=== lilstevie is now known as lilstevie|ZNC
Neko__ogra, does anyone know why mono doesn't install on armel (especially in qemu?)03:05
persiaIt's a qemu bug.03:18
persiadirecthex got Mono working cleanly on the EfikaMX for maverick.03:18
persia(even moonlight)03:18
persiaIf you're using rootstock, don't install Mono until you've booted on the real hardware.03:19
persia(unless someone made rootstock work unemulated for native builds, in which case it ought be no issue)03:21
persiakmargar, Hey.03:21
Neko__persia, I just saw there were tons of bugs in Mono for mx51 (but not dove!) for some reason, and ogra was involved03:22
Neko__testing at least03:23
Neko__directhex is one of the guys we sent boards to?03:23
Neko__the mono guy?03:23
Neko__did he have to fix anything or was it just working out of the box at release?03:23
persiaIndeed.  mono on the i.MX51x was exceedingly painful for a while, but is now sorted.03:23
Neko__okay great03:23
Neko__so what's the qemu bug :D03:23
persiaHe had to fix something.  There was a blog post.03:23
Neko__it would be great if that was fixed03:23
persiaWell, qemu doesn't support some huge chunk of stuff (you'll see lots of unsupported syscall and unsupported ioctls pass in stdout).03:24
Neko__once I finish this next rootstock I am done with it and will start messing with livecd-rootfs03:24
persiaAnd something in that unsupported stuff is enough to wedge Mono.03:24
persiaThe impression I have is that nobody wants to fix it, instead switching from qemu-versatile to qemu-omap to collect a separate set of issues, but one that can easily be compared against hardware.03:25
persiaWhether using qemu-omap is enough to install Mono is not clear to me: there may be more involved.03:25
Neko__I saw his post about moonlight and firefox ABIs today03:25
Neko__nothing about Mono proper though03:25
persiaHe got Mono working back in Jaunty, but only for some things.  It's a steady progress to test all the corner cases.03:26
Neko__I'm impressed that Moonlight is working..03:26
persiaMoonlight is rather demanding on the CLI, and exercises it fairly well.03:26
Neko__I don't understand about this codec pack though03:27
Neko__anyway brb Patch Tuesday03:28
persiaFor those lurking, the above is another demonstration why it's handy to run Ubuntu :)03:31
=== JaMa|Zzzz is now known as JaMa|Wrk
=== kmargar is now known as markos_
doko_ogra: the hint for tar --numeric-owner is missing in your README08:22
persiadoko_, Please file a bug :)08:23
lagogra: Are you around yet?08:33
=== hrw|gone is now known as hrw
vstehleogra, ogra_ac: Hi, how are you? I am "fighting" with the "install extras" icon right now; we think that universe & multiverse have not been enabled in the preinstalled image, which make installing our "extras" a bit difficult...09:12
ndecogra: lag: hey! questions for the PPA: I see that the builders are idle, however our builds don't get started. is that expected?09:19
persiavstehle, Take a look at your /etc/apt/sources.list to verify universe and multiverse are enabled.09:20
lagndec: This isn't something I deal with - perhaps ogra or persia can help?09:21
persiaBest to ask the launchpad folk (in #launchpad).  My guess would be that it had something to do with natty opening, but there's a high chance I'm wrong.09:21
persiandec, I'm happy to lead the discussion, if you'd prefer, but please also join to provide any support information :)09:24
ndecpersia: ok09:24
persiandec, Just to confirm, which PPA?09:25
ndecpersia: well, our private OMAP ppa09:25
ndecpersia: and I got a 'waiting 11 hours' for an arch all package using an i386 builder...09:26
=== doko_ is now known as doko
vstehlepersia: Well, /etc/apt/sources.list _is_ the problem :) We have those of the preinstalled image. And they don't include multiverse and universe "out of the box".09:26
persiavstehle, Aha!  I suspect it's one of the preinstalled vs. performing the install things.  Please file a bug against jasper-initramfs, and we'll sort it.09:27
vstehlepersia: Hum. Can you "fix" the images on the website? I thought they were released and "frozen" now.09:27
vstehlepersia: I thought our only options were fixes in the PPA packages, and documentation.09:28
persiaNo, but we can 1) fix jasper to not do this next time images are produced, and 2) brainstorm about ways to work around it.09:28
persiaThere's also bugfix uploads: won't affect the images, but can affect user experience once the user updates.09:29
vstehlepersia: Ok, I'll file a bug on LP for (1). I am very interested in (2) right now :)09:29
persiaWe'll use the one bug for both.09:29
persiaJust with two tasks.09:30
persiaogra, Did you implement the special software channel as a separate package, or as part of jasper?09:31
vstehlepersia: Bug #65975409:33
ubot2Launchpad bug 659754 in jasper-initramfs (Ubuntu) "Universe & multiverse are not enabled on OMAP4 preinstalled image (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/65975409:33
persiaGreat.  Now to wait for the confirmation of the implementaiton, and we can get started on fixing it :)09:34
vstehlepersia: I tell you: I am not really worried about the fix for future distributions. What worries me is that users will d/l the OMAP4 preinstalled image, click on the icon and get an error message.09:37
vstehlepersia: We are thinking about documenting an extra step like "enable multiverse/universe manually" on a website somewhere. Or even more nasty: do some hacks in the ubuntu-omap4-extras package (break the deps on other real packages, perform hacks in postinst scripts). What do you think?09:38
persiaThere's an API to enable universe/multiverse, so we oughtn't have to go to the extreme of a nasty hack.09:48
persiaJust a matter of determining where/how to drop it in place.09:49
persiaBut I really don't think it's wise not to worry about the fix for future releases: firstly, the SRU process requires "fixed in current development" as a prequisite for an SRU.  Secondly, if we don't fix it in both places, we'll just end up in this state again in six months.09:50
persiaThat said, I very strongly suspect we can get a solution ready for testing within the next 24 hours, and probably within the next 6-8.09:50
vstehlepersia: Would your solution require an extra user step? Or are you thinking of doing stuff in preinst?09:55
persiaI was thinking a postinst: just have to think about where.09:56
persiaBut, ideally, the user experience would be nothing other than a regular update.09:56
persiaFor the future, the planned jasper rewrite *should* cover it, and we can do a quick python-apt hack to force-enable for the short-term.09:59
persiaIdeally, ogra implemented the omap4-special software channel in a separate package, so we can just update that package: this has the lowest chance of affecting others.10:00
persiaBut I'm not precisely certain how that was implemented, which is the main deciding factor for not fixing it now.10:00
ograpersia, it is just implemented through a .desktop file containing an aptulr10:01
persiaogra, No package then?  Darn.10:01
persiaAny suggestions on where we can stick a postinst fragment to enable universe/multiverse?10:01
ograthe postinst of the meta needs to be easily able to just rebmove it10:01
ograa package would have gotten in our way10:02
persiaAha.  I'll call that excessively layered hackery, but I understand why it was done that way.10:02
persiaSo, what do we have that would only hit preinstalled folk and can enable universe?10:03
ograa package would either have to be seeded or be a dep of jasper10:03
ograin the seed way it would always come back10:03
ograand with jasper it would have been removed with oem-config10:03
persiaLet's not discuss that implementation now: we can do that in the jasper-rewrite discussion.10:03
ograthanks to the live seed10:03
persiaLet's talk about how we can SRU a fix.10:03
ograwell, first of all livecd-rootfs needs a fix10:04
persiaBecause a pre-jasper-rewrite natty fix is a trivial python-apt call.10:04
ograit apparently misbehaves10:04
persiaThe model is that the live environment has universe disabled, and it's enabled during install.10:04
persiaThe rationale is to save space for the universe Packages file in the live environment.10:05
ograit writes the sources.list and obviously doesnt write the same default one that ubiquity sets up by default10:05
persiaAnd livecd-rootfs does this.10:05
persiaRight.  Jasper should take care of that.10:05
persia(and it's trivial).10:05
ograjasper isnt SRUable10:05
persiaBut the hard question is what to do for maverick.  What can we update that is only going to affect jasper-installs.10:05
persiaI know/.10:05
persiaThe only reason to add the hack to jasper is to justify the SRU of something else.10:06
ograand we have no package that would just do it10:06
persiaWe have nothing that is only on preinstalled images?10:06
ogranothing apart from jasper10:06
persiaThat makes this tricky :)10:06
persiaDo we have any programmatic way we can distinguish a preinstall from an install?10:07
ogranot from an oem install, no10:07
ograwell, you can check if jasper is installed, but only until oem-config was removed10:07
ograyou can also check if the ppa enablement files exist10:08
persiaDo we do an apt-get upgrade before that happens?10:08
ograupdate-manager is supposed to care for upgrades10:08
persiaAh, right.  The PPA enablement files are probably a good test.10:08
persiaLet me rephrase then: do we invoke a python-apt cache update prior to oem-config?10:09
* persia really doesn't care *how* the apt-cache is updated, but whether it is updated10:09
ograthe prob here is that either oem-confir doesnt run apt-setup at all or that livecd-rootfs doesnt enable what it should10:09
ograi'm still tending to blame the latter10:09
persialivecd-rootfs is supposed to leave universe and multiverse disabled.10:09
ograwhy ? thats nonsense10:10
persiaThis is by design for all live environments building from main.10:10
persiaIt saves several megabytes from the CDs.10:10
ograsince we enable it by default and give no opportunity to override that10:10
ograwhy would it10:10
ograsources.list is setup at a point where that doesnt matter10:10
ograhmm, k10:11
persiaYes, but you need an accurate apt-cache to be able to run things like the codec installer from within the live environment...10:11
ograi wouldnt think Packages.gz is that big10:11
persiaAnyway, it7s not a livecd-rootfs bug.10:11
ograit is10:11
ografor future images we should just enable whats needed in the preinstalled path10:11
ograthat wont touch the livefs10:11
persia5MB on my mirror for maverick.10:12
ograit gets compressed again10:12
ograwont be 5MB10:12
persiaYes, for packages.bz2  Packages.gz is seven and a half.10:12
persiaYes it will.10:12
ograwell, if you say so10:12
persiaMaybe 4.5, because squashfs is lzma, but anyway...10:13
* ogra wont debate that since we'll be adding several 100 anyway in natty10:13
ograno squashfs involved for preinstalled10:13
persiaOh, then 7.5 because that's gz.10:13
persiaAnyway, doesn't matter.  Let's get back to the point.10:13
ograright, whats your suggestion ?10:14
persiaSo, it's trivial to enable universe/multiverse in current jasper by calling into python-apt.10:14
persiaAnd we'll use apt-setup to do it with rewritten jasper.10:14
persiaSo natty is sorted.10:14
ograjasper is not SRUable10:14
ograwhat about maverick10:14
ograi know what to do about natty10:14
* persia is getting there, and requests a bit of patience for a summary10:14
ograand i wont do it in jasper there since that enfoces a network connection at oem-config time10:15
persiaNo, it doesn't.10:15
persiaThe idea is to use the same sequence of stuff used for normal installs.10:15
ograhow do you get the package cachje without network if not at image buildtime ?10:15
persiaYou fail to get it if there's no network, and you set it up to be a request for it in software-sources, so that next time update-manager does a regularly scheduled run, it pulls them.10:16
persiaAnd update-manager automatically doesn't do that if there's no network, so you end up doing it the first time the user *has* a network, which is the right time.10:16
persiaAnyway, can we get back to maverick?  We've agreed natty is easy.10:17
persiaSo, we can detect that we're in a preinstall because of the PPA enablement files.  And we can detect that universe/multiverse are not enabled, and use python-apt to request enablement.10:17
persiaWhat we need is some package to which we can attach the script that does that.10:18
persiaAnd that has to *not* be jasper, but something that we can usefully SRU.10:18
persiaAnd I don't care about delayed-enablement, because the user has a network connection to download the SRU anyway.10:18
ograis there a way to enforce u-m directly after login ?10:19
ograwithout much hackery indeed10:19
persiaIt's already set to run the first time there is both a logged in admin user and a network.10:19
persiaSo we get that for free.10:19
* ogra has never seen that10:20
ograand by rules of mpt it should only run after 7 days for the first time10:20
persiaTry doing a networkless install about three months after release sometime, and you'll see it.10:20
ograbut i might be wrong10:20
persiaYou can test with a networkless install of lucid today, if you like.10:20
ograit changed between lucid and maverick10:20
ogramvo told me it has to wait 7 days now10:21
ograno matter what10:21
persiaIt runs 7 days after the last run.  ubiquity tries to do an update at install time.  If ubiquity fails to do an update, the last run timestamp is the image creation date, which is more than 7 days ago.10:21
persiaUnless the implementation changed massively, we ought still get free updates.10:21
ogranot my experience and not what i was told for maverick, but if you say so10:21
ograso where would we put it ? we dont have any special packages installed10:22
persiaMaverick might be different.  Dunno.  We'll check if we find a victim.10:23
persiaYeah, the where to put it is the tricky bit.10:23
persiaDon't we have a kernel SRU pending that's specific to certain hardware?  Could we add it to that?10:23
persia(noting that anywhere we put it will annoy the SRU team, so there's no better/worse place)10:24
ograubuntu-netbook-efl-default-settings ... ?10:24
ograthats quite freely hackable10:24
ografor us at least10:25
ograit will only be installed on arm by default (there is a switch in the netbook seed)10:25
persiaI think the default-settings package is probably better.  Same argument with the SRU team, but won't block the kernel if it gets sticky.10:25
ograyou will never get it on anything but arm10:26
persiaThat's not true.10:26
ogra(nothing depends on it on other arches)10:26
persiaBut you won't get it *by default*, and we can guard it with the appropriate preinstall check.10:26
ograonly if you explicitly install it by hand you will get it10:26
ograubuntu-netbook doesnt pull it in, nor does any other dep10:27
persiaRight.  We have to also consider that case, but I think we're safe if we check for the PPA enablement files in the postinst.10:27
ograwell, that wont solve omap310:27
ograi'd say we do an arch check10:27
ogracheck arch, check sources.list, if universe is missing and arch matches, enable universe10:28
ograbecause omap3 doesnt have a ppa10:28
persiaIf someone decides, for reasons we'll never understand, to use livecd-rootfs to create an i386 preinstall and has this issue, we ought sort it for them as well.10:28
ograthat would mean he has to change seeds too10:29
persiaUnless you want to SRU jasper.10:29
ograan x86 preinstall would still not pull in -efl10:29
ograand jasper on x86 would do nothing wrt ppa10:29
persiaOh right, and at that point, it becomes a case of "the bug was reported, and you should have backported the fix from natty".10:29
ograit does an arch check10:29
persiaOK.  I hate arch checks on general principles, but I think it's probably the least bad option at this point.10:30
persiaAnd it's only the omap3 and omap4 images that are broken.10:30
ograit wouldnt do much at all in fact, just enable oem-config and the default session (the latter would be a bug on x86)10:30
persiaWhy would it be a bug on x86?10:30
ograbecause the efl session cant exist there10:31
persiaWhy not?10:31
ogranot from livecd-rootfs10:31
ograbecause the seed doesnt pull it in, on x86 it defaults to unity10:31
persiaDifferent definition of "can't".  Right.  I agree.10:31
persiaDo you want to summarise the above in the bug, or shall I?10:31
ograit would require massive seed hackery that would make the netbook CD size explode10:31
ograwould you ?10:32
persiaNo, but we don't care, because it's large enough to be derivative at that point.10:32
persiaI'm happy to: let's just recap quick to make sure we agree.10:32
ogratell that to didrocks :P10:32
persia1) Long-term fix is jasper-rewrite10:32
ograhe forced -efl off the x86 installs because of size10:32
persia2) short-term natty fix is python-apt hack in jasper10:32
persia3) maverick fix is -default-settings postinst hack with arch guard.10:33
persiaLook right to you?10:33
ogra(which got us in the awkward situation of unity sessioons coming back all the time, since he hardcodes a switch of the session towards unity in his -settings postinst)10:33
persiaRight, netbook-efl for x86 is only interesting if one drops unity.10:33
ogralooks good to me10:34
* persia updates the bug10:34
* ogra just did an update on x86 and glares at the (felt) 2mx2m sized button in update-manager "restart now"10:35
ogrageez, thats huge !10:35
ograi wonder how that looks on a small screen, it must have 60px padding around the text10:35
persiaogra, Could you quickly accept jasper/natty and -default-settings/maverick tasks and reject the other two for bug #659754 ?10:36
ubot2Launchpad bug 659754 in ubuntu-netbook-efl-default-settings (Ubuntu) (and 1 other project) "Universe & multiverse are not enabled on OMAP4 preinstalled image (affects: 1) (heat: 10)" [Undecided,New] https://launchpad.net/bugs/65975410:36
ograoh, why are you not allowed to ?10:36
* persia isn't core-dev10:36
persiaI'm still writing up the issue, etc.10:37
mopdenackerHi! I'm trying to make a few changes to the uInitrd of pre-installed images, to support installation on the Blaze's eMMC storage. However, the kernel says "failed to execute /init" when my modified uInitrd is used. Here are the steps I took:10:37
mopdenackerdd if=../uInitrd of=initrd.lzma bs=1 skip=6410:38
mopdenackermkdir initrd10:38
mopdenackercd initrd10:38
mopdenackerlzcat ../initrd.lzma | cpio -id10:38
mopdenacker<make changes: just adding "set -x" in some scripts>10:38
mopdenackerfind . -print -depth | cpio --quiet --dereference -o -H newc | lzma -c > ../initrdnew.lzma10:38
ograpersia, hrm, doesnt work10:38
mopdenackercd ..10:38
mopdenackermkimage -A arm -O linux -T ramdisk -C none -a 0 -e 0 -n initramfs -d ./initrdnew.lzma ./uInitrdnew10:38
mopdenackerDo you see anything wrong?10:38
persiaogra, Ugh.  Maybe accept everything, and we can set somethings to Invalid?10:38
persia!paste | mopdenacker10:38
ubot2mopdenacker: For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://tinyurl.com/imagebin | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic.10:38
ograpersia, sounds better10:38
ogramopdenacker, thats not the right way, chroot into the filesystem and run update-initramfs (and make sure /proc is mounted ... and unmounted before you leave the chroot)10:39
mopdenackerubot2: ok, I thought this was short enough, but I got it wrong. Will do it in the future...10:39
ubot2mopdenacker: Error: I am only a bot, please don't think I'm intelligent :)10:39
lagogra: persia: Explain to me what the default.pa file does for _us_10:39
lag! mopdenacker | lol10:40
ubot2Factoid 'mopdenacker' not found10:40
persialag, Ask me again in 10 minutes.10:40
lagI just want a short explanation for my write-up10:41
ogralag, imho it gets constantly in our way :P10:41
ogra(but dont put that in your summary )10:41
lagogra: That's not helpful10:41
lagWhy do we need it?10:42
lagogra: You contradicted yourself10:43
lag1. HDMI doesnt get initialized and exposed to pulse10:43
lag2. its fine and i saw hdmi exposed in it10:43
ogralag, when hacking default.pa10:43
ograyou pulled that out of context :)10:43
ograwe cant hack default.pa because it breaks all other arches10:44
persiaogra, Note that Kubuntu is also affected.  Do you know if the PPA enablement stuff works there?10:44
ograpersia, no idea10:44
mopdenackerogra: I kind of disagree. It never hurts to understand the lowlevel formats. Would you still go through elaborate scripts (which could go wrong in multiple ways) to make a quick change to an initramfs if it was just a plain compressed tar archive?10:44
lagogra: But I thought that was the solution until UCM comes along10:44
ogramopdenacker, yes, always10:44
ogralag, what ? editing default.pa ? no, thats no solution to anything10:45
ogralag,  as i undrestood we'll get a pulse profile that will fix that10:45
lagWhat's the difference between default.pa and the Pulse profile?10:45
ograthey live in different places and the profile can be card specific10:46
persiaogra, Please *accept* the maverick task.10:46
lagAh, that clears a lot up10:46
lagI thought they were the same thing10:46
lagWhat's the Pulse profile called? And where does it live?10:46
vstehlepersia: I am not sure I understand all the outcomes of your discussion with ogra; what will the end-user see after all? D/l preinstalled image, receive "magical" update due to "older than 7 days", then click icon and everything ok?10:46
ograpersia, invalidated the jasper bug for maverick and accepted both tasks10:47
lagOh, that's what the default.conf is?10:47
persiavstehle, That's the idea.  If nothing else, the apt cache update to get the PPA stuff ought get the universe stuff as well.10:47
mopdenackerogra: I'm still not sure I will go this way. But thanks for your answers anyway :-)10:47
ubot2Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-)10:47
ogramopdenacker, at *least* make sure you pack the initrd the exact same way update-initramfs uses10:47
ubot2The linux terminal or command-line interface is very powerful. Open a terminal via Applications -> Accessories -> Terminal (Gnome), K-menu -> System -> Konsole (KDE), or Menu -> Accessories -> LXTerminal (LXDE). Guide: https://help.ubuntu.com/community/UsingTheTerminal10:48
vstehlepersia: But how could you change the apt-cache update in the preinstalled image?10:48
ubot2Factoid 'stupid-bot' not found10:48
ubot2Hi! I'm #ubuntu-arm's favorite infobot, you can search my brain yourself at http://ubottu.com/factoids.cgi | Usage info: http://ubottu.com/devel/wiki/Plugins | Bot channels and general info: https://wiki.ubuntu.com/IRC/Bots10:48
ogravstehle, with the update to the -sessings package10:48
persiavstehle, We don't: when the user updates the apt-cache, the update will show: when they install the update, it will then cause universe to be enabled for the *next* apt-cache update.10:48
ogravstehle, on first boot, forst thing the user sees will be update-manager10:49
vstehleogra: I don't see that.10:49
ograwith installing the update, universe and multivers get magically enabled10:49
ogravstehle, thats hardcoded10:49
vstehleogra: I am testing the preinstalled image on a Panda right now. I never saw the update manager.10:49
persiaMight have to wait until the 17th to see it.10:50
persiaThat 7 days thing.10:50
ubot2If you're having problems with sound, click the Volume applet, then Sound Preferences, and check your Volume, Hardware, Input, and Output settings.  If that fails, see https://help.ubuntu.com/community/Sound - https://help.ubuntu.com/community/SoundTroubleshooting - http://alsa.opensrc.org/DmixPlugin - For playing audio files,  see !players and !mp3.10:50
ograu-m checks for timestamps10:50
lagThere's our answers10:50
ograyou will see it earlier10:50
ograthe final image for omap4 was built on the 7th10:50
ogratomorrow or thursday it shoudl start to show up10:51
persiaBut if the apt-cache gets updated for any reason (and there are several things that do this), u-m will notify the user of the updates without waiting.10:51
vstehleOk, suppose I am now 7 days later. I install, I am prompted to do the updates, I accept, this adds universe+multivers. Then I click the icon, hopefully this does update the sources _again_ and it works. right?10:51
ubot2Yum! Err, I mean, APT!10:51
ograit does run apt-get update10:52
ograit has to, since it enabled a new source10:52
persiavstehle, That's the idea.10:52
vstehleOk, I get it. Thanks!10:52
persiaogra, No it doesn't: it will only update the apt-cache if the network is on (although the chances of this are astronomically high during SRU application)10:52
ograwell, it will update the cache once you enable universe10:53
ograand it will do that again once you enable the ppa10:53
persialag, So, the big things we get out of our specialised default.pa are 1) automatic configuration restoration, 2) automatic detection of devices, 3) bluetooth support, 4) legacy overrides to work around the idea that pulse and alsa and oss somehow compete,10:54
lagI thought we weren't touching *.pa?10:55
lagI thought we were using *.conf10:55
ograwe dont touch default.pa beyond whats there already10:55
persia5) cleaner audio load/unload for suspend/resume, 6) integration with console-kit so we don't have to fuss with the audio group and other stuff, 7) positionally meaningful events for cooler integration with libcanberra10:55
ograall arch specific changes we do now have to happen elsewhere10:56
persiaRight.  We don't touch default.pa for the special sdp4430 stuff, but that's what the Ubuntu default.pa is configured to "give us"10:56
persiaogra, Are you preparing fixes for this, or do you want patches to review?11:14
ogrado you have patches ? i'm currently trying to get crimsuns fixes to work11:14
ograwe need to get rid of the alsactl init11:14
persiaYou go work on sound: that's HW dependent, so I can't do it.11:14
persiaI'll prepare some patches.  Worst case, they'll be ready when you wake tomorrow.11:15
ograbut first ... coffee11:15
ograthis week is key11:15
persiaI'm not going to do the final natty jasper fix now: that needs to be part of the rewrite.11:15
persiaThis week is not an issue.11:15
ograwell, i wouldnt actually call it a rewrite :)11:15
persiaJust "tonight" is uncertain.11:15
ograthe main issue is splitting out most of jasper into packages11:16
ograthe things it does are mostly ok, just not the way11:16
persiaYou can continue not to call it a rewrite as long as you like.  I have a huge desire to rearchitect it to such a degree that I expect to retain 0 lines of code.11:16
ograthats impossible11:16
ograthe resizing bit wont be doable anywhere else11:17
ograthe setup script i agree .... but even that will still need one line (teh enablement of oem-config cant be done anywheer else)11:17
persiaOh, I might reuse some of the resizing code, actually, but likely nothing else.  Enabling oem-config will happen there, but in an entirely different way.11:18
ograyou need the rootfs mounted to enable oem-config11:19
ogracant happen in jasper_growroot11:19
persiaI know.11:19
persiaDoesn't mean I won't try to do it as d-i controllers anyway.11:20
persiaHey berco11:20
bercowhen an upstart script is installed from a deb package, does it require a posinst script?11:20
bercoI get lintian warning script-in-etc-init.d-not-registered-via-update-rc.d11:20
bercobut i'm confused as the upstart script goes into /etc/init and not /etc/init.d11:21
persiaupstart scripts don't belong in /etc/init.d/ : they belong in /etc/init11:21
bercoso lintian is complaining for nothing?11:21
persiaAre you using dh_installinit?11:21
berconot specifically11:22
persiaAre you using debian/package.upstart?11:22
bercopersia: yes11:22
bercobut in the rules file, I see "dh $@"11:22
persiaThat's fine.11:22
persiaDo you have a manual postinst script for the package?11:23
bercopersia: no11:24
ndecpersia: berco: isn't that better to use cdbs in this case? less chance to get errors, no?11:25
ogradh should do the right thing11:25
persiandec, No.  Actually, just means one has to know lots more make and slightly less perl to track down the same errors.11:25
bercondec: that's what I usually use but not my initial package. I can change it though if needed11:25
persiaBut the extra perl for dh(1) is kinda trivial when compared with the perl one needs for dh_* anyway.11:26
bercopersia: ogra: so is this fine to ignore this lintian warning?11:27
persiaberco, Could you paste the output of dpkg --contents foo.deb ("No" is an acceptable answer, but then I want the package name)11:27
ndecok. wasn't sure if dh_installinit would be called with just dh $@11:27
ograndec, yes, it is if the file is properly named11:28
persiadh $@ should call all of them.11:28
ograit looks for .init or .upstart11:28
persiaIf there's something missing from the sequence, well want it added.11:28
ograand puts them in the right places11:28
persiaogra, Or .default11:28
ogra.default doesnt go into an init dir11:28
persiaNo, but dh_installinit is responsible for installing it anyway.11:29
persiaPuts it in /etc/default/package11:29
persiaAnyway, berco, about those contents?11:29
lagbug 65203511:44
ubot2Launchpad bug 652035 in alsa-lib (Ubuntu) "libasound2 not finding usb sound card (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/65203511:44
bercoogra: persia: any issue with launchpad? Impossible to upload or fetch from TI 3PA14:54
ograworks fine for me14:55
bercomaybe network issue on our side then :(14:56
bercondec: does it work for you? try apt-get source tisyslink for e.g14:58
ndecberco: it's downloading.14:59
ndecberco: what's the exact problem? any log?14:59
=== xfaf is now known as zul
bercondec: Err https://private-ppa.launchpad.net/tiomap-dev/private-release/ubuntu/ maverick/main tisyslink (tar)15:03
berco  gnutls_handshake() failed: A TLS packet with unexpected length was received.15:03
bercoFailed to fetch https://berco:xxx@private-ppa.launchpad.net/tiomap-dev/private-release/ubuntu/pool/main/t/tisyslink/tisyslink_0.24.9.2.orig.tar.gz  gnutls_handshake() failed: A TLS packet with unexpected length was received.15:03
bercoE: Failed to fetch some archives.15:03
bercondec: was working a few min ago15:03
* berco should have pastebin this log15:03
loolberco: Looks like a network issue15:06
ndeclool: but I am sitting less than 3 meters away from berco and it works for me ;-)15:07
rsalvetindec: do you need proxy in your network?15:09
rsalvetimaybe it's missing something15:09
loolndec: At so it's a PBCAK15:10
* berco damned. corkscrew killed and everything goes back to normal15:15
=== lilstevie|ZNC is now known as lilstevie
=== rgreening_ is now known as rgreening
mpoirierGrueMaster: good morning15:51
GrueMastermpoirier: good morning.15:54
mpoirierGrueMaster: a couple of weeks ago you gave me a command that plays sound, alternating from one speaker to another.15:55
mpoirierI noted it somewhere but to save my life, can't find it anymore.15:55
GrueMasterspeaker-test -c 2 -t wav15:55
mpoirieryes, thanks.15:55
GrueMasterThis will play directly through the hardware.15:56
GrueMasterNo pulse.15:56
mpoirierI'm getting *nothing*...15:57
GrueMasterOn which platform?15:57
mpoirierI'm panda, I'm testing the new patches sent by Irg15:58
ogrampoirier, they are not complete15:58
ograand you also need to make sure to have no state file15:58
mpoirierstate file ? please expand.15:59
GrueMaster/var/lib/alsa/asound.state I believe.15:59
mpoirierhold on, checking.15:59
ograright, you need to disable its creation15:59
ograelse you wil always get the same state15:59
mpoirierindeed I have such file.16:00
mpoirierhow to I prevent its creation ?16:00
mpoirierIrg: good morning - I have your patches live on my card.16:02
ografor the init file you need to call alsactl init manually after boot16:02
lrgmpoirier: morning16:03
mpoirierogra: i did.16:03
mpoirierIrg: is there something you'd like me to check ?16:03
mpoirierIrg: I have all 4 patches, the new 00main and omap4.16:04
lrgmpoirier: ok, have you ran alsactl init ?16:04
mpoirierIrg: I also manually ran alsactl init16:04
lrgok, does audio work at this point16:05
mpoirierI tried "speaker-test -c 2 -t wav" without results.16:05
mpoirierI get clear silence.16:05
lrgok, what does cat /proc/asound/cards show ?16:05
mpoirierthat's the first thing I checked and it seems right to me.  Hold on.16:06
mpoiriermpoirier@panda:~$ cat /proc/asound/cards16:06
mpoirier 0 [Panda          ]: OMAP4 - Panda16:06
mpoirier                      TI OMAP4 Board16:06
lrgok, cool :)16:06
mpoirieryes, at least that works.16:06
mpoirierany sound control you're interested in ?16:06
ograthat lookd fine16:07
lrgwhat about aplay -Dplughw:0,0 -f dat /dev/urandom16:07
mpoirierIrg: yes, success.16:07
lrgok, the ABE uses some non standard formats16:08
mpoirierplease expand.16:08
lrgeverything has to go via alsaplugins atm16:08
lrgwell it uses S32_LE instead of S16_LE for PCM format16:08
mpoirierIrg: educate me: what makes you say that ?16:09
* rsalveti lunch16:10
lrgmpoirier: it's all in the TRM16:10
mpoirierIrg: ok, I'll check it out.16:11
lrgthe plugin converts wav S16_LE to ABE S32_LE16:11
lrgfor playback16:11
lrgand vice versa for capture16:11
mpoirierIrg: should I understand my test results aren't a surprise to you ?16:13
ogralrg, i dont think your omap4 file can work16:13
ogralrg, as i understand, the init files need to use dB values, you use absolute numbers everywhere16:13
lrgogra: iirc, thay can use both since some drivers do not export dB info16:14
ograwell, when i tested, absolute values didnt work16:15
ograbut i'm happy to be proven wrong :)16:15
lrgogra: ok, I'll ping Abraham as he craeted the file16:15
ogradid he test it ?16:15
lrgI did, on SDPand it worked for me16:15
ograthen it shoudl work on panda too16:16
lrgI'm pretty sure I removed the old state file too - but it was quite late at night though16:16
ograyou also need to disable its recreation16:16
ograit comes back with every boot and is executed immediately on device init by a udev rule16:17
ogramove /etc/init/alsa-mixer-save.conf to /etc/init/alsa-mixer-save.conf.old16:17
lrgogra: ok thanks, I'll check this after some food16:18
lrgabduenas: morning Abraham16:18
lrgabduenas: ogra has some questions about the config file16:19
ograabduenas, hey, i didnt get the init file to work when i used absolute numbers and had to use dB instead16:19
lrgabduenas: I have to go, will be back in 30m16:19
abduenashmm really? it work for me with those numbers16:20
ograback last week when i created the first version of it which is attached to the bug16:20
ograliam said it works for him too16:20
ograso it was probably a temporary glitch on my side16:20
ograjust wanted to know if you are sure one can use absolute numbers too16:20
abduenasyes i used to manually set to zero then alsactl init and then check the status on alsamixer it was resotred properley16:22
* ogra takes a break too and will then test the alsa-lib fix16:22
abduenasno prob... i can try your dB also if you want16:22
ograi havent tested the new file yet16:23
ograso dont worry16:23
berco pushed my source package 30min ago and since then launchpad says it will start in 1min... is it blocked in some ways?16:25
=== zyga_ is now known as zyga
bercoabduenas: haven't tested yet your file. still busy w/ today release16:35
abduenasberco: no prob16:38
oOSkarI need little help on Ubuntu on my Beagleboard16:49
oOSkarI don't have any username-password after using https://wiki.ubuntu.com/ARM/OMAPMaverickInstall16:49
oOSkarsystem boots perfectly but i'm blocked at login16:49
GrueMasteroOSkar: Which image did you use?16:50
oOSkarI used ubuntu-netbook-10.10-preinstalled-netbook-armel+omap.img.gz16:51
GrueMasterDid it run the firstboot scripts that resize the root partition?  It should have been on the display.16:52
oOSkari did not see any firstboot script16:53
oOSkardisplay was : orange (beagleboard boot) then black (linux booting on serial) and finally ubuntu purple login16:53
oOSkari've rewritten the SD card to try again but still no script running16:54
GrueMasterReboot but this time hold down the user button while the board resets.  This will load the boot loader from the sd card.16:54
GrueMasterYou should not see an orange background while this boots.16:55
oOSkarno change16:57
GrueMasterIs this a beagle C4?16:57
GrueMasterNote that this image will be at the edge of what can run on that board.  256M memory is very limiting.16:58
=== kmargar is now known as markos_
GrueMasterThere are special directions for C4 on the wiki, below the BeagleXM & Panda instructions.16:59
GrueMasterYou need to have a serial console open to run a uboot script.16:59
oOSkari have the serial console17:00
oOSkari don't see any instruction for C4 on the wiki17:00
ogra_acit falls under "older beageboards"17:01
GrueMasterLook for the "older beagleboards" section.17:01
oOSkar(note that this step is only necessary if you have a NAND and the system does not default to reading boot.scr)17:01
oOSkarhum, i may have read it too fast17:01
oOSkaras the system was booting i tought it would have been unnecessary17:01
ogra_acthe image wont be much fun on an old C417:02
oOSkarit's not that old17:02
ogra_acit will swap a lot and be reather slow17:02
oOSkarbut i agree that boot is quite slow17:03
ogra_acthe images are built for HW with 512M17:03
oOSkari've used debian on it previously but wanted to try some more hardware17:03
ogra_acit will run indeed but dont expect speed17:03
oOSkarbut i can remove a lot of the stuff from that ubuntu version manually17:03
ogra_acif you are patient enough to finish the install :)17:04
oOSkarscript is running now, thx :)17:04
ogra_acits all graphical17:04
oOSkar"setting up swap" oO17:05
ogra_acyeah, that will take a while17:05
oOSkari don't want swap on a 4Go microSD card :/17:05
ogra_acyou do17:06
ogra_acelse the installer wont run17:06
oOSkarthat's quite a strenge choice17:06
oOSkaryou mean that the installer would need more than 256Mo of RAM ?17:07
ogra_acthe image is built for beagle XM or panda boards17:07
GrueMasterX and gnome use a lot.17:07
ogra_acwell, actually they dont once everything is up17:07
ogra_achtop on my ac100 here shows only 162M used17:08
ogra_acand i have firefox and xchat open17:08
oOSkardid i read "ac100" ?17:09
oOSkari was thinking about buying one17:09
oOSkarbut still hesitating vs tablets17:09
ogra_acwhat would you do with a tablet ?17:09
* ogra_ac really doubts their usability if yuo actually want to type etc ....17:10
oOSkarsurfing, email, programming, sys admin over ssh17:10
ogra_acprogramming ?17:10
ogra_acon a virtual kbd ?17:10
oOSkarwith usb keyboard they might be quite usable no ?17:10
ogra_acyeah, that might work if you like to carry around a stand to put the tablet on17:10
oOSkarwhat do you think about the ac100 ?17:11
oOSkareverything working on ubuntu ?17:11
ogra_acbut before i carry a USB kbd plus a tablet plus a stand i rather use an ac100 :)17:11
ogra_acno, not everything17:11
ogra_acbut getting there17:11
ogra_acthe only kernel they provide is 2.6.29 so much stuff isnt as gooc supported as it could be17:12
ogra_acand things like powermanagement are completely done by a proprietrary tool17:13
ogra_acsame goes for sound17:13
ogra_acits nvidia after all17:13
oOSkarthat's sad17:14
ogra_acwell, its good enough for working17:15
GrueMasteroOSkar: If you want to customize the beagle image, you could use rootstock or you could try installing from the net with the untested netboot install image from http://ports.ubuntu.com/dists/maverick/main/installer-armel/20100211ubuntu29/images/omap/netboot/omap/17:15
oOSkarthx GrueMaster ;)17:16
oOSkari finally see the config graphic screen17:16
oOSkaryou would have guessed it, it's quite very very slow17:16
ogra_acGrueMaster, btw, there is a shorter url for netboot now :)17:16
GrueMasterogra_ac: That lists links to what I just pasted.17:17
oOSkarthx for these links, look quite interesting :)17:18
oOSkardo you know where i can change the bootargs environment variable ? is it on the SD card or using the bios from the beagleboard ?17:23
GrueMasterIt is in the boot.scr on the first partition.  You can change it by editing the /boot/boot.script on the rootfs and running sudo flash-kernel.17:24
GrueMasterTHis is the easiest method.17:25
oOSkarok thx17:25
oOSkar@ogra_ac : do you still have good autonomy without power management ? is there only sound missing ?17:26
ogra_acwell, under android the ac100 has about 7h, on ubuntu its 4-5 depending on what i do17:27
oOSkarand how is the android ? still usable for a hacker ?17:28
ogra_achavent tried how long it lasts with wifi off and 3g on yet17:28
ogra_acandroid is completely unusable imho17:28
ogra_acits a great phone OS but should stay there17:29
ogra_acmight be usable on tablets but surely not for netbooks17:29
ogra_actry an ac100 at a dealer and you will understand :)17:30
oOSkarif i find one ^^17:30
=== sumitsemwal is now known as sumits
oOSkari've seen the videos and it looked strange, but wasn't sure17:30
oOSkarthanx for all your advices ;)17:39
oOSkarsee ya !17:39
GrueMasterogra_ac: http://www.engadget.com/2010/10/13/specs-released-for-advent-vega-the-249-android-tegra-tablet/17:42
ogra_aclooks like an ac100 without kbd :P17:44
bercoogra_ac: what's wrong with the build machines?17:57
bercoI'm waiting for more than 2 hours...17:57
ogra_acberco, ask #soyuz or #launchpad17:58
bercoogra_ac: thx17:58
=== JaMa|Wrk is now known as JaMa
=== XorA is now known as XorA|gone
ogralrg, lag, berco ... (and who else is involved in panda sound) i tested the alsa-lib fix and updated bug 637947 ... please someone provide a kernel package with the other changes so i can test the other init file19:28
ubot2Launchpad bug 637947 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "no sound devices on current ES2.0 boards (affects: 2) (dups: 1) (heat: 26)" [High,Confirmed] https://launchpad.net/bugs/63794719:28
* ogra vanishes into his evening again19:28
rsalvetiGrueMaster: can you check the x-loader build date from the released image?19:30
GrueMasterSure in a few minutes.  (currently hobbled with bad knee).19:31
rsalvetiouch, np19:31
rsalvetiogra_ac: at http://cdimage.ubuntu.com/ubuntu-netbook/ports/releases/10.10/release/, preinstalled USB image19:33
rsalvetiit should be sd image, I believe19:34
ogra_acrsalveti, yeah19:34
ogra_acwe need to fix that for natty19:34
ogra_ac.img files are by default set to USB image in the scripts19:35
rsalvetiogra_ac: oh, ok19:35
rsalvetiogra_ac: and when are we getting natty images?19:35
ogra_aci want an SD card icon too :)19:36
ogra_acnot sure19:36
rsalvetiyup :-)19:36
ogra_acthey wont be usable anywy before alpha 119:36
rsalvetitrue, but just to know19:36
ogra_acso dont worry about image now19:36
ogra_acwe'll start building them after UDS19:36
ogra_acas soon as there is an archive :)19:37
ogra_actoolchain and basic packages usually take a week to ten days19:37
ogra_acthe sound seems to sort itself now too19:39
ogra_acseems we have all bits19:39
rsalvetinice, still need to read the backlog and emails about it19:40
ogra_acwe had a very fruitful meeting yesterday thanks to lag :)19:40
ogra_ache got all involved parties together19:40
ogra_acand even more imprtant, the issue with alsactl init was actually a alsa-lib bug19:41
ogra_acfix is already pending for SRU19:41
rsalvetioh, so it was really a bug :-)19:41
ogra_acBug #65203519:42
ubot2Launchpad bug 652035 in alsa-lib (Ubuntu) "libasound2 not finding usb sound card (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/65203519:42
GrueMasterrsalveti: Texas Instruments X-Loader 1.41 (Oct  6 2010 - 17:27:48)19:44
GrueMasterI don't think that is the right one.19:45
lagI'll post my email onto the LP bug so people not involved can read19:45
lagGive me a moment19:45
ogra_aclag, no hurry19:45
ogra_acwould be good to get lrg's patches too there19:46
lagThere's nothing stopping you ;)19:46
ogra_acyeah, there is alwas tomorrow ... i wont go upstaris again today, testing the alsa-lib stuff was enough at that time19:46
rsalvetiGrueMaster: cool, thanks19:47
rsalvetiGrueMaster: that should be the one19:47
GrueMasterI'm redownloading from the above link for comparison.19:47
rsalvetifor some reason I flashed the wrong image19:47
lagrsalveti: bug 63794719:47
ubot2Launchpad bug 637947 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "no sound devices on current ES2.0 boards (affects: 2) (dups: 1) (heat: 26)" [High,Confirmed] https://launchpad.net/bugs/63794719:47
rsalvetiGrueMaster: for some reason when I got the rc instead of the release when using zsync19:50
rsalvetiprobably lack of coffee19:50
rsalvetiso I was surprised when I saw x-loader from september hehe19:50
ogra_acyeah, we need to make sure the coffee is compiled in in the future so you dont need it additionally19:51
ogra_acsounds like it would depend on QT19:52
rsalvetiqt doesn't taste good19:53
mpoirierogra_ac: I just read your email.19:55
mpoirierogra_ac: you need a kernel with the 4 patches suplied by Irg ?19:55
ogra_acmpoirier, yeah, tomorrow though19:56
ogra_acwont test today anymore19:56
ogra_acbut his patch names the card based on the board19:56
mpoirierogra_ac: yes,19:56
ogra_acto test the switching init file we need the name19:56
mpoirierogra_ac: cat /proc/asound/cards does yield panda19:57
lagmpoirier: I'm assuming you mean lrg? :)19:57
mpoirierlag: Irg yes19:57
lagNo lrg19:57
ogra_aclike a small L19:58
mpoirierogra_ac: are you looking for a .deb ?19:58
* GrueMaster hobbles back out to the couch where laptop & ice pack awaits.19:59
ogra_acmpoirier, right19:59
mpoirieryou'lll have a link in you inbox tomorrow.19:59
mpoirierI'll get to it by the end of my day.19:59
ogra_acsweet !20:00
ogra_acthanks a lot20:00
=== ian_brasil_ is now known as ian_brasil
sguiriecmpoirier:if you want to have more data about OMAP4 ABE paths you can look at the next link  (http://www.omapzoom.org/wiki/File:Omap_abe.jpg http://www.omapzoom.org/wiki/File:ASoC_OMAP4.jpg). first one is mainly Audio Back End path. The second one is mapping the kernel audio controls on the graph.21:02
sguiriecmpoirier: I hope it will be the missing part for your understanding21:03
mpoiriersguiriec: this is very cool thanks - at first glance there is a fair a mount of things that are new to me.21:04
sguiriecmpoirier: OMAP4 audio is very different from OMAP1/2/3. ABE and McPDM introduce new functionalities.21:05
mpoirierya - someone else (can't remember who) in TI told me the same.21:06
sguiriecmpoirier: May be berco21:07
mpoirierprobably yes.21:07
mpoirierhe was the first person to enlight me in this matter21:07
lagmpoirier: I didn't think you were working on this anymore?21:11
sguiriecmpoirier: I am still working and helping on the driver21:12
mpoirierlag: I was waiting for that info - now I can move forward.21:12
mpoirierlag: it has a lot of the missing peices we were talking about.21:12
lagI was under the impression that all the outstanding tasks have been already issued21:17
lagWhat needs to be done that I am unaware of?21:17
lagmpoirier: ?21:19
mpoirierlag: yep21:19
mpoirierlag: sorry - I had not seen your post.21:20
mpoirierlag: everything has been covered.21:20
mpoirierI peeled the sdp4430.c and twl6040.c a couple of weeks ago.21:20
mpoirierbut there was something missing,21:21
mpoirierthose two files we setting forth the BE and FE but again, the path between them was incomplete.21:22
mpoirierI was specifically interested to know how berco came up with the setting of the alsamixer.sh file21:22
mpoirierthat file was bridgin ABE, BE and FE together.21:22
mpoirierI was missing part of the puzzle.21:23
sguiriecmpoirier: Nomally should be more clear for you now.21:23
mpoiriersguiriec just painted the missing part.21:23
mpoirierlag: there was a grand scheme I couldn't see - and you could not have deduce the information from the TRM or the schematics.21:24
mpoirierdrove me nuts.21:24
lagSo it was more a quest for information21:25
mpoirieryes, 'cause we are bound to do this again...21:25
mpoirierwhenever another omap4 board gets spun off.21:25
lagI don't believe that to be the case21:26
mpoirierthe more we know, the better we can point out problems and work with TI.21:26
lagWell ...21:26
lagSo, are you happy that all the tasks have been addressed for Panda -> Maverick?21:27
mpoirierI have nothing more to requets21:27
lagTry not to spend any more time on it, testing not withstanding21:28
lagIf you believe we have missed something, let me know and I will endeavour to address it asap21:29
=== III is now known as orbarron
=== JaMa is now known as JaMa|Zzzz
=== jcrigby_ is now known as jcrigby

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