Re: [Cooker] [Fwd: muse] Odd QT problem
On Tuesday 11 November 2003 13:42, Austin wrote: Maybe some cooker QT expert can help me. I assure you, there is no .MusE file. Thanks, i had the same problem with psi. i just upgraded kdebase yesterday, and now, it work fine. i didn't see i was not running the latest version because of space problem. can you try with latest kdebase ? -- Mickaël Scherer
Re: [Cooker] kernel source requires libgpm1-devel ?
On Tuesday 04 November 2003 20:33, Brad Felmey wrote: This is not necessary, methinks. Is this an issue in urpmi or kernel-source? kernel-source need libncurses-devel libncurses requires libgpm1-devel . -- Michaël Scherer
Re: [Cooker] kernel source requires libgpm1-devel ?
On Tuesday 04 November 2003 20:38, Brad Felmey wrote: On Tue, 2003-11-04 at 13:33, Brad Felmey wrote: This is not necessary, methinks. Is this an issue in urpmi or kernel-source? And requires libncurses5-devel? yes, for make menuconfig. In a perfect world, it should be possible to have one package for each configuration :) In fact, since kernel 2.6 offer qt and gtk, something should be done here. -- Michaël Scherer
Re: [Cooker] jigdo cannot be installed on mdk92
On Friday 31 October 2003 14:44, Pascal Cavy wrote: # urpmi jigdo Some package requested cannot be installed: jigdo-0.7.0-1mdk.i586 (due to unsatisfied libwwwftp.so.0) (Y/n) n something is wrong on the mirror, because it work fines on other 9.2 libwwwftp.so.0 is provided by w3c-libwww can you check on other mirror ? -- Mickaël Scherer
Re: [Cooker] When can we finally get X to fall back to XFdrake??
On Thursday 30 October 2003 16:00, Jan Ciger wrote: | Le jeu 30/10/2003 à 14:20, Buchan Milne a écrit : |Sorry to be against you Buchan, but as he knew there is a console |program in RH, he should knew there is also one in Mdk aka XFdrake | ... | |And what if the user had never used Redhat before??? This guy has | been using Linux on hist desktop for over a year, do we expect | newbies to know more than this? I have better idea than to dump the newbie to XFdrake (that is finally a security risk, since the tool runs as root - i.e. you reboot, X fails, XFdrake runs, you are away from the machine and somebody screws your computer in the meantime). and what about asking the root password before ? -- Michaël Scherer
Re: [Cooker] Setting up a virtual provides to all console apps
On Tuesday 28 October 2003 20:05, Olivier Thauvin wrote: It would be nice to add a provides like 'console' or 'terminal' to console apps, xvt is link to some console apps but all those rpms haven't a common provides: while we are on it , something for pdf reader. it may be useful for documentation. -- Michaël Scherer
Re: [Cooker] Boot-up speed
On Tuesday 28 October 2003 20:14, Vedran Ljubovic wrote: Hi, Recently there were some articles on Slashdot proposing alternative boot procedures. Most of them advocate parallelizing of SysV-init tasks. Motivated by these I've decided to make a detailed analysis of boot procedure and see how can I make it faster. I have an oldish PC that can run most modern software fast enough. This is a vanilla Mdk 9.2 install and I use the auto-login feature with KDE. Complete time from pressing enter in lilo to a usable desktop is exactly 81 seconds. On this same computer Win98 is usable in around 35 seconds (I did some tweaking though). Here is a detailed account of the boot procedure. I made it using logs and a stopwatch. It should clearly point the likely targets for optimization. [ skip ] - Do we need to run depmod if no new modules are installed (many ways to check)? i think it use -A switch and so is runned only if it should. - I just can't understand why kde takes so long to start!??!? This has to do with the disk. a faster disk help moe than a faster processor. Maybe using -Os could help, but recompiling to test is maybe too much. - Idea: postpone cups and scannerdrake startup for 60 seconds to give desktop a chance to start and avoid unnecessary context switching. It's very unlikely that someone would like to print or scan something in the first 60 seconds? The same trick could be used for crond and atd I believe. So, the first 60 secondes would be as unusable as on windows ? -- Michaël Scherer
Re: [Cooker] urpmi basesystem
On Sunday 26 October 2003 01:58, Greg Meyer wrote: On Saturday 25 October 2003 06:39 pm, Michael Scherer wrote: On Sunday 26 October 2003 00:12, Greg Meyer wrote: This appears to be an undocumented feature. Are there any others that are interesting? a lot. but since they are undocumented, we do not always know them. have you ever heard of rpm --repackage ? or fileshareset ? No. Do I have to read the source to learn about these? (since I am not a programmer, that would be an arduous task), well, i would say yes :( or is there a project page somewhere that talks about what is in development? no. not all features are mandrake specific ( ie rpm --repackage ) but they lack documentation. This should be put on the wiki ( the user one ). -- Michaël Scherer
Re: [Cooker] Re: LG Drives
On Sunday 26 October 2003 09:51, Jos wrote: On Sunday 26 Oct 2003 01:03, Greg Meyer wrote: On Saturday 25 October 2003 07:20 pm, Juan Quintela wrote: marc == Marc Guise [EMAIL PROTECTED] writes: marc I read the post about Mdk 9.2 and LG on mandrakeusers.org. I have cd-rw drive, marc model HL-DT-ST GCE 8400b and I have Mdk 9.2rc2 running. There are no problems marc with my LG drive 21mdk just updated (vdanen should be doing the official update) fixes that problem. Only LG plain CD-ROMS are affected. Later, Juan. PD. Yep, whoeved decided at LG that reusing for UPLOAD_FIRMAWARE command FLUSH_CACHE comand should be shoot. Twice. SO it turns out to be a firmware bug after all that. I really hope you guys don't take the heat for this in the court of public opinion. Yes, it is a firmware bug, and yes, the LG drives are responsible for this. But, it is Mandrakes own fault that this happens. If you take beta / heavy modified kernels instead of kernels that have been tested by the entire linux community, you can expect things like these to happen. True, strange that this bug was able to tunnel trough all RCs, but this once again proves that it's better to use the entire linux community as testers instead of a few beta testers, i.e. use only stock kernels. We should drop support for all hardware not in the kernel ? To give a example, my usb modem, and so we should let the newbies patch their own kernel. Yes why not. After all, they will not know how to do this, but, if they want something that simply works, they should take windows instead. This is simply not the way mandrake works. If you want a default kernel, you can take the kernel-marcelo package. If you want something with 6 month of test on a 3 years old kernel, take debian. And now imagine this : we use stock kernel as all distibution in the world, and this bug now remains undiscovered. It will be here for 3 years, fry some cd readers, but, since they do not have anything in common, all people just say that lg drive are crap. and the bug continue to fry cdrom drive for years. In fact, since nobody would expect a kernel to fry a cdrom drive, i think that no one would have fill a bug report. It may even remain undiscovered until kernel 4.2, when ide will maybe no longer supported. I have 3 dead drives, and one of them simply stop working one day ( creative dvd ). Maybe it was something like this. Maybe not. What if the patch goes in the main stream kernel, and the bug still remains undiscovered, for the same reason ? It will do more harm. But since no newbie will ever post on lkml to say 'kernel fried my cdrom drive', it will not be discovered. Or not in 1 week. Maybe one year, until one distro ship a kernel and people start complaining. So, using stock kernel wouldn't have change anything. I do not know if the error is recoverable ( ie reflashing the firmware ), but, this is clearly not a problem on mandrake side. They have done more than one month testing, and the bug passed all RC and betas. In fact, maybe someone fried his drive and think it burned because it worked too long. All i hope is that LG give updated firmware, and instruction on how to recover the disk. Mandrake is not a linux distribution known for stability, let's please take stability as top priority for Mandrake 10.0. This way Mandrake will get the good name that belongs to such a cool distribution. And so, what do you propose ? To have a 3 month freeze period ? To use one year old software ? We do as much as we can for stability, but, as you have seen, some bugs remains undiscovered even with 4 or 5 beta releases. And i do not think that adding 10 releases would really help. -- Michaël Scherer
Re: [Cooker] frozen-bubble competition?
On Sunday 26 October 2003 18:07, [EMAIL PROTECTED] wrote: Wondering if gc has seen this: http://monkey-bubble.tuxfamily.org/screenshots.html yes, a news already passed on a french linux website 2 month ago. -- Michaël Scherer
Re: [Cooker] urpmi basesystem
On Sunday 26 October 2003 00:12, Greg Meyer wrote: On Saturday 25 October 2003 04:32 pm, Pascal Terjan wrote: I just tryied urpmi --root /tmp/mdk basesytem, it worked really nicer than the first time I did it months ago. Only got 2 errors (except for kernel but that's not a problem) : 8:glibc ## /var/tmp/rpm-tmp.27792: line 24: rm: command not found This one is easy to fix but has no impact... 55:rpm-helper ## /usr/share/rpm-helper/add-group: line 23: /dev/null: No such file or directory This one may turn something wrong but I didn't look into it more. This appears to be an undocumented feature. Are there any others that are interesting? a lot. but since they are undocumented, we do not always know them. have you ever heard of rpm --repackage ? or fileshareset ? -- Michaël Scherer
Re: [Cooker] Broken KDE in Cooker
On Thursday 23 October 2003 11:10, Quel Qun wrote: On Wed, 2003-10-22 at 20:57, Han Boetes wrote: David Kobler wrote: Can't we just set a simple standard to TEST YOUR PACKAGES before uploading ^^ them. It seems that simple things are constantly ignored by those who maintain the cooker RPMS. How hard would it be to test your rpms on a recently installed cooker box before uploading them? Show us the code :) Maintain a few rpms yourself and your talk a lot different. Hmmm, I'd bet all (most of) the contrib rpms are at least installed and run. I can't remember the last time fluxbox never started at all ;) kde have been tested, since it was build on the wrong library, so, the problem is not the test. The only way to test it was to install, so the packager installed on his own computer, and gues what, it work, because it seemed to used a new version of qt ( you know to test kde3.2 before uploading it ). so, to me, this is not a big problem, and after all, cooker is meant to be broken. trying to keep things always working will slow down the developpement too much, look at debian testing. This would allow people to test cooker for system specific bugs instead of distribution specific bugs that exist because people do not do BASIC testing. Nope, you guys are our test subjects. Our labrats. :) This is the deal. We make packages and you make decent bugreports or you learn to live with the fact cooker is broken. This is somehow the deal, but the point is valid. There is no way to find a bug if the program cannot be started at all on any platform. Well, this will be corrected as soon as cooker is unfrozen if you wan a working kde, use the 9.2 rpms, as i do. Simple test cases, a minimal test plan and a few release procedures could be automated and should not be so much of a burden once they are established. It's called investment, and I thought that was what a commercial company had to do. so, you propose to launch konqueror before uploading and having some automated script to see if it work ? have you sometime to test kde ? Note: These are just 2¢ thoughts. I am only sad to see that nothing changes despite years and years of disappointment. well, i do not see anything to do, so, yes, this is always the same problem, but, we have no easy solution. -- Mickaël Scherer
Re: [Cooker] Mandrake uses Photoshop... What a pity!
On Wednesday 22 October 2003 18:03, Pierre Jarillon wrote: But some people have found that photoshop was used in Mandrake. Olivier Thauvin has a quick way to solve the problem and avoid any remark. But the problem will be hidden only. I do not see any problem with that, as long they have a valid excuse. I fear that the bios on their pc is not free, so, this is the same. we cannot blame them to use a progam whn nothing good and free exist. Trying to hide this is like to recognize that mdk have done something bad. People should be realist, and stop dreaming. -- Michaël Scherer
Re: [Cooker] Mescalero
On Tuesday 21 October 2003 20:39, Sascha Noyes wrote: On Tuesday 21 October 2003 02:20 pm, Diego Iastrubni wrote: are you asking why does mandrake do not let me steal this album? lol Please look up the word steal in the dictionary. lol means lot of laugh in english, so i guess it was a joke. at least, i think -- Mickaël Scherer
Re: [Cooker] How to package plugins?
On Monday 20 October 2003 20:22, Buchan Milne wrote: I am packaging some amateur radio software, a number of the packages depend on hamlib, so I have packaged that too. However, hamlib ships with some plugins, and I don't seem to find a solution that rpmlint agrees with. So, I have these packages: $ rpm -qlp libhamlib1-1.1.4-1mdk.i586.rpm [ .. skip .. ] It looks like rpmlint is wrong, since: -this is not a library package -the so names look ok? So, is this a problem with the source package (should these be /usr/lib/hamlib-version/plugin.so, or is rpmlint just being too picky? i guess it is too picky. E: hamlib-plugins invalid-soname /usr/lib/hamlib-dummy.so rpmlint check that file under /lib, ending in .so, are real library. a library must provide a soname, ( obtained by objdump, i think ). Without it, this is a error, since the dynamic loader won't be able to find it. But, since this is not a library, this is wrong. E: hamlib-plugins non-versioned-file-in-library-package /usr/lib/hamlib-yaesu.la Since the plugin are not executable or valid library ( .so ), rpmlint think it is some kind of data, and so, they should be in a specific versioned dir, in order to have library installed side by side. rpmlint think this is a library, because it put some file in */lib, ending with .so. You cannot put the plugin under share, since share should be used for architecture independent files. So, i think this should be added as an exception, or to have some directory for .so files without soname. by the way, are .a needed for plugins usage ? I thinked that it was used for static linking, and so, this belong to -devel package, no ? -- Michaël Scherer
Re: [Cooker] 9.2 disasters list (continuing)
On Sunday 19 October 2003 02:05, Ron Stodden wrote: Michael Scherer wrote: On Saturday 18 October 2003 20:56, Ron Stodden wrote: Michael Scherer wrote: ok, this maybe a coredump. can you place it on the web somewhere, so someone with enough skill to run file core.5465 can enlighten us with the name of the crashed program ? http://members.optusnet.com.au/ronst/download/ (use Save Link Target as) thanks. so according to file, this is : [EMAIL PROTECTED] michael] $ file core.5465 core.5465: ELF 32-bit LSB core file Intel 80386, version 1 (SYSV), SVR4-style, SVR4-style, from 'X' [EMAIL PROTECTED] michael] $ gdb -core core.5465 Core was generated by `/etc/X11/X :1 -deferglyphs 16'. Program terminated with signal 6, Aborted. #0 0x4008bd71 in ?? () (gdb) so, this is something with X. can you describe the video card ? Thanks. Riva TNT2 Model 64. Tnt 2 ? That's a nvidia card, no ? did you use the driver provided by nvidia ? or the one that comes with xfree ? In either case this seems to be a application ( or a driver ) problem beyond the scope of mandrake linux, so you should directly contact the upstream authors. Since they are the developpers, they will be able to use the core provided more efficiently than i am. you can also put some of the file that filed your hard disk before. ?? sorry, i meant that filled the file you were talking about, in /root/, but, i think you removed everything... As I remember, all files were zero length except for the first two (neither of which was anything like the 7.2MB size of a core dump), and all were purged from another partition to release space on / and its directory, which restored Mandrake operability. Ok, so, unless you find a way to reproduce it easily, i think we are forced to class it. -- Mickaël Scherer
Re: [Cooker] !! 9.2 issues -- help
On Sunday 19 October 2003 19:59, Adrian Rodriguez wrote: Hi I just installed 9.2 and I have a few problems. 3. My Linksys WPC11 v3 wireless card does not get configured properly. It does recognize the orinoco_cs module that it needs and even make the ifcfg-eth1 script but I still had to do the hermes.conf hack to get it to work. --- this step is on my website under tutorials. maybe take care of this for mdk 10? i own the same, but i didn't tested it under linux, apart from plugging it and see the ifconfig output. it runs dhcp, so, what is the problem ? all i can say is that orinoco_cs doesn't offer all that linux-wlan-ng give. with wlan-ng, i can use kismet and airsnort, but i never been able to connect to a access point. 4. I have an inspiron 4000 with a rage mobility 128 w/ 8MB or ram. It says i can choose X with 3d hardware acceleration but when i try playing a game it is slow...very slow so tuxracer and chromium or just not playable. i tried running other games like frozen bubble. Although they do play I'm not sure why they don't fullscreen correctly. I see the screen change modes but the game picture is in the center and not full screened. It happens with alot of applications that I fullscreen. Mplayer when I fullscreen does the same thing... screen changes modes but the picture is in the center still small. However... KMplayer uses mplayer and fullscreens correctly. any ideas about the 3d acceleration and fullscreening? i have a ati rage mobility, and, to use fullscreen in mplayer, i use xv output, and ati.2 package. there is no 3d acceleration, or only in cvs. -- Mickaël Scherer
Re: [Cooker] 9.2 disasters list (continuing)
On Saturday 18 October 2003 09:39, Ron Stodden wrote: Simon Oosthoek wrote: On Sat, Oct 18, 2003 at 11:04:01AM +1000, Ron Stodden wrote: Another 9.2 showstopper! Post it at: bugs.mandrakelinux.com No. I have legally carried out my only duty - to notify MandrakeSoft, the owner of the cooker mailing list. To do that is all that can possibly be required. Maybe in Australia, but, here, i never heard of someone being forced by the law to signal software bugs. It is not required of the public to know anything about MandrakeSoft´s internal structure or methods of working - it is their job to efficiently manage their internal communications. To the public, MandrakeSoft is a single legal entity, IOW in effect a single person. That why they ask you to use a standard web form to ease the dispatching of the bugs. Unless you prove that the email have been acknowledged by a mandrakesoft employee. So, please do what you should for everybody benefits, post a detailed bugreport on bugs.mandrakelinux.com. That is true of all corporate entities, including the various government entities. So if I write a letter to George W Bush, then I am correct in assuming that all relevant groups and individuals operating below him will have received it and will co-ordinate a single comprehensive timely reply. Only if the post office doesn't lose your precious letter. The same would be true if I had written personally to the most humble counter clerk in a remote branch on a subject far removed from his/her daily operations - the corporation has been notified. I know it´s not easy on corporations, but that´s the law, you know, and the law encapsulates accrued wisdom. Again, what law ? Some unknow international law ? It is hard to see how things could possibly be otherwise. The way things are wisely prevents the common accountability escape of deniability - no officer can deny knowledge of anything that happens within the corporation or level of government (¨I did not know¨, ¨Nobody told me¨ are not acceptable responses - they indicate a mentally diseased corporation - i.e. schizophrenic). I do not see how they can be liable for this, because you clicked on yes in the box saying roughlt 'there is no garantees for this software, and i agree to use it knowing this'. If you think this box is illegal, then, you are still bound by the GPL and all others license. So, even with your mail, mandrake is not liable. -- Michaël Scherer
Re: [Cooker] 9.2 disasters list (continuing)
On Saturday 18 October 2003 16:27, Ron Stodden wrote: I noticed today the presence of a new file under the cleaned up /root directory: [EMAIL PROTECTED] root]# ls -l total 7608 -rw---1 root root 7827456 Oct 18 23:35 core.5465 drwx--3 root root 4096 Oct 18 02:47 Desktop/ drwx--2 root root 4096 Oct 16 07:44 drakx/ drwx--2 root root 4096 Oct 18 18:56 tmp/ [EMAIL PROTECTED] root]# Quite possibly all the files that eventually filled up the /root directory were all such core dumps @ 7.8 MB each. Inspection of /var/log/messages showed nothing helpful. What does the 5465 indicate? the pid of the process crashed The only 3rd party software is my gShield firewall and privoxy called by mozilla as 127.0.0.1. ok, this maybe a coredump. can you place it on the web somewhere, so someone with enough skill to run file core.5465 can enlighten us with the name of the crashed program ? you can also put some of the file that filed your hard disk before. -- Michaël Scherer
Re: [Cooker] 9.2 disasters list (continuing)
On Saturday 18 October 2003 20:56, Ron Stodden wrote: Michael Scherer wrote: ok, this maybe a coredump. can you place it on the web somewhere, so someone with enough skill to run file core.5465 can enlighten us with the name of the crashed program ? http://members.optusnet.com.au/ronst/download/ (use Save Link Target as) thanks. so according to file, this is : [EMAIL PROTECTED] michael] $ file core.5465 core.5465: ELF 32-bit LSB core file Intel 80386, version 1 (SYSV), SVR4-style, SVR4-style, from 'X' [EMAIL PROTECTED] michael] $ gdb -core core.5465 Core was generated by `/etc/X11/X :1 -deferglyphs 16'. Program terminated with signal 6, Aborted. #0 0x4008bd71 in ?? () (gdb) so, this is something with X. can you describe the video card ? you can also put some of the file that filed your hard disk before. ?? sorry, i meant that filled the file you were talking about, in /root/, but, i think you removed everything... -- Michaël Scherer
Re: [Cooker] IDN (3)
On Saturday 18 October 2003 00:18, Oden Eriksson wrote: fredagen den 17 oktober 2003 15.43 skrev John Keller: Oden Eriksson wrote: There could be patents regarding this... http://www.computerworld.com/managementtopics/ebusiness/story/0,108 01,59043 ,00.html http://www.i-d-n.net/#patents The dateline in the first article is 28 March 2001, and the article mentions that the IETF would start over from scratch, if needed, to avoid infringement. The second has a link to the patent claim, and its withdrawal on 15 March: http://www.ietf.org/ietf/IPR/WALID-IDN From what I remember with the W3C's tangled mess(es) stemming from companies not filing until late in processes, companies are asked to come forward at the start so whatever standards body can avoid problems. I am far from being a lawyer, but if there is an IETF has an actual recommendation then I would expect that it's because it's confidant that legal issues have been cleared up. I could be wrong, so this is an pen invitation to slap me down. In light of Sweden's adoption at the tld level, I would expect that Mandrake has a pretty clear path in the wake of others' due diligence. Thank you. As Verisign has done an plugin for MSIE I guess it's safe for Mandrake to implement this as well. well, verisgn could have paid to use the patents. did someone tried to contact the patch author to see what he think ? -- Michaël Scherer
Re: [Cooker] IDN (in KDE)
On Monday 13 October 2003 20:29, Oden Eriksson wrote: http://dot.kde.org/1057799730/ this sound interesting, this could be good for the next mandrake, but right now, i think that concerned peoples are on holiday, you should fill a bug report on bugzilla for this enhancement. -- Mickaël Scherer
Re: [Cooker] [Mandrake 10] Ideas for RpmDrake [long]
On Wednesday 08 October 2003 18:10, FACORAT Fabrice wrote: Le mer 08/10/2003 à 14:02, Eric Fernandez a écrit : Yes, your message below had been delayed on my box :))) For the work that it would take, why not asking Club members ? Let's decide a format, resolution and compression level, and open a thread so that people can post their shots. very good point, that's sound sensible ! This will increase the community part of mandrake ... no cost for them ... very interesting indeed. so why not having apps of the month on the club, and test and so on. this way people will have it in their language, with more bigger screenshot and more interaction and so on. -- Michaël Scherer
Re: [Cooker] [Mandrake 10] Ideas for RpmDrake [long]
On Wednesday 08 October 2003 21:42, Pierre Jarillon wrote: Le Mercredi 8 Octobre 2003 20:27, FACORAT Fabrice a écrit : Le mer 08/10/2003 à 15:05, Pierre Jarillon a écrit : I like evolution instead of revolution. I don't know how is made the rpm database. I suggest to improve this management step by step. The first goal is to get a localization of the rpm. But to avoid to modify the rpms, this could be a separate database in a first step. i think there is something already for this, according to mdk-rpm-howto, and this part of the cvs : http://cvs.mandrakesoft.com/cgi-bin/cvsweb.cgi/poSPECS/po/ according to the cvs, there is not much activities on these translations. it worked, i have seen some rpm in french, long ago. this should be discussed on cooker-i18n ML. I spoke at Metz with Martin Michlmayer (Debian project). They have the same problem. I dream of a common localization... why not ? because we do not use the same string for description :) maybe in 10 years, of course, we may have the same description. But until a debian mdk fusion, this seems utopic to me. -- Michaël Scherer
Re: [Cooker] Mandrake 9.2 update policy
On Tuesday 07 October 2003 02:59, Adam Williamson wrote: On Mon, 2003-10-06 at 14:46, Eric Fernandez wrote: Very soon MSN is no option at all for non Windows users anyway (15th October). And Yahoo is going to block 3rd party softwares too. Then I Heh... they're going to TRY. I think that yahoo only blocked other software by accident, because, according to some articles, they worked with cerulean studio ( who made trillian ) to restore connectivity, and CS send some explanation to gaim people. So, since i never read any official statements from yahoo saying they will lock other client, i guess they won't ( at least for the next 3 month ). -- Mickaël Scherer
Re: [Cooker] Mandrake 9.2 update policy (was: gaim 0.69 and 0.70 are out )
On Sunday 05 October 2003 16:42, Austin wrote: 2. jabber is not an option for most people since their friends don't use it Wrong. Jabber is the best option because, they can talk to people on others network., by using a gateway on the server. they will not need to upgrade, because this will be done on the server side. I do not have a upgraded gaim, but i am still able to talk with friend on yahoo, with amessage.info server. So, i think that using jabber is a more than valid alternatives, for the moment. Take a look at the public server list on http://www.jabber.org/user/publicservers.php As you can see, there is more than 40 servers with these gateways, and it works great. The only problem i see is that gaim is unable to register on a gateway, i had to use kopete for this but kopete is too buggy for daily use ( at least on my computer ). but, they can also use gabber, or psi, i didn't tested them. -- Mickaël Scherer
Re: [Cooker] alsaconf won't start?
On Wednesday 01 October 2003 21:59, Vincent Meyer, MD wrote: Trying to start alsaconf gives: [EMAIL PROTECTED] meyerv]# alsaconf device_mode int, description Device file permission mask for devfs. which: no dialog in (/sbin:/usr/sbin:/bin:/usr/bin:/usr/X11R6/bin:/usr/local/bin:/usr/loc al/sbin) which: no whiptail in (/sbin:/usr/sbin:/bin:/usr/bin:/usr/X11R6/bin:/usr/local/bin:/usr/loc al/sbin) Error, dialog or whiptail not found. So what the heck's a whiptail? a software used to show dialog box. /home/misc $ urpmf whiptail$ newt:/usr/bin/whiptail so, alsaconf should depend on newt, can you file a bug report on the package ? -- Mickaël Scherer
Re: [Cooker] [Mandrake 10] Replacing proftpd by pureftpd ?
On Wednesday 01 October 2003 21:40, Guillaume Rousse wrote: Even if no wiki is currently available, i really think we should better use domain name urpmi.org, and make it the mdk equivalent of apt-get.org. apt-get.org is an aberation. having to search on a webpage each time people want a package is one step behind, in my opinion. -- Mickaël Scherer
Re: [Cooker] And next ?
On Tuesday 30 September 2003 05:38, Brad Felmey wrote: On Sun, 2003-09-28 at 03:19, Warly wrote: - What was wrong in 9.2 development process? The same thing wrong with them all - too little QA/freeze, although this is somewhat better now. Also ludicrous changelogs like fixed something or rebuild. Laurent is the worst about this. Fixed *WHAT*? Rebuild WHY? - We though a bit late in the 9.2 developement process to split cooker ml, we should do it now. Debian is split to death, and it's a mess. I'd send the bugs to a different list, with followups set to cooker, and leave the rest. Well, all others projects are split, not only debian ( think freebsd, gentoo, etc ). So, i do not fill this is bad. The time lost due to a post on the wrong list will still be inferior to the time gained by having people focused on the subject they want. And anything related to the mandrakelinux distro. For God's sake, a urpmi proxy, like apt-proxy. Corporations are not going to want each box pulling packages separately, and they don't want to mirror, either. They just want to pull the stuff they need - once. wel, they can either setup a mirror, or use squid with correct configuration. I do not know squid too much, but, it should be possible to cache everything that come from a certain host, whatever the size is ? Having a doc explaineing this method on the user wiki will be sufficent, i think. Also, a place where folks can go get urpmi lines. Not just for stuff like MdkClub and mirrors, but kind of like Debian has a list of misc. repositories available that make it easy to locate Mdk-specific packages for whatever the user is looking for. Almost all unsupported and unaffiliated, but at least a place where a user can go look for homegrown packages of stuff (a la Texstar, etc.). Do you mean something like http://plf.zarb.org/~nanardon/ -- Michaël Scherer
Re: [Cooker] And next ?
On Monday 29 September 2003 23:14, Vincent Danen wrote: c On Mon Sep 29, 2003 at 10:50:48PM +0200, Michael Scherer wrote: - How to have more contributors? while browsing the web, i have seen that gentoo and netbsd announce their new developpers. this may be a good idea,in order to show that new contributers comes. And, even if this sound a little childish, having a @contributer.mandrake.org mail address for contributer could do some subliminal advertisement, showing that people can contribute without having to be employed by mdk. This is what @linux-mandrake.com addresses are for. Yes, but someone said that they were no longer given, and this address sound too much like the one of the employees. That's why i have proposed something that clearly show that outside people can contribute. -- Michaël Scherer
Re: [Cooker] gideon from cvs, linmodems, linmodems + seti stuff
On Monday 29 September 2003 23:50, Diego Iastrubni wrote: , 30 2003, 00:34,Buchan Milne: dosemu (compiled in 9.0, runs in 9.1, 9.2), please run first time in a console, then you can use the menu item: http://iglu.org.il/pub/Hebrew/diego/dosemu/ dosbox is in contrib to replace this, since dosemu can't be compiled in it's entirity (or the dos image) with free software in the distribution. I'm not sure if PLF has a package? dosbox sux, dosemu ownz. :) I needed dosemu for NASM which ide is in PMODE, and not supported by dosbox. as I understand, in the definition of free s/w here is it can be compiled using free tools. I think that freedos can be compiled with some free (no cost and open source compiler) the name ran out of my head sorry. If anyone wants dos emu, and he cannot have it from the official sites, he can have it here. well, i think that dosemu is split in two package. the dosemu binary, and the dos image. Only the dos image is considered to be not free. So, only the dos image should go to plf... If nobody is against having dosemu without freedos in contribs, i think that plf will accept to have the freedos image. We have basiliskII wthout the rom in contrib, so that shouldn't be a problem. -- Michal Scherer
Re: [Cooker] And next ?
On Tuesday 30 September 2003 19:08, jokerman64 wrote: On Sunday 28 September 2003 04:19 am, Warly wrote: It may be a good idea, before cooker opens again, to take these days to have some brainstorm. May you give your opinion on : - What was wrong in 9.2 development process? [worksforme] I hate it when people close bugs w/ works for me instead of actually trying to fix the prob. I think it's a major copout. ok, so what should they do ? Trying to fix a bug thay cannot see or reproduce ? - What should we do to improve the Wiki. wtf is the wiki. i've only seen reference to it on cooker. http://qa.mandrakesoft.com/wiki/ if you do not know what a wiki is, please see http://www.wikipedia.org/wiki/Wiki ( wikipedia is a wiki, of course :) ) And anything related to the mandrakelinux distro. How 'bout adding sites like mandrake user sites on another page (mandrake community or something) Not like a webring or anything (those are gay). I suppose that, like all the others the last time, you have nothing against gay people and you didn't want to offense them, as usual when this kind of problem occurs ? -- Michaël Scherer
Re: [Cooker] And next ?
On Monday 29 September 2003 00:26, Diego Iastrubni wrote: , 28 2003, 21:44,Michael Scherer: * use also apt-get (with the same repositories, IMHO this can be done no?), let the user which app use to install rpms. IMHO it just wastes mirror space ... how many use apt? it doesn't matter, apt is patched to use hdlist in mdk :) does it mean i CAN use apt in mdk? yes. will it be able to recompile packages? no. this is apt-build who is in charge of this, and it is not packaged. howto? manuals? just look at the configure file, it should work out of the box, or with very few configuration. -- Mickal Scherer
Re: [Cooker] And next ?
On Monday 29 September 2003 13:38, Eddie wrote: Do not split KDE into so many silly packages. No distro does this and it is ridiculous and too confusing. debian does it, suse too. Some one will say that you can install what you want and leave out other stuff, but if you use urpmi and/or gurpmi it still installs all the packages as before because of dependencies. Please go back to the way it was, it was s easier to keep track of. Eddie Mihalow Jr-Silver Club Member Splitting is good. I do not own a scanner so why should i install sane required to use kooka, beacuse i use the old kdegraphics package ? i only use kmail and konqueror, why should i have 5 more useless kdenetwork application that i will never use. -- Mickaël Scherer
Re: [Cooker] And next ?
On Sunday 28 September 2003 10:19, Warly wrote: - How to have more contributors? while browsing the web, i have seen that gentoo and netbsd announce their new developpers. this may be a good idea,in order to show that new contributers comes. And, even if this sound a little childish, having a @contributer.mandrake.org mail address for contributer could do some subliminal advertisement, showing that people can contribute without having to be employed by mdk. -- Mickaël Scherer
Re: [Cooker] And next ?
On Sunday 28 September 2003 13:17, Michael Reinsch wrote: Hi! - some packages have a maintainer who is not really the maintainer of this package (i.e. most changes are done by someone else), so a bug report isn't assigned to the correct person. we all receive bugreport for all packages, but, this is right, this could be improved. A proposal how development/mdk could be a bit different: - change release cycle for full blown releases to about once a year no, because hardware change so often that 1 year is too long to have the hardware supported. 6 month allow people to have their new hardware without too much problem. This question has been so discussed that i think it should go in a FAQ on the wiki :) -- Michaël Scherer
Re: [Cooker] And next ?
* use also apt-get (with the same repositories, IMHO this can be done no?), let the user which app use to install rpms. IMHO it just wastes mirror space ... how many use apt? it doesn't matter, apt is patched to use hdlist in mdk :) -- Michaël Scherer
Re: [Cooker] [Mandrake 10] Replacing proftpd by pureftpd ?
On Friday 26 September 2003 22:10, Vincent Danen wrote: On Fri Sep 26, 2003 at 08:23:24PM +0200, Michael Scherer wrote: can you put exim in contribs at least ? There's never been an interest before. I've been maintining exim packages for about 1.5-2 years now, and I never even knew anyone was interested. IIRC, I had asked about putting it in contribs when I first built the package (since it wasn't wanted in main) and the answer I got then was two MTAs were enough. Never bothered since. sorry, i was not here at this time. I should have search the archives before asking. If someone wants to put it in contribs, that's fine. I'll still maintain it on rpmhelp.net because, right now, you can grab exim for 9.0-9.2, and instead of maintaining for cooker I maintain for stable releases. do you think it is some much troubles to maintain it in contribs ? I mean, this is just rebuilding it in a cooker environnement. And it doesn't prevent you to rebuild in 9.0-9.2 this will be easier to find it . Dispertion of rpm is not good, this force people to search too much and this gives them bad habits ( downloading semi official rpm ). Putting it in contribs will allow it to be tested by more people, and integrated to cooker. I disagree. PLF doesn't have a problem with people using their packages. yes, but plf is bigger than rpmhelp. if you take a look at easy urpmi, you will see that a lot of repositorie exist, some of them providing rpms already in contribs, and this is not easy to keep track of who do what. And i am pretty sure that plf would not exist if we were able to put the package in contribs. IIRC, if you were to search in Club's rpm database or rpm2html listing or whatever it is, you'll find exim listed under rpmhelp.net. yes, but, i think we should not encourage people to search rpm on a website. urpmq and rpmdrake are here for this reason. How can people choose between your great package, checked with rpmlint, and others packagers made by alien ? Are you saying that everything on rpmhelp.net, PLF, Texstar's stuf, etc. should go in contribs? yes, and no. Texstar is building backport. Plf is for stuff that cannot go in contribs. I think that a repository than can be put in contribs should be in contribs , in order to not have the apt-get.org dispertion effect. I think separate rpm repositories that specialize in certain things is good. It depends on what is the specialisation. But, even if exim is put in contribs, nothing stop you to provides backport for people who want it. In fact, more and more people are interested in providing backport , so, there is a place for a global repository offering sane backport. But, having it in contribs will give more visibility. And, this is the first step toward inclusion in main. And, we will also be able to integrate some tools ( spamassassin, amavis ) more closely with exim. There really is nothing stopping you from doing that now. It doesn't need to be in contribs for that. if someone do a spamassassin-exim package, this person cannot put it in contribs, since it has a unresolved dependancy ( exim ). A package in contribs can only depend on main and contrib. now, i fully undestand that you may have too much work to maintains exim in 2 places. -- Michaël Scherer
Re: [Cooker] [Mandrake 10] Replacing proftpd by pureftpd ?
On Friday 26 September 2003 18:28, Vincent Danen wrote: On Fri Sep 26, 2003 at 03:29:28AM -0400, Brook Humphrey wrote: Heck, I'm all for it and agree with all your reasons. But the example is a touch out... wu-ftpd hasn't been in main since 8.2 (last version it shipped in main). Hey, while we're at it, can we throw sendmail in contribs? =) (Serious about killing wu-ftpd altogether, semi-serious about sendmail) I'm with you vincent sendmail is well it's. I wont say it would not be nice but I would not mind seeing it go. Postfix is more than good enough to take it's place for those that choose it as are many other options out there. Well, let's put it this way. For about 2 years now I've wanted sendmail in contribs and if we really needed another MTA alternative to postfix, I have very nice exim packages on rpmhelp.net and exim is a really nice MTA. Not that this is a plug to put exim in main (doesn't matter to me... it's available via rpmhelp.net), but a definite looking forward to not having sendmail in main anymore. can you put exim in contribs at least ? this will be easier to find it . Dispertion of rpm is not good, this force people to search too much and this gives them bad habits ( downloading semi official rpm ). Putting it in contribs will allow it to be tested by more people, and integrated to cooker. And, we will also be able to integrate some tools ( spamassassin, amavis ) more closely with exim. -- Michaël Scherer
Re: [Cooker] [Mandrake 10] Replacing proftpd by pureftpd ?
On Friday 26 September 2003 18:11, Vincent Danen wrote: Very true. But every openssh vuln hasn't been a crash or DoS. Mind you, with openssh a DoS is bad enough. Need to remote admin servers? What happens if the server goes down and you're stuck driving a long time to get to the machine? (It's happened). A DoS in apache is one thing... ssh in and restart. A DoS in ssh is another... how do you ssh in to fix it if ssh doesn't work? isn't there sshd-monitor for restarting ssh ? -- Michaël Scherer
Re: [Cooker] Cooperation with RedHat Linux project?
On Monday 22 September 2003 12:48, Adam Williamson wrote: On Mon, 2003-09-22 at 01:06, Pierre Jarillon wrote: Le Dimanche 21 Septembre 2003 16:41, David Walser a écrit : http://rhl.redhat.com/ From reading this, it sounds like the possibility of some cooperation between Cooker and the new RHL project is even more possible. RedHat said: We are excited to announce that we are working on an alliance with another well-known provider of Red-Hat compatible packages. I try to guess who is this well-known provider of Red-Hat compatible packages. IMHO, not so many ! Sounds like Freshrpms to me. what about fedora ? ( http://fedora.us ) Maybe SCO ? No just kidding. -- Michaël Scherer
Re: [Cooker] So, which is broken?
On Monday 22 September 2003 19:36, Felix Miata wrote: Three candidates: 1-grep 2-rpm -qa 3-pipe filtering On Redhat 9: # rpm -qa | grep gpm gpm-... libgpm1-... # On mdk cooker: # rpm -qa | grep gpm gpm-... # /home/misc $ rpm -qa | grep gpm libgpm1-1.20.1-9mdk gpm-1.20.1-9mdk libgpm1-devel-1.20.1-9mdk are you really sure that mdk is broken ? didn't you set some alias to grep ? -- Michaël Scherer
Re: [Cooker] Make xconfig fails on kernel-source-2.4.22-10
On Monday 22 September 2003 19:46, Gary Walsh wrote: I get an error when running make xconfig on kernel-source-2.4.22-10mdk. ~ Am I missing something? cat header.tk ./kconfig.tk ./tkparse ../arch/i386/config.in kconfig.tk 3rdparty/lufs/Config.in: 2: unknown command make[1]: *** [kconfig.tk] Error 1 make[1]: Leaving directory `/usr/src/linux-2.4.22-10mdk/scripts' make: *** [xconfig] Error 2 bug was fixed in the kernel, but it reappeared. http://qa.mandrakesoft.com/show_bug.cgi?id=5827 -- Michaël Scherer
Re: [Cooker] Anti-VeriSign patch for BIND
On Thursday 18 September 2003 03:48, Leon Brooks wrote: Since it's small and by default doesn't do anything, is there room for ISC's anti-wildcarding BIND patch at this late date? If BIND has to be rebuilt for anything else...? I do not think so, because it still need to be tested. And, if verisign decide to react and change something, this patch will be useless. So, i really prefer to have package avaliable on oden website, for people really intested by the patch than having a untested patch by default in a untested package. Even if it is small, it can break. -- Michaël Scherer
Re: [Cooker] [9.3/10.0 THOUGHTS] Drakconnect and network configuration
On Tuesday 16 September 2003 16:29, Simon Oosthoek wrote: Wow! That's a very detailed description and it looks like it would improve network configuration (at least for experts) by unmeasurable amounts! It would be much better on the wiki. Because, keeping track of email is not good. The wiki can help discuss idea, you can post pictore ( and not ascii art :) , and is really suited for these kind of discussion. -- Michaël Scherer
Re: [Cooker] Re: [Contrib-Rpm] zope-2.7.0b1-2mdk
On Tuesday 16 September 2003 20:09, Luca Olivetti wrote: Lenny Cartier wrote: Lenny Cartier [EMAIL PROTECTED] 2.7.0b1-2mdk - requires python not python2.2 Considering that each zope release is tested and requires a specific python version (2.2.3 for 2.7.0b1) is this a good idea? (not that I'd ever consider deploying zope through an rpm but..) i have forced the installation, it worked fine, two weeks ago. I guess that python2.3 is still compatible. but, 2.70b2 is out, but i did not update this, it was too complex. maybe a quick look trough changelog will say if it work or not ? -- Michaël Scherer
Re: [Cooker] New SSH bug ?
On Tuesday 16 September 2003 22:15, Jan Ciger wrote: Han Boetes wrote: | Always fun in the #openbsd channel. Always some people who want to | make it seems like the end of the world and the next worldwar. Ehm, there are reports that it lead to root compromise already, so I would execute extreme caution about this one. Considering that SSH is on almost every Unix system, this may be a major issue. well, after reading the diff http://www.freebsd.org/cgi/cvsweb.cgi/src/crypto/openssh/buffer.c.diff?r1=1.1.1.6r2=1.1.1.7f=h i see that some memory that shouldn't be freed is freed, thus probably crashing sshs ( which is annoying, if you do not use ssh_monitor ). But, i do not see how someone can use this to inject a shellcode, but maybe time will prove i am wrong. | The text is very clear though: | | | All versions of OpenSSH's sshd prior to 3.7 contain a buffer | management error. It is uncertain whether this error is | potentially exploitable, however, we prefer to see bugs | fixed proactively. This just means, that they do not know about the exploit yet :-( Not that your machine cannot be compromised. the same can be say about any server. -- Michaël Scherer
Re: [Cooker] um.
So what's made Mandrake not open? The fact that there will be one ad in the installer or the fact that I spoke? the installer is just a step, like the bookmarks, the question is what will follow well, the same kind of reproch were made to the club, and now what ? well nothing special happened. the way you speak, and mainly (once again) the way it was found out everyone here would understand and this thread would never exist, if you had possted a note because of the hard financial situation, we are forced to do it i hope that everyone here realize that this thread is here because mandrake added a new section on his website, and some guys posted a thing to osnews without asking to mandrake what they planned to do. in case if you wonder, mandrake already had a partnership with oreilly. and they displayed publicity for the gnu project. now, what happenned. they setup a new section to their website, dedicated to their partner. they also put a link to a page, stating that partner can have their links in the browser and so on. this was never announced nowhere, but somehow, someone on osnews see this. they read the site, remove everything which is not sensational, and post it. and everything begin. people start to complain that mandrake didn't ask to them what to do with their own product, and say that they will all swith to debian and this is unacceptable. and they complained until mandrake post a clarification notes. and everything in a timeframe of 8 h. Given the fact that mandrake direction does not read slashdot every minutes, this is acceptable. In fact, mandrake listened to their complaints. And this is good. now, if you thinked thay should have asked to you before if they can have partnership like they did before, ok, I guess your boss ask you before choosing a new employee, and everything related to financial operations ? but why boder informing cooker, we value them, but who cares, they'll live with it, or some of them might go away, who cares Well, as stated before, this has been done before, and nobody send a mail to mandrakesoft. If you feel they should have asked and want to leave, fine, we cannot force you to stay with us. This is your decision. -- Mickaël Scherer
Re: [Cooker] mod_python applications?
On Saturday 13 September 2003 17:19, Oden Eriksson wrote: Hi. I'm trying to find web applications for apache2-mod_python, but the only thing I've found is http://pywm.org/, and it's the only one I have been using while testing the module. Surely there must be others out there, but I just can't seem to find any, so I need help finding some. glasnost. use urpmf --require mod_python -- Mickaël Scherer
Re: [Cooker] mod_python applications?
On Saturday 13 September 2003 18:10, Oden Eriksson wrote: lördagen den 13 september 2003 17.48 skrev Michael Scherer: On Saturday 13 September 2003 17:19, Oden Eriksson wrote: Hi. I'm trying to find web applications for apache2-mod_python, but the only thing I've found is http://pywm.org/, and it's the only one I have been using while testing the module. Surely there must be others out there, but I just can't seem to find any, so I need help finding some. glasnost. use urpmf --require mod_python Cool. But I get: Mod_python error: PythonHandler code.webhandler Traceback (most recent call last): File /usr/lib/python2.3/site-packages/mod_python/apache.py, line 326, in HandlerDispatch log=debug) File /usr/lib/python2.3/site-packages/mod_python/apache.py, line 496, in import_module f, p, d = imp.find_module(parts[i], path) ImportError: No module named webhandler Which means? ouch. i have this : Mod_python error: PythonHandler code.webhandler Traceback (most recent call last): File /usr/lib/python2.3/site-packages/mod_python/apache.py, line 326, in HandlerDispatch log=debug) File /usr/lib/python2.3/site-packages/mod_python/apache.py, line 499, in import_module module = imp.load_module(mname, f, p, d) File /var/www/html/glasnost/code/webhandler.py, line 78, in ? class BufferedRequest(apache.Request): AttributeError: 'module' object has no attribute 'Request' i will try to check this :/ tomorow. -- Mickaël Scherer
Re: [Cooker] urpmi / apt4rpm - not for 9.2
On Friday 12 September 2003 13:02, vbnh fdgfd wrote: Hi, Apt4rpm (http://apt4rpm.sourceforge.net/) have the same fonctionality as URPMI. URPMI is a greet tool but why reinvent the wheel ? yes. so, can you tell to people who ported apt that urpmi was here before apt-rpm, and then, they should work on urpmi ? Mdk team is not big at this time, why loose time with a redundant software devel ? because urpmi has some fonctionnality not found in apt-rpm upgrading a whole distro whithout filling the /var is a nice one. And Rpmdrake could be changed to use apt-get, no ? or you can use synaptics. it reminds me that it is not packaged for contribs :/ -- Michaël Scherer
Re: [Cooker] urpmi / apt4rpm - not for 9.2
On Friday 12 September 2003 16:19, Pierre Jarillon wrote: Le Vendredi 12 Septembre 2003 15:11, Teletchéa Stéphane a écrit : Le ven 12/09/2003 à 14:49, Pierre Jarillon a écrit : Le Vendredi 12 Septembre 2003 13:27, Teletchéa Stéphane a écrit : urpmi is THE killer feature of Mandrake above all others distributions. Apt was a must, but for now, it seems urpmi is better (from my point of view ...) When common practices can be found they became standards. But to enforce a standard a priori without a good experience reduces the creativity. I could agree if apt and urpmi were only for the same purpose, but did urpmi-parallel (possibility to install on many differents machines of a cluster -may be a subnetwork-) exists for apt for example ? A full merging is not a good wish. But all differences which are not still useful. For example you can launch a gnome program in a KDE environment. It would be nice to make a package for Mandrake and Debian with the same tool. what would be the benefit ? being able to install the same package on the 2 distros ? And having a package not integrated with the 2 distro, or at least, with one and not the others ? I don't know the debian policy, but i know that our binary are not fully working on debian, and that mandrake naming scheme is different from the debian ones, so, dependancy will be unsatisfied. So, maybe you think of something which can produce the files neeeded to produce a package for each distro, which would be quite interesting, but, not so useful, IMHO. -- Michaël Scherer
Re: [Cooker] urpmi / apt4rpm - not for 9.2
On Friday 12 September 2003 22:09, Christiaan Welvaart wrote: On Fri, 12 Sep 2003, Michael Scherer wrote: or you can use synaptics. it reminds me that it is not packaged for contribs :/ I packaged it months ago, after I fixed apt to work with cooker mirrors directly. But it looks like I can't get it into contribs. can you give me the url, i will try to upload it during the week end ( if it work, of course ) -- Mickaël Scherer
Re: [Cooker] um.
On Friday 12 September 2003 22:07, Jan Ciger wrote: Ok, breaking my own wow not to reply to this thread anymore, *sic*. [EMAIL PROTECTED] wrote: | I think this may help: | | http://www1.mandrakelinux.com/en/mdkads.php3 | | it clears up some misconceptions and definitely eases my mind regarding the | situation. Thanks for pointing it out, it does clear up the issues. But why this page had to appear only after the uproar erupted ? Maybe because they didn't planned to have this uproar ? You know, it was discovered this morning. And nobody at mandrake send the pages to osnews. Maybe it was planned to be announced today, or it was a accident. We will never know, but i am sure that mandrake did nothing to actively show this to the world, _precisely_ because they didn't want this to be know right now. They cannot react as fast as people can discuss, on this liste or on slashdot. You know, before talking to everybody, you should carefully think about what you will say and this is exactly what they did. I think that the speed of today information didn't ameliorate the situation... -- Mickaël Scherer
Re: [Cooker] [Bug 5465] [gcompris] New: Additional package is chosen during install
On Wednesday 10 September 2003 20:52, Bruno Coudoin wrote: I confirm that gcompris needs gnuchess. There is a chess training board in gcompris. On the other hand, what happen if it doesn't found it ? Is there a simple way to split the board from the the main program ? I wanted to split for python stuff, but i guess i am trying to split too much :) -- Michaël Scherer
Re: [Cooker] Final in Septembre???
On Thursday 11 September 2003 18:09, Pierre Jarillon wrote: Le Jeudi 11 Septembre 2003 10:56, Simon Oosthoek a écrit : On Thu, Sep 11, 2003 at 09:10:51AM +0100, Emmanuel wrote: Hi all, Quick question for all of you guys: will Mandrakesoft release 9.2 even if it is not ready??? i.e. is Mandrakesoft desperate to get a version out before the end of Septembre or is there any chance for a RC3 ?? i.e. does management or development rule Mandrakesoft??? Based on past experience, 9.2 will be released on schedule on 20th of September even when it doesn't install for some people who have reported serious issues for their particular hardware. I suppose this is a company's way of making compromises :-( I hope a RC3. There are too much bugs in RC2 to call this a Release Candidate. If we compare with OOo, they made RC which was really RC. It is obvious that a distro is more complicated because of the great diversity of hardware. But the number of misses is not a problem of hardware. In fact RC2 is a beta 4. Please, make a RC3 ! Don't waste the good reputation of Mandrake with a lack of finition. This is very funny, because, _each_ release, this is the same story, people say that rcX is not good enough and we need a rcX+1. ( and there is always someone to make the statement i am doing ) There is always too much bugs. -- Mickaël Scherer
Re: [Cooker] [Bug 5566] [kdebase] New: use of proprietary menu location defeats the purpose of a standard Linux desktop
On Thursday 11 September 2003 23:01, Alex Chudnovsky wrote: On Thursday 11 September 2003 23:31, Buchan Milne wrote: Until then, it is impossible to satisfy everyone completely. It is. Switch to the exact Debian way - add entries to KDE menu, don't substitute it. Mandrake already use Debian menu. When something add a menu file for kde, does it appear on gnome ? No. Does it appear on fluxbox ? And, as said before, you can change it with menudrake. It's not only for third party applications, it's for installing from source too. It's not always possible to find Mandrake-specific RPM for some KDE application, especially for a fresh one. Then you should learn how menu work. http://qa.mandrakesoft.com/twiki/bin/view/Main/MandrakeMenu It is very easy to add a menu file. You can even add it in your home, in ~/.menu -- Mickaël Scherer
Re: [Cooker] Re: What happened to am-utils?
On Sunday 07 September 2003 12:20, Luca Berra wrote: On Sat, Sep 06, 2003 at 01:14:59PM +0200, Buchan Milne wrote: Well, if am-utils is better, it should replace autofs in main (autofs can go to contrib). But, AFAIK, there is no alternative for uucp, and packages in main (postfix) contain entries in the default config for it (thus you kind of expect it to be available ...). i'd love to see those entries in postfix commented by default, or not there at all. i would like to be able to remove procmail, that i do not use. having a setuid package who serve noting on my system is bad. but, since postfix depend on it, i cannot remove without forcing or rebuilding the rpm. does it really mandratory to have procmail as a default ? -- Mickaël Scherer
Re: [Cooker] Re: What happened to am-utils?
On Sunday 07 September 2003 03:37, Vincent Danen wrote: So, by your point above, we should have one of everything in main? Well, where's my better implementation of gkrellm-plugins (in contribs). Oh, well, just for kicks, let's throw a few others out there. Where's my alternative to zssh or zope or xscorch or webalizer/analog or uudeview or seahorse... ad naseum. What makes uucp better than any of these programs that don't have a better alternative in main? we should first think what does having a package in main implies : Package in main are on cds. This is important, because it means that they can be installed without internet connection. This also means that some packages, like uucp, who are mostly used with a internet connection, could be moved to contribs without too many problems. Packages on cds are more visible. But this is a problem of contribs visibility, and it should be solved by making more publicity for contribs, and having urpmi.setup on cd. Packages on cds can be installed automatically by drakX. It can still work with a tweaked install cd, thanks to MakeCD, but not everybody know this possibility. Having the capacity to script install is importatn for professionnal, i guess. Packages in main are updated for security, and contribs are not. This is bad, since, to give a example, some packages like apache2 modules are in main, and others are in contribs. But, on mdk9.1, apache was updated, and only main packages got rebuild. We should have a way to propose some update for contribs, even non official ones. I know that security team is quite over booked, but contributors could handle the problem. At least, it is better than nothing. So, based on this difference, we should think about what should be in main, and what should go in contribs. We should strive to reduce the difference between contrib and main. This would stop some of the endless discussion on what in main and what in contribs. We can even dream of a distribution where there is no difference between main and contribs ... -- Mickaël Scherer
Re: [Cooker] Re: What happened to am-utils?
On Sunday 07 September 2003 13:27, Buchan Milne wrote: On Sun, 7 Sep 2003, Michael Scherer wrote: we should first think what does having a package in main implies : Package in main are on cds. This is important, because it means that they can be installed without internet connection. This also means that some packages, like uucp, who are mostly used with a internet connection, could be moved to contribs without too many problems. Hmm, do you know that uucp is about the only solution for mail transfer *without* an internet connection? no. I was thinking of this related on having somehow a connection to the internet, in order to have a mail server, but obiously, it was the wrong example :/ Anyway, IMHO, uucp is a more feature-critical package than a log analyzer (if you need one, you *must* be connected to the internet, you can always run your log analysis later once you have installed it). Right. but, as i said before, if people want to script install, it should on cd for ease of use. Even if we know it could be done without it. Packages on cds are more visible. But this is a problem of contribs visibility, and it should be solved by making more publicity for contribs, and having urpmi.setup on cd. It also needs a menu entry ... Yes, i will do it. I will try to put some icons, i need to see with mdk employees. So, based on this difference, we should think about what should be in main, and what should go in contribs. Well, we need a wider audience than just us. Something like popularity_contest would be a way to see which packages in main are effectively obsolete (or need other justification to stay in). Yes. and, while collecting some information, we should also get the hardware. BTW, regarding what roles Mandrake is used in, the poll on MandrakeClub shows 50% use Mandrake on the desktop, 50% use Mandrake on the desktop and server, which that server use *is* very popular (imagine the chaos if 50% of Windows users used Windows on servers). Well, even without it, this is the chaos... -- Mickaël Scherer
Re: [Cooker] Re: synchroning ftp
On Sunday 07 September 2003 17:25, Thierry Vignaud wrote: Warly [EMAIL PROTECTED] writes: Question really is who set the 6 month limit? If the mad rush to keep up with the cycle causes the problems that are occurring at the moment then it must be questioned. If the problem isn't faced up to, there may not be a requirement for a six month cycle in the future, there may not be a future. You have a question about that? I think the answer is obvious. Mandrakesoft. There was some discussion about backing off on the rate of releases a while back. Mandrakesoft may be the obvious reason, but would it be a better release cycle? Presently I would even favor a 4 months release cycle. And there are far enough changes in the linux world in 4 months to justify a new release. I am convinced of several things : - a non fix release date is bad, because a release is never finished and we need to move on. - a too long release cycle is bad, because new computers does not work correctly without new XFree, new kernel and so on. - a too long release is bad because it means less pressure, ond good things only happen under pressure. i agree, not so many things happens on the first monthes after the release. most things are done in the last monthes before the release. I think that most people would not want to update every 4 months. This is like forcing the upgrade, which is something that everyone blame microsoft for. Even if upgrade are simple and doesn't pose any problem. And how do you want to be entreprise ready if it change so often ? On a side note, you need some rest from time to time. For me, 6 month is fine. -- Mickaël Scherer
Re: [Cooker] packager email?
On Sunday 07 September 2003 20:54, Oden Eriksson wrote: Hi. Since when did this change? W: libdspam3-devel invalid-packager Oden Eriksson [EMAIL PROTECTED] The packager email must finish with @mandrakesoft.com or must be [EMAIL PROTECTED] rpmlint complains about this, what should I do? nothing. this is a simple warning, people can still write, and it does not block the upload. -- Mickaël Scherer
Re: [Cooker] adiusbadsl should be a real kernel module
On Saturday 06 September 2003 20:10, cpjc wrote: Hi Thomas, I mean that when you install a kernel package (like kernel-2.4.22.5mdk-1-1mdk.i586.rpm), adiusbadsl.o.gz is not included as a pre-compiled module in the /lib/modules/etc and one's have to install the kernel sources, build a new kernel and install the module in /lib/modules/... which could really discourage newbies and upset others... ;-) That's also why the drakeconf wizard used to set adsl connection with Sagem 800 cannot succeed. But you're right, it could stay in 3rdparty, I would be ok just with the module included in the kernel binary package. /home/misc $ urpmf adiusbadsl kernel-2.4.22.5mdk:/lib/modules/2.4.22-5mdk/kernel/3rdparty/adiusbadsl/adiusbadsl.o.gz it is included. but, it need some user land tools to load his firmware, if i remember correctly. this is what the adiusbadsl rpm is used for. -- Michaël Scherer
[Cooker] Re: [Bug 5195] [subversion-client-local] subversion repository format had changed
On Friday 05 September 2003 06:10, [ben] wrote: I was talking about the sever/client interoperability. For that they as a matter of policy only guarantee compatability between one major revision back (e.g. .27 will be compatable with .26 but .28 will not necessarily be guaranteed). ye, and this means that a 9.1 server is still copatible with a 9.1 client. You can find their policy at: http://subversion.tigris.org/project_faq.html#interop I was thinking this applied to the schema format but apparently they don't apply it to that. yes, this is clearly stated in the FAQ : The repository db schema is stable now and should only undergo one more major change before 1.0. These migrations have happened before, and we already have dump/load utilities available to aid you in the migration. However, because of this policy and subversions instability we should encourage people to do dump/load at upgrade time everytime. i didn't say that people should not be encouraged to do this, but, if we can help them, then we should. This is all the better reason to urge people to use svnadmin dump *BEFORE* upgrading. Always doing a svnadmin dump before an upgrade and a svnadmin load after, will always avoid issues like this. My point is including an old svnadmin is only a solution for this particular issue. I think that having a solution is better than having nothing. If it can help 50% of people, it is better than nothing. And, having a solution does not prevent us to warn people. In fact, you can patch svnadmin to display the messages if you want. or wrap it in a shell script that does everything. Doing the dump before an upgrade and a load after is the only reliable way to ensure a good clean upgrade. It's unreasonable to carry around dozens of old svnadmin versions to help people upgrading to deal with specific issues... dozen of old svnadmin. Do you really think that subversion people are insane ? They use their own software, and they will not break it every month. The project changed only once in a way that break everything since the beginning, in 2 years. we will need a dozen of svnadmin in 24 years, if they keep the same rate of breakage. I guess that subversion will be stable in 2 years, so, unless they break for fun, we will never have dozen of old binary. Please stop exagerating in order to convince people. i do not suggest to have it for years, only to help the upgrade from the current mandrake release, ie 9.1. IMHO the 0.27 svnadmin should be removed from the package. I understand it helped you get out of a sticky situation, but having it implies that we'll provide that sort of thing in the future. yes, it helped me. and it should not help other people ? what will be the avantage of not helping people ? forcing them to learn the hard way and to lose time, because it will be good for them. what will people learn with this ? that they should not use subversion, because they break thing ? they should not use mandrake because someone didn't want to ship a simple solution ? openldap2.1 changed the way it handle the schema, but, buchan tried to provides some migration script. I am sure it will help lot of people. This is *our* job to provide smooth upgrade. The problem is they can choose to change the schema anytime they want. Do we carry a svnadmin version for every schema version? I do not think that subversion project is going to change schema every month. This break too much thing for being done. I do not say we should ensure a upgrade from any release to any release. but from the 9.1 to 9.2. I didn't intend it to be the only message. Just an extra notice... I figure most people using subversion as a server are likely to be using urpmi not rpmdrake... even if they do not use rpmdrake, they can miss the message. and what is the message good for ? I upgrade subversion, and it will say : oh, you should have dumped before upgrading, sorry, that's too bad. and what will i do ? i will recompile from source. People will not blame mandrake for the error, but, if we provides them with a solution, they will appreciate the effort. Now, of course, you can do nothing. Maybe they will switch to a distro that care for them and try to make their life easier. And, if we provides this only for the release, we can drop it once cooker is unfrozen, if you think it would take too much time to maintain it. i just want to ensure that people using subversion will have a easier path for upgrading. -- Mickaël Scherer
Re: [Cooker] pkgconfig not required by -devel
On Thursday 04 September 2003 00:32, Guillaume Rousse wrote: Ainsi parlait Michael Scherer : i suggest the last solution, but, maybe someone see a problem ? Yes, coherency with other development tools. How many packages really requires autoconf/automake ? I guess less than 1%. However, autoconf/automake are rpm-build dependencies. Here we have more packages requiring what appears some new build tool, and we are gonna add it as an explicit require for each of them. This is plainly silly. this will affect maybe 10 packages, which is not too much. i guess that only 10 packages, because nobody see the problem until now. this 10 package will be used to : - build rpm, and it will use pkgconfig, because of autoconf, most of the time. - build from source, and it will use pkgconfig, for configure script, because library authors put pkgconfig in their macros. - develop a software, which means that people will have to figure what are the good argument for gcc, and this is where they will use pkgconfig. I guess that documentation will also say to use pkgconfig. but, of course, it can be used without it. autoconf/automake is only used when developping a application, and when we patch a configure script for rpm building. Patching a script to suit our needs is almost the same as developping a application. That's why packages should not requires autotools. Most of the time, people don't need it. pkgconfig is different, because, most of the time, people will need it for a proper use of the library ( -devel rpm of the library ). I do not say to add pkgconfig requires on each library, but only to the ones that have almost 90% chances of using it. -- Michal Scherer
Re: [Cooker] [RC1 bug] apache-conf and esperanto
On Thursday 04 September 2003 16:09, Grgoire Colbert wrote: Saluton! There is an internationalisation problem with apache-conf-2.0.47-3mdk, which is very easy to fix. In /etc/httpd/conf/commonhttpd.conf, you cannot find the line : AddLanguage eo .eo which would allow to use content negociation with esperanto web pages. This is not a surprise, because the bug is corrected *only* in the current Apache CVS, not in 2.0.47 which appeared earlier : http://nagoya.apache.org/bugzilla/show_bug.cgi?id=21685 So, can you fix this file before Mandrake 9.2 final? can you please fill a bug report for it on qa.mandrakesoft.com ? it allows us to keep track of what should be done, and will send the bugreport to the right person, instead of having the mail lost in the 1000 mails per day of the ml. thanks for your testing and your attention -- Michal Scherer
Re: [Cooker] Re: [CHRPM] gnome-media-2.3.90-2mdk
On Thursday 04 September 2003 19:59, Olivier Blin wrote: On Thu, 4 Sep 2003 19:48:00 +0200 (CEST) Frederic Crozat [EMAIL PROTECTED] wrote: -=-=-=- Name: gnome-media Version : 2.3.90 Release : 2mdk Frederic Crozat [EMAIL PROTECTED] 2.3.90-1mdk - Release 2.3.90 - Fix BuildRequires The release tag doesn't match the release in changelog, isn't rpmlint supposed to detect that ? yes, it should issue a warning. incoherent-version-in-changelog, checked in Tags.py -- Michaël Scherer
Re: [Cooker] Suggestion for improved user experience
On Wednesday 03 September 2003 00:14, Adam Williamson wrote: I ran urpmi.setup the other day and I didn't see the opportunity to setup a plf source anywhere... it is a secret trick. you should run it from the command line, and use --allsources. -- Mickaël Scherer
Re: [Cooker] pkgconfig not required by -devel
On Tuesday 02 September 2003 08:42, Michael Scherer wrote:
The patch is not perfect. If applied right now,linphone would
requires pkgconfig, which is useless.
So, what i propose is to add pkgconfig as a requires if :
- a file is dropped in /usr/lib/pkgconfig
and if
- a file is dropped in /usr/share/aclocal and this file test the
presence of pkg-config. It should be engouh to grep
AC_PATH_PROG(PKG_CONFIG, pkg-config, no)
in the file.
here is new patch, that implement what i have explained.
right now, as stefan stated , we have 3 solutions.
adding pkgconfig to rpm-build.
this will solve the problem for rpm building without needing to add
dependancy to all rpm.
but people using libfoo-devel, who use pkgconfig for ./configure script
will have to download it afterward. And, i do not think this is clan.
adding pkgconfig to each library
this is also bad because it would be added by hand. this is error prone,
and almost as bad as current solution. But this would solve the problem
of ./configure. I guess a rpmlint warning ( pc-files-without-pkgrequire
) would be enough for this one.
using the patch and automaticaly adding it to the package that need it
this would only add pkgconfig for library that really use it when
aclocal is used, and is automated, so no need to change library, just
rebuild them.
all file dropped in /usr/share/aclocal are included in ./configure
script whan aclocal is used, so, if the file check pkgconfig existence,
with a macro, it mean that all configure script trying to detect this
library will use pkgconfig, and so it is good to add it as requires.
i suggest the last solution, but, maybe someone see a problem ?
if not, i will submit a bugreport during the weekend.
--
Mickal Scherer
--- find-requires.bak 2003-09-01 09:49:23.0 +0200
+++ find-requires 2003-09-03 20:42:35.0 +0200
@@ -165,6 +165,13 @@
echo $tcllist | tr '[:blank:]' \\n | /usr/lib/rpm/tcl.req | sort -u
#
+# --- pkgconfig .pc files
+( echo $filelist | tr '[:blank:]' '\n' | grep -q '/usr/lib/pkgconfig/' )
+( aclocal_files=`echo $filelist | tr '[:blank:]' '\n' | grep '/usr/share/aclocal/'`;
+[ -n $aclocal_files ] grep -q 'AC_PATH_PROG(PKG_CONFIG, pkg-config,' $aclocal_files 2/dev/null echo 'pkgconfig'
+)
+
+#
# --- .so files.
for i in `echo $filelist | tr '[:blank:]' \n | egrep (/usr(/X11R6)?)?/lib(|64)/[^/]+\.so$`; do
objd=`objdump -p ${i} | grep SONAME`
Re: [Cooker] pkgconfig not required by -devel
On Monday 01 September 2003 21:16, Guillaume Rousse wrote: Ainsi parlait Oden Eriksson : I agree, rpm-devel should require pkgconfig. Same for me. That's silly to add it to all packages individually. But some package use pkgconfig in their autoconf macro, as a grep pkg /usr/share/aclocal * will show you. -- Mickaël Scherer
Re: [Cooker] pkgconfig not required by -devel
On Monday 01 September 2003 23:31, Abel Cheung wrote: On 2003-09-01(Mon) 22:29:52 +0200, Stefan van der Eijk wrote: I agree, rpm-devel should require pkgconfig. Same for me. That's silly to add it to all packages individually. So..., what do you all suggest? File a bugreport or bug the rpm package maintainer? I personally think this solution would benefit the most... I guess there are 3 options: - add it as a Requires to rpm-build (not rpm-devel!) It is no really mandatory to use it for rpm building, only for some packages, so, there is no need to a requires for some packages. - add some functionality to the find-requires script - hunt down the packages and add it to the spec file Michael has already done the 2nd option as a patch to find-requires, posted in cooker. The patch is not perfect. If applied right now,linphone would requires pkgconfig, which is useless. So, what i propose is to add pkgconfig as a requires if : - a file is dropped in /usr/lib/pkgconfig and if - a file is dropped in /usr/share/aclocal and this file test the presence of pkg-config. It should be engouh to grep AC_PATH_PROG(PKG_CONFIG, pkg-config, no) in the file. -- Michal Scherer
Re: [Cooker] pam_console_apply Segmentation fault during rc.sysinit
Hi. On Tuesday 02 September 2003 21:08, Tony Rick wrote: Just joined, first post: is it a bug? Yes, it seems I downloaded and installed 9.2RC1 on 28 August. During startup, the trace reports that getgrname failed to find group video, and then pam_console_apply reports a segmentation fault. I reproduced this by simply calling 'pam_console_apply -r' as root (this is how it appears in rc.sysinit, line 863 or there about). [ snip ] I can't find this in the cooker mailing list archives or in Bugzilla, but I have only been looking for references to pam and segmentation faults. then you should enter it in bugzilla. you should first upgrade if possible to cooker, to check, if , by chance, it has been resolved, and if not, create a bugzilla account and add it. since pam is developped by rh, you should look on their bugzilla, and report the info in the bug report. -- Michaël Scherer
Re: [Cooker] pkgconfig not required by -devel
On Monday 01 September 2003 07:54, Stefan van der Eijk wrote:
Michael,
Would you consider making this into a find-requires script? In that
case the dependencies won't need to be added manually.
done.
--
Mickaël Scherer
--- find-requires.bak 2003-09-01 09:49:23.0 +0200
+++ find-requires 2003-09-01 10:52:13.0 +0200
@@ -165,6 +165,10 @@
echo $tcllist | tr '[:blank:]' \\n | /usr/lib/rpm/tcl.req | sort -u
#
+# --- pkgconfig .pc files
+( echo $filelist | tr '[:blank:]' \\n | grep -q '/usr/lib/pkgconfig/' ) echo 'pkgconfig'
+
+#
# --- .so files.
for i in `echo $filelist | tr '[:blank:]' \n | egrep (/usr(/X11R6)?)?/lib(|64)/[^/]+\.so$`; do
objd=`objdump -p ${i} | grep SONAME`
[Cooker] pkgconfig not required by -devel
Hi.
will trying to fix buildrequires, i have come on this one
http://eijk.homelinux.org/build/contrib/i586/problem/ices-0.3-2mdk
the compilation stopped when configure tried to detect libshout3-devel,
listed as buildrequires. It stopped because pkgconfig was not found.
libshout3-devel should require pkgconfig, since it is needed for the
proper auto detection of the include.
there is a lot of package who drop a file in /usr/lib/pkgconfig/ without
having a requires on pkgconfig.
this script [1] gives most of the package who misses the requires on
pkgconfig [2].
Am i wrong when i say they should be fixed ?
Because there is a lot of wrong requires, i may have missed a detail
Some of the packages are not -devel library, i guess they should be
fixed, because pkgconfig files belong to the -devel.
And some are not library, such as linphone, who does not provides any
include or libray, so i guess there is a problem.
[1]
for i in `urpmf '/usr/lib/pkgconfig/' | awk -F: '{print $1}' | sort |
uniq `;
do
if ! urpmq -d $i | grep -q pkgconfig ; then
echo $i ;
fi;
done;
[2]
autotrace
epiphany-devel
gdesklets
gedit
gkrellm-devel
gnome-mime-data
gnome-python
gnome-system-tools
gok
gretl
gstreamer-python
gtk-doc
gtk-engines2
gtkmathview
ImageMagick
libaiksaurus-1.0_0-devel
libalsa2-devel
libalsaplayer0-devel
libaudiofile0-devel
libavifile0.7-devel
libdia-newcanvas0
libdirectfb0.9_18-devel
libebg1
libecore0
libedb1
libeet0
libefs1-devel
libesound0-devel
libevas1
libexif9-devel
libflatzebra0.1
libfontconfig1-devel
libgmime2.0-devel
libgnokii0
libgphoto2-devel
libgsl0-devel
libgtkglarea2.0
libgtksharpglue-devel
libladcca1-devel
liblrdf0-devel
libmnote7-devel
libmpeg2dec0-devel
libmusicbrainz2-devel
libneon0.24-devel
libnspr4-devel
libnss3-devel
libogre0-devel
libopenbabel0-devel
libopenssl0.9.7-devel
libots-1_0-devel
libparagui1.0-devel
libpng3-devel
libscaffold-1_0-devel
libshout3-devel
libsmi2-devel
libsqlite0-devel
libstartup-notification-1_0-devel
libticables3-devel
libticalcs4-devel
libtifiles0-devel
libtre3-devel
libuser1-devel
libxfree86-devel
libxine1-devel
libxml++10-devel
libxml1-devel
libxml2-devel
libxslt1-devel
linphone
mozilla-devel
mozilla-firebird-devel
pstoedit
pyorbit-devel
rhythmbox
spirit
streamtuner-devel
zziplib0-devel
--
Mickaël Scherer
Re: [Cooker] [PATCH] torrentsniff spec file
On Friday 29 August 2003 08:00, Todd Lyons wrote: It works well, but it failed at first because it was missing a perl file Digest/SHA1.pm, provided by perl-digest-SHA1. Add a simple Requires. since perl-Requires should be automatically discovered, there is a bug somewhere, no ? maybe a simple rebuild is sufficient ? -- Mickaël Scherer
Re: [Cooker] Main mirror BROKEN AGAIN!!!!!
On Thursday 28 August 2003 22:57, dave wrote: The main mirror sunsite.uio.no is not current again. It has not received an update in over 2 days!!! if the main mirror is broken, can you explain why I have yesterday update on the one i am using ( sunsite.informatik.rwth-aachen.de ) how do you know it is the main mirror ? and what does main mirror mean for you ? if this is the mirror from where all other update, then this doesn't seems to be the case anymore. What is causing this? I do not remember having this many mirror problems with cooker before 9.1 was released. How can cooker be tested when the main source for download is constantly having problems. by changing sources, and by being sure that you do not have a problem on your side. you can also use rc1. LETS make is priority number 1 to fix the mirror problems!!! ok, what is your proposition ? if there is a problem with a mirror, then you should write to mirror administrator, not to mandrake developpers. i agree when you say it slowdown the whole process of testing, but, what is the problem of having a mirror 2 day behind ? usually, when you signal a bug in bugzilla, you check to see if it has already been reported before, no ? and you use the fields version to signal your version, so, what is really the problem ? -- Mickaël Scherer
Re: [Cooker] More mirror madness
On Thursday 28 August 2003 19:24, BLINDAUER Emmanuel wrote: For my part, I must be in the mountain, because, since 3 ou 4 weeks, I don't have access to a mirror which is complete. How ca we want to help at cooker, if we even can get the packages ? The original unstable side of cooker is on packages, not on mirror isues! Today I wanted to look back at kopete. No mirror (I use 4 differents) was able to provide me a package, via urpmi or apt-get. so, this is a problem on the mirror. i use ftp://sunsite.informatik.rwth-aachen.de, and i have all update. I use for 4 months without too much problem. -- Michaël Scherer
Re: [Cooker] More mirror madness
On Wednesday 27 August 2003 09:32, Dave Cotton wrote: On Tuesday 26 August 2003 22:48, Paul Dorman wrote: but I'm peeved You are not alone. This happens too often (like, every day). Can the maintainers **pah-lease** sort out the mirror situation. That gets my vote. I have posted on this list and 1 or 2 others have. Perhaps everyone in control things everything is hunky dory because so few complain. Maybe because there actually are so few who have not given up totally. Frequently when it all blows up in their faces after release we get the well nobody told me so I assumed it was all right type of posting. How many people are actually trying to test on as near to production as possible machines, I doubt more than a handful. There cannot be many because the mirror debacle over the last few days has not created a mass of postings. Mirrors all over the world have not been updating, mirrors have multiple versions of the same package. I receive changelogs that never reach a mirror. Mirrors have stale Carroll archiving messages. The primary mirror link on Cooker's home page hadn't updated for days. The two highest bandwidth mirrors in France club-internet and proxad had not been updated for days, proxad in fact had zero length files. The plea is always please test, give us the means to test and we will. Well, i think that people forget that mandrake employees are working. They do not break mirrors for pleasure, and they are fully aware of the problem. Do you really think they try to slow down the test ? to annoy their testers ? no, they are working on it, and all we have to do is to wait. Sometimes, a solution is not easy to find, or take time to be applied. -- Mickaël Scherer
Re: [Cooker] Bugzilla down?
On Wednesday 27 August 2003 14:38, Steffen Barszus wrote: Am Mittwoch, 27. August 2003 13:58 schrieb Buchan Milne: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Götz Waschk wrote: Am Mittwoch, 27. August 2003, 00:40:40 Uhr MET, schrieb Warly: Gettting forbidden error on qa.mandrakesoft.com fixed :) It's not fixed for me: http://qa.mandrakesoft.com/show_bug.cgi?id=4918 It's not possible to view/search for bug reports, but it is not possible to post, and probably not to modify: http://qa.mandrakesoft.com/short_desc=Zenity%20emulation%20of%20gd ia log%20is%20broken...keywords=form_name=enter_bug Forbidden You don't have permission to access /short_desc=Zenity emulation of gdialog is broken (at least for --menu)comment ... Additionally, a 403 Forbidden error was encountered while trying to use an ErrorDocument to handle the request. Yep that is what i get since weeks now, but nobody seems to have recognized that yet. It works for me, and i guess for other too so, there is something strange. are you using a proxy ? on which navigator ? can you tell me what did you try exactly to do ? -- Mickaël Scherer
Re: [Cooker] Re: [Contrib-Rpm] wmtv-0.6.5-4mdk
On Monday 25 August 2003 19:19, andre wrote: On Monday 25 August 2003 18:10, Michael Scherer wrote: [Contrib-RPM] -=-=-=- Name: wmtv Relocations: (not relocateable) Version : 0.6.5 Vendor: MandrakeSoft Release : 4mdk Build Date: Mon Aug 25 17:32:37 2003 Install Date: (not installed) Build Host: klama.mandrake.org Group : Video Source RPM: (none) Size: 48646License: GPL Signature : (none) Packager: Antoine Ginies [EMAIL PROTECTED] URL : http://www.student.uwa.edu.au/~wliang Summary : WMTV is a WindowMaker dock.app that controls TV Description : wmtv is a WindowMaker dock.app that controls TV Cards which are supported under the Linux Kernel 2.2.x series (mainly for cards based on BrookTree BT848/848a/84 9a/878/879 chipsets: There are lots of cards out there which uses these chipsets such as the Hauppauge WinTV, Aimslab VideoExtreme, miroVIDEO PCTV, AverMedia TV Phone, etc). Linux Kernel 2.2.x series? I assume it also supports 2.4 i do not have a TV card to test, i just added the build depend. If someone is using it, or even having a card, i would like to have some feedback, to be sure it work. If it does not, we should remove it, or port it. I have uploaded cpufreqd 3 weeks ago , and, i just saw it doesn't work, even if 2 or 3 people said they were impatient to test it just before i have made the package. I wasn't able to test it sooner , since my laptop lacked the powernow support so, i guess we should have something to see packages used and tested, and this kind of information. it has been already discussed, IIRC. -- Mickaël Scherer
Re: [Cooker] non-builders
On Tuesday 26 August 2003 02:28, Austin wrote: I have several packages that either don't build or don't run with the current cooker. I've contacted most of the authors, but they're either clueless (like me), or non-existent. So if anyone is interested, they are welcome to help me fix these packages. Otherwise I guess I should delete them from the mirrors before 9.2 goes final right? viewmol (won't build) it just builded fine on klama this morning . Since i do not know how to use the software, i cannot test. you can found it on http://klama/~michael/rpm I will try to see the other problem if i found the time, can you put the trivial modification ( ie increase the python version ) you have made somewhere ? -- Michaël Scherer
Re: [Cooker] [REQUEST] OpenOffice.org 1.1 RCx inclusion
On Tuesday 26 August 2003 11:27, Buchan Milne wrote: Austin wrote: On 08/25/03 17:34:00, Buchan Milne wrote: So we have a 1.1 coming? Any chance at SDK also (or OpenOffice.org-devel package or similar)? There have been requests for cuckoo[1] (kpart embedding OO.o) on MandrakeClub, and it looks very cool (and functoinal), but requires the SDK to build. Hell, we don't even have the XFree86 SDK packaged yet. ??? I have to build XFree86 every time I want to link against it... which also means rebuild bots fail for other platforms. Whatever toolkit you are linking against should require XFree86-devel (recently renamed to libxfree86-devel), and pull it in. Or am I missing something? he is talking about the ati.2 drivers from the gatos project, i think. -- Michaël Scherer
Re: [Cooker] /proc/cpufreq/
On Sunday 24 August 2003 18:39, Laurent Culioli wrote: Le Samedi 23 Août 2003 16:46, Helge Hielscher a écrit : Hello, are there any additional steps to get /proc/cpufreq (or the same in the /sys/ filesystem) to work? Hi , 1/ use acpi ( if not , delete acpi=off in your lilo.conf ) 2/ add speedstep-centrino ( i think this , cause you have a pentium-M , if the module failed try with speedstep-ich ) in /etc/modules 3/ reboot or modprobe speedstep-centrino 4/ then you can setup cpu performance with : echo -n X /proc/acpi/processor/CPU0/throttling mhh, on my laptop, i must use powernow-k7, and the commande the slow the processor is echo '0:100:powersave' /proc/cpufreq as root. you can use cpufreqd in contribs to achieve this automaticaly -- Michaël Scherer
Re: [Cooker] Re: [Contrib-Rpm] amaya-7.0-2mdk
On Monday 25 August 2003 12:23, Thierry Vignaud wrote: Michael Scherer [EMAIL PROTECTED] writes: Name: amayaRelocations: (not relocateable) Version : 7.0 Vendor: MandrakeSoft Release : 2mdk Build Date: Sun Aug 24 21:27:20 2003 latest availlable release is 8.x that's right. I just checked what the bots reported and, they didn't mention anything about amaya, so, i didn't checked on the website, my goal was to correct the buildrequires . the database behind the bot should be fixed, where should i submit a bugreport ? -- Mickaël Scherer
Re: [Cooker] gdm eats 100% CPU
On Friday 22 August 2003 16:23, J.A. Magallon wrote: Hi. After latest update, gdm eats 100% CPU when I am loged in (not when nobody is logged). Anybody else suffers this ? same problem here. gdm-2.4.2.101-1mdk -- Michaël Scherer
Re: [Cooker] Software submission for the Mandrake distribution
On Thursday 21 August 2003 14:12, Duncan wrote: On Wed 20 Aug 2003 07:25, Buchan Milne posted as excerpted below: Well, the first question I have is, can this software be compiled from source using only free software available in Mandrake (main + contrib)? If not, then it can't really be included. What about PLF? I know they handle quite a bit of stuff that contrib can't, due to the licensing requirements. I don't quite understand all they do and where they draw the lines, but Manoj, if you haven't, consider investigating PLF, as it may be just the type of place for such a thing. As for the driver, I see a very practical use for it. However, one of the reasons I'm on Mandrake is because I support their software libre philosophy, and the others are correct -- given the situation with the DDK, this doesn't belong in the distrib itself. However, PLF? PLF only distribute rpm. Rpms of linux software, for mandrake. There is no reason to distribute windows program. if you want a cd of free software tools, you can check this website. http://gnuwin.epfl.ch/en/index.html or this one ( in french ) http://www.framasoft.net/ -- Mickaël Scherer
Re: [Cooker] [HELP WANTED] model name of your /proc/cpuinfo
On Tuesday 19 August 2003 16:01, Guillaume Cottenceau wrote: Hi, Could you please copy the model name line of the /proc/cpuinfo of your laptop computer(s)? sony pcg-fx702 : model name : mobile AMD Athlon(tm) XP 1400+ sony pcg-fx707 : model name : Pentium III (Coppermine) sony pcg-fx403 : model name : Pentium III (Coppermine) -- Michaël Scherer
Re: [Cooker] Software submission for the Mandrake distribution
On Wednesday 20 August 2003 16:54, Olivier Blin wrote: 2. Microsoft's DDK which is _free_. Free as in free beer or free speech ? i never understood the difference between the two, it should be free as youpaynothing. -- Michaël Scherer
Re: [Cooker] 9.2 beta 2 Konqueror Bookmark Bar bug/freeze
On Monday 18 August 2003 11:47, Mark Watts wrote: Steps to reproduce: 1) Install 9.2b2 2) Start Konqueror from the Kicker. 3) Browse to a selection of sites and add each site as a bookmark. Add enough to make the bookmark bar fill and display the icon. (If the bookmark bar isnt visable, make it, using the Settings Toolbars menu) 4) Close konqueror and restart it. 5) Use the bookmark bar icon to click on a bookmark link that is currently not visible. Clicking on a link on the bookmark bar that _is_ visable does not trigger this bug. Konqueror will probably freeze now, and take 100% cpu usage. I have seen this exact behaviour on my 9.1 box as well. Note: This only seems to be triggered if the first thing you do after starting konqueror is to open a bookmark, using the bookmark bar, which is not currently visable. can you please enter the bug in the bug database ? or see if it handn't been already entered ? http://qa.mandrakesoft.com/ it allows us to keep track of the bugs and to classify them more easily. -- Michaël Scherer
Re: [Cooker] ProPolice stack protection for next Mandrake release?
On Sunday 17 August 2003 17:47, Aleksander Adamowski wrote: How about adding ProPolice stack protection to stock Mandrake GCC? For discussion, see this Mandrake Wiki topic: http://qa.mandrakesoft.com/twiki/bin/view/Main/ProPolice OpenBSD already did that. having the patch would be nice, but not applied as default. we already have libsafe for a runtime protection, which is more flexible than compile time protection. in fact with the patch, you only would need to rebuild and change the .rpmmacro files of the user who is used to build. -- Michaël Scherer
Re: [Cooker] dynamic DNS and /etc/ppp/ip-up (need to be rethinked ?)
On Sunday 17 August 2003 00:44, Steffen Barszus wrote: Am Sonntag, 17. August 2003 00:23 schrieb Steffen Barszus: [:snip a lot:] else unset PPP_TEMP_ENTRY fi Are there cases there a dynamic DNS don't want to be used ? which ? From what I have read on the net, if the option usepeerdns is used the variable USEPEERDNS is set, so this could be used, am I right ? right. to answer your question about dynamic DNS I think that pppd can be used for weird thing, such a a pppd tunnel over ssh. example : http://www.linux.org/docs/ldp/howto/mini/VPN.html i have also seen that some people use alternative DNS in order to access some non standard domain, such as http://www.name-space.com/ and, i may want to use my own dns servers, for security, for example ( dns sec ). all of these would requires to not use the option usepeerdns, so, i guess you are right. In order to work on a real script, i propose to use the wiki to write the files : http://qa.mandrakesoft.com/twiki/bin/edit/Main/DraftPppScript By the way, i have put ppp maintener in Cc, please use answer to all, thanks :) and, for completeness, i have take a look at /etc/ppp/ipup on debian, it only contains the runparts line of the proposal i have made. -- Mickaël Scherer
Re: [Cooker] dynamic DNS and /etc/ppp/ip-up (need to be rethinked ?)
On Saturday 16 August 2003 18:35, Olivier Blin wrote:
Hi,
I've found some weird things in /etc/ppp/ip-up
1) what's this trick for PPP_TEMP_ENTRY ?
# for dynamic DNS support with gnome-ppp and kppp and draknet
if grep -i '#.*ppp temp entry' /etc/resolv.conf /dev/null ; then
PPP_TEMP_ENTRY=`grep '#.*ppp temp entry' /etc/resolv.conf | \
tail -1 | sed 's/.*ppp temp entry/# ppp temp entry/' `
else
unset PPP_TEMP_ENTRY
fi
Why use this grep | tail | sed ? The result should always be # ppp
temp entry
2) kppp now writes temp entries in /etc/resolv.conf after
/etc/ppp/ip-up is run, gnome-ppp isn't supported anymore, and
drakconnect uses kppp (it even run it as root ...), so there is no
need to check for #.*ppp temp entry here ...
grep -i '#.*ppp temp entry' /etc/resolv.conf will never match here.
So PPP_TEMP_ENTRY will never be set.
3) Since PPP_TEMP_ENTRY is always unset, the script never writes DNS
in /etc/resolv.conf
This is a problem when pppd or wvdial are used to establish the
connection.
if [ -n $PPP_TEMP_ENTRY ]; then
[ -n $DNS1 ] \
echo -e nameserver $DNS1 $PPP_TEMP_ENTRY /etc/resolv.conf
[ -n $DNS2 ] \
echo -e nameserver $DNS2 $PPP_TEMP_ENTRY /etc/resolv.conf
fi
I propose to drop all this if grep -i '#.*ppp temp entry'
/etc/resolv.conf and if [ -n $PPP_TEMP_ENTRY ] blocks and to
replace them with :
PPP_TEMP_ENTRY=# ppp temp entry
[ -n $DNS1 ] \
echo -e nameserver $DNS1 $PPP_TEMP_ENTRY /etc/resolv.conf
[ -n $DNS2 ] \
echo -e nameserver $DNS2 $PPP_TEMP_ENTRY /etc/resolv.conf
Have I missed something ?
For dialup users, with the current ip-up script, only kppp sets
dynamic DNS in /etc/resolv.conf (it doesn't use the ip-up to do that)
I would also add something to replace if-up.local
[ -x /etc/ppp/ip-up.local ] /etc/ppp/ip-up.local $@
replaced by
if [ -d /etc/ppp/ifup-ppp.d -a -x /usr/bin/run-parts ];
then
/usr/bin/run-parts --arg $@ /etc/ppp/ipup.d/
fi
then just drop a script to call /etc/ppp/ipup.local in
/etc/ppp/ipup-ppp.d, for backward compatibility.
the same applis for ipdown.local .
This would also mean we can remove the piece of code to flush sendmail
queue, and put it in a separate script, to be cleaner.
the second problem i see is that each file dropped in
/etc/sysconfig/network-scripts/ifup.d/, are not executed with pppd.
to give a small example ( and the only one i see ), wshaperx does not
work out of the box, on a adsl connection.
it seems that this line :
[ -f /etc/sysconfig/network-scripts/ifcfg-${LOGDEVICE} ]
/etc/sysconfig/network-scripts/ifup-post ifcfg-${LOGDEVICE}
never call ifup-post, because $LOGDEVICE is always unset.
at least for adsl and RTC connection. I know that it also handle ppp
connection for Isdn and maybe for tunnel, so this line is maybe for
this kind of connection.
removing this line and dropping a script in /etc/ppp/ipup.d/ could be a
solution.
I think this solution can be more flexible than the current one.
did i forget someting ?
is it possible to replace the script in the current package ?
--
Michaël Scherer
Re: [Cooker] kppp and suid
On Saturday 16 August 2003 19:51, Diego Iastrubni wrote: Hi kppp has root suid, is it really necessary ? pppd has already root suid, shouldn't it be be enough ? no, since kppp needs to write into /etc/ppp/* what about using a group for this ? /etc/ppp/ could be group writeable, and kppp sgid to this group. -- Michal Scherer
Re: [Cooker] Re: [CHRPM] perl-String-ShellQuote-1.00-7mdk
On Friday 15 August 2003 02:58, Per Øyvind Karlsen wrote: On Friday 15 August 2003 02:45, Guillaume Rousse wrote: Ainsi parlait Guillaume Cottenceau : - fix DIRM: file ShellQuote.pm /usr/lib/perl5/vendor_perl/5.8.1/String/ Not sure it is really a good idea however: they are many perl String::Foo subpackages, even if not already packaged, are they all going to own the String directory ? Unfortunaltly, there is no top-level String module :-( hm, it does'nt do any harm, and if they all own the String directory, it will only be rm'ed when the last package owning it is removed.. IMHO it's a good idea in such cases.. so, why do not we say that a package own alls directory ? A package using /usr/bin would own /usr and /usr/bin. This way, /usr/bin will be removed when there is nothing left. And the same goes for all packages. -- Mickaël Scherer
Re: [Cooker] OT - can someone pass this on to Andrey?
On Friday 15 August 2003 15:49, Dave Cotton wrote: On Fri, 2003-08-15 at 14:10, John Keller wrote: If you or anyone else figures out the basics of the form, please let me know. My ISP's SMTP is also blacklisted (on an on-and-off basis), and it'd be nice to address it when it happens. Too bad what little Russian I learned was pushed out of my brain by my French... Oleane is blacklisted? matchbox.fr resolves to an address from the Oleane block, do you have a fixed IP? If so perhaps you were the open relay :) If not you're another victim of the a previous user of the IP had an open relay syndrome. I just realised why providers like Wanadoo/Oleane never check for open relays, if you complain about the problem that others are causing you they'll say they can always sell you a fixed IP at an inflated price. indeed and the price of the fixed IP alone is the same as a adsl connection with a fixed IP at their competitors... I will not give any name... -- Mickaël Scherer
Re: [Cooker] makewhatis can't find man pages
On Sunday 10 August 2003 13:28, guran wrote: Hi From: [EMAIL PROTECTED] guran]# /usr/sbin/makewhatis /usr/sbin/makewhatis: line 157: cd: /usr/man: No such file or directory [EMAIL PROTECTED] guran]# line 157 should contain /usr/share/man or from: /usr/share/man/man8/makewhatis.8.bz2 /usr/share/man/man1/whatis.1.bz2 /usr/share/man/cs/man1/whatis.1.bz2 /usr/share/man/da/man1/whatis.1.bz2 /usr/share/man/es/man1/whatis.1.bz2 /usr/share/man/fi/man1/whatis.1.bz2 /usr/share/man/hr/man1/whatis.1.bz2 /usr/share/man/ja/man1/whatis.1.bz2 /usr/share/man/nl/man1/whatis.1.bz2 /usr/share/man/pt/man1/whatis.1.bz2 /usr/share/man/sl/man1/whatis.1.bz2 can you please use bugzilla to report bug ? you go to qa.mandrakesoft.com, and enter the bug. if will be easier for us to correct it, since the bug will be send to the mainteners of the rpm. -- Michaël Scherer
Re: [Cooker] Just asking- Open Office 1.1
On Thursday 07 August 2003 11:47, Robert Fox wrote: Is there any reason why OpenOffice 1.1 is not in Cooker? I see lot's of people asking about it for 9.2 - and it's supposed to be quite stable (RC2) beta 2, not rc2 this is not the same. -- Mickaël Scherer
Re: [Cooker] 9.2 Req. Laptop Support
On Tuesday 05 August 2003 13:26, Emmanuel Moll wrote: Features that would be interesting in a laptop installation class though are: -enable cpufreqd by default and make it easy to configure it ok, packaged, and uploaded. can you please test it ? I was unable to load the powernow modules on my laptop ( sony pcg fx702 ), so, i guess the daemon work, but i was unable to use it. -- Mickaël Scherer
[Cooker] Re: [Maintainers] rpmlint: perl5-naming-policy-not-applied
On Friday 08 August 2003 17:17, Guillaume Rousse wrote: Ainsi parlait Michael Scherer : if this is not a pure perl module, you can ignore it.is we or, if you really want to be clean, just split the package, in linuxdoc-tools and perl-Text, or something like that. Many perl script have part of their code into modules, but as long as these modules are not standard, and won't be requested by other packages, there is no need to split just to comply to a misapplied naming policy. That's right. The same could be applied to library, but there is the 64 bit library problem, i guess, and there is nothing like this for perl and python. The problem now is to fix the test. Having a useless warning is bad, it add noise to the output. We have 2 choices : add a exception, or change the files regexp. adding exception is not a valid solution, because having too much exceptions is a proof they are not exceptions. the test is based on the presence on files in some directory. /usr/lib/xmms = xmms naming policy is tested /usr/lib/python-* = python so, i propose to ignore the test if there is a file in some directory ( mainly bin/ or sbin/ ). WDYT ? -- Michaël Scherer
Re: [Cooker] Packaging errors as of today
On Monday 11 August 2003 22:15, Frank Griffin wrote: The following occur in today's cooker: thanks for your report, but we have distriblint, made by olivier thauvin to check this kind of error. eveything is detailed on the wiki, and each maintener is mailed by a bot when a error is found. -- Mickaël Scherer
