=== Teiana is now known as Exaltia [02:35] i need some help about ...how to say... redistributing some modifications i've made to ubuntu [02:36] dunno how to do, and what i must do regardingthe gpl [07:52] morning [07:54] hi [10:07] ogra: Hmm in the power menu, it looks like the suspend icon is actually missing, [10:08] hrm [10:09] lool, tedg uploaded a change on monday or so i hope he didnt remove it in favour of fusa inclusion but have to wait for him to get up [10:09] So currently, booting -mobile will cause madwifi to be used, that works, then if I suspend + resume, ath5k gets used instead and doesn't manage to associate with my wpa [10:10] In fact, I get half of the madwifi drivers loaded and ath5k at the same tmie [10:11] yeah [10:16] you can still get the suspend item by hitting the power button though [10:16] hmm, i got the suspend/hibernate items here though [10:17] njpatel: morning. [10:17] hey james_w [10:17] njpatel: there is another change in netbook-launcher, the addition of the quit window, I'd forgotten [10:17] that's why I didn't upload last night [10:18] james_w: Oh right, I forgot about that [10:18] james_w: Is that too much of a change to upload? (I guess it is) [10:18] so, we can still get a feature freeze exception if you think it is the right thing to do [10:18] having "This should work..." in the changelog entry might not help there though :-) [10:19] :-) [10:19] Hmm, I was waiting for reports on if it works or not, and it seems to [10:20] james_w: I still need to fix it for smaller screens so is it possible to wait until tonight? [10:20] and I can guarantee a better commit message ;-) [10:20] sure, tonight will be the last chance I think [10:21] and we should discuss it with ogra now anyway [10:21] the problem isn't so bad anymore, as ted has patched gnome-power-manager to pop up the shutdown dialog when you press the power button [10:21] i'm open for everthing as long as you can guarantee there are no regressions :) [10:21] cool, I'll try and get it done as soon as I get a few moments. I'm currently engulfed by oem-unr work [10:21] ogra: No regressions, promise :-) [10:22] I'll test it once you have it available === asac__ is now known as asac [11:01] ogra: I get the suspend item when I press power button, but it's icon looks like the "no icon" broke nicon [11:02] i get a black screen icon with a slightly lifted corner [11:02] (which is the suspend icong imho) [11:08] james_w, njpatel, please make sure to have proper bugs opened for the fixes [11:08] sure [11:08] ogra: will do [11:10] ogra: It's not the good icon, check on your ubuntu intrepid install on your desktop :) [11:10] ogra: I get the same icon at the bottom of "dates" which is the icon missing icon [11:11] weird [11:11] note that we dont use human icons in mobile [11:11] 90% of them are not scaling to 48px in the panel [11:12] (they all are hardcoded to 24) [11:12] if you switch your desktop to the gnome icon theme, do you get the same ? [11:13] (i do) [11:14] seems proper here ... do you run from usb key or is it installed ? [11:15] I run from USB key ATM [11:15] I have something installed, not sure what [11:15] Anyway, not enough time to look into this today, neither for you nor me [11:16] well, i see the proper gnome suspend icon on the installed image from yesterday [11:16] and my gpm menu works [11:17] i think the livecd scrips from casper suppress hibernate, but you should see suspend in yours [11:18] lool, you dont happen to talk about -mid, do you ? :) [11:30] No, -mobile [11:31] ogra: I see Suspend, it just has a black icon with grey border and one border cut icon, which is the missing icon icon I think [11:36] mumble ... if only eog wouldnt crash all the time [11:36] ogra: touchscreen doesn't work from live ubuntu intrepid i386 CD on Q1U [11:37] hal race ? [11:37] I have no idea [11:37] does mouse or keyboard work ? [11:37] USB ones, yes [11:37] the one one the Q1 as well [11:37] builtin i mean [11:38] hmm, wors fine here [11:38] mouse works fine here as well [11:38] and keyboard too [11:38] thats weird [11:39] do we have the same Q1 ? [11:39] :) [11:39] Perhaps a race [11:39] yeah [11:39] I'm booting from live CD plugged as USB drive [11:39] It worked from the daily -mobile USB image though [11:39] oh [11:39] well, evtouch isnt included in default ubuntu [11:39] its a universe package [11:39] That's probably why [11:39] :) [11:40] apt-get install it and restart hal and your xsession [11:40] * ogra curses ... eog definately only works the first time i start it after login [11:40] ogra: Shouldn't we aim at promoting evtouch and pulling it in input-all? [11:41] Late for intrepid, but perhaps for next cycle [11:41] no, the plan is to generalize all touchscreens on evdev upstream for next release [11:42] Ok; what's the issue with evdev on the Q1U? [11:42] evdev ? none [11:43] we currently have six touchscreen X drivers, upstream belives they could all be handled by evdev with some changes to evdev, so the others (like evtouch) can go away [11:44] hmm, you are right, its the missing image, i wonder why upstream doesnt properly set a suspend icon here [11:44] ogra: I suspect it might not be the pristine upstream menu [11:45] afaik that dialog is the original upstream gnome dialog, using the gnome iconset it should use a proper icon [11:45] oh [11:45] So it might rely on some ubuntu-ish icon [11:45] But ICBW [11:46] well, easily solved by providing a link from -settings if thats the case [11:46] hmm, no seb128 there ... he would know it i guess [11:46] Something to look at, but more urgent stuff today [11:46] yeah [11:47] ogra: Shall we set a time for reviewing the cdimage scripts situation tonight? [11:47] Or should I just wait for an update this afternoon? [11:47] sooo ... [11:47] njpatel: oh, also bug 277133, is that fix in these packages, is it supposed to be? [11:47] wget ftp://archive.ubuntu.com/ubuntu/pool/main/l/linux-backports-modules-2.6.27/linux-backports-modules-*-generic_*_i386.deb [11:47] Launchpad bug 277133 in netbook-remix "ume-launcher application name wordwrap issue" [High,Confirmed] https://launchpad.net/bugs/277133 [11:47] does that look sane ? [11:48] its quite ugly, but i dont see a awy to find out the used kernel version *inside* the image [11:48] ftp? [11:48] since we only use mobile-kernel-generic in the build system and vmlinuz in the image [11:48] http doesnt support wildcards [11:48] I see; wget is da suck [11:49] the naming in the image has no versions [11:49] thats not wget's fault [11:49] james_w: the notification-window word-wrap issue needs to be ported from ume-launcher still afaik. I'll add it to the todo [11:49] the protocol doesnt support wildcrs [11:49] njpatel: cool, it's not critical. [11:49] *wildcards [11:49] ogra: This might result in fetching multiple packages when we have some intrepid-updates ones in the archive [11:49] lool: re madwifi stuff. problem with wpasupp + wext + madwifi driver is that it doesnt work for all cards. however it worked for some i nhardy and maybe driver has improved so it works for even more now [11:50] lool, well, thats a special corner case ... we *should* always only have one lbm for 2.6.27 ...this case might show up only between two publisher runs [11:50] lool: not sure how good mobile can tweak this. for the "general" distro we dont have enough data to decide [11:51] asac: My point really boils down to that the fact that the madwifi backend was removed from wpasupplicant doesn't prevent madwifi + wpasupplicant to work at all [11:51] ekk, no, indeed, -updates uses the same pool [11:51] ogra: Anyway, this might as well be a feature :) [11:51] well, makes the howto more complex [11:52] and might make the image explode at some point, StevenK only leaves 50M spare space by default iirc [11:52] lool: yes. but that is known. it is just that there are cards for which wext didnt work in hardy [11:52] though its unlikely that we build images after release i guess [11:52] at least for non LTS ... [11:52] (and before) ... and the madwifi plugin for wpasupp was the only way to get all working [11:52] It's still 20M [11:53] and in jaunty we should have a pool directory inside the image [11:53] ogra: But I thought some images are already pulling some .debs in there; can't we do the same? [11:53] liveCD does [11:53] but that change would be a lot bigger [11:53] lool: Not easily [11:53] my current ide only involves 5 lines [11:53] *idea [11:53] we dont have pbm for lpia btw [11:53] *lbm [11:54] so it will only help the mobile image atm [11:54] * ogra goes to make some coffee [11:55] asac: There are so many claims from so many people concerning so many different hardware with madwifi and ath5k and wpasupplicant wext versus madwifi and various versions of kernel and ath5k and madwifi that I can't possibly feel confident to assert anything :) [11:55] What I can say is latest version of what we have in ubuntu works on this hardware that I just tested it on [12:01] lool: yeah. lucky you ;). my advice for kernel team was: get a list of cards that dont work through community testing and blacklist them for ath5k so they fall back to ath_pci (which hopefully works)) [12:03] Exactly what I suggested... [12:03] and which I'd still consider sane [12:16] but the list is missing :P [12:54] So one has to modprobe -r ath_* before suspend and load them after resume, then wifi works again [12:54] ath5k is just a red herring, it's always loaded but seems ineffective; didn't try to keep it too hard though [12:55] Now trying the reverse [13:01] arf replacing blub blub [13:01] +with [13:08] ogra: It's not trivial to use the lbm ath5k [13:09] you need to modprobe -r ath5k, mac80211, and cfg80211 as they conflict with the lbm versions [13:09] for me just installin lbm in the installed image worked [13:09] mac80211 and cfg80211 are replaced by the lbm modules [13:10] did you try that on the live image ? [13:10] or in an installed system [13:13] ogra: I tried that on installed system, but I had to modprobe -r manually; or reboot [13:13] yes, you need to reboot [13:14] lbm should set the reboot notification iirc it regenerates the initramfs in postinst [13:14] ogra: So you don't get wifi during install if you offer lbm on live images [13:14] no [13:14] unless you manually remove modules like I did [13:14] thats not why i want lbm o the image [13:14] *on === davmor2 is now known as davmor2_lunch [13:15] Ok; you just want it for installed systems [13:15] Upon installation, people can install lbm [13:15] if you installed and have no wired connection (i.e. no NIC) you dont have any opporunity to make it work [13:16] so i want the package available for them by plugging in the usb key after the first reboot and be able to gdebi install it [13:16] if lbm doesnt show the reboot notification, thats a bug imhp [13:16] *imho [13:18] I updated bug #284354 with in depth testing [13:18] Launchpad bug 284354 in linux-lpia "AR2424 on Samsung Q1 loads both ath_pci and ath5k modules" [High,Triaged] https://launchpad.net/bugs/284354 [13:18] ogra: What I'd prefer is networking working straight from the live session, e.g. to install langpacks or additional packages [13:18] right [13:18] Anyway, I finished investigations [13:19] that will make the howto bigger indeed, but is documentable [13:19] The best way out of this mess is probably to fix ath5k without swallowing the 80211 ABI changes [13:19] ogra: We can offer a script along the .debs? [13:19] enable-lbm-ath5k [13:19] which does install, unload, reload [13:20] It can unload madwifi and ath5k [13:20] Didn't try this out BTW [13:20] likely, let me finish my build-mobile-img changes first, cjwatson wants me to use grep-dctrl for which i have to find the right runes [13:20] Only tried between old and new ath5k [13:20] ogra: I think using grep-dctrl was the most sensible things to do :) [13:20] yes, but i rarely use it :) [13:21] ogra: wget -O - http://../Packages | grep-dctrl -F Source -X -s Package [13:21] Or rather grep-dctrl -S -X -s Package [13:22] and -n to remove Package: from the output [13:22] i need Filename [13:22] Oh right, then -s Filename :) [13:22] but your pattern already is better than what i had yet [13:22] gets me -server though [13:22] and headers [13:22] That's just a template for matching this source [13:23] You can also match package name [13:23] or match only that [13:23] right, but package name includes versioning [13:23] so thats ut of discussion [13:23] ogra: You can use regexps [13:23] grep-dctrl -F Source -X linux-backports-modules-2.6.27 -s Filename|grep -v server|grep -v headers [13:24] that seems to get me the right think [13:24] *thing [13:26] ogra: grep-aptavail -n -s Filename -X -S linux-backports-modules-2.6.27 -a -F Package -e '^linux-backports-modules-2.6.27-[0-9]+-generic$' [13:26] or if you don't want to check source, just grep-aptavail -n -s Filename -F Package -e '^linux-backports-modules-2.6.27-[0-9]+-generic$' [13:27] (grep-aptavail is just for local testing) [13:27] ogra: urgh grep-ctrl | grep -v | grep -v [13:27] yeah [13:27] I just had lunch === davmor2_lunch is now known as davmor2 [14:02] ogra, hi. do you need the gconf paths for the compiz configuration? [14:02] oh, yeah, that would rock ! [14:09] lool, should we link to https://wiki.ubuntu.com/MobileTeam/Mobile/HowTo/ImageWriting from the release.u.c index files ? [14:11] set /apps/compiz/plugins/move/allscreens/options/constrain_y = false and you should move the the window where eer you want [14:12] cool, thanks, i'll put that in after RC [14:16] ogra: Yeah [14:16] good, i'll try to get that in then [14:17] waiting for cjwatson feedback on the final fix for build-mobile-img first , then i'll move on to the publishing scripts [14:19] anyone tested if the gsm modem could be used for calls? [14:19] nope [14:19] we had some more serious probs with wlan which i'm currently trying to add a workaround for [14:20] ipw2200? [14:20] the gsm isnt really in focus atm [14:20] we'll get to that in jaunty [14:20] no, madwifi vs ath5k modules [14:21] * ogra points to bug 284354 [14:21] Launchpad bug 284354 in linux-lpia "AR2424 on Samsung Q1 loads both ath_pci and ath5k modules" [High,Triaged] https://launchpad.net/bugs/284354 [14:46] ogra, imho snapping windows are anoying too on small displays. just a experience from my eee. i saw somw problems in your talk. [14:46] sanpping ? [14:46] *snapping [14:52] the windows glued to the desktop borders and "snap" when moved away [14:52] ah [14:52] they are actually maximized by devilspie [14:53] on small display the might snap on/off all the time [14:53] * ogra would prefer to use maximus, but thats sadly only working with metacity atm [14:53] just saw it on some dialogs on your q1 [14:53] right, that happens with apps that dont have the right minimal window size [14:54] on the TODO list for jaunty :) [14:59] * ogra rsyncs what hopefully will be the RC of ubuntu-mobile [16:48] hi just browsed through the ubuntu mobile page [20:05] hi there [20:06] the ath5k wifi driver in ubuntu mobiles are the madwifi's one isn't it ? [20:09] Sometimes. [20:09] hu? :) [20:09] There's a bug that some devices need the ath5k driver, and some devices need the madwifi driver. [20:10] ha, ok, so, the ath5k driver doesn't come from the madwifi drivers suite ? [20:10] This is true in one case where the device uses the same identifier. As a result, depending on your specific hardware, you may have to force either ath5k or madwifi. [20:10] No, it's the one from the stock kernel tree. The madwifi version is also available. [20:11] I believe that most of this issue has been worked around in the latest images, but depending on your hardware, you may have to do some configuration. [20:11] no i understand why i see, despite the madwifi mention of "we don't do roaming until complete signal loss" of roaming :) [20:12] now* [20:14] time to search how to desactivate it, as after a couple of crazy roamings beetween 2 routers, the card purely desactivate [20:15] me kicks himself for filing a bug about loss of wifi when the hardware kill switch was on [20:16] ian_brasil, Well, now you get to triage the bug :) [20:16] * Exaltia shall look, do you have an url ? :) [20:16] URL for? [20:17] ian brasil's bug :) [20:18] Exaltia, It only needs to get invalidated as the cause is known and it's not a software bug. Doesn't need more eyes. [20:18] okay :) [20:18] * Exaltia blindes herself ;) [20:19] blinds* [20:29] i just quit somewhere [20:31] ian_brasil_, Hrm? [20:37] persia: i must have been logged into IRC elsewhere and forgotten so I was rather shocked to see myself quit [20:38] Oh. I see. Is your nick registered with freenode? If so, you might find /msg nickserv ghost a useful command. [20:39] nice command..it looks like my home network is down [20:43] Usually one doesn't notice the network going down a few minutes *before* losing the connection :) [20:44] me polishes third eye [20:48] heh [21:38] * popey hugs ogra [21:38] :) [21:38] heya [21:38] just trying out ubuntu mobile on my eee, to take the the linux expo and show off tomorrow [21:39] cool [21:39] i thought intrepid was supposed to natively support the wifi in this thing (with linux-restricted-modules) but it seems not to - need more digging [21:39] did you get the most recent image ? [21:40] yes [21:40] see bug 284354 [21:40] Launchpad bug 284354 in linux-lpia "AR2424 on Samsung Q1 loads both ath_pci and ath5k modules" [High,Triaged] https://launchpad.net/bugs/284354 [21:40] specifically the last comment i made [21:40] looks good [21:43] ogra: gdebi is telling me that "you likely do not want to install this package directly", i can ignore that can't I? this supplements existing modules? [21:43] right, it replaces ath5k [21:43] might or might not work on eee [21:43] well i do have the problem on this that both ath5k and the ath_pci driver loaded at boot [21:43] but the one in l-b-m is definately better than the one shipped by default [21:44] so it certainly _sounds_ plausible :) [21:44] we have an eeepc running mobile here [21:44] we set the wifi up using https://help.ubuntu.com/community/EeePC/Fixes [21:44] ian_brasil, they differ from model to model [21:45] the 701 and 900 are the same from a wifi point of view [21:45] its only from the 901 up they changed it [21:45] its a 900 [21:45] so is mine [21:45] right, according to james_w either a 700 or 900 should work [21:45] he tested the mobile image recently and reported that at least [21:45] yay [21:46] that fixed it [21:46] great [21:46] * popey hugs ogra again [21:46] * popey comments on the bug [21:46] i'll write a proper howto tomorrow [21:46] currently i'm up to my ears in the release publishing scripts, -mobile and -mid need to become official releases with RC [21:47] back from blackout ... === Teiana is now known as Exaltia [22:24] ogra: hey, this is raw netbook-launcher diff: http://paste.ubuntu.com/61219/ [22:24] ogra: it's pretty big due to autotools changes [22:25] ogra: are you happy with that, would you like me to make any more changes before upload? [22:25] two simple enough bug fixes: bug 283374 [22:25] Launchpad bug 283374 in netbook-remix-launcher "Shrinks desktop background" [Undecided,Fix committed] https://launchpad.net/bugs/283374 [22:26] bug 277133 [22:26] Launchpad bug 277133 in netbook-remix "ume-launcher application name wordwrap issue" [High,Confirmed] https://launchpad.net/bugs/277133 [22:26] and a quit dialog, bug 275644 [22:26] Launchpad bug 275644 in netbook-remix "With Intrepid, Quit icons don't allow to turn off the PC anymore" [High,Fix committed] https://launchpad.net/bugs/275644 [22:26] james_w, urgh, 62000 lines ? [22:27] yeah :-/ [22:27] how did you convince the pastebin to not explode ? [22:27] diffstat http://paste.ubuntu.com/61222/ [22:27] james_w, This is for post-RC, right? [22:27] persia: for the -mobile image? [22:28] james_w, Yes. [22:28] yeah [22:28] yeah, sorry, I'm not used to working on CD packages [22:28] I've only been able to register one valid test so far, as the images keep getting respun, and the deadline is looming, so I fear anything that changes things. [22:28] I was just thinking about final freeze [22:28] ok, if its for post RC it needs to wait, we're all busy with RC atm [22:28] Milestone freeze trumps Final Freeze. [22:28] Come back Friday :) [22:29] sure, I can wait [22:29] * ogra needs to get the publish scripts on cdimage working [22:29] * persia needs to get around to writing the script to identify packages that break images [22:30] anything I could lend a hand with? [22:31] james_w, Actually, sure. Basically, there should be a script that takes a package name as an argument, and performs a comparison to report whether it would be part of an image build, and if so, for which flavours. [22:32] Preferably this would happen in a way that didn't require either a germinate run or full install of the packages concerned. [22:32] hmm, no idea where to get that information from, care to give me a pointer? [22:32] I suspect that setting up something that pulled the relevant tasks to identify the packages would be best. [22:33] apt-cache show has Task: when they are part of a task, is that sufficient? [22:36] james_w, Maybe. Let me try to find an exception. [22:36] also, is there germinate output stored online? [22:36] Maybe under p.u.c/~u-a/ somewhere. I forget [22:37] Hmm... [22:39] james_w, Seems like apt-cache's Task: works. [22:39] http://people.ubuntu.com/~ubuntu-archive/germinate-output/ubuntu.intrepid/all [22:40] so you want list of packages->task->image? [22:41] No need to show the task. The important part is to create a tool that allows developers to know what is safe to upload during a milestone freeze, and allow the release managers to know what is safe to accept without querying a flavour (and which flavour to query) during final-freeze. [22:42] yeah, I meant logically, but I guess a developer can work it out from the task name. [22:44] hmm, but developers think in source package terms [22:44] no trouble [22:45] Right. It's the source-package-name -> image mapping that is tricky for developers. [23:07] persia: first stab: http://people.ubuntu.com/~jamesw/which-images.py [23:08] the mapping from task -> image is completely incomplete, as I don't have either an exhaustive list of tasks, or a knowledge of which tasks ends up on which CD [23:09] http://paste.ubuntu.com/61244/ [23:09] see :-) [23:12] which tasks end up on which CD is in a combination of debian-cd and livecd-rootfs. [23:12] I'm not sure how to get an exhaustive list of available tasks. [23:13] I'll get one by brute force [23:14] james_w, Hrm. Found another one : jackd is on the Ubuntu Studio CD, but doesn't show a task for some reason. Might be a bug in ubuntustudio seeds though. [23:19] hey, just a bit of oftopic chat : Cheers and good luck for the future update ! [23:19] \o/ \\o o// o/ \o [23:19] james_w, Actually, I've not yet found a package for which I don't get that traceback. [23:22] persia: yeah, just fixing it up now [23:26] persia: better version available, still far from complete though [23:39] james_w, works for man-db, f-spot, human-theme. Fails for phonon, jackd. [23:39] Getting close :) [23:42] yeah, not sure what's going on with jackd [23:44] Yeah. phonon is just a mapping issue? [23:44] Also, from where are you pulling that mapping? Is it a separate source, or are you leveraging some other tool? [23:45] read the source :-) [23:46] if you know of a better way then I will happily implement it [23:46] KeyError means it's not been added to the mapping dict yet [23:47] I see. Hrm. Probably sensible to find a way to have a common place for this that would be sourced by debian-cd, livecd-rootfs, and which-image. [23:48] But that's deep architecture. Once this works, and a first draft is in ubuntu-dev-tools, let's hit up cjwatson to see how we might better integrate. [23:49] certainly [23:51] Also, seems the issue with jackd (or mythnews) is an LP bug, so may as well ignore that for now (although ignoring it makes it more complicated for some folk)