=== bjf is now known as bjf[afk]
=== bjf[afk] is now known as bjf
rsalveticooloney: hey!02:32
rsalveticooloney: any news about the mem issue bug?02:32
rsalvetirobclark: do you know if we're going to get the fb resize feature from the omap 4 x11 video driver?02:44
robclarkrsalveti: I guess the new pvr driver eventually.. although it would help to get DRM support into the kernel in a sane way for non PCI systems..02:46
cooloneyrsalveti: nothing new here.02:48
cooloneyrsalveti: i got the same result like you02:48
cooloney2G:2G split to disable highmem02:49
rsalvetirobclark: that's for sure..02:49
rsalvetirobclark: was thinking for maverick, how we're going to deal with people changing monitors, and even when we can't probe the correct edid when booting02:49
rsalvetithe case of my monitor02:49
rsalveticooloney: for me it's quite stable with 2G:2G, no highmem and using just one cpu, or no L202:50
cooloneyyeah, that's ture02:50
GrueMasterNeed at least a sane default.02:50
GrueMasterfor monitor resolutions.02:50
cooloneyrsalveti: but one cpu might not be acceptable, i think02:50
rsalvetiGrueMaster: for default there's a bug currently for monitors that turns off and on during probe (my case)02:50
GrueMasterI hit the same thing with my HDMI switchbox.02:51
rsalvetithat the screens turns 640x480 initially and then it changes to the correct one02:51
rsalvetiprobably the same bug02:51
rsalvetiI'm looking at it now..02:51
GrueMasterExcept mine stays at 640x480.02:51
robclarkyeah, probably same bug02:51
rsalveticooloney: probably not, even without L202:51
robclarkGrueMaster: pls send your edid.. /sys/devices/omapdss/display0/edid02:52
cooloneyL2 will have some impact to the performance02:52
rsalvetithat's the problem ...02:52
cooloneyrsalveti: so there is not good solution now,02:52
robclarkyeah, disabling L2$ or 1 cpu is not ideal.. but 768m has been relatively stable for me so far..02:52
rsalvetiGrueMaster: yeah, can you paste your edid when using the switchbox?02:52
cooloneyrsalveti: i also applied the patches in, and
cooloneyfrom stable tree02:53
cooloneygot the same result02:53
rsalvetiif we get a correct edid then it's just timing issues when probing the edid02:53
IIIhey all... is current build having any issues?02:53
rsalvetijust if you're using omap 402:53
GrueMasterErm, is there a tool for reading this?  It looks like garbage.02:53
* III is using omap402:53
rsalvetiGrueMaster: yep, binary02:53
rsalvetiIII: currently when using 1GB we get build issues when building the kernel02:54
rsalvetiwith highmem you can easily get it, without highmem you get it some times02:54
rsalvetiif we disable L2 or use just one cpu, everything works fine02:54
rsalvetianother workaround is to use 768MB (the one we're using now)02:54
* III moves to older build :D02:55
rsalvetiand if we don't fix it for the release, this will probably be the way to go02:55
rsalveticooloney: the problem is that it's quite hard to trace it, I'm out of ideas02:56
cooloneyrsalveti: me either, i tried to enable ftrace with dynamic tracing02:56
cooloneybut it took me lots of time to get some useless information02:57
rsalvetiit works fine with all mem stress software I use, but breaks when building the kernel02:57
rsalveticooloney: but were you able to get something?02:57
cooloneyrsalveti: i get function tracing information from the abort handler02:59
cooloneyand it changes every time before the abort handler02:59
rsalveticooloney: hm, can you paste it?02:59
cooloneyi think we need another simpler test case03:00
cooloneyrsalveti: np, man. let me find it. paste it to you guys03:00
rsalveticooloney: sure, but I tested a lot of mem stress software and unable to reproduce the issue03:00
robclarkGrueMaster: can you email me the binary?03:00
rsalvetimem stress with disk io and etc, and nothing03:00
cooloneyrsalveti: cool, what are those mem stress test cases?03:01
GrueMasterrobclark: working on it.03:01
robclarkk, thx03:01
rsalvetithe problem probably happens with mem + i/o + dma03:01
GrueMastersystem is being real sluggish.03:01
GrueMaster(actually, all networked systems are).03:02
cooloneyrsalveti: that combination is too complicated. sign03:04
* cooloney brb03:04
GrueMasterhttp://paste.ubuntu.com/502398 - text version.03:04
rsalveticooloney: maybe something at arch/arm/plat-omap/iodmm.c, uses l2...03:04
GrueMasterGotta run.  I'm being drug away.03:06
rsalvetiGrueMaster: the problem at bug 644714 is that the resolution is not changed at the x11 driver03:06
ubot2Launchpad bug 644714 in linux-ti-omap4 (Ubuntu) "Screen corruption waking from screen blank if no monitor present (affects: 1) (heat: 526)" [Undecided,New] https://launchpad.net/bugs/64471403:06
rsalvetithe fb is changed with latest code from robclark, but not x1103:07
* rsalveti out04:07
IIIdoes anyone know of a good working build I can use? I just tried  20100926 and 20100927 but I am not able to boot.. anyone else seen this issue?05:41
IIIsorry that is for OMAP405:41
persiaWhat issue do you see booting?05:48
IIIseems to hang on boot05:52
IIItake that back... seems to boot in to busybox05:54
persiaAh, good.  Not booting is complex.  Any messages before being dumped to a busybox prompt?05:57
IIIno init found... I'm trying again but with a new card.05:59
persiaI had that on an upgrade, and was certain it was PEBKAC (and it was omap3), but maybe it wasn't.05:59
persiaogra, So, it might be nice to have manifest files also for non-omap4 preinstalls :)06:02
persiaIII, Hrm.  Please let me know if it works/doesn't work with the new card.  Seems 20100927 has the latest upstart, sysvinit, initramfs-tools which are the bits that are immediately suspicious (but the changelogs don't show anything worrisome).06:07
* persia waits for rsync for 20100927/omap for a parallel test06:07
=== JaMa|Zzz is now known as JaMa
ograpersia, ??08:19
persiahttp://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/ : note the number of .manifest files08:20
ogradid NCommander kill acorn again with extenbsive dove builds ? :=08:20
persiaI thought you were the person who put them in, and don't happen to know which bit to frob for that08:20
ograpersia, manifers files are installed some point around bootloader install ;)08:21
persiaMy issue isn't the lack of ...0928 images (although that's frustrating), but the lack of a manifest for the images that are posted.08:21
persiaAnd bootloader is broken for omap?08:21
ograthe images without bootloader logically have no manifest (and no partition table either)08:21
ograwas broken08:21
ograwithout a livefs i cant tell08:21
ograi would expect that archive to have settled by now though08:22
persiaOn the 27th?  There *is* an omap livefs (or something using 509M compressed)08:22
ograwintout bootloader partition08:22
=== hrw|gone is now known as hrw
ograjust on a sidenote i have debian lenny running natively on the ac100 (with the existing kernly only somethin that old (and without update) works)08:23
persiaWould that also be why III couldn't boot?08:23
ograno, didnt affect omap408:24
persiaDunno then.08:24
ograi cant ifup the wlan card though08:24
ograsomething is blocking08:24
persiadriver issue?08:24
ograno, premission issue08:24
ogragod, i cant type tofay08:24
NCommanderogra: it was still running when it spit out hte last image08:25
NCommanderogra: I need a package merged :-)08:25
ograNCommander, i was making a bad joke :)08:26
NCommandercan you review the patch so I can get the feature freeze exception?08:26
ograNCommander, post RC ?08:26
NCommanderogra: bah, now I know this is reality. the real ogra doesn't joke :-/08:26
NCommanderogra: its critical enough that I want to at least run it by the release team08:26
NCommanderogra: if it flops, it flops, but I have to at least *try*08:26
* NCommander wants his ubiquity bug fixed :-/08:27
ograyou can start uibiquity from the terminal, right ?08:27
NCommanderogra: well, yes, but that's less than ideal08:27
* ogra wants his "rootfs fillt to 100% wint encrypted home" bug fixed ... but i wouldnt risk RC for that08:27
persiaogra, Did you make a patch for that yet?08:28
ograthats really something you can release note08:28
ograpersia, nope08:28
persiaIt's a trivial patch, if you want to test it.08:28
ograbut i expect to have one before end of the week08:28
persiaAh, OK.08:28
ograoh,you worked on it already ?08:29
persiaOnly in the sense of the IRC session a few days ago.08:29
ograsure i'll test :)08:29
ograah, k08:29
persiaBut it's < 15 lines of code :)08:29
ograNCommander, i'll review it (and nod it off if its ok) later today ... but i wouldnt push for getting it into RC (get it to the queue though)08:32
ograNCommander, the changelog disagrees with the change in your branch08:35
ograNCommander, scripts/casper/47une_ubiquity is missing completely08:35
NCommanderogra: argh, i readded tha tand push it08:35
NCommanderHold on08:35
* NCommander goes to get pants so I can turn the Dove on08:36
ograNCommander, also dont forget to check for executability (i cant see that in the LP UI) i dont remember which files need to be executable and which dont08:36
ograis your dove on the balcony ?08:36
NCommanderogra: no, I'm just living in a place that if I walk to the room where the board is without pants, someone will scream08:40
ograthe board is without pants too ?!?08:40
ogrageez !08:40
* NCommander decides to discontinue this line of conversation08:40
ogracheck the quotes page :P08:41
NCommanderdamn it08:42
NCommanderogra: re-pushing. bzr hates me08:42
NCommander(I miss git)08:42
ograbzr add isnt that hard :)08:43
NCommanderogra: I just forgot to bzr push08:44
NCommanderogra: its there and executable08:44
hrwNCommander: there is git-bzr-ng project on github to provide "git bzr" plugin08:44
hrwNCommander: I use it to manipulate bzr repos but so far failed to push with it08:44
ograNCommander, so why do you apt-get remove ?08:47
ograNCommander, i would bet it slows down booting a lot08:47
* ogra would just have removed the diversion instead of waiting for the package DB08:47
NCommanderogra: because it also removes the desktop file oem-config installs asking to finalization the installation08:48
NCommanderogra: well, I wanted to remove it during image building :-P08:48
ograNCommander, well, your call, butu i'D say it surely adds 10-20sec to the boot08:49
ograand ubiquity will remove it anyway08:49
NCommanderogra: didn't see that bad, and ubiquity doesn't remove it until AFTER installation08:49
NCommanderKinda confusing to have it there before you install08:49
ograwhats confusing about it ?08:50
NCommanderif you want a livecd-rootfs patch, I already have one of those, though its not quite as tested (I can fix that though now)08:50
ograthe user never sees it08:50
NCommanderogra: er, yes theydo?08:50
NCommanderIts right therein Settings.08:50
ograno, i dont want to tinker with livecd-rootfs08:50
ograNCommander, where do you test for the arch ?08:50
ograi dont see any code for that08:50
ogracurrently your patch breaks all oem installs08:51
bercogood morning all08:51
NCommanderogra: no it won't, because OEM installs don't install eom-config until after installation which doesn't use casper.08:51
ograthe action should be arch (or even subarch) specific08:52
ograNCommander, i dont think thats true08:52
bercoanyone knows if there's a particular issue with building webkit on arm?08:52
ograNCommander, did you talk to superm1 about that ? i bet you break all dell installations with that08:52
persiaberco, Shouldn't be.  What issue are you seeing?08:53
ogra(note i dont talk about the OSG team here but about oem's using modified images)08:53
bercopersia: It fails all the time08:53
bercopersia: wondering if not enough memory would be the issue08:53
NCommanderogra: you can't hav eoem-config in the same image as installed ubiquity with the later working properly since it will hide the installer icon08:54
ogradoesnt seem to fail for us ... http://qa.ubuntuwire.org/ftbfs/08:54
NCommanderogra: I *wanted* to remove in livecd-rootfs, you said remove it in casper08:54
ograNCommander, well, every thought why superm1 added that diversion ?08:54
ograNCommander, no, i said remove the diversion08:54
NCommanderogra: er, I think ev wrote that diversion08:54
NCommanderogra: and that's the wrong bloody solution08:54
ograwell, from casper, but you know what i mean08:54
NCommanderThe package shouldn't be there08:54
ograthe packages are there in some setups08:55
ograthere is an OEM who uses an automated preseeded ubiquity to create images with oem-install08:55
NCommanderogra: oem-config however is installed on fht elfy if its needed08:56
NCommanderThat's why its in ship08:56
ograanyway, make it arch specific and i'll approve08:56
NCommanderLooking at oem-config code, the postinst doesn't handle the case if the diversion is already there08:56
ograi dont care if you fiddle with the diversion ... but i do care if you make it a default for everyone08:57
* NCommander will sit down at do it a bit later then08:57
ograremove it or remove the diversion ... as you like08:57
ograbut make sure people explicitly installing oem-config dont get screwed by it08:58
* persia vaguely wonders why there are both "webkit" and "webkitkde" packages09:00
persiaberco, Seems to build on the buildds: https://launchpad.net/ubuntu/+source/webkit/1.2.4-1ubuntu1/+build/194888009:00
persiaMight be RAM, might be missing patch.  Might be different version.09:00
ogralooks like KDE needs an older version09:02
persiaThat's not a good reason, but potentially.09:03
ograNCommander, was a meta uploaded for the dove kernel ?09:04
NCommanderogra: no idea09:07
ograi though you'd care ... thats why i talked about it in the meeting yesterday09:07
ograthough if you dont ...09:07
* ogra shrugs09:07
bercopersia: I was looking at that page too :). Argh!09:08
ograwill probably break d-i though09:08
ograon dove09:08
persiaberco, My recommendation would be to compare your build log to the build log on LP and find the point at which they differ: this will probably be a hugely valuable hint to getting the build to work.09:09
bercopersia: do you know the board hubbard?09:09
persiaNever heard of it.09:09
bercoarmel package was built on this board09:09
bercoon Launchpad09:10
persiaOh, you mean https://launchpad.net/builders/hubbard ?09:10
bercoyes sorry09:10
persiaI believe that's a Freescale Babbage 3.0 board, but I could be mistaken.  Should be ~800MHz, ~512MB unless someone did something special compared with other implementations of that SoC I've seen.09:11
ograyes, all buildds are babbage 3.009:12
NCommanderogra: patch added and pushed via bzr. I can test a respun image you want to confirm the updated code09:20
ograNCommander, if Riddell allows ... triger a full set of armel then please09:23
NCommanderogra: Riddell gave me permission to respin ARM at will09:24
=== asac_ is now known as asac
NCommanderogra: I meant a locally respun image09:24
ograoh, yeah, test it09:24
* NCommander groans09:25
hrwogra: yep.12:37
NCommanderogra: the patch seems valid, but I'm having testing issues due to instability13:24
NCommanderogra: I am going to commit a slightly newer version, and ask you to merge and handle getting the RC freeze exception, but hold off on the actual upload until I can get someone beside myself to reconfirm the test results13:25
persiaDoes anyone still use redboot-imx?  Can it be dropped from the archive?14:09
=== ApOgEE__ is now known as ApOgEE
=== zyga is now known as zyga-cold
GrueMasterpersia: our Lucid imx51 images use it.15:46
persiaGrueMaster, But do we need/want it in maverick?15:48
persiaIII, Just FYI, there's a 0929 image that ought be in better shape.15:49
GrueMasterTrue.  imx51 was disabled sometime around alpha 2-3.15:49
IIIpersia: thanks... just downloaded it15:50
mpoirierndec: do you think you could get me access to some ABE routing documentation ?16:00
rlameirompoirier: you know, now ther is a #igep channel, could be easier to find igep testers now16:02
ndecmpoirier: hi... sorry I saw your email, but didn't have too much time to digg into this.16:03
mpoirierrlameiro: indeed, but I would really need to get a board.16:03
ndecmpoirier: why do you need this info?16:03
rlameirompoirier: If i was rich, i would buy you one :D16:03
ndecmpoirier: I am not sure which document you would need in fact16:03
mpoirierndec: we are bound to repeat the current sound investigation on our other omap3 boards.16:05
mpoirierndec: omap4 too for that matter.16:05
ndecmpoirier: ABE is OMAP4 only, this is a complete different solution for audio h/w on OMAP316:05
mpoirierndec: the more I understand, the better I can give meaningful information to your team.16:05
ndecmpoirier: sure...16:05
mpoirierndec: I suspected that much.16:05
mpoirierndec: I'm mostly interested by the process berco took to produce the omap4_sound_config.sh script.16:06
mpoirierthose commands come from some sound data somewhere and that's what I'm after.16:07
ndecmpoirier: this file is produced by the audio dev team, not us. I will reply to your email.16:07
mpoirierndec: cool - you're my contact into that world.16:08
mpoirierndec: if I'd have some clue on how omap4_sound_config.sh and the SDP4430.conf were generated, I could help on my side.16:09
GrueMasterrsalveti: RE:  Bug 644714  I have another edid output that essentially is blank.  THis is what I get when the system wakes and no monitor is attached.  The bad thing is, Xwindows is garbage, which is why I recommended 1024x768 as a fallback.16:27
ubot2Launchpad bug 644714 in linux-ti-omap4 (Ubuntu) "Screen corruption waking from screen blank if no monitor present (affects: 1) (heat: 526)" [Undecided,New] https://launchpad.net/bugs/64471416:27
GrueMasterI'll try to add the data after RC testing.16:27
rsalvetiGrueMaster: yep, this would explain the issue16:30
rsalvetiboot with edid -> 1440x90016:30
rsalvetithen when it gets back, finds no edid -> 640x48016:31
rsalvetibut the x11 continues to think the fb is 1440x90016:31
GrueMasterWell, 640x480 is absolute garbage.  looks like 4 screens side by side only partially drawn.16:31
GrueMasterCan't get a screen capture atm.16:31
rsalvetiGrueMaster: I get this with my monitor: http://www.flickr.com/photos/rsalveti/4932601965/16:32
GrueMasterThat is actually usable.  Mine is much worse16:33
GrueMasterI'll try to get a capture during testing.  Heavily into RC atm.16:33
rsalvetiGrueMaster: np16:34
=== kmargar is now known as markos_
robclarkhrw: is there some trick to install your cross compiler packages on a lucid box?16:53
robclarkPackage libmpfr4 has no installation candidate16:53
hrwrobclark: http://people.canonical.com/~hrw/ubuntu-lucid-armel-cross-compilers/ for amd6416:53
hrwrobclark: i386 in next few days16:54
robclarkperfect, amd64 is what I need16:54
hrwrobclark: libmpfr4 is in maverick16:54
hrwrobclark: maverick pacakges on p.c.c. are obsolete anyway16:54
robclarkyeah.. but I was trying to use your maverick PPA on lucid..16:54
robclarklet me try this other site instead16:54
hrwppa is also obsolete16:55
hrwI just keep both because some toold use them16:55
* robclark can't keep track of what isn't obsolete16:55
hrwfor maverick: use maverick archive. for lucid: use what i just gave you16:56
robclarkk, will do16:56
robclarkohh..  I think it's working :-)16:58
mpoirierrsalveti: you have a significant head start on b644714 due to your previous work on EDID.  Do you have the cycles to work on it or you'd be happy to offload it ?16:59
rsalvetimpoirier: I'm working on it today17:21
rsalvetifirst debugging the issue when booting and activating/deactivating the hdmi with my monitor17:21
rsalvetiand then the x11 part17:21
=== hrw is now known as hrw|gone
GrueMastermpoirier: on the omap4 sound issue, this should probably be changed to libasound2 instead of the kernel, as libasound2 contains all the /usr/share/alsa/cards/*.conf files.17:31
GrueMasterI have the same issues with Beagle (omap) and now Dove A0.17:32
mpoirierhumm... there is also a kernel portion of it.17:32
mpoirierthe TI patch for omap4, and some work to differentiate card on beagle and others.17:32
mpoirieras you said, there will be a libasound2 part of it.17:33
GrueMasterOk.  Guess I should file separate bugs for each system then.17:33
GrueMasterOr should we just track them all centrally?17:34
persiaBetter to have multiple tasks for one bug.17:34
persia"Also affects distribution ..."17:34
mpoirieris there a way do link two bugs in lp ?  i.e this bug will only be completed if that bug is completed...17:34
persiaI should say "Please file one bug per platform, with as many tasks for each bug as are required to address the issue."17:34
persiampoirier, Intentionally not (although debate continues on this subject)17:35
GrueMasterWe either have one main bug and list all the kernels & asound, or one bug for each kernel.17:35
GrueMasterOh, ok17:35
persiaI think we should have one bug for each *board*, and list asound & relevant kernel for each.17:35
GrueMasterpersia: That makes the most sense.17:35
persiaPotentially list multiple kernels for a single board if it's supported by multiple kernels and multiple kernel maintainers want to apply the same class of patch (e.g. for omap3)17:35
mpoirierGrueMaster: I also think we should have a bug per board.  Otherwise bugs never get closed.17:36
mpoirieras persia indicate, the fixes will span multiple kernels.17:36
persiampoirier, Only in cases where multiple kernels support a given board.  For some hardware (e.g. imx51), there may only be one kernel available.17:37
persiaIdeally, this is a short-term thing, until linaro fixes things so that there is only one true kernel.17:38
prpplaguefor anyone that is interested, the deadline for features for the bamboo board is october 8th17:54
robclarkprpplague: IR receiver?17:55
robclark(and lasers... sharks with lasers)17:55
prpplaguerobclark: hehe17:55
prpplaguerobclark: i have IR on my list to look at17:55
persiaTalk of SATA was for a different extender, right?18:03
prpplaguepersia: still trying to figure that out18:05
prpplaguepersia: i'm considering dropping the second sd/mmc in favor of a 32GB eMMC18:05
persiaUnless you can find a way to boot off that, I'd rather see a second SD/MMC.18:06
prpplaguepersia: boots fine18:06
persiaOff a secondary eMMC?18:06
prpplagueyea, same as blaze18:06
persiaOh, please do that then.  I'd dearly like to return to having a sane "install" path.18:07
persiaAnd I'll have good arguments to do that if the Bamboo works with it :)18:07
persia(extra benefit: the same codepath would work for both panda and blaze, etc.)18:07
bjfGrueMaster, did that fsl-imx51 kernel show up for you yet?18:08
GrueMasterNot sure, will check.18:08
bjfGrueMaster, https://launchpad.net/ubuntu/+source/linux-fsl-imx51/2.6.31-608.20/+build/197818418:09
GrueMasterI know it is listed on launchpad.  But until it shows up in http://ports.ubuntu.com/pool/main/l/linux-fsl-imx51/, it is more difficult to deal with, and it still isn't there.18:11
bjfGrueMaster, ok, trying to figure out where it's hung up18:11
persiaIf it's not published by the turn of the next hour, might ask if the release team has the publisher on manual.18:11
GrueMasterpersia: This is for lucid-proposed.  Shouldn't be affected afaik.18:12
persiabjf, I haven't checked current performance, but it used to take 43-103 minutes to get from "built" to "available in the archive"18:12
persiaGrueMaster, it's the *same* publisher.18:12
bjfpersia, thanks for the info18:12
GrueMasterWell, since I'm deep into Maverick RC testing, it will have to wait until Late tomorrow/Friday for testing.18:13
bjfGrueMaster, np, mostly i'm just trying to make sure its available to you when you are ready to work it18:14
loologra: Around?19:05
=== amitk is now known as amitk-afk
GrueMasterWell, now this is odd.  Attempted to change only the font in une-efl, the system seems to have completely changed themes.  Very unusual.19:31
hans2Question: Is there any progress being made with suspend on imx51?19:39
GrueMasterA new kernel was uploaded yesterday, but I'm not sure what was changed in it.  It is currently pending publication after Maverick RC freeze is lifted.19:45
ogra_acrc freeze for a lucid SRU ?19:45
ogra_acGrueMaster, it should just go in19:45
ogra_acshould be in -proposed for testing if it has build19:46
GrueMasterWell, that's what I have been told.  I'm too busy with RC to go hunting for it.19:48
GrueMasterAnd the kernel team just keeps sending me links to launchpad where the packages are sitting.  Need it in lucid-proposed for apt-get to pull.19:49
persiaGrueMaster, The acceptance queue is different from the publisher.20:05
ogra_aclool, yes (as long as this tegra WLAN stays stable)20:05
GrueMasterpersia: I don't really care about the background processes and the mechanisms that they work through.  What I do care about is when I get an email requesting me to test a kernel update in lucid-proposed, and people asking me why it isn't there for apt-get to work.20:07
persiaGrueMaster, That's fair: it's sensible for the uploader to take responsibility for ensuring the upload ends up being distributed in the repositories.20:08
GrueMastermy point exactly.  :P20:09
persiaAnyway, I'll continue to explain workflows and background processes and mechanisms as long as people ask questions.  Appropriate folk (e.g. bjf) might want to chase up on things.20:11
GrueMasterexplaining them is fine, as long as they are also addressed to the people that should be able to do something about the process when it doesn't work.20:13
bjfpersia, i'm interested in the workflows and background processes so i appreciate the info20:13
persiaGrueMaster, Sorry then, I'll try to read more carefully and not mis-highlight.20:14
GrueMasterDon't get me wrong, I too am interested.  Just not when I have 1 day to test multiple releases on multiple platforms.  :P20:14
persiabjf, So, anyway, once LP gets a build, it submits it to the queue (e.g. https://launchpad.net/ubuntu/lucid/+queue ).  The queue has a few states, and things can be in one or another depending.  Nothing will land in the archives until it gets to Accepted (NEW and UNAPPROVED need manual action by an archive-admin).  Stuff in Accepted will be pushed to the archive each publisher run (typically once an ho20:17
persiaur: sometimes set to manual for release management purposes).20:17
persiabjf, So, find out where your package is: if it's not anywhere in the queue (including DONE), then something is odd with the LP build.  If it's in the queue, but not ACCEPTED or DONE, you need an archive-admin (and likely some associated paperwork).20:18
persiabest place to find archive admins is #ubuntu-devel20:18
persiaIf it's been in ACCEPTED for a couple of hours or more, and it's around some special time in the Ubuntu Release schedule, you might check with the folk in #ubuntu-release to make sure the publisher is running.20:19
bjfpersia, there isn't a way for me to check if the publisher is running without asking?20:19
persiaI think only LOSAs can check, but there may be some other interface exposed.  You could ask in #launchpad, but I generally just assume it's running except when it clearly hasn't for about 3 hours.20:21
dcordes-libHi. Is anybody aware of fast mirrors to the arm preinstalled daily builds ?20:46
persiaI believe they are all private.20:48
dcordes-libHm. Currently only getting around 50kB/s from cdima.ubuntu.com20:50
dcordes-libbut it might be the library hating me20:50
dcordes-libah seems to be a local problem really20:53
dcordes-lib3,3MB/s now20:53
persiaThat's a much better speed :)20:57
Nekopersia, all ubuntu boxes shipped, exactly the numbers from the email21:17
persiaNeko, Cool.  Thanks for the confirmation.  Please let me know if anything isn't happening to your satisfaction.21:18
NekoI expect about a month while people get used to it21:19
Nekothey all shipped with maverick+xfce but I assume the kde guys etc. will wipe it immediately21:19
persiaShipping with maverick will make lots of folk happy.  I hear markos_ is coming to UDS, and I suspect people will be asking questions there.21:21
=== linuxlover is now known as Riotta
robclarkso quick question.. does dkms log to somewhere, so I can see *why* it failed to rebuild some module?21:39
loolrobclark: perhaps in /usr/src?21:45
* robclark looks21:46
robclarkhmm.. src is there.. but don't see a log..21:49
robclarkok.. well, I think I figured out how dkms is invoked, so I can do this manually..21:51
rsalvetirobclark: yep, /var something, let me see it21:52
rsalvetihad the same question when I created the sgx package21:52
rsalvetimake.log, something like that21:53
robclarkrsalveti: the sgx package is the one I'm having trouble with ;-)21:53
rsalvetiin my case it was /var/lib/dkms/powervr-omap3/
rsalvetirobclark: well, I created only the omap 3 one :-)21:54
robclarkahh,  perfect.. found the log.. thx rsalveti21:54
robclarkhmm.. tho the log has no errors..  I wonder why dpkg thought it failed when I tried to install new kernel?21:55
rsalvetirobclark: can you paste the error you got?21:56
rsalveticould be that it couldn't find the proper headers for your kernel21:56
ndecrobclark: it looks like you are installing the wrong headers. the OMAP4 kernel should be linux-headers-2.6.35-90321:59
ogra_acrobclark, you are building omap3 stuff21:59
ogra_acapt-get install linux-headers-omap4 iirc22:00
rsalvetirobclark: yep, that should be your problem22:00
ogra_ac(there is a metapackgae for everything ;) )22:00
rsalvetieven a meta of the meta22:00
ndecrobclark: you probably want to remove the omap3 kernel and headers packages as well.. that said it's probably a bug in our sgx omap4 package... it should not try to build/install on non supported kernel22:02
robclarkndec: well.. I was just trying to install a self-built kernel, and keep the sgx stuff working..22:02
ogra_acndec, well, it also was a bug that foreign headers were installed at all22:03
ogra_acthats fixed since a few days22:03
ndecrobclark: argh... self built, you mean as ubuntu packages? or just uImage.22:03
ndecogra_ac: cool22:04
robclarkndec: .deb...  make CROSS_COMPILE=... ... deb-pkg22:04
ndecrobclark: ok. you will need to generate the headers too.22:05
robclarkwell... I guess headers should be same as what I already have... I just git pull'd the kernel-ubuntu tree from d.oz.o..22:05
robclarkhmm.. I guess I shouldn't have tried to reboot22:06
ndecrobclark: the next problem... will be that sgx module will likely not natively build if your kernel has been cross compiled... just fyi: native kernel compilation as .deb should take ~2.5 h on panda22:07
robclarkfwiw, I am using the linaro cross compile toolchain to build kernel.. so same gcc version22:08
ndecrobclark: the compiler is not the problem... but the generated tools and makefile during kernel build will assume cross compile. when compiling module natively with dkms, it's causing issues. I don't know all the details, but rsalveti or sebjan might know22:12
robclarkmodules are such a pain22:13
robclarkbut ok... I can cross compile the modules too22:13
Nekovar/cache/fontconfig/99e8ed0e538f840c565b6ed5dad60d56-mipsel.cache-2 <-- what does mipsel mean?22:16
Nekohttp://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501700 REALLY? "fixed in latest fontconfig"??22:17
ubot2Debian bug 501700 in fontconfig "writes cache files with "mipsel" in name .. on armel" [Minor,Fixed]22:17
Nekowell now it says le32d8 or so which I guess is less retarded but.. why would the cache file need to be so arch specific like that.22:22
persia"mipsel" is little-endian MIPS22:49

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!