feeling with dbus dependencies which seem to reduce the capabilities of otherwise perfectly working software. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/516450e9.8000...@rail.eu.org
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, 2013-04-09 at 19:33 +0200, Erwan David wrote: WHen reporting erreor messages when emacs-gtk from a ssh -X session, I was told that it's normal that it won't work since it uses dbus. Since then I have a bad feeling with dbus dependencies which seem to reduce the capabilities of otherwise perfectly working software. AFAIK jack2 does work perfectly with dbus too, but using it without dbus is KISS/idiot-proof and using it with dbus is rocket science. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365529320.2607.260.camel@archlinux
Re: dbus - Was: A thread that shouldn't be mentioned anymore
Ralf Mardorf wrote: On Tue, 2013-04-09 at 19:29 +0200, Ralf Mardorf wrote: On Tue, 2013-04-09 at 07:21 -1000, Joel Roth wrote: Ralf Mardorf wrote: dbus often is a PITA! Do you have experience with dbus's predecessors, such as CORBA? No. I guess a predecessor won't help, if applications depend on dbus, such as jackd/jackdmp. I'm aware that I can use jackdmp without dbus, since I'm already doing this and I read that others are able to handle this dbus issue even when they run jack with dbus for sessions without X, IIUC. Oops, perhaps you mean that they were less good :D. I confused it with the word successor and noticed it after I sent the mail. Perhaps D-Bus is not good, but maybe it is less bad? I'll go on the record in favor of configurability: good to be able to opt out if you don't need it. ( /me doesn't know if he needs it or not. ) -- Joel Roth -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130409180409.GA16123@sprite
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, Apr 9, 2013 at 11:04 AM, Joel Roth jo...@pobox.com wrote: Ralf Mardorf wrote: On Tue, 2013-04-09 at 19:29 +0200, Ralf Mardorf wrote: On Tue, 2013-04-09 at 07:21 -1000, Joel Roth wrote: Ralf Mardorf wrote: dbus often is a PITA! Do you have experience with dbus's predecessors, such as CORBA? No. I guess a predecessor won't help, if applications depend on dbus, such as jackd/jackdmp. I'm aware that I can use jackdmp without dbus, since I'm already doing this and I read that others are able to handle this dbus issue even when they run jack with dbus for sessions without X, IIUC. Oops, perhaps you mean that they were less good :D. I confused it with the word successor and noticed it after I sent the mail. Perhaps D-Bus is not good, but maybe it is less bad? I'll go on the record in favor of configurability: good to be able to opt out if you don't need it. ( /me doesn't know if he needs it or not. ) D-Bus is good overall... There could definitely be improvements in remote connections though. I think there are workarounds... I use only CLI over SSH, though, so I never messed with it. CORBA was just terrible. I have encountered it a bit in the old Gnome days, and on AIX with CDE. Believe me, you don't want to deal with its bullshit. And that is coming from a user/admin POV. From everything I have heard it was worse for a programmer. I have done some simple Dbus stuff in Python and such, it seems simple enough. I never want to have to program any Corba... Cheers, Kelly Clowers -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAFoWM=_kHdQj3_R0905GVac7wm=rcaredqoe9ytke0zior0...@mail.gmail.com
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, 2013-04-09 at 11:47 -0700, Kelly Clowers wrote: D-Bus is good overall... There could definitely be improvements in remote connections though. I think there are workarounds... [snip] DBus isn't an issue for applications you'll use with a desktop environment, when those apps should communicate with each other, but it could become an issue, if apps should run on other setups (too) and it's an issue if simple commands that worked before, then won't work anymore, the user has to read tons of explanations and to do complicated things. Clueless users run into issues with jackd(mp), since they were not aware that they can use jackd without DBus, resp. they perhaps were not aware that it does run with DBus on their machines. Last year everything was ok, I updated Jack and this or that doesn't work anymore. - Imaginary User There are examples for other software, than DBus, where dependencies made production environments unusable, e.g. when the hard dependency to pulseaudio was added and disabling or coexisting didn't work. For other stuff package maintainers have to do a hard job, somebody seemingly does extract udev from systemd for Debian. I'm on a distro that follows upstream and there were many issues when they switched to systemd. IMO unneeded hard dependencies are an issue for several applications. http://en.wikipedia.org/wiki/KISS_principle Often new policies tend to break PC environments and are only an advantage for mobiles and tablet PCs. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365534505.2607.284.camel@archlinux
Re: dbus - Was: A thread that shouldn't be mentioned anymore
OT regarding to dbus: I wrote: For other stuff package maintainers have to do a hard job, somebody seemingly does extract udev from systemd for Debian. I'm on a distro that follows upstream and there were many issues when they switched to systemd. http://packages.debian.org/search?keywords=udev Perhaps nobody needs to do a hard job until now. I'm not aware when they merged udev and systemd, but all Debian versions of udev are completely outdated. Latest version of udev for Debian is 175-7.1. Arch Linux: $ pacman -Q systemd systemd 200-1 -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365535215.2607.289.camel@archlinux
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, Apr 09, 2013 at 09:20:15PM +0200, Ralf Mardorf wrote: OT regarding to dbus: I wrote: For other stuff package maintainers have to do a hard job, somebody seemingly does extract udev from systemd for Debian. I'm on a distro that follows upstream and there were many issues when they switched to systemd. According to the announcement at the time, udev is designed to build separately, without systemd for use under conditions (such as initrd) where systemd is not available. https://lwn.net/Articles/490413/ http://packages.debian.org/search?keywords=udev Perhaps nobody needs to do a hard job until now. I'm not aware when they merged udev and systemd, but all Debian versions of udev are completely outdated. Latest version of udev for Debian is 175-7.1. Arch Linux: $ pacman -Q systemd systemd 200-1 -- Joel Roth -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130409193241.GB17280@sprite
Re: dbus - Was: A thread that shouldn't be mentioned anymore
D-Bus is good overall... The good thing about standard IPC was that you would have to develop the protocol etc.. which means if your app used it. 1./ You needed to use it otherwise you wouldn't. 2./ You made an app specific mechanism (very good if your good but could be bad, the latter is what dbus tackles) The problem with dbus isn't dbus, it is that developers are becoming a big problem because they are using it way too much as a first choice. You should only use dbus when you need to. Some software is unfortunately encouraging this and in turn other bad practices. Take Windows, atleast XP (I haven't looked so close since but I expect little has changed in this department), Scripts have to be enabled to activate XP and IPC is required for almost everything. Try switching off RPC (prepare to reboot to re-enable it), and guess what, Windows is completely unreliable and insecure. Polkit instead of sudo, IPC and scripts when neither are required or a good choice. The reason for IPC, because it wasn't designed for just the task that sudo does. Why polkit is used rather than sudo because polkits author helped write the odd script like pm-suspend and is in with the udisks author. What I really don't get though is why there are so many easily avoidable hard dependencies. -- ___ 'Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface' (Doug McIlroy) ___ -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130409225956.46ecf...@kc-sys.chadwicks.me.uk
Re: dbus - Was: A thread that shouldn't be mentioned anymore
Too funny, under the jokes is one, asking to merge it with dbus. Udev and systemd to merge Posted Apr 3, 2012 22:31 UTC Next step is of course to integrate D-Bus in systemd, no? On Tue, 2013-04-09 at 09:32 -1000, Joel Roth wrote: https://lwn.net/Articles/490413/ I don't know if systemd and udev still can be easily separated. I don't think so. Does anybody know? I guess what does cause all that frustration is, that nobody can follow all those new changes and instead of maintaining one project, they (or we, the community) too often drop and replaced projects. Btw. does Debian still auto-mount to /media or did it also adapt the new file system hierarchy? I guess regarding to DBus everything was said? DBus isn't a problem per se, it just can cause issues, when implemented without thinking about the needs of all users? -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365542919.2607.307.camel@archlinux
Re: dbus - Was: A thread that shouldn't be mentioned anymore
On Tue, 2013-04-09 at 22:59 +0100, Kevin Chadwick wrote: What I really don't get though is why there are so many easily avoidable hard dependencies. +1 We can see what projects come with the most hilarious dependencies and that's why we should avoid EMONG, or similar, have forgotten the name. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1365544778.2607.324.camel@archlinux