=== Topic unset by herb on #ubuntu-devel === herb changed the topic of #ubuntu-devel to: Intrepid created, frozen for toolchain | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/feisty/gutsy/hardy, #ubuntu+1 for intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs === herb changed the topic of #ubuntu-devel to: Launchpad is going down from 00:00 UTC until 02:00 UTC for a code update. Intrepid created, frozen for toolchain | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/feisty/gutsy/hardy, #ubuntu+1 for intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for === cjwatson changed the topic of #ubuntu-devel to: Launchpad down 00:00-02:00 UTC for a code update | Intrepid created, frozen for toolchain | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/feisty/gutsy/hardy, #ubuntu+1 for intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com === cjwatson changed the topic of #ubuntu-devel to: Launchpad down 00:00-02:00 UTC for a code update | Intrepid frozen for toolchain | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/feisty/gutsy/hardy, #ubuntu+1 for intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingW [00:09] seven more characters ... === cjwatson changed the topic of #ubuntu-devel to: Launchpad down 00:00-02:00 UTC for a code update | Intrepid coming soon | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/feisty/gutsy/hardy, #ubuntu+1 for intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs [00:11] cjwatson: Canonical should buy the w.ubun.tu domain for the wiki, to safe /topic pace [00:11] space* [00:11] heh [00:12] you could do s#dapper/feisty/gutsy## now, maybe [00:12] crimsun: Ok, with the changes in the config file for pulse, using dmix, the startup sound starts playing, but then stops. [00:12] crimsun: This is on my amd64 box. [00:13] TheMuso: hmm, so "cut off abruptly"? Interesting. Any ~/.xsession-errors or syslog debugging? [00:13] crimsun: Actually, pulseaudio crashed. [00:14] Hang on, I only used the config change. I'll ahve to rebuild it. [00:16] hopefully you don't [still, possibly] have pcm & ctl pulse set in an asoundrc :-) [00:17] crimsun: Nothing is set asoundrc wise. This is a fresh install of hardy amd64, as of last night. [00:17] ok [00:22] crimsun: Ok, same behavior. Pulse/alsa via dmix work fine, but no startup sound. [00:22] Checking gnome-sound-properties now to see if I still get the caching error. [00:22] Yep, I do. [00:22] does it show it in ~/.xsession-errors or elsewhere? [00:23] Let me check. [00:26] crimsun: Only that the sample is too long/sample caching errors from the startup sound. [00:26] that's odd. [00:26] I'll try with a fresh user. [00:27] All my tests have been conducted on fresh installs. [00:28] crimsun: http://www.pastebin.ca/1003438 from syslog [00:31] TheMuso: ok, for sanity-check, please pastebin your /etc/pulse/default.pa [00:32] crimsun: http://www.pastebin.ca/1003445 [00:35] ok, that looks good [00:35] azeem: but I already got enough space so there's no need to optimise further [00:36] TheMuso: I can't reproduce it on any hardware locally, but I'll try to chase it down with a desktop cd [00:36] TheMuso: thanks for testing [00:36] crimsun: I'll test on other boxes with other sound hardware as well. [00:36] crimsun: np. [00:37] cjwatson: oh come on, it's only 0:30AM and you're done with optimzations already!?! [00:41] I wish there were a way to tell valgrind not to bother about child processes [00:41] of course there's a zillion reachable bytes when I exec(), I'm hardly going to bother freeing them, am I?! [00:42] you should obviously be freeing them in a fini [00:43] I personally consider exec() to be a great big free() from god [00:43] I have similar feelings about exit() [00:44] youtube does JS detection of a flash plugin and directs users to Adobe's web if not found [00:44] and it's quite common expectation that "go to adobe.com" is the way to install flash if a user doesn't have it installed [00:44] i was thinking we could subvert the whole thing by installing a sort of stub flash plugin [00:45] alex-weej: I've been nagging adobe about that [00:45] about providing instructions on using APT? [00:45] or about providing their own debs? [00:45] about improving that general experiance [00:45] if you land at that page in Ubuntu, it should just work [00:45] well i think the blame lies partly on web developers [00:45] ironically including a flash animation on that page would make it work [00:46] yeah [00:46] so how about we create a stub flash plugin that will respond positively to the detection routines [00:46] yet will trigger the proper installation [00:46] is that possible? [00:47] and would there be odd side effects? [00:47] i don't know why it shouldn't be [00:47] Youtube seems to detect Gnash correctly, so I'm sure it's possible. [00:47] and swfdec [00:47] i think it just detects if there's a handler for the MIME object type [00:48] it may be easier to circumvent that inside mozilla [00:48] but doing it with a stub plugin may be less intrusive [00:48] And I suspect it might be more fruitful to nag Google than Adobe. [00:49] why Google? [00:49] Well, their Youtube division. [00:49] oh, of course [00:49] but there is still a lot of other websites that do detection [00:49] they think they're being clever [00:49] and all make the fatal misassumption that "go to adobe.com" is a universal instruction [00:51] so if it's possible, what's the consensus? good idea? [00:52] * Nafallo votes yes [00:58] or can anyone think of a better way of dealing with the JS detection? [01:00] alex-weej: find some individual example pages, and just see what they're doing [01:01] alex-weej: they're all going to copying each other, so after 5-6 combinations it should be possible to have covered the bases [01:02] IsObject(CreateObject("ShockwaveFlash.ShockwaveFlash.6")) [01:03] maybe that's user functions... [01:03] oh wait it's VBScript, haha [01:03] for IE/Win on one particular site (www.anjunabeats.com) [01:04] simplest way is

Go to Adobe.com which is what the Beeb appear to be doing [01:04] navigator.plugins["Shockwave Flash"] [01:05] alex-weej: Or, provide the plugin. And as soon as it gets called in anger, call out tothe install assistent [01:05] sladen: doesn't that method trigger the plugin installation tho? [01:06] sladen: what, my original suggestion? :) [01:06] alex-weej: The example above will display in the case that the embed is not run [01:07] yes but the

will trigger the install assistant anyway, right? [01:07] alex-weej: I guess so :) [01:07] even if it says "go to adobe.com" :( [01:07] don't know. try lots of things [01:07] apturl and collaboration from adobe may be the best long term strategy tbh [01:08] there is still some worth in not jumping the gun when the plugins array is looked ats [01:08] *at [01:12] ok, 3300-odd lines of dpkg diff left to deal with one way or another; bedtime [01:12] actually, more like 2800 when you discount the stuff at the top that I know I'm keeping [01:13] cjwatson: You're doing the dpkg merge? [01:13] yes [01:13] I mentioned it earlier [01:13] Ah, I must have missed it as I ran out the door. [01:13] you haven't started, have you? [01:14] Oh, dear god, no. [01:14] ok, good, I didn't fancy merging two merges [01:14] I'm surprised there's that much of a diff, though? [01:14] Triggers landed in Debian now, right? [01:15] That used to account for most of the merge. [01:16] packages.d.o confirms my suspicion. [01:16] they reindented it on the way in so it's taking a while to verify [01:16] I don't want to just assume it's the same, that no bug-fixes got lost, or that there's no upgrade handling to do [01:18] YEah, I've run into that a few times, too. What I've done a few times was to run the two files through "diff -w" and let that weed out the formatting differences. [01:19] I'm doing that kind of thing, sure [01:19] *shrug* I'm sure you'll work it out. [01:19] it's not all just whitespace though [01:19] * ScottK is glad to find out the whole triggers mess did not in fact cause a fork between Debian and Ubuntu in the package. [01:20] Well... It did for while. [01:22] ugh, merges ;) [01:25] was it something I said? :) [01:27] Heh [01:27] How is the merger of MoM and DaD coming along? [01:28] I have some patches [01:28] I'm still getting MoM back on her feet [01:29] 20:27 < ScottK> How is the merger of MoM and DaD coming along? [01:29] *grin* [01:29] you walked right into that one ;-) [01:30] jdong: Did you find out if today's LP upgrade includes fixing the breakage in Main backports? [01:30] ScottK: sorry, no time [01:30] I guess we'll know shortly. [01:32] cjwatson: as saifouch [01:32] ouch [01:39] kees: why is bug 223865 a compiz bug? looks like gnome-screensaver just isn't locking fast enough [01:39] Launchpad bug 223865 in compiz "On awaking from hibernation screen is shown before password is entered" [Low,Confirmed] https://launchpad.net/bugs/223865 [01:40] also, why is ubottu allowed to see security bugs? [01:40] Amaranth: it's not a private bug [01:41] I thought security meant private [01:41] Amaranth: no, security is security, private is private. :) [01:41] Amaranth: anyway, I haven't investigated it, but I can't reproduce it _without_ compiz [01:41] Can you reproduce it _with_ compiz? [01:42] g-s should be locked prior to hibernation. yeah, my hardy laptop flashes the screen for me too [01:42] I can only successfully hibernate when I get rid of the nvidia driver so I can't test hibernation with compiz [01:42] Amaranth: ah, I'm on ati [01:42] Does it happen when you use metacity compositing? [01:43] oh is that the bug that locking the screen isn't actually a blocking command? [01:43] no, that was fixed long ago [01:43] first with a hack in compiz, then properly in the X server [01:43] ah [01:48] I do hope that my little bzr push didn't take down LP ;) [03:22] Keybuk: oh, so you're the one to blame! [03:23] * ion_ sings "Blame Keybuk" to the melody of "Blame Canada" [03:29] A curse on projects with no officially known vcs repo, no bug tracker, and no one place to find patches. It is not fun jumping from distro to distro just to get patches for a package. [03:29] TheMuso: sadly that means you just cursed most of base ;) [03:30] Haha [03:30] Keybuk: I only had one package in mind when I said that, but I'll take your word for it. [03:30] TheMuso: which? [03:30] dmraid [03:30] eww ;) [03:31] e.g. where's the fsck home page? :) [03:31] Yeah well we have Microsoft to thank for that mess. [03:31] In not allowing software RAID 1-5 to be used in consumer OSs. [03:32] I could have sworn e2fsprogs has a home page [03:32] StevenK: you mean some antique page on gnu.org or similar? [03:32] Ah, I was thinking of fsck.ext[23], not fsck itself [03:51] How do I report a bug against the ubuntu release notes? [03:51] It's a major omission, that hda>sda from dapper to hardy. [03:54] slangasek: ^ [03:56] hmm, they enabled libata by default? [03:56] * bd_ has been on custom kernels precisely to use libata for a while now [03:56] bd_: Libata has been enabled since edgy. [03:57] TheMuso: really? I could've sworn it was using legacy IDE for my cd-rom not long ago... [03:57] bd_: I think it depends on the chipset. [03:57] possibly [03:58] I mean, my sata drive was on libata ofc [04:03] so.. [04:10] ffm: https://bugs.launchpad.net/ubuntu/+filebug choose "I don't know" for the package [04:10] bd_: I'm triaging, already filed. [04:11] ah [04:11] I would think a bug report would not be the best place to do that [04:14] where else would it be reported? [04:14] ubuntu-devel mailing list would be the most appropriate I would think [04:14] LaserJock: You can report bugs in ubuntu-website [04:14] ffm: I'm fully aware of that [04:15] but the release notes are not written by the ubuntu-website people [04:15] LaserJock: Unfortunately I'm not at a full session right now, and email is down. [04:15] LaserJock: Who writes thems? [04:15] then [04:15] them [04:15] the release manager [04:15] LaserJock: Who is... [04:15] slangasek [04:15] which is why Hobbsee pinged him [04:15] (I think ubuntu-devel-discuss@ is more appropriate, actually.) [04:16] crimsun: yeah [04:16] Do I have to subscribe to post to the lists? [04:16] I thought launchpad was down... [04:16] I'm already swamped fronm wiki-en-l. [04:16] Oh it should be back up by now [04:16] ffm: no, the post will just be moderated [04:21] Ok, any of you people here mods that can get me through the posting queue quickly? [04:23] (Thanks for all the help, btw) [04:29] Gah. My city is flooded :( [04:30] This is a picture of downtown: http://photos-b.ak.facebook.com/photos-ak-sf2p/v238/99/87/556140141/n556140141_2901289_3517.jpg [04:30] I wonder how I'm going to get to work in the morning. [04:31] cody-somerville: Swim [04:31] cody-somerville: :( [04:32] we're having an earthquake swarm here [04:32] LaserJock: it gives you something to talk about, at least :) [04:32] ;-) === herb changed the topic of #ubuntu-devel to: Intrepid coming soon | Ubuntu 8.04 LTS released! | Development of Ubuntu (not support, not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/feisty/gutsy/hardy, #ubuntu+1 for intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs [04:33] I dunno if you can see those facebook pictures but here is another startling one: http://photos-g.ak.facebook.com/photos-ak-sf2p/v238/99/87/556140141/n556140141_2901238_9117.jpg [04:34] holy cow [04:34] Here is a funny one: http://photos-f.ak.facebook.com/photos-ak-sf2p/v238/99/87/556140141/n556140141_2901205_7289.jpg [04:35] intrepid coming soon? [04:35] :) [04:47] jdong: Colin shortened it earlier to make sure all the needed bits would fit into the topic., === toresbe_ is now known as toresbe [05:49] Is Intrepid open for upload? [05:54] Not yet. [05:54] I'm seeing Intrepid packages in the Ubuntu package search... [05:54] RyanPrior: yes, it's currently being prepared [05:54] Alright, cool. [05:54] right now it's mostly the same as hardy, but several core toolchain packages are being updated [05:55] once they're in, it'll be opened up for universe sync and normal uploads [05:55] Exciting. :-) === Martinp24 is now known as Martinp23 === hunger_t is now known as hunger === sabdf1 is now known as sabdfl === rZr is now known as rZZZr [12:20] why does apt-get upgrade wait for a friggin minute or two before understanding "this host is down, lets try another" ?? [12:21] this is very annyoing.. [12:22] standard TCP timeout I imagine [12:22] you don't need to add colourful language either [12:24] cjwatson: how anamic [12:24] referring to what? (I assume you meant anaemic) [12:24] s/anamic/anaemic/ [12:25] in some environments it does take quite a while to make a TCP connection, you know [12:25] cjwatson: care to provide an example? [12:26] so timeouts are always a trade-off between speed of detecting when a host is really down, and coping with slow links [12:26] err, very heavily loaded networks of course :) [12:26] don't be too colorful and add smileys when you speak, you might raise the channel temperature (laugh) [12:27] if you object to the timeouts in TCP, you can address your complaints to its authors; it doesn't really make sense to fiddle it application-by-application === davmor2 is now known as davmor2_lunch [12:30] don't know if there's somewhere else that would be better for this, but doesn't seem to be anyone on #ubuntu able to help. security.ubuntu.com, although pinging ok, isn't responding to http [12:32] knewt: I *suspect* it's just very slow, but at any rate #canonical-sysadmin would be the best place to enquire about sysadmin issues [12:32] cjwatson: a better way to deal with that is to do multipe runs. eg. you try all hosts for archive.ubuntu.com and wait only 2 seconds between each. then next round you increase that delay drastically, to allow for slower links to get a chance.. see? it's not a good idea in general to shove responsability over to someone else saying "this is a fact of TCP" etc. i'm sure it could be handled if one bothered. [12:32] cjwatson: ta [12:33] squid_: which would substantially *increase* general network load and seems rather anti-social [12:33] it's sort of the equivalent of a child bouncing up and down in the back seat shouting "are we there yet?" [12:34] cjwatson: perhaps so, but would allow for better usability for a lot of people on fast lines [12:34] cjwatson: you are most colorful. [12:34] you're welcome to select a different mirror which is less heavily loaded [12:35] i don't know how, and i'm somewhat familiar to ubuntu [12:35] expecting that from your users is not a good solution [12:35] System -> Administration -> Software Sources [12:35] also that info is not given while i'm waiting for a connection. [12:36] it might allow for better usability if it just did it for you, but I think you'll find that if it did it for everyone it would make matters worse rather than better [12:36] you seem to not listen and rather provide quick fixes - i may be talking to myself. anyways i have things to do [12:36] what a nice chap [13:07] when i insert a cd/dvd in a second dvd drive, the system don't mount it and in dmesg, i've got this: http://pastebin.com/m1cc3d71c === davmor2_lunch is now known as davmor2 === fta_ is now known as fta [13:46] * lamont wonders why his ctl and shift keys magically quit working under X [13:46] lamont: oh, you're still getting that? [13:47] well, not sure when the previous reboot was... [13:48] Sun Apr 27 19:20 [13:48] hmmm [13:48] my version of that got fixed. [13:48] mind you, it actually was working inside of the windoze client in vmware [13:48] and ctl-alt-backspace provided no love, since it included ctl [13:49] my motherboard is making beeps like space invaders running on the livecd [13:52] lamont: hit hte logout icon. [13:52] yeah - that's how I eventually escaped... [13:52] and it keeps logging out of the gui like ctrl+alt+backspace or something [13:52] since ctl-alt-f1 suffered a very similary uh, issue [13:53] hwilde: neat [13:55] yes it would be interesting if my current development did not depend on this functioning normally [13:56] I didn't even know the onboard buzzer could make such an array of beeps [14:02] hwilde: that's happening on one of my systems, but under general load as well as just on the live CD; I blame a loose connection somewhere but haven't investigated [14:03] cjwatson, in the terminal it is registering keypress charcters ^[[C^[[C^[[D [14:03] but I unplugged the kb and mouse and it continued [14:04] soren: do you happen to remember what was actually wrong with dpkg when you fixed it like this? [14:04] * Fix wrong call to open in Dpkg/Control.pm that makes using a different [14:04] control file than debian/control fail horribly. [14:04] soren: your change seems to have been: [14:04] - open(CDATA, "<", $file) || syserr(_g("cannot read %s"), $file); [14:04] + open(CDATA, "< $file") || syserr(_g("cannot read %s"), $file); [14:04] soren: but three-arg open should work perfectly well, so I don't quite understand [14:08] cjwatson: that was as dpkg broke create-dbg-syms [14:08] geser: I'm aware of that, but *why* did that break? As far as I know those two lines are equivalent [14:09] iirc create-dbg-syms used stdin as $file [14:09] ah [14:10] and the three-arg-syntax didn't parse - as stdin but as a filename [14:10] so the changelog is misleading [14:12] cjwatson, do you have a serial or usb tty connected to that machine? try disabling brltty. the beeps do not occur until that process loads on startup [14:12] hwilde: no, I don't [14:13] cjwatson, "detected unknown HandyTech model with ID 28" says brltty [14:13] it's not anywhere near that predictable, it seems to be heat-related [14:13] sounds like we have wildly different problems [14:13] cjwatson, brltty is grabbing one of my serial->usb inputs on the tty and trying to read its input like it's abraille touchpad [14:37] * lamont wonders where the 'smake' package went in hardy? [14:37] are you missing it? [14:37] lamont: Licensing issues with upstream? [14:38] broonie: I'd have expected it to land in multiverse? [14:38] lamont: Debian dropped it as well [14:38] Would it not get automagically kicked when it vanished from Debian? [14:39] cody-somerville: Head for higher ground :) [14:40] * cody-somerville wades towards Amaranth riding a wild mountain goat. [14:42] Hobbsee: and my USB thumb drive still doesn't automount, since consolekit has decided that I don't need any sessions [14:55] broonie: ish [14:55] semi-automagically anyway === evalles_ is now known as effie_jayx [14:58] * ogra wonders why he has a ubuntu package search item in his FF searchbar but no LP bugs item :/ [15:46] hi there [15:46] I've a problem, I'm trying to backport a package from debian lenny to ubuntu 6.06LTS [15:47] So I take all the .dsc .diff.gz and .orig.tar.gz [15:47] dpkg-source -x package.dsc [15:47] dpkg-buildpackage and dpkg -i package [15:47] everything works perfectly untill the package installatio [15:48] what's the error? [15:48] when I install the pacakge I get this error : Template #1 in /var/lib/dpkg/info/mimedefang.templates does not contain a 'Template:' line [15:48] that's probably slight changes in po-debconf [15:48] It apparently because the file has 4 blank lines before to begin [15:48] let me just have a quick look and confirm [15:48] when I remove the lines, it works [15:49] johmathe: right, to backport it, after dpkg-source -x, edit debian/templates and remove the comment block and the subsequent blank line from the start of the file [15:49] then build [15:49] ok nice [15:49] * johmathe tries [15:50] debconf/po-debconf (I forget exactly where the change was, sorry) have got a bit more liberal since 6.06 and people have started taking advantage of that fairly widely [15:52] ok it appears to works [15:52] cjwatson: thanx a lot [15:52] you're welcome [15:57] Keybuk: can MoM be taught to look at experimental for certain nominated packages? [15:58] it can [16:01] Keybuk: can we do it for dpkg (once there's a dpkg in experimental, but there will be)? [16:04] yes [16:04] tell me when I need to add it [16:06] ok, will do, thanks === kitterma is now known as ScottK2 [16:26] !help [16:26] I am ubottu, the all-knowing infobot, standing in for ubotu while he's getting his haircut done, nose pwodered, updated and transitioned to his new gorgeus looks in the near future ;) [16:26] !ask [16:26] Please don't ask to ask a question, ask the question (all on ONE line, so others can read and follow it easily). If anyone knows the answer they will most likely answer. :-) [16:26] there we go [16:27] !HALP [16:27] I am ubottu, the all-knowing infobot, standing in for ubotu while he's getting his haircut done, nose pwodered, updated and transitioned to his new gorgeus looks in the near future ;) [16:27] lol [16:29] cjwatson, is vfat supported automatically in the install cd when detecting partitions? [16:29] hwilde: should be, yes [16:33] hi... I am trying to install ubuntu from a USB stick, but it's a tough ride... [16:34] https://help.ubuntu.com/community/Installation/FromUSBStick fails entirely, at least with 8.04. the boot image is not even found by syslinux, it seems to try to put full paths into 8.3 filenames, resuling in a "install./vm" instead of "install/vmlinuz" [16:34] got past that, but more problems later on. gave up. [16:34] now I try to use https://help.ubuntu.com/8.04/installation-guide/i386/boot-usb-files.html [16:35] but the ISO file copied to the USB stick (FAT), as instructed there, is not found. [16:35] ... because the installer does not have automatic vfat support [16:35] when I got into a shell, and did mkdir /source; mount /dev/sda1 /source/, it says "Invalid argument". same with -t vfat. when I do modprobe vfat first, it works. I'm told that's a bug? [16:36] hwilde: *blink* more details? [16:36] there BenB finally got around to it :) [16:36] oh, er, you two are working together [16:36] even then, the installer does not find and use the iso file. I have no options in the installer menu to make it use that, only options about mirror servers. [16:36] install wouldn't work until he modprobed vfat [16:36] ok, I'm on the phone, can you please stick around and I'll get back to you? [16:36] cjwatson: sure [16:37] * cjwatson notes that iso-scan does at least try modprobe vfat [16:37] I'd be really nice to have ready-made USB stick images in addition to the ISO images, given that USB sticks are very common these days and appropriate for initial install. [16:38] * BenB will try iso-scan on the shell whiel cjwatson is on the phone [16:38] erm [16:38] iso-scan isn't a program you can run easily unless you're familiar with d-i, so don't do that :) [16:38] I'm just thinking out loud [16:39] and whatever you do, don't call it gparted :) [16:39] yeah, noticed now :) [16:40] hwilde: right. I'm not a friend of anything that parts my harddrisks. [16:41] tkamppeter! [16:41] ;p [16:41] you scared him off [16:41] definitely [16:54] BenB: ok, so firstly the quick answer: I definitely agree that we should have an easier way to get USB images. I don't think it can really be ready-made images, because that means we essentially double the number of 700MB blobs we're shipping around, which would be difficult from an admin point of view, but I want to ship an application for Linux and Windows that does the conversion in one step [16:55] BenB: secondly, boot-usb-files is unfortunately not all that clear, so let me check what you're doing. I assume that you're following the "flexible way" approach (or you wouldn't be copying in an ISO file). It unfortunately doesn't link to the proper places to retrieve vmlinuz and initrd.gz; where did you fetch those from? [16:56] how about a java applet that you click on that takes over your windows machine and installs ubuntu? :) [16:56] hwilde: haha [16:56] cjwatson: I did: [16:57] 1. zcat boot.img.gz > /dev/sdb1; mounted it and noticed that the filesystem is only 16 MB large. so, next try: [16:58] 2. gunzip boot.img; mount -o loop boot.img /mnt/1/; mkfs.vfat /dev/sdb1; mount /dev/sdb1 /mnt/2/; cp -a /mnt/1/ /mnt/2/ [16:58] (1. would actually have worked, it'd just have retrieved *everything* over the network) [16:58] cp ubuntu-...804....iso /mnt/2/; [16:58] which boot.img? [16:59] http://archive.ubuntu.com/ubuntu/dists/hardy/main/installer-i386/current//images/netboot/boot.img.gz [16:59] (linked on https://help.ubuntu.com/8.04/installation-guide/i386/boot-usb-files.html ) [16:59] ok, you have to use hd-media in order to add an .iso image [16:59] I apologise that the documentation is not clear here [17:00] you either use netboot and no .iso image (in which case it retrieves everything over the network) or hd-media and an .iso image (in which case it uses it) [17:00] fine with me, as long as the next person won't run into the same problem :) [17:00] ah, clear [17:00] I'll update the documentation, though I'd appreciate it if you can try this out to confirm that there's nothing else I've missed [17:00] of course [17:01] though it would help me if you could file a bug on https://bugs.launchpad.net/ubuntu/+source/installation-guide/+filebug about it so that I have a handy justification for the stable update [17:01] ok [17:02] * BenB tried that, will take a while [17:02] tries [17:02] hmm, hd-media is a 256MB image [17:02] that's not especially convenient [17:02] * ogra thinks we should just have an option in gfxboot in the future: create USB installer from CD :) [17:02] you have to make a filesystem of the right size with mkdosfs and copy the files in [17:02] ogra: how were you planning to boot as far as gfxboot without a CD? :-) [17:03] cjwatson, you need a machine with CD indeed, but then it will just dump the squashfs onto an usb key and make it bootable :) [17:03] let's just leave it at "an application" for now. :-) [17:04] well, i guess we'll discuss it in detail in prague :) [17:04] * ogra is just brainstorming [17:05] ogra: I'd prefer something where I don't have to run any executables on the preparing machine. [17:05] (apart from stock ones like zcat) [17:05] I think that will be rather difficult for us unfortunately, although it could easily be set up such that somebody else can build it for you [17:06] ogra: requiring a CD-ROM is definitely a stop-gap. [17:06] cjwatson: that wouldn't help (security). but I guess a script that can be run on any random linux would be fine, too (preferably as user, not root). [17:07] BenB, i'm totally on your side here, i'd lobe to provide ready made dd'able images for usb keys :) but colin has a very valid point about size and space [17:07] *love [17:07] security> no different from downloading the image from us [17:07] I understand that most mirrors care more about disk space than traffic :) [17:07] assuming it were properly signed and such [17:07] yeah, if signed, sure [17:07] it's really just that we're almost on the edge of the amount of data we can cope with on releases.u.c right now [17:08] (signed by ubuntu) [17:08] and often in excess of that amount on cdimage.u.c [17:08] udnerstood [17:08] when I boot up it gets to grub, then it says "mount: only root can do that" over and over. then it says "mkdir cannot create directory /var/run/network: Read only filesystem" [17:09] then it boots up [17:09] the files appear to have the correct permissions and ownership... it's not a read-only filesystem... what am I missing [17:18] why does it tell me read-only filesystem ://// [17:20] BenB: so, I ran mkdosfs and syslinux on my USB stick, mounted it, copied /dists/hardy/main/installer-i386/current/images/hd-media/vmlinuz and /dists/hardy/main/installer-i386/current/images/hd-media/initrd.gz from the archive onto it, created syslinux.cfg with the two lines mentioned in https://help.ubuntu.com/8.04/installation-guide/i386/boot-usb-files.html, copied ubuntu-8.04-alternate-i386.iso in, booted it, and it ... [17:20] ... works fine [17:20] BenB: (actually, I lied slightly there, the "USB stick" was really a loop-mounted file that I booted in kvm) [17:21] hwilde: perhaps the kernel encountered errors and remounted the root filesystem read-only as an emergency measure? check dmesg [17:22] hwilde: (note that /etc/mtab might still show it as read-write) [17:30] ah I guess that's it.... why does grub use UUID which can't be ported from one system to another instead of just /dev/hda1 ? [17:30]