Re: [Mageia-dev] [RFC] urpmq --not-available (was renaming -debug packages as -debuginfo)
Pascal Terjan a écrit : On Mon, Dec 3, 2012 at 2:09 PM, Pascal Terjan pter...@gmail.com wrote: On Mon, Dec 3, 2012 at 1:13 PM, andre999 andre999...@laposte.net wrote: Colin Guthrie a écrit : ... general advice of running urpmq --not-available would be a good enough fallback IMO. By chance is that compared to cauldron ? To the configured media Or is some other parameter required ? None should be needed On my system, it lists 981 packages (more than 2/3 of those installed), on an up to date mga2. The vast majority are tagged mga2 (A few are tagged mga1, required by some mga2 package, as well as various upstream packages.) Scanning my mirror 2/core/updates, a lot are in my list. (The list gives the full package file names, except the .rpm suffix.) Do you have some example? Can you give your urpmi configuration? Testing on various machines, it seems wrong when there are several versions of the package available (updates or core+tainted) Ok, most of my problem was due to my inadvertantly removing the official mirrors when I had a problem booting a few days ago. The only repo defined was the mga2 dvd. (I had tried simply reinstalling mga2, a trick which sometimes works. I ended up having to scan for remove bad blocks, which did work.) Sorry for the noise. After activating official mga2 repos, adding + activating my usual local repos of updates I've installed, I got a list of 133 packages. A few tests showed that the packages listed were only those *not* in activated repos, whether local or not. Duplicate packages or updates in activated repos didn't produce any false positives. (It is important to ensure that local repos are up to date.) In other words, the list was correct. So to ensure that the urpmq --non-available lists all non-supported (including obsoleted) packages, in addition to setting up official repos, local repos which might contain such packages should be deactivated. -- André
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
Le 03/12/2012 23:42, zezinho a écrit : Trying to launch openalchemist gave me a crash because Clanlib2 uses SSE2 on SSE2 systems for software rendering, while we disabled it at configure. What should we do? I think enabling it is better, as all i586 CPU with less than 10 years come with SSE2. But our policy forbids it ;-) Cool. Now we have software crashing because they don't use available features, what a nice idea... -- BOFH excuse #186: permission denied
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
On Mon, Dec 03, 2012 at 11:42:12PM +0100, zezinho wrote: ERROR FOLLOWS : Exception caught: Sorry, this compiled clanSWRender does not support SSE2, but your CPU does support SSE2. (Update clanSWRender to contain SSE2) Add a patch to remove the check. -- Regards, Olav
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
'Twas brillig, and zezinho at 03/12/12 22:42 did gyre and gimble: Trying to launch openalchemist gave me a crash because Clanlib2 uses SSE2 on SSE2 systems for software rendering, while we disabled it at configure. What should we do? I think enabling it is better, as all i586 CPU with less than 10 years come with SSE2. But our policy forbids it ;-) ERROR FOLLOWS : Exception caught: Sorry, this compiled clanSWRender does not support SSE2, but your CPU does support SSE2. (Update clanSWRender to contain SSE2) The code should really be complied but only enabled at runtime if it's available. PulseAudio does this. It's more of a pain to code for but does make like easier for distributions. But yeah, short term fix, nuke the test, and as a meme says, keep calm and carry on. Col -- Colin Guthrie colin(at)mageia.org http://colin.guthr.ie/ Day Job: Tribalogic Limited http://www.tribalogic.net/ Open Source: Mageia Contributor http://www.mageia.org/ PulseAudio Hacker http://www.pulseaudio.org/ Trac Hacker http://trac.edgewall.org/
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
'Twas brillig, and zezinho at 03/12/12 22:42 did gyre and gimble: Trying to launch openalchemist gave me a crash because Clanlib2 uses SSE2 on SSE2 systems for software rendering, while we disabled it at configure. What should we do? I think enabling it is better, as all i586 CPU with less than 10 years come with SSE2. But our policy forbids it ;-) ERROR FOLLOWS : Exception caught: Sorry, this compiled clanSWRender does not support SSE2, but your CPU does support SSE2. (Update clanSWRender to contain SSE2) The code should really be complied but only enabled at runtime if it's available. PulseAudio does this. It's more of a pain to code for but does make like easier for distributions. But yeah, short term fix, nuke the test, and as a meme says, keep calm and carry on. but, do ALL recent i586 CPU's have SSE2? what about the embedded and VIA cpu's and the likes?
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
On Tue, Dec 4, 2012 at 12:10 PM, AL13N al...@rmail.be wrote: 'Twas brillig, and zezinho at 03/12/12 22:42 did gyre and gimble: Trying to launch openalchemist gave me a crash because Clanlib2 uses SSE2 on SSE2 systems for software rendering, while we disabled it at configure. What should we do? I think enabling it is better, as all i586 CPU with less than 10 years come with SSE2. But our policy forbids it ;-) ERROR FOLLOWS : Exception caught: Sorry, this compiled clanSWRender does not support SSE2, but your CPU does support SSE2. (Update clanSWRender to contain SSE2) The code should really be complied but only enabled at runtime if it's available. PulseAudio does this. It's more of a pain to code for but does make like easier for distributions. But yeah, short term fix, nuke the test, and as a meme says, keep calm and carry on. but, do ALL recent i586 CPU's have SSE2? what about the embedded and VIA cpu's and the likes? The test is that it refuses to run if the lib is not using sse2 while it could (slowly). The clanSWRender lib needs to be fixed to have sse2 support on sse2 system and not on non sse2 systems so that we can enable it but that's something unrelated to dropping this test.
[Mageia-dev] strange GTK theme issue under KDE
When using Evolution under KDE - certain elements do not show up until I move the mouse over them (see snapshot) Is this a theme issue? This also happens in Libreoffice during the printing dialog - many elements in the window re missing until I move the mouse over them . . . Cheers, R.Fox attachment: snap1.png
Re: [Mageia-dev] strange GTK theme issue under KDE
On 12/4/2012 1:25 PM, Robert Fox wrote: When using Evolution under KDE - certain elements do not show up until I move the mouse over them (see snapshot) Is this a theme issue? This also happens in Libreoffice during the printing dialog - many elements in the window re missing until I move the mouse over them . . . Cheers, R.Fox Could you post the version you have of - evolution - oxygen-gtk3 - gtk3 Thanks ! Hugo
Re: [Mageia-dev] strange GTK theme issue under KDE
On 12/04/2012 01:25 PM, Robert Fox wrote: When using Evolution under KDE - certain elements do not show up until I move the mouse over them (see snapshot) Is this a theme issue? This also happens in Libreoffice during the printing dialog - many elements in the window re missing until I move the mouse over them . . . Libreoffice, as far as I know, uses gtk2 theming (or possibly Qt, not sure about what's packaged in latest mageia). Can someone confirm ? But in all cases this should be a different issue. Can you also test whether other gtk3 applications are affected, such as gedit, or nautilus ? And gtk2 applications ? (such as gimp) ? Thanks in advance, Hugo Cheers, R.Fox
Re: [Mageia-dev] [RFC] renaming -debug packages as -debuginfo
On Monday 3. December 2012 21.07, AL13N wrote: Which problem? the one which people complain about where the --orphans option removes too much... I've stopped using the orphan option. A few days ago when I uninstalled something I tested before submiting, and urpme said afterwards that several houndred packages where orphaned. One of them was kdebase4-common and every kernel except the running one. :-)= -- Johnny A. Solbu PGP key ID: 0xFA687324 signature.asc Description: This is a digitally signed message part.
Re: [Mageia-dev] strange GTK theme issue under KDE
On 12/04/2012 01:25 PM, Robert Fox wrote: When using Evolution under KDE - certain elements do not show up until I move the mouse over them (see snapshot) Is this a theme issue? This also happens in Libreoffice during the printing dialog - many elements in the window re missing until I move the mouse over them . . . Last question: can you describe which part of the window is in the screenshot ? I'm trying to reproduce here (with somewhat older versions of evolution than you though), and find no issue so far. I'll udpate later on. Hugo Cheers, R.Fox
Re: [Mageia-dev] [RFC] renaming -debug packages as -debuginfo
On Tue, Dec 4, 2012 at 1:55 PM, Johnny A. Solbu coo...@solbu.net wrote: On Monday 3. December 2012 21.07, AL13N wrote: Which problem? the one which people complain about where the --orphans option removes too much... I've stopped using the orphan option. A few days ago when I uninstalled something I tested before submiting, and urpme said afterwards that several houndred packages where orphaned. One of them was kdebase4-common and I guess you don't have task-kde4 installed anymore (which mean you don't want kde anymore). every kernel except the running one. :-) It should keep the running one and the latest one (if they are different), this is intended behaviour unless you explicitly installed them because you wanted those specific versions
Re: [Mageia-dev] [RFC] renaming -debug packages as -debuginfo
'Twas brillig, and Pascal Terjan at 04/12/12 14:16 did gyre and gimble: On Tue, Dec 4, 2012 at 1:55 PM, Johnny A. Solbu coo...@solbu.net wrote: On Monday 3. December 2012 21.07, AL13N wrote: Which problem? the one which people complain about where the --orphans option removes too much... I've stopped using the orphan option. A few days ago when I uninstalled something I tested before submiting, and urpme said afterwards that several houndred packages where orphaned. One of them was kdebase4-common and I guess you don't have task-kde4 installed anymore (which mean you don't want kde anymore). I think I remember it being removed as a result of some kde-games thing semi-recently. That's the fun of running Cauldron :D I see the usefulness of the orphan option, but I like a bit more control which is why I pushed (many years ago now) for the --not-available option that Pascal made not-rubbish :) Gives me a little more control over things. Col -- Colin Guthrie colin(at)mageia.org http://colin.guthr.ie/ Day Job: Tribalogic Limited http://www.tribalogic.net/ Open Source: Mageia Contributor http://www.mageia.org/ PulseAudio Hacker http://www.pulseaudio.org/ Trac Hacker http://trac.edgewall.org/
Re: [Mageia-dev] Will the 'official' Mageia-3 installer have a 'GRUB Legacy' option?
On Tue, 04 Dec 2012 12:19:21 +1100, blind Pete wrote: What I've been doing lately is to install everything with its default boot loader into its root partition and have a tiny partition at the end of the disk that contains just a lilo configuration file and map file. I have similar 'Plan B', where I would have GRUB-Legacy in the MBR pointing at a minimal GRUB-Legacy install, whose menu.lst could then be edited to chain into any other install (GRUB2 or GRUB-legacy). -- /\/\aurice
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
Le 04/12/2012 13:21, Pascal Terjan a écrit : On Tue, Dec 4, 2012 at 12:10 PM, AL13Nal...@rmail.be wrote: The code should really be complied but only enabled at runtime if it's available. PulseAudio does this. It's more of a pain to code for but does make like easier for distributions. But yeah, short term fix, nuke the test, and as a meme says, keep calm and carry on. It is not only a test, it is a crash. but, do ALL recent i586 CPU's have SSE2? what about the embedded and VIA cpu's and the likes? http://en.wikipedia.org/wiki/SSE2#CPUs_supporting_SSE2 The test is that it refuses to run if the lib is not using sse2 while it could (slowly). The clanSWRender lib needs to be fixed to have sse2 support on sse2 system and not on non sse2 systems so that we can enable it but that's something unrelated to dropping this test. No, the clanSWRender has only a SSE2 implementation because no one wants to write a SW render which is too slow. http://lists.debian.org/debian-devel-games/2012/03/msg00076.html I wonder if we could enable for now SSE2, till upstream/somebody has a better fix : for now, all SSE2 cpus crash. Enabling SSE2 it will only crash on non-SSE2 cpus (older than 10 years)
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
On Tue, Dec 4, 2012 at 6:00 PM, Jose Jorge lists.jjo...@free.fr wrote: Le 04/12/2012 13:21, Pascal Terjan a écrit : On Tue, Dec 4, 2012 at 12:10 PM, AL13Nal...@rmail.be wrote: The code should really be complied but only enabled at runtime if it's available. PulseAudio does this. It's more of a pain to code for but does make like easier for distributions. But yeah, short term fix, nuke the test, and as a meme says, keep calm and carry on. It is not only a test, it is a crash. It is a crash because the code says to crash if there is no sse2. but, do ALL recent i586 CPU's have SSE2? what about the embedded and VIA cpu's and the likes? http://en.wikipedia.org/wiki/SSE2#CPUs_supporting_SSE2 The test is that it refuses to run if the lib is not using sse2 while it could (slowly). The clanSWRender lib needs to be fixed to have sse2 support on sse2 system and not on non sse2 systems so that we can enable it but that's something unrelated to dropping this test. No, the clanSWRender has only a SSE2 implementation because no one wants to write a SW render which is too slow. http://lists.debian.org/debian-devel-games/2012/03/msg00076.html The message complains that your CPU has SSE2 but it can not use it as clanSWRender was built without support from it. How can it be built without support for SSE2 if it is not optional ? I wonder if we could enable for now SSE2, till upstream/somebody has a better fix : for now, all SSE2 cpus crash. Enabling SSE2 it will only crash on non-SSE2 cpus (older than 10 years) The crash is in the app, are there other apps using clanSWRender which are working fine without SSE2 but would crash if clanSWRender had SSE2 support?
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
On Tue, Dec 4, 2012 at 6:08 PM, Pascal Terjan pter...@gmail.com wrote: On Tue, Dec 4, 2012 at 6:00 PM, Jose Jorge lists.jjo...@free.fr wrote: Le 04/12/2012 13:21, Pascal Terjan a écrit : On Tue, Dec 4, 2012 at 12:10 PM, AL13Nal...@rmail.be wrote: The code should really be complied but only enabled at runtime if it's available. PulseAudio does this. It's more of a pain to code for but does make like easier for distributions. But yeah, short term fix, nuke the test, and as a meme says, keep calm and carry on. It is not only a test, it is a crash. It is a crash because the code says to crash if there is no sse2. but, do ALL recent i586 CPU's have SSE2? what about the embedded and VIA cpu's and the likes? http://en.wikipedia.org/wiki/SSE2#CPUs_supporting_SSE2 The test is that it refuses to run if the lib is not using sse2 while it could (slowly). The clanSWRender lib needs to be fixed to have sse2 support on sse2 system and not on non sse2 systems so that we can enable it but that's something unrelated to dropping this test. No, the clanSWRender has only a SSE2 implementation because no one wants to write a SW render which is too slow. http://lists.debian.org/debian-devel-games/2012/03/msg00076.html The message complains that your CPU has SSE2 but it can not use it as clanSWRender was built without support from it. How can it be built without support for SSE2 if it is not optional ? I wonder if we could enable for now SSE2, till upstream/somebody has a better fix : for now, all SSE2 cpus crash. Enabling SSE2 it will only crash on non-SSE2 cpus (older than 10 years) The crash is in the app, are there other apps using clanSWRender which are working fine without SSE2 but would crash if clanSWRender had SSE2 support? So, reading more about it, clanSWRender is a software renderer instead of using one of the opengl ones Why would an application need it? Is it faster than using for example mesa's software rendering?
Re: [Mageia-dev] Clanlib2 and SSE2 on i586
Guillaume Rousse guillomovi...@gmail.com writes: Le 03/12/2012 23:42, zezinho a écrit : Trying to launch openalchemist gave me a crash because Clanlib2 uses SSE2 on SSE2 systems for software rendering, while we disabled it at configure. What should we do? I think enabling it is better, as all i586 CPU with less than 10 years come with SSE2. But our policy forbids it ;-) Cool. Now we have software crashing because they don't use available features, what a nice idea... Recent CPUs also all come out with x86_64 support, so the i586 port is basically only used for older CPUs. It makes sense to not enable SSE* optimizations for i586, to be compatible with Geode/VIA/etc processors. -- Olivier Blin - blino
Re: [Mageia-dev] [changelog] [RPM] cauldron core/release bind-9.9.2-1.mga3
Le 20/11/2012 19:10, Guillaume Rousse a écrit : Le 20/11/2012 01:07, Christiaan Welvaart a écrit : guillomovitch guillomovitch 9.9.2-1.mga3: + Revision: 316758 - sync with fedora spec, for easier maintainance: the real files now live outside the chroot It looks like this overwrote my /var/lib/named/etc/named.conf with a default version so all zones were disabled. I had to restore the config file from backup. Not sure how this happened, though. The real file should have been moved to /etc/named.conf, whereas it was previously just a symlink. As it happens during %pre stage, tough, I'm not sure it will prevent overrun by default configuration files, indeed. If that's the case, I should rather use a versioned trigger to handle this change. I just turned the pre-installation scriptlet into a versioned trigger on old release, so as to fix the issue. -- BOFH excuse #274: It was OK before you touched it.
Re: [Mageia-dev] [soft-commits] [6562] Allow setting some env variables to define and default to setting PERL_EXTUTILS_AUTOINSTALL and PERL_AUTOINSTALL to --skipdeps
r...@mageia.org writes: Revision: 6562 Author: pterjan Date: 2012-12-04 14:22:38 +0100 (Tue, 04 Dec 2012) Log Message: --- Allow setting some env variables to define and default to setting PERL_EXTUTILS_AUTOINSTALL and PERL_AUTOINSTALL to --skipdeps Why is this in iurt? Shouldn't it be the default in our default rpmbuild setup? -- Olivier Blin - blino
Re: [Mageia-dev] [soft-commits] [6562] Allow setting some env variables to define and default to setting PERL_EXTUTILS_AUTOINSTALL and PERL_AUTOINSTALL to --skipdeps
On Tue, Dec 4, 2012 at 11:10 PM, Olivier Blin mag...@blino.org wrote: r...@mageia.org writes: Revision: 6562 Author: pterjan Date: 2012-12-04 14:22:38 +0100 (Tue, 04 Dec 2012) Log Message: --- Allow setting some env variables to define and default to setting PERL_EXTUTILS_AUTOINSTALL and PERL_AUTOINSTALL to --skipdeps Why is this in iurt? Because rebuild of the distro was strongly slowed down due to this (current build is running with default timeout meaning that each perl package which had this problem caused us to lose 6 hours) so preventing it from happening in iurt is something I want whatever is done in general. I have also added some code in iurt to detect stalled builds. Shouldn't it be the default in our default rpmbuild setup? Probably
Re: [Mageia-dev] [soft-commits] [6562] Allow setting some env variables to define and default to setting PERL_EXTUTILS_AUTOINSTALL and PERL_AUTOINSTALL to --skipdeps
Pascal Terjan pter...@gmail.com writes: On Tue, Dec 4, 2012 at 11:10 PM, Olivier Blin mag...@blino.org wrote: r...@mageia.org writes: Revision: 6562 Author: pterjan Date: 2012-12-04 14:22:38 +0100 (Tue, 04 Dec 2012) Log Message: --- Allow setting some env variables to define and default to setting PERL_EXTUTILS_AUTOINSTALL and PERL_AUTOINSTALL to --skipdeps Why is this in iurt? Because rebuild of the distro was strongly slowed down due to this (current build is running with default timeout meaning that each perl package which had this problem caused us to lose 6 hours) so preventing it from happening in iurt is something I want whatever is done in general. I have also added some code in iurt to detect stalled builds. But will the timeout be big enough to let WebKit link go through? :-) Shouldn't it be the default in our default rpmbuild setup? Probably Maybe we should have a macro for perl Makefile.PL with your variables exported, if this is where the hang happened. perl-HTML-Parser even has the following... # compile with default options (prompt() checks for STDIN being a terminal). # yes to not ask for automate rebuild yes | %{__perl} Makefile.PL INSTALLDIRS=vendor %make OPTIMIZE=%{optflags} -- Olivier Blin - blino
Re: [Mageia-dev] Video players
On 11/24/2012 09:49 PM, Pierre Jarillon wrote: Le samedi 24 novembre 2012 17:25:02, JA Magallón a écrit : If you can post your video test collection somewhere, I can try to see if totem works for me with them. Sorry for the delay. I have uploaded a kit (33 files 500MB) of videos on http://pjarillon.free.fr/docs/demos/ Some types of videos are given with the command file $ ls | while read f;do file $f;done|awk -F: '{print $2 : $1}' | sort Some formats 19/9 are displayed at 4/3 ratio. As I said previously VLC works fine in every case. Sorry for the delay... In my case, all play in totem except SWF files. The only problem I saw was some chopping with audio in a couple videos, the ballmer_dance one and the casseaur.mov. -- J.A. Magallon jamagallon()ono!com\ Winter is coming...