[12:01] seb128: yes, thanks [12:01] np [12:01] check if they're sane, of course ;) [12:01] but the current releases look ok [12:02] jdub: ah, and a guy asked for a gmane archive of ubuntu-fr, I guess that's ok ? :) [12:02] don't worry, for big package I run them one day one my box before uploading most of the time [12:02] or at least time to reboot and make basic tests [12:02] seb128: i think it's already on gmane [12:02] some dude mailed me to say some of the posts were missing ;) [12:02] oh, ok, cool :) [12:03] 37 people on the -fr list, that's a good start :) [12:10] cool [12:10] you are beating spain and germany ;) [12:13] he he [12:32] Kamion: still here? [12:38] T-Bone: you modify the source... [12:39] this is what i did, eventually ;) [12:39] unfortunately gcc takes _time_ to build [12:47] but build deps get dealt with early... [12:48] Rejected: file 'linux-686-smp_2.6.8.1-10_i386.deb' has invalid priority 'restricted/optional' [contains '/'] . [12:48] sigh [12:49] mdz: current daily is candidate [12:50] Kamion: ok, I'll download and test [12:50] you can probably make amd64 your bottom priority, I've tested that [12:54] lamont: can you and/or elmo arrange for those messages to go someplace useful? [12:55] mdz: will do [12:55] lamont: warty-changes should be a reasonable place [12:56] you want all REJECT mail to go there? [12:56] yes, unless there's a bunch of spam I don't know about [01:03] hm, WinXP boots for me after installing Ubuntu === Kamion wonders what the variable is [01:03] mdz: all katie mail that would have gone direct to me now goes to warty-changes. [01:05] lamont: thanks much [01:05] lamont: new linux-meta happier? === lamont checks [01:06] WHOA NEW EXCITING KERNEL TO INSTALL [01:07] mdz: give it 2.6 minutes. [01:07] jdub: new exciting sounder CD candidate to test [01:07] either warty-changes should get the reject mail shortly after 16:10 your time, [01:07] sweeet [01:07] i386 WFM [01:07] or the new linux-meta -11 package will hit the archive at 16:33 [01:08] ran out of disk space, restarted rsync [01:08] heh [01:08] disk is cheap :) [01:08] yeah, but adding disks is not [01:09] just testing daily+new-parted for good measure, even though I can't reproduce the problem with old-parted [01:09] all my disk is being eaten by mythtv [01:09] mdz: btw, what's the status of mythtv in sid or universe? [01:09] Free PE / Size 14499 / 56.64 GB [01:09] Kamion: what was the package update cutoff point for the new sounder? [01:10] did you guys see the latest gnoppix announcement? [01:10] jdub: waiting for elmo to bring in lame [01:10] jdub: about 1300 GMT or something [01:10] "You'll find a new 0.8.1beta ( codename Ubuntu ) with bugfixes and Gnome 2.8 at: ftp://194.94.72.226/linux/gnoppix/gnoppix_0.8.1b4.iso" [01:10] mdz: "oh." [01:10] azeem: haha [01:10] Kamion: cool, ta [01:11] mdz: so the non-free universe is still going to be called 'multiverse'? [01:11] we *so* need livecds for testing :| [01:11] mdz: i punted alex bugging to you, btw. [01:11] jdub: noticed, followed up [01:12] cool [01:12] perhaps gnoppix is the first case of a CUD, as opposed to CDDs [01:12] azeem: it is all in the CDD family :) [01:12] jdub: actually later than that [01:12] cjwatson@little:~$ head -n 2 cdimage/log/daily-20040928.1.log [01:12] ===== Syncing Ubuntu mirror ===== [01:12] Tue Sep 28 15:55:48 BST 2004 [01:12] azeem: but yes, CUD explosion is very much in the plans ;) [01:13] I talked to amu about gnoppix for quite a while at LinuxTag, and he said they entirely repackage GNOME for gnoppix. Perhaps now with ubuntu he finally got driven to sanity :) [01:15] mako: will there be an ubuntu traffic coming soon? [01:21] Kamion: what do you think we should do about #1608 (some CD-ROM drives screw up DMA)? [01:22] Kamion: disabling DMA on all CD-ROM drives seems awfully heavy-handed, but it's simple and safe to implement [01:22] there was a bug in Debian about this ... [01:22] I think [01:22] * Osamu Aoki [01:22] - When CD mount fails, unmount it and try again, to work around DMA [01:22] issues. Closes: #255128 [01:23] that was an error on mount, though === sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-devel [01:24] Kamion: ok, nothing in the syslog about firmware loading [01:25] mdz: killing DMA on failure works for me, don't like the idea of doing it unconditionally [01:26] sabdfl: have you run the hw-detect stage yet? [01:27] Kamion: interesting, detect hardware comes after configure network [01:27] but even after jumping through it, nothing about the firmware load [01:28] sabdfl: um ... configure the network definitely has a higher menu-item-number than detect network hardware [01:29] sabdfl: ok, well, it was untested code, guess it doesn't quite work :-/ === T-Bone hopes to start ia64 stage2 by tomorrow, calls it a night in the meantime [01:29] mdz: didn't you have an ipw2100 you could play with too? [01:30] T-Bone: wow, good progress? [01:30] Kamion: on my system it is showing: [01:30] ... [01:30] Load installer components from CD [01:31] Detect network hardware [01:31] Configure the network [01:31] that's not what you said above ... [01:31] 00:27 < sabdfl> Kamion: interesting, detect hardware comes after configure network [01:31] then [01:31] Detect hardware [01:31] oh, that's detect disks [01:31] nice :-) [01:32] "configure the network" shows the wifi card [01:32] should fix the string there [01:32] Template: debian-installer/hw-detect-full/title [01:32] Type: text [01:32] # Main menu item [01:32] _Description: Detect hardware [01:32] but it doesn't seem to work [01:32] still, it's not screaming spam log messages :-) [01:33] to console [01:33] sabdfl: i'll post the m-l as soon as i start "stage2" (ie: building warty main against warty binaries, stage1 was building warty main against debian binaries) [01:33] sabdfl: this should hopefully happen tomorrow [01:33] awesome [01:34] sabdfl: guess it just egregiously failed to load firmware then. does that machine have PCMCIA? [01:34] yes [01:34] can i force a firmware load? [01:34] I hacked the firmware loading stuff onto the side of PCMCIA, which is kind of broken but I figured would work for now [01:34] Kamion: I have only a 2200 [01:34] not sure; mdz, do you know? [01:34] mdz: that would do [01:35] Kamion: I'm not sure that there's a condition we can detect to automatically disable DMA [01:35] sabdfl: another that should happen soon is my beginning to work on the installer for ia64. probably before the end of the week, or in the worst case next week [01:35] you can force a firmware load by loading the driver [01:35] +thing [01:35] chinstrap:~cjwatson/warty-i386-hacked.iso if you want to give this a go, no firmware on that image so you'll have to bring it in by USB stick or whatever [01:35] mdz: already did that, but nothing showed up in syslog [01:35] T-Bone: give me a shout if you need help, some ia64 love with the Ubuntu customizations will be needed [01:36] Kamion: that's noted, and I will undoubtly bother you alot ;) [01:36] sabdfl: it doesn't log anything when it's successful [01:36] mdz: it so does [01:36] mine doesn't [01:36] log "Firmware for $FIRMWARE loaded successfully." [01:36] mdz: it does in d-i [01:37] sabdfl: side note, i've started "for fun" an hppa builder. But that's "for fun". For now ;) [01:37] sabdfl: what's in /proc/sys/kernel/hotplug at the moment? [01:37] Kamion: oh, that [01:37] I'm pretty sure it loads the firmware every time the ipw2200 module is loaded [01:38] it says /sbin/hotplug [01:38] aha [01:38] shouldn't that be /sbin/hotplug-special-d-i-variant? [01:39] which doesn't exist [01:39] I think it's hotplug-pcmcia [01:39] neither of those exist [01:39] the special d-i variant is /bin/hotplug-pcmcia, and that's only loaded for brief periods while the PCMCIA queue is run [01:39] I think that's probably the bug [01:39] can i reset that and try to reload the driver? [01:39] there's no rmmod in d-i, you'd have to reboot [01:40] ewww [01:40] oooookkkkkkk [01:40] that's probably the biggest pain in d-i, actually [01:40] i really need to spend some face time with my launchpad code [01:40] sabdfl: after reboot, you could try copying /bin/hotplug-pcmcia to /sbin/hotplug and removing the non-firmware parts (since I think those will have unintended side-effects if run with non-PCMCIA) [01:42] sorry for awkwardness, will clean it up once an approach has been demonstrated to work [01:43] Kamion: I did this experiment a while back; it does work [01:44] if I poked all the bits in the right places, the firmware was loaded [01:44] mdz: in d-i? [01:44] Kamion: yes [01:44] mdz: is your ipw2200 card PCMCIA? [01:44] Kamion: no [01:44] miniPCI [01:44] mdz: then you must have had to set up /sbin/hotplug? [01:44] or possibly make /bin/hotplug-pcmcia the hotplug handler temporarily [01:45] Kamion: I did echo /bin/hotplug-pcmcia > /proc/sys/kernel/hotplug or whatever [01:45] aha [01:45] that was one of the bits that needed poking :-) [01:46] how about I create a hotplug-misc which sits in /sbin/hotplug for the rest of the time, then? [01:46] we don't have time to port proper hotplug now, will be doing that for hoary [01:48] sounds fine to me [01:49] what are the non-standard firmware parts [01:49] patch is trivial, I'll upload if there are no objections [01:49] non-standard? [01:50] your comment above... remove the "non-firmware parts".... [01:50] sorry, not non-standard, non-firmware [01:50] so i've got the firmware in /lib/firmware and /usr/lib/hotplug/firmware for good measure [01:50] network events and cardbus events [01:50] Kamion: is #1586 RC for Warty? [01:51] i've echo >/bin/hotplug-pcmcia /procy/sys/kernel/hotplug [01:51] Kamion: no objections to having an /sbin/hotplug [01:51] Kamion: sounder + current daily are the same today? [01:51] mdz: it's RC for me to look at it at the very least [01:51] now can i just modprobe the ipw2200? [01:51] I just haven't had time :( [01:51] what about the depmod? [01:51] if modprobe doesn't work then depmod ... [01:52] the build was hacked-up, hopefully shouldn't act that way if built normally [01:52] FATAL: module ipw2200 not found [01:52] jdub: aye, not that I've actually blessed S9 yet [01:52] ok, now: [01:52] burning powerpc now, I've tested the others [01:52] hotplug pcmcia says it detected pcmcia network interface eth0 [01:53] then ieee80211_crypt gets loaded [01:53] then the driver comes on stream [01:53] but says nothing about firmware being loaded [01:53] ah, hold on this is interesting [01:53] 00:51 < sabdfl> i've echo >/bin/hotplug-pcmcia /procy/sys/kernel/hotplug [01:53] if i dmesg i see a message at the end: [01:54] hope the position of the > there was a typo [01:54] Kamion: yes it was, that's all correct afaics [01:54] ok [01:54] but listen to this [01:54] Kamion: cool, thanks [01:54] it says "Radio frequency kill switch is on" [01:54] which was also in the syslog [01:54] then it says: [01:55] Kill switch must be turned off for wireless networking to work" [01:55] hmm [01:55] that'll be the wifi button/switch on the laptop somewhere [01:55] +To determine if you have such a switch, you can check the contents of: [01:55] + [01:55] + /proc/net/ipw2100/eth1/state [01:56] i don't have /proc/net/ipw2100 [01:56] this is a 2200 [01:56] probably same difference [01:57] is there a formal mechanism to suggesting a package for inclusion in ubuntu? [01:58] or maybe not, the ipw2200 driver's /proc-handling code seems different, oh well [01:58] Kamion: is there an option to debootstrap to tell it that the archive is not in a {dists,pool} form? [01:58] hazmat: email to ubuntu-devel [01:58] okay.. thanks. [01:59] lamont: debootstrap couldn't care less about pool, it uses Filename: fields [01:59] it cares about Release, though. :-( [01:59] lamont: it does indelibly want dists/$SUITE/Release [02:00] which needs to have the packages md5sums? [02:00] and dists/$SUITE/$COMPONENT/binary-$ARCH/Packages* [02:00] right [02:00] ok [02:00] I usually just fill the right values into the Release file from whatever CD image I have handy [02:01] sabdfl: /sys/bus/pci/devices/xxxx/rf_kill [02:01] seems like the right place [02:01] mdz: just found it too [02:01] using the fn-f8 sequence doesn't seem to affect that [02:01] 00:51 * joeyh notices we just got ntfsresize support [02:01] nice [02:02] BUT i can echo 1 > /sys/..../rf_kill [02:02] or vice versa [02:02] will play with that [02:03] I don't think my laptop has a switch; the sysfs flie says 0 [02:03] several of the test laptops in oxford had switches [02:03] some of them in unlikely places [02:08] Kamion: I think we need to change the default sources.list again [02:08] https://bugzilla.ubuntu.com/show_bug.cgi?id=1469 [02:08] should have one line with main and restricted [02:08] and then another line, commented out, with only universe [02:08] I thought I already fixed that [02:09] oh? [02:09] I'll find out soon, doing a new install [02:09] it was part of the universe changes a couple of days ago [02:09] * Remove 'main restricted' from the universe line, so people don't end up [02:09] with duplicate sources.list entries. [02:10] cunnin [02:10] g [02:10] the switch was well hidden, in the off position [02:10] WELL hidden [02:11] Kamion: please could you file a bug to detect those rf-kill switch situations, and alert the user, because i *bet* a lot get bitten by that [02:11] Kamion: oh, ok [02:11] you even updated the bug [02:11] oh, you did that just now. cheat :-) [02:11] yes :) [02:11] sabdfl: if you configure postfix, edit main.cf, and say 'dpkg-reconfigure postfix', do you (a) expect it to overwrite your changes, or (b) preserve them? (for questions that we ask in configure, that is) [02:11] sabdfl: ok [02:12] lamont: it's too late for me to give you a reliable answer to that :-) [02:12] sabdfl: ok. something to ponder in the morning then? [02:13] perhaps it should figure out you've made changes, offer to (a) show you a diff, (b) forget about it, (c) forge ahead [02:13] and backup the old config? [02:13] today we preserve them. which is why postfix is broken (see 1711) [02:13] sabdfl: sounds like ucf [02:13] that'd be ucf [02:14] what postfix actually does is to set the defaults to what you have configured currently. [02:14] so that hitting return all the way through produces no change. [02:14] that's pretty neat [02:14] and then debconf and base-config run: debconf does the wrong things because we haven't set things up yet, and base-config does nothing because the defaults are all what you said in debootstrap time... [02:15] elmo: PowerEdge 2650 is on the HardwareSupport page, but without a comment about whether it worked. did it? [02:15] elmo: we have a success report from a user on one [02:15] sabdfl: neat, but breaks ubuntu install of postfix [02:15] http://www.es.gnome.org/~telemaco/ [02:15] ^ :-) [02:16] lamont: the debconf/debootstrap fix would work around this, AIUI ... [02:16] I worked out the problem with my earlier change after talking to joeyh; I'd just implemented noninteractive multiselect wrong [02:16] ah, ok. that was the default vs seen thing? [02:17] trying to get debconf to set the seen flag on questions that would be asked if the frontend weren't noninteractive [02:17] my earlier fix caused all systems to have the aa_DJ locale configured [02:17] didn't upload that one :) [02:18] WOOHOOO! [02:18] custom install just went all the way through with wifi on ipw2200 and just worked [02:19] Kamion: it would be nice if debootstrap would accept Packages.gz in place of Packages. :-( [02:19] messages read in 2004-08: 6288, messages read so far in 2004-09: 13296 [02:19] actually, that's fnny [02:19] I was wondering why it took mutt so much longer than usual to open this month's read-mail folder [02:19] during the pre-boot it configure eth0 (wired) [02:19] then it rebooted, and ifup'd the wifi [02:19] which Just Worked [02:19] sabdfl: I guess you got lucky and the ipw2200 is later on the pci bus [02:20] no, it's earlier [02:20] really? [02:20] mdz: it's almost done [02:20] mdz: UT taht is [02:20] pre-boot, it didn't detect it, and configured eth0 (wired) for dhcp [02:20] mdz: where'd you get that stat? [02:20] mako: cool, that's a ton of mail to process [02:20] lamont: bzgrep -c '^From ' [02:20] heh [02:20] ~17k messages :) [02:20] then post-boot, the kernel finds ipw2200 and makes that eth0 [02:20] but now, with the rf_kill switch on, it actually boots [02:21] oh, heh [02:21] so the solution to the bug is turn off the rf_kill, or turn on the radio [02:21] so it wrote a config for eth0 for the wired net [02:21] and then used that config for the wireless? [02:21] yes [02:21] which just worked [02:21] the friday-friday stat is 17k [02:21] it's picked up [02:21] interestingly it doesn't need the essid [02:21] mako: 17k what? [02:21] it seems to just go with the closest / best ap [02:21] dhcp [02:21] and go [02:21] ergh.. 1.7k [02:21] sabdfl: the 0.8 driver grabs whatever has the strongest signal if it's set to scan [02:21] decimal error [02:22] hmm... [02:22] 1.7k mesasges [02:22] i'll update the bug,and file upstream [02:22] holy crap you guys need a more time-efficient sabdfl [02:22] sabdfl: kill-switch bug is #1877 [02:23] i can't belive i spent two whole nights on this at this stage :-) [02:23] sabdfl: i'm just going through user now and picking up the last few threads i flagged for possible inclusion [02:23] ergh.. for mdz [02:23] but it feels good to find the bug ;-) [02:23] possible inclusion? [02:23] sabdfl: don't worry about it, it was misdirected [02:23] mako: btw, thanks for the lurker tip [02:24] http://lists.ubuntu.com/lurker/ === mako nods [02:24] awesome [02:27] sometimes I want to strangle the genius at apple who eliminated the cd-rom eject button [02:27] the system is powered off, and I want to boot it from CD [02:28] I have to boot it up, software-eject the CD, and then reboot it [02:28] mdz: or have the always-useful paperclip handy. [02:29] you mean you want to boot it from a different CD from the one already in the drive? [02:29] m_tthew: they even put a door in front of the thing, making the hole difficult to get to (it's below the lip) [02:29] Kamion: the drive is empty, and I want to put in a CD and boot from it [02:29] oh your drive isn't slot-loading [02:29] correct [02:29] mdz: try 'eject cd' in Open Firmware [02:29] it's a tray with a door in front of it [02:30] mdz: a door? ugh [02:30] ah, I'll keep that in mind next time [02:30] then mac-boot [02:30] never tried this myself but it's what the net claims [02:30] is ndiswrapper in the current -3 kernel? [02:31] mdz: one rumour claims holding down the left mouse button during boot also ejects the CD ... [02:31] ooh, DMA is working on the hard disk now [02:32] I didn't think that was working before [02:32] bummer [02:32] jdub: yes it is [02:32] mdz: you won't believe this [02:32] with hotplug setup, and the firmware there [02:32] the installer sees the wifi and lets you use it [02:32] hey, warty-powerpc.iso burnt, time to reboot and try this stuff out myself [02:32] but it configures it as eth1 [02:32] Kamion: i386 and powerpc in progress now [02:32] i suspect, after reboot, it will be eth0 [02:32] why? [02:33] discover vs. hotplug [02:33] could it be because discover ran first? [02:33] argh. [02:33] discover must scan in a different order [02:33] maybe even exactly the reverse; wouldn't that just be grand? [02:34] well, discover goes first, doesn't find the wifi [02:34] sabdfl: give it a try; it really should preserve them [02:34] then hotplug could go, and find it [02:34] ifrename? [02:34] if they're both detected in the installer, ifrename should be able to do its job [02:34] ok [02:34] check that they're both in /target/etc/iftab [02:34] kamion, this is looking good to me (hotplug) [02:34] mdz: yes [02:35] hmm, so let's just switch d-i over to use hotplug exclusively, shall we? === mdz runs for cover [02:35] mdz: you can stay. you've clearly got the cohones for the position :-) [02:36] mdz: (also, you're ifrename magic seems good so far, they are both in there) [02:36] in theory, when hotplug finds them after the reboot, it'll look in that file and switch the name before bringing it up [02:36] so it'll find the wifi first, but call it eth1, then find the wired and name it eth0 [02:36] we'll see in about 15 minutes :-) [02:36] this may be the first valid test of that code, ever :-) [02:37] that was an "oh, look, all the bits are already there, we just need to write iftab and it'll work" sort of job [02:37] so if it works this time, we ship it? [02:37] duck [02:37] well it wrote iftab [02:37] i386 stage 1 successful [02:40] gah, 12M of updates on top of the daily CD already [02:43] mdz: do you know which was the bug i filed on the ipw200 loop of death? can't find it [02:43] Kamion: final dialo pre-boot still says "Finishing the installation" [02:44] can it say "Rebooting to finish installation"? [02:44] sabdfl: #1751 [02:45] mdz: thanks [02:45] jdub: can I get the name and address of the person who tied gnome-system-tools to wvdial, so I can send them a lump of coal? [02:52] heh [02:52] you trying to untie them? [02:57] hi [02:57] yo tseng [03:17] sabdfl: all depends how many translations you want to trash :) [03:17] sabdfl: but sure, can do [03:17] mdz: powerpc worked for me, xresprobe apparently hung solid on i386 ... [03:18] suspect it's specific to that machine, it's one of the Vias [03:19]