**** BEGIN LOGGING AT Mon May 19 02:59:57 2008 May 19 07:08:05 * * OE Bug 4277 has been created by lynn.lin(AT)avocent.com May 19 07:08:07 morning May 19 07:08:07 * * new package of dietlibc May 19 07:08:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4277 May 19 07:44:03 hi all someone here play with ep93xx ??? May 19 07:44:51 i'm going to fetch ts7260 today from the customs May 19 07:57:56 gm May 19 08:22:59 gremlin[it]: yes, using OE on ts7250 May 19 08:30:58 kario, where i can find latest ep93xx patch? on line May 19 08:32:48 you can try the unofficial technologic systems yahoo group May 19 08:33:19 http://tech.groups.yahoo.com/group/ts-7000/ May 19 08:33:46 kario, thanks! May 19 08:34:46 there are also some patches available in the oe dev branch for ep93xx May 19 08:34:58 kario, ynezz i suppose no one of you use framebuffer and/or touchscreen May 19 08:35:16 nope May 19 08:35:20 kario, thnaks now i pull latest dev to check ... May 19 08:35:45 i have touble with the usb host controller and a FTDI rsb-serial converter ... :S May 19 08:37:06 http://tech.groups.yahoo.com/group/ts-7000/message/11062 May 19 08:37:18 "USB serial kernel module bugs" thread May 19 08:38:29 gremlin[it]: no, I'm not using FB/TS May 19 08:40:10 mhh thanks ... but i'm using a FTDI chip ... the bug seem quite strange ... bany urb error sometimes on changing RTS signal pin and some strange behavior sending burst of serial packets ... May 19 08:44:00 mhh about OE for ep93xx ... you use soft or hard FPU setup ??? May 19 08:55:56 hi all May 19 08:56:07 hi May 19 09:05:13 hi May 19 09:05:13 moin moin May 19 09:32:42 good morning May 19 09:40:44 morning May 19 09:41:08 yop hrw :) May 19 09:44:56 * hrw just bricked one device May 19 09:48:30 Morning all May 19 09:48:57 hrw, thanks for the alix kernel fix :) May 19 09:51:45 Notenliste_TGdI1_WS0708.ods May 19 09:51:50 ups May 19 09:51:52 sry. May 19 09:52:14 Gadget-Work: np May 19 09:52:21 Gadget-Work: how goes alix2 use? May 19 09:53:59 hrw, slowly, just getting into it, using WRAP with m0n0, but want to use alix2 to do some specific linux stuff. May 19 09:54:33 OE is all a bit new to me so still getting to grips. May 19 09:54:51 I've got an image, just need to actually test it :) May 19 09:57:23 Gadget-Work: I think that we need to enable more modules for alix May 19 09:57:29 wifi ones etc May 19 09:58:28 hrw, sounds good. Although for me, i2c, and USB<->serial is all I'm really interested in May 19 10:05:32 provide patch? May 19 10:06:24 hehe, not got that far into it yet :) May 19 10:29:05 hello May 19 10:29:24 hi jeremy_laine May 19 10:29:53 hi hrw! May 19 10:30:07 I have made some progress on the "oestats" build statistics tool I'm working on May 19 10:30:43 see http://oestats.jerryweb.org/ for the updated demo May 19 10:31:11 now tasks are grouped for a given package / version / revision and details are collapsed May 19 10:31:25 you can get the details by clicking on the status (failed or succeeded) May 19 10:46:18 jeremy_laine: impressive May 19 10:46:59 jeremy_laine: 'show log' function planned or not? May 19 10:47:41 jeremy_laine: is there something more then one class needed to use it? May 19 10:57:12 gremlin[it]: the .config is in OE tree also May 19 10:59:42 hrw: sorry, I'm back now May 19 11:00:01 hrw: the logs are only uploaded when a task fails for now May 19 11:00:15 jeremy_laine: nice May 19 11:00:17 hrw: but we can have a config variable for this easily May 19 11:00:27 jeremy_laine: OESTATS_BUILDER can contain spaces? May 19 11:00:32 hrw: take a peek at oestats-client.bbclass May 19 11:00:53 hrw: ehm.. it should do, I might need some URI escaping on the server side May 19 11:01:42 hrw: (for the case where you ask for a specific builder as /builds/builder/) May 19 11:02:00 lets check then May 19 11:02:31 hrw: for your tests you can use oestats.jerryweb.org, after that you can always set up your own server, it's a snap May 19 11:02:39 sure May 19 11:03:48 INHERIT += "oestats-client" May 19 11:03:48 OESTATS_SERVER = "oestats.jerryweb.org:8080" May 19 11:03:48 OESTATS_BUILDER = "Hrw_Poky_build" May 19 11:03:49 ok? May 19 11:04:02 hrw: nope, drop the port from OESTATS_SERVER May 19 11:04:24 ok May 19 11:04:54 hrw: it would be cool if someone could patch bitbake to report when a build fails May 19 11:05:15 report on ml? May 19 11:05:29 hrw: currently, at the "build" level we get BuildStarted / BuildCompleted events but no BuildFailed May 19 11:05:48 hrw: unlike for tasks where we get TaskSucceeded and TaskFailed May 19 11:05:49 jeremy_laine: http://oestats.jerryweb.org/builds/builder/4/ - thats with Poky not OE May 19 11:06:29 hrw: hm, the METADATA_REVISION doesn't get reported May 19 11:06:50 nor tasks info May 19 11:06:57 hrw: no packages are showing up, weird May 19 11:06:59 this build ended already May 19 11:07:48 hrw: could you make sure your OE is up to date? May 19 11:08:31 hrw: oestats-client changed a bit over the weekend, and I see your requests don't use the latest version (there is no User-Agent: oestats-client/0.1) May 19 11:08:56 will clear out your builds in the meantime May 19 11:09:16 phone May 19 11:09:42 hrw: going out to grab food, back soon May 19 11:10:24 ok May 19 11:10:32 updating then May 19 11:21:57 hrw: ah looks better :) May 19 11:22:17 hrw: still a problem with METADATA_REVISION, what does bitbake report when you run the build? May 19 11:22:33 re May 19 11:23:04 * * OE Bug 4278 has been created by elena.valhalla(AT)gmail.com May 19 11:23:06 * * BB for enhanced ctorrent and its controller May 19 11:23:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4278 May 19 11:26:47 ah.. right May 19 11:26:55 as this is git-svn May 19 11:31:24 hrw, is there an easy way to see what's supported with the alix platform in OE ? ie where should I be looking for resolving say i2c ? May 19 11:31:29 and this is shown but not with htmlentities() May 19 11:32:53 http://oestats.jerryweb.org/builds/build/10/ May 19 11:41:14 jeremy_laine: can packages be sorted not alphabetically but by time of their first task? or let it be sortable by clicking on headers? May 19 11:42:44 hrw: ok will look into it May 19 11:43:39 jeremy_laine: now it shows what is/was cooking but no idea about what when May 19 11:43:52 jeremy_laine: with large builds it can show nothing May 19 11:44:03 hrw: well you do get the timestamp for each task May 19 11:44:23 jeremy_laine: when I click on status to open info May 19 11:45:08 http://oestats.jerryweb.org/builds/build/10/ shows that I have lot done/in progress but not what is going on ~now May 19 11:49:05 jeremy_laine: also summary like "X packages started, Y do_fetch done, Z do_patch done" etc would be handy May 19 11:50:57 jeremy_laine: I would also revert order of tasks to make them chronological May 19 11:59:31 hrw: ok for the order of the tasks, have reversed them May 19 12:03:48 jeremy_laine: if package got to BB_DEFAULT_TASK will it be marked as completed? May 19 12:04:28 hrw: I didn't know BB_DEFAULT_TASK May 19 12:04:42 jeremy_laine: http://oestats.jerryweb.org/builds/build/9/ say that automake-native succeeded but it also say that for packages which are between do_fetch and do_patch May 19 12:04:52 jeremy_laine: so now you know;) May 19 12:04:54 hrw: the answer is no in any case, there is no "package" structure as such in the database, just a collection of tasks May 19 12:05:22 hrw: I group the tasks if they have the same PN / PR / PV and build id May 19 12:05:42 hrw: and I put "succeeded" if all tasks succeeded, "failed" if one or more failed May 19 12:06:02 jeremy_laine: export BB_DEFAULT_TASK and then: if package[list of tasks].contains(BB_DEFAULT_TASK) then mark package as completed. May 19 12:06:02 for a failed example, see: http://oestats.jerryweb.org/builds/build/15/ May 19 12:06:35 hrw: (notice how the tasks are expanded for the failed "package") May 19 12:06:44 hrw: (and the link to the log) May 19 12:06:50 noticed May 19 12:07:17 jeremy_laine: change background of failed package would be nice too May 19 12:07:57 jeremy_laine: go to build/15/ and click on 'failed' to hide infos. that line should be marked somehow more visibly then just 'failed' May 19 12:09:37 hrw: how's the package coloring now? May 19 12:13:27 jeremy_laine: I would not change succeeded ones May 19 12:13:59 to keep it sane May 19 12:17:30 hrw: ok done May 19 12:17:37 jeremy_laine: http://oestats.jerryweb.org/builds/builds/ is linked from http://oestats.jerryweb.org/builds/build/10/ but is 404 May 19 12:17:44 good morning everyone May 19 12:18:37 Laibsch: thanks, fixed May 19 12:23:12 Cool, thanks May 19 12:24:40 jeremy_laine: remove 'width:70%' from task-groups May 19 12:25:14 Is this http://oss.leggewie.org/nas/ OE-material? I don't see a serial port for flashing, but first of all, I wonder what kind of CPU this is. May 19 12:25:43 I'll be happy with it just as a NAS, but if it can be hacked, it has to be done ;-) May 19 12:26:28 brb May 19 12:26:35 * Laibsch thinks about adding ext2 support for example, which is probably lacking May 19 12:33:47 jeremy_laine: on 'build page' remove buildtime and builder from table - they are already present on top of page May 19 12:37:40 Laibsch: that's a crippled x86 AFAIK May 19 12:38:02 hrw: how's the sorting of the packages now? May 19 12:38:24 jeremy_laine: http://oestats.jerryweb.org/builds/package/hostap-utils/ - where is do_fetch? May 19 12:42:07 jeremy_laine: Thanks for taking a look. Crippled in what way? And what part exactly made you think it was an x86? Do you think there is any way to flash or reprogram this thing? May 19 12:43:18 jeremy_laine: Du sprichst Deutsch? ;-) May 19 12:43:58 jeremy_laine: headers on build page... May 19 12:44:00 jeremy_laine: that is a django app? May 19 12:44:28 jeremy_laine: Are you a DD or are your debian packages sponsored? May 19 12:44:29 zecke: yes May 19 12:44:49 Laibsch: yup I'm a DD, why? May 19 12:45:29 I have long been looking for a sponsor for some packages May 19 12:45:34 He's after peopel to sponsor packages for him :) May 19 12:45:41 Recently, broonie was kind enough to help me May 19 12:45:48 hrw: ok, have removed builder / build id in the build page May 19 12:45:54 broonie: Not anymore, I already found you ;-) May 19 12:46:23 And I don't intend to have a ton of packages, just a few I like and which still lack from debian May 19 12:46:45 Laibsch: what packages are you thinking of? May 19 12:48:24 well, we just got pastebinit ready. And I hope it will progress through ftpmasters soon. It is a cli pastebin client. It is already in ubuntu and I use it for pasting stuff when logged in over ssh. Very handy and afaik nothing like it in debian, yet. All pastebin clients require X, AFAIK. May 19 12:49:27 Another one is grecipe-manager.sf.net, a mealmaster-compatible recipe collection manager. That one will require a little bit more work. May 19 12:49:39 Generally, it can come in handy at times to know a DD ;-) May 19 12:50:30 zecke: yep it's django, see http://opensource.bolloretelecom.eu/projects/oestats/ May 19 12:50:41 jeremy_laine: is there a limit of processed tasks per build? I see ncurses/do_install in my build but not on build page May 19 12:50:45 cool May 19 12:51:13 hrw: there shouldn't be, maybe there's a bug in my task grouping function May 19 12:52:32 jeremy_laine: how much mem/cpu does it takes? May 19 13:02:51 hrw: not sure of the exact mem / cpu use, I have a bunch of other sites on the same machine May 19 13:03:17 hrw: but my server has 512MB RAM and the load average is really low May 19 13:03:40 hrw: can you grep for "disabling stats" in your build output? May 19 13:04:15 hrw: if at any point the server cannot be reached I disable stats so as not to slow down the build, and I've restarted apache a bunch of times in the past hour May 19 13:09:25 I stopped and restarted then May 19 13:31:51 c7x0 in qvga is strange May 19 13:35:24 Laibsch: it was me saying that chip is x86, not jeremy. May 19 13:35:49 Yes, I just saw that May 19 13:35:53 Laibsch: I have come across a wiki page once that specified in what way that chip manufacturer is doing (crippled) x86's May 19 13:36:03 Laibsch: haven't bookmarked it so it seems... May 19 13:36:18 :-( May 19 13:37:09 likewise: I just found an option in the HTML interface of this thing (NAS for the price of an external HD case!!) to upload new firmware May 19 13:37:15 so maybe I am in luck May 19 13:37:35 Laibsch: yes, the chip seems *very* cheap but with a reason May 19 13:37:37 That would be the coolest thing May 19 13:37:57 Well, I was only looking for an external HD case May 19 13:38:14 I already got a NAS server free on top May 19 13:38:38 If that NAS server turned out to be hackable, I guess that would indeed be asking too much ;-) May 19 13:38:52 I don't mind the cheap chipset as long as it does not eat my data May 19 13:38:52 Laibsch: what does it do by default? May 19 13:38:57 samba + http? May 19 13:39:00 Laibsch: http://www.devtools.com/rdc.htm bottom of page May 19 13:39:03 samba + ftp May 19 13:39:15 http only for admin May 19 13:39:26 it also has a built-in DHCP server May 19 13:39:40 Laibsch: http://www.rdc.com.tw May 19 13:39:42 I guess every LAN device has that these days ;-) May 19 13:42:39 Laibsch: Your blog is missing its database... May 19 13:43:07 florian: indeed it is May 19 13:43:12 thanks for pointing that out May 19 13:43:12 bitte, Laibsch May 19 13:43:20 Laibsch: yw May 19 13:43:56 jeremy_laine: build finished May 19 13:44:10 hi, how can I add a keyboard layout to OE image building? May 19 13:44:16 florian: apparently, the mysql server had died May 19 13:44:31 or maybe it did not come up after the reboot yesterday May 19 13:45:54 * florian is happy about software that does not need a database server :) May 19 13:46:13 understood May 19 13:46:33 * Laibsch is happy about software that is nice to look at and does not need too much twiddling May 19 13:52:44 likewise: If I understand correctly, I have the R2882-G chip, so http://www.paradigmtools.com/docs/R2880.PDF from http://www.devtools.com/rdc.htm would be the closest match, right? May 19 13:53:18 And I am probably right in assuming that this thing in running on Linux? I don't think there is room for the royalties of a proprietary OS. May 19 13:53:37 henomis79: what kind of keyboard map? May 19 13:54:12 Laibsch: that's what I found in a few seconds. May 19 13:54:25 Laibsch: more googling might bring up the specific info May 19 13:54:26 zecke: it-it May 19 13:56:07 henomis79: kernel keymap? xkb? May 19 13:56:19 zecke: xkb May 19 13:56:31 likewise: Thanks for helping me identify the chip May 19 13:57:04 There is indeed quite a bit of info to dig through on the net once the chip has been identified May 19 13:57:36 Hey Guys, I've a Problem with Crosscompiling apache2 on a ubuntu 8.10 for slugos 4.8 May 19 13:57:42 it fails with: http://pastebin.com/m4a9d2cf9 May 19 13:57:46 can anybody help me? May 19 13:57:56 henomis79: do you know if it is packaged in any package? May 19 13:58:22 hrw: I added a recap of the packages built so far: http://oestats.jerryweb.org/builds/packages/ May 19 13:59:06 http://oestats.jerryweb.org/builds/package/ is not nice page May 19 13:59:31 hrw: crap broken links.. May 19 13:59:32 jeremy_laine: add letters on top and group packages by letter there May 19 13:59:45 jeremy_laine: it is not broken link but link done by me May 19 14:00:04 hrw: no the links in the "package" view are busted, need to fix May 19 14:00:06 jeremy_laine: /package should redirect to /packages/ May 19 14:00:27 aargh... 53M for rootfs hit me ;( May 19 14:02:17 http://oestats.jerryweb.org/builds/package/gcc-cross/ does not link to builders May 19 14:02:31 or builds May 19 14:04:01 zecke: no i don't May 19 14:04:53 henomis79: find that out :) May 19 14:07:03 hrw: it does now May 19 14:07:53 "OE build stats" should link to / May 19 14:09:07 hrw: actually I'm kind of wondering whether I'm not going to get rid of /, everything now lives under /builds May 19 14:10:48 rather get rid of /builds/ May 19 14:11:26 ROOT/builds ROOT/packages ROOT/builders May 19 14:19:05 no way to set up locale + keymap through IMAGE_LINGUAS? May 19 14:22:12 hrw: ok, am working on the URLs so the stats may be down for a bit May 19 14:25:17 henomis79: no way May 19 14:37:57 looks like this is a very common chip and I might be in luck with alternative firmware May 19 14:37:58 http://www.aroundmyroom.com/tag/landisk/ May 19 14:53:01 Laibsch: what NAS device is that? May 19 14:54:04 that rdc stuff sounds familiar but I think it was not linux May 19 14:54:55 Jin^eLD: I think http://www.sysopt.com/features/storage/article.php/12033_3622186_3 is the one I've got May 19 14:55:27 at least very similar May 19 14:55:34 There are tons of clones, it seems May 19 14:56:31 ah, I thought about a different one May 19 14:56:40 http://www.pyrogate.de/catalog/product_info.php?products_id=62 seems looks exactly like the one I have May 19 14:56:47 Laibsch: nas-central.org is a good place to look for GPL sources May 19 14:56:59 Nice, thanks May 19 14:59:10 kgilmer: ping May 19 14:59:21 Laibsch: let me know once you did a cat /proc/cpuinfo ;) I'm trying to provide static binaries of my UPnP server as an addon to existing firmware for various devices May 19 14:59:54 That would be nice May 19 14:59:58 :-) May 19 15:00:00 bye May 19 15:00:22 Jin^eLD: Not sure if the images I found so far provide a shell May 19 15:00:43 Laibsch: usually you will find serial connectors or at least points for soldering on the baord May 19 15:00:50 then you just need the RS232 converter May 19 15:00:56 uart to serial May 19 15:01:11 or whatever these things are called May 19 15:01:15 * Laibsch has VERY little experience soldering May 19 15:01:22 But I'll look around May 19 15:01:38 Laibsch: me too :) if you are lucky the pins may be there already, then you just need the right connector May 19 15:01:46 It looks like this is hackable and if it is it seems common enough that somebody else has probably done it already May 19 15:01:53 most probably May 19 15:02:03 Jin^eLD: Have you looked at the pictures? May 19 15:02:17 http://oss.leggewie.org/nas/ May 19 15:02:26 yes, but I am not familiar with this one, the case looked like one of the units that I knew, but inside it was different May 19 15:02:48 ah these pictures, not yet, let me check May 19 15:03:02 I don't see a an obvious connector May 19 15:03:53 wow that's all? very small board May 19 15:04:29 are you sure it's running linux? May 19 15:05:37 no, I can't say for sure May 19 15:05:51 the board is the size of a PCI network card May 19 15:05:53 I also do not see anything that would look like a uart May 19 15:05:57 very similar dimensions May 19 15:06:16 rdc sounds familiar though, but rdc was not linux May 19 15:06:38 there are not much choices for linux anyway, all NAS reference hardware I have seen is either ARM9, MIPSel or PPC based May 19 15:06:38 Do they have their own OS? May 19 15:06:52 I have no idea May 19 15:06:55 nothing x86? May 19 15:07:14 not really, most NAS'es are one of the three above archs May 19 15:07:29 I think intel may have some NAS chipsets but I guess thats too expensive May 19 15:07:31 the price for this NAS will more or less exclude royalties for a separate OS May 19 15:07:42 which si why I thought chances are high it is Linux May 19 15:07:53 lots of NAS'es are based on Marvell Orion reference design which is ARM9 based May 19 15:08:10 well, let me dig through the mails May 19 15:08:14 http://www.paradigmtools.com/docs/R2880.PDF is probably the chip May 19 15:08:28 I once had contact to a user who hat some rdc based stuff as well, and it was not linux May 19 15:08:58 or at least a close cousing May 19 15:09:00 and likewise mentioned it was a crippled x86 May 19 15:09:15 well, I'll do some more reading and I'll find out May 19 15:10:31 can't find that mail May 19 15:39:41 I am seeing that while trying to build with eglibc its pulling in glibc-2.7 too may be angstrom needs a package that's not being provided by eglibc how can I find that dependency I tried bitbake -vvv -DDD -n May 19 15:54:30 03likewise 07org.oe.dev * rb1e59e39... 10/ (3 files in 2 dirs): rt-tests: Bump to release 0.21. May 19 16:18:42 03jeremy_laine 07org.oe.dev * r4b2e5f44... 10/ (1 classes/oestats-client.bbclass): oestats-client.bbclass: switch to new URL scheme May 19 16:27:01 hi all May 19 16:27:33 khem: any preferred revision for eglibc that you'ld recommend? I'm currently using the Angstrom sane revision May 19 16:28:21 khem: which is eglibc-2.8+svnr5963 May 19 16:37:39 Laibsch: do we need to build mysql-native for php-native? Can't PHP be configured without sql support? May 19 16:37:55 likewise: maybe it could May 19 16:38:08 I was running into trouble building php-native from scratch May 19 16:38:31 I am certain, there are still some inefficiencies in the DEPENDS for PHP May 19 16:38:56 In fact, I introduced a very minor one with my last commit, but I thought it did not warrant another commit to correct things May 19 16:39:02 Laibsch: It's just that, before we know, we are building a complete native distro before we can build for the target. May 19 16:39:11 likewise: Understood May 19 16:39:36 Laibsch: we only need the minimum of minimally configured tools. Do you know what brings in the SQL dependency? One of the PHP modules? May 19 16:39:37 I was more concerned with getting php-native to compile at all May 19 16:39:42 Laibsch: ok May 19 16:39:48 likewise: no, I have no idea May 19 16:40:04 hrw|gone: I have updated the URLs used by oestats, so you'll need to do an "mtn pull && mtn up", oestats-client.bbclass has changed May 19 16:40:13 likewise: Feel free to improve things May 19 16:40:46 likewise: when you do, drop the mysql-native DEPENDS in php.inc May 19 16:40:53 it certainly does not belong in there. May 19 16:41:22 Laibsch: I think I am improving things elsewhere (real-time linux kernel), lack of time.... :-) May 19 16:41:31 it is php-native that depends on mysql-native (I am also wary of the mysql DEPENDS in php.inc, but was too lazy to test/inspect) May 19 16:41:37 likewise: same here May 19 16:41:52 But since you raised the subject, I thought you were going to tackle it May 19 16:42:15 No, I just wondered when I was browsing the commits. No biggie, just wondered . May 19 16:42:27 yeah, it is suboptimal May 19 16:42:34 but it should work May 19 16:44:24 khem: if you have the time, could you look at this? I remember this issue being mentioned (by you?) once. http://www.pastebin.ca/1022885 May 19 16:51:18 a May 19 16:52:04 + May 19 16:55:18 likewise: http://permalink.gmane.org/gmane.comp.handhelds.openembedded.scm/3502 May 19 17:04:05 * * OE Bug 4279 has been created by  May 19 17:04:07 * * glib-2.0 fails do_compile May 19 17:04:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4279 May 19 18:01:05 * * OE Bug 4262 has been RESOLVED (FIXED) by May 19 18:01:07 * * The /usr/include/scsi headers are missing in some cases May 19 18:01:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4262 May 19 18:01:14 yay May 19 18:01:52 thanks for the patch, Tartarus May 19 18:02:20 03TomRini 07org.oe.dev * r54f92ebb... 10/ (25 files in 3 dirs): May 19 18:02:20 glibc/linux-libc-headers: make sure that scsi headers are available when needed. Fixes 4262. May 19 18:02:20 * commit by Laibsch May 19 18:04:26 Laibsch, will 3146 (and maybe 3880) be committed soon? I'm all happy about getting some of my stuff commited and closed :) May 19 18:04:40 !oebug 3146 May 19 18:04:41 * * Bug 3146, Status: NEW, Created: 2007-10-11 00:49 May 19 18:04:42 * * eha(AT)doredevelopment.dk: Samba smbfs and cifs packages May 19 18:04:43 * * http://bugs.openembedded.net/show_bug.cgi?id=3146 May 19 18:04:46 !oebug 3880 May 19 18:04:48 * * Bug 3880, Status: NEW, Created: 2008-02-25 20:58 May 19 18:04:49 * * trini(AT)embeddedalley.com: The samba package is not having a correct mount.smb/mount.smbfs link created May 19 18:04:50 * * http://bugs.openembedded.net/show_bug.cgi?id=3880 May 19 18:05:14 3146 still has open questions AFAIR May 19 18:05:23 so the same goes for 3880 May 19 18:05:34 I think Esben answered 'em last night for 3146 May 19 18:05:35 but you will have witnessed the recent activity in both May 19 18:05:39 Maybe May 19 18:05:42 and no ones asked me about 3880 yet May 19 18:05:46 I have not looked at them yet May 19 18:05:48 k May 19 18:06:07 I am going through the bugs with open patches May 19 18:06:19 and I sure would hope other devs joined May 19 18:06:30 I have already brought the count down from about 200 to 160 May 19 18:06:33 still too many May 19 18:06:35 good work May 19 18:07:34 * Laibsch once again slaps everybody around with http://bugs.openembedded.net/buglist.cgi?keywords=patch&resolution=--- May 19 18:13:29 Hi there. I am new to OE and the bitbake build system.. I am looking right now to get used to bitbake. but where should I look in OE website for instructions about how-to change the kernel .config that will be used when building de "generic" distro. thx May 19 18:13:41 is it visible in the bugtracker if patches in the bugtracker have been acked on the ml? May 19 18:14:01 hernan, that's not an easy question :) May 19 18:14:07 hernan, it starts with what machine did you set? May 19 18:14:12 x86 May 19 18:14:16 the machine conf file says what kernel recipe it wants to use May 19 18:14:18 the i486 one May 19 18:14:27 that in turn determines where it gets a defconfig May 19 18:14:29 ok.. i will ok for it.. May 19 18:14:33 didnt know the name "recipe" May 19 18:14:39 hernan, to modify (NOT to save these mods, just modify) May 19 18:14:52 hernan, try bitbake -c configure virtual/kernel then bitbake -c devshell virtual/kernel May 19 18:15:05 it will compile just the kernel? May 19 18:15:16 hernan, recipe == metadata file, ie the file in openembedded/packages/linux that sets down all of the rules May 19 18:15:37 hernan, the first command will setup the whole OE env up to the point of running make oldconfig the first time May 19 18:15:44 the 2nd will set you up to go and run menuconfig to change things May 19 18:16:16 ahhh , ok.. i need to get used to the recipe syntax.. but thanks alot May 19 18:19:32 hernan: Are you using a distro with deb packages? May 19 18:19:40 no May 19 18:19:45 the .stable from oe May 19 18:19:53 the host May 19 18:19:59 ubuntu/debian/... May 19 18:20:03 something like that? May 19 18:20:04 ubuntu May 19 18:20:06 8.04 May 19 18:20:10 good choice May 19 18:20:31 RP: ping? May 19 18:20:58 hernan: http://blog.leggewie.org/?p=39 May 19 18:21:24 Maybe that is interesting for you May 19 18:21:39 thanks May 19 18:24:06 * * OE Bug 3146 has been RESOLVED (FIXED) by May 19 18:24:08 * * Samba smbfs and cifs packages May 19 18:24:10 * * http://bugs.openembedded.net/show_bug.cgi?id=3146 May 19 18:24:28 i found a couple of mail threads and there were questions and answers like these: "i want to change my kernel config" and the ans: "why would you want to change your distross kernel config?" hehe May 19 18:24:49 i just want to change my nic driver from Module to builtin so I can pxe+nfs boot my itx May 19 18:25:56 Laibsch: is it somehow visible in the bugtracker if a patch has been acked on the ML? May 19 18:26:09 Jin^eLD: .stable? May 19 18:26:15 I implemented that May 19 18:26:24 Laibsch: yes, .stable May 19 18:26:43 Laibsch: I got one ack, need a 2nd, but I wonder if that is ever visible in the bt May 19 18:26:48 but for political reasons, some (should I say one?) of the .stable guys don't seem to want to get involved with the BTS May 19 18:27:14 Jin^eLD: bug the .stable people to fully embrace the tracker for what it is May 19 18:27:24 khem repllied to my bug suggesting a different way, at the same time I have one ack on the ML, and basically it seems that the bug will stay there forever anyway May 19 18:27:25 They might actually find out that it makes their life easier May 19 18:27:43 Jin^eLD: I am staying out of .stable May 19 18:27:51 "for political reasons" May 19 18:27:55 doh May 19 18:28:17 well, this will basically lead to the situation that ppl will not see a point in submitting patches/bugs vs. stable May 19 18:28:25 exactly May 19 18:28:35 Very unfortunate May 19 18:28:40 But I can't change it May 19 18:28:59 But we are seeing some activity from some .stable heroes in the BTS May 19 18:29:04 I did go the way via ML but the bug is still there for the 2nd week, it is a small and trivial and probably not so critical fix, nevertheless - if it will stay around forever, what's the point May 19 18:29:07 Maybe and hopefully it will eventually work out May 19 18:29:30 so do I need to set any flags if I get an ack and am waiting for a 2nd ack? May 19 18:29:39 Ask those questions on the stablebranch ml May 19 18:29:51 k May 19 18:29:52 you are running into open doors with me May 19 18:30:15 forget about the bug tracker for now if you think .stable May 19 18:30:27 sad, but that is the way I would describe it May 19 18:30:49 There has been *no* activity in the BTS for months when it comes to .stable May 19 18:30:56 well, likewise was saying that I should still submit the bug into BST in addition to mailing to the stable ML May 19 18:31:07 and he was also saying that the one who applies the patch is supposed to close the bug May 19 18:31:15 Again, I am glad to see it changed a little over the last few days after I raised hell (and got burned) on oe-devel May 19 18:31:27 so there seems to be some process/idea behind it May 19 18:31:43 yes and no May 19 18:31:46 well.. i see some stable branch as the last glow of hope for professional oe users May 19 18:31:50 and currently it is not working (yet) May 19 18:31:59 in combination with the BTS, that is May 19 18:32:17 _users_ means people who only add some packages and develop a product like dream, openmoko and some others May 19 18:32:45 roh: I was seeing it like that too, but I somehow think that it was a misconception May 19 18:32:59 likewise: was kind enough to try and get some integration of the BTS with .stable processes May 19 18:33:04 I hope he is successful May 19 18:33:20 Jin^eLD well.. i am at OM for over a year now, and it took me about 6 month to know 'where to kick' oe to get it builing ONCE. May 19 18:33:36 heh May 19 18:33:38 roh, that is odd May 19 18:33:46 what did not work right? May 19 18:33:53 should not be *that* bad :) unless you do some uclibc/mips stuff :) May 19 18:33:54 but its still far from something i can tell a user to 'just build it in configuration foo' and then add your packages May 19 18:34:00 Laibsch it breaks all the time May 19 18:34:04 If you add a machine or a complete distro, I'd understand May 19 18:34:14 roh, OK, yes that too ;-) May 19 18:34:34 any bitbake hackers around? May 19 18:34:40 and given alone the changes on the compiler or the libc every now and then makes it unbearable to develop a stable distro upon, where you can start fixing application/integration bugs. May 19 18:35:01 problem with .stable is that if you do find an issue you either have to fix it yourself or hope that someone helps you (so far RP saved me a number of times), but otherwise you are on your own May 19 18:35:01 means people waste 80% of the time with getting the meta-foo to work and not on the real problems. May 19 18:35:18 meta-foo as in meta-toolchain? ;) that cost me *a lot* May 19 18:35:34 roh and Jin^eLD, I agree that stable is very important. I urge the two of you to pitch in with .stable and pressure for better integration with the BTS. May 19 18:35:49 thats why i agree that there is a real reason and use for some stable branch. and yes it has to be maintained. (and i guess that will be loads easier as soon as oe gets rig of the monotone-showstopper) May 19 18:35:54 I am willing to help with setting up good processes May 19 18:36:09 Laibsch: the problem is - I can't really demand anything, ppl looking after it do not get payed for that May 19 18:36:20 it's the same as with my project, sometimes I get feature requests but I can't do them May 19 18:36:25 either because I have no time/do not want to/etc. May 19 18:36:37 1) everybody can pitch in May 19 18:36:45 Laibsch welll.. mtn was so bad we basically branched/forked (that monotone.openmoko.org which was 'behind' .dev May 19 18:36:46 actually its even wrong to say "the problem is", it's the nature of OSS stuff May 19 18:36:58 2) you can still apply presure by giving your opinion May 19 18:37:23 and merging back breaks stuff all the time on both sides. .. well.. now where all people start switching to git we will see what happens. May 19 18:38:07 roh: Please help with testing out git, so we can accelerate the evaluation process May 19 18:38:08 not that i like git for its use-syntax at all ;) but at least it doesnt hog the fastest available workstations for hours just for some pull of a few days May 19 18:38:12 gitweb.openembedded.net May 19 18:38:59 Laibsch i am not really a devel.. i just wanted to give my 2 euro cents from my point of view because ive seen so many fuckups and suffering colleagues as well as users from the community May 19 18:39:04 I find the idea of MTN nice, that everything is in the database, this way I can make it coexist with SVN easily May 19 18:39:19 and i am really grateful for mokomakefile by rod whitby. May 19 18:39:22 actually I commit the mtn db into svn and keep my local changes separate via a custom site config and collections May 19 18:39:48 and leave the org.oe.* untouched May 19 18:40:39 Laibsch: sweet, will join in right now on the git trial May 19 18:40:40 Jin^eLD the idea is ok. to be fair i do not care how a scm works as long as it does. but the NCC on merges and the usage-fuckups due to confused devels as well as the laughable performance didnt really help us get things done. May 19 18:40:53 Laibsch: I take it it's still read-only? May 19 18:41:09 roh: there are also some reservations about that makefile. Are you aware of http://blog.leggewie.org/?p=39. I try to synthesize the best of both worlds (at least for those on deb-based distros) May 19 18:41:28 jeremy_laine: You were still unaware of it? May 19 18:41:34 roh: well, my approach limitd the use of mtn to pull && update, and I did that only if I tested it outside my trunk and was sure that it will not break things May 19 18:41:56 jeremy_laine: I believe it shall eventually be done for some write tests as well May 19 18:41:58 so I had no merging etc, was just overriding stock oe stuff if I had to May 19 18:42:28 jeremy_laine: But expect any particular branch to go offline at any time possibly, until an official position has been decided May 19 18:42:33 This *is* for testing May 19 18:42:54 Laibsch the point is that people need 'something' which they know how to use as tool to get started. to get a working repo which builds. when you are there you can start teaching people stuff. May 19 18:43:30 ACK May 19 18:43:37 but to be fair, a buildsys which depends on exported magic environments.. well.. thats somehow wrong yearthousand. thats what configfiles are for in my opinion. May 19 18:43:56 roh: isn't OE building most of the -native stuff? May 19 18:43:58 Getting things "dumbed down" (in a very positive sense) is a very valuable thing to do May 19 18:44:04 Jin^eLD yes and no. May 19 18:44:05 roh: yes, but bitbake is not so complex. The problem is more in the 'breaks all the time' May 19 18:44:46 cyrilRomain still, one doesnt even get a glimpse what happens till you got it working somehow reproduceable May 19 18:45:05 roh: and local.conf + documentation.conf are well documented. I mean I'm not sure the makefile actually resolves underlying problems May 19 18:45:07 same with the tree.. its NOT obvious your build ends up in ./tmp/.. ;) May 19 18:46:07 it took me a while to figure out a "setup", with separate site, collections, build directory etc May 19 18:46:16 hrw's blog helped me there May 19 18:46:30 cyrilRomain it solves the problem to get all parts in some 'standard' tree layout and build it atleast once May 19 18:46:35 after I got that - building was not a real problem anymore May 19 18:50:42 roh: bad reproducability is *the* real problem I always feel to have. Because there are too much 'side effects' on builds. insane.bbclass and the like are an attempt to solve that. May 19 18:51:15 03Laibsch 07org.oe.dev * r56e10655... 10/ (9 files in 2 dirs): samba: fix packaging and close 3146. May 19 18:52:31 cyrilRomain i now know ;) May 19 18:52:34 sometimes removing /tmp and restarting 'fix' a problem .. sometimes it ends up with frustration (a package that build previously doesn't build anymore ..) May 19 18:53:13 re May 19 18:53:27 re hrw May 19 18:54:12 hi hrw May 19 18:56:28 jeremy_laine: pông May 19 18:59:18 roh, Laibsch, Jin^eLD: the most frustrating is to know being able to help in the oe effort and enjoy using a distro build with oe, but to actually not living it because it requires working on oe fulltime :-/ May 19 18:59:41 Should 56e1065540e7dba03559387ef0025d4a89c4fe3e be pullable by now? May 19 19:00:09 cyrilRomain: that's true :P May 19 19:00:19 I mean it is really difficult to try to follow oe development on spare time :-/ May 19 19:01:44 at the same time, oe is still wonderfull framework that is not so bad, especially being maintained by only few people May 19 19:01:46 i beg to differ! May 19 19:02:04 what it smallest webkit based browser buildable with current trunk? May 19 19:02:18 hrw: midori ? May 19 19:02:21 * timtimred works for 12-13 hours a day including travel and maintains 3 distros in a fork May 19 19:02:57 ok, so you cant learn everything, but like, its not that bad really May 19 19:03:54 re May 19 19:03:59 timtimred: congrats. I wish I could do the same May 19 19:04:02 hi flo May 19 19:04:10 hi flo_lap May 19 19:04:23 hi florian May 19 19:04:43 (im not trying to be mean here but) oe isnt a distro May 19 19:04:57 it feels like your expecting one May 19 19:05:02 hrw: qt4 comes with a tiny webkit-based browser in examples May 19 19:05:15 hrw, any hints on getting a rootfs onto CF for alix ? May 19 19:05:18 cyrilRomain: I need gtk - no space on device May 19 19:05:19 morning all May 19 19:05:25 hrw: ok :-/ May 19 19:05:29 hi roh May 19 19:05:32 hi RP May 19 19:05:40 Gadget-UK: partition cf card and as root unpack rootfs.tar on it May 19 19:06:10 Ah,k always the simple stuff :) , than apply the required bootloader I guess May 19 19:06:17 Gadget-UK: yep May 19 19:06:46 Gadget-UK: I booted qemu with cf as hda, installed debian, booted alix. then installed angstrom on other partition May 19 19:07:01 hrw: hiyza May 19 19:07:22 hrw, neat. May 19 19:07:27 hrw: I've finished cleaning up the urls for oebuilder, theyr're now much cleaner May 19 19:07:34 s/oebuilder/oestats/ May 19 19:07:58 hrw: you need to pull oestats-client.bbclass May 19 19:08:41 hrw: also, I have a patch for bitbake which would allow us to pick up failed builds (so they don't stay in 'started' state) May 19 19:09:14 jeremy_laine: send patch to bugtracker with description May 19 19:10:53 hrw: will do, mind looking the patch over? http://pastebin.com/m803ab6a May 19 19:11:15 hrw: it's 2 lines actually :) May 19 19:11:40 hrw: you can get an example of a failed build here http://oestats.jerryweb.org/builds/ May 19 19:11:52 RP: what do you think about that patch? May 19 19:14:28 Laibsch: i'd like to try oe with git. how can I get a clone of the repository ? May 19 19:16:06 hrw: We probably should do something like that, yes May 19 19:18:57 RP: I'll go ahead and post to the tracker unless you feel you can commit the change May 19 19:19:25 jeremy_laine: Give me a minute and I'll see if I can commit something May 19 19:19:36 RP: ok May 19 19:22:00 hrw: heh, another interesting failure example: http://oestats.jerryweb.org/builds/11/ May 19 19:23:16 Laibsch: 'git clone git://git.openembedded.net/org.openembedded.dev' gives me a connection refused. Should I use git+ssh with a special port ? May 19 19:23:23 nice jeremy_laine... May 19 19:25:27 jeremy_laine: We have a small problem here since bitbake trunk handles this differently :/ May 19 19:26:00 In trunk you will get a BuildCompleted event with the number of failures that occured May 19 19:30:16 cyrilRomain: It works here, but RP also me somehow. 'git clone git://git.openembedded.net/org.openembedded.dev.git' instead of 'git clone git://git.openembedded.net/org.openembedded.dev'? May 19 19:30:55 s/also me/also added my key/ May 19 19:31:18 Laibsch: same error with both :-/ May 19 19:31:51 ask RP May 19 19:31:55 Laibsch: maybe do I need to open some port .. May 19 19:32:05 I don't think so May 19 19:32:06 let me poky amethyst, its broken May 19 19:32:23 Oh, that would explain it, too ;-) May 19 19:32:25 timtimred: I added a link to the OE web interface to MTN: http://oestats.jerryweb.org/builds/11/ May 19 19:33:00 nice, going to add a blame-guesser? :) May 19 19:34:15 * hrw pulls OE to check oestats updates May 19 19:34:28 timtimred: possibilities are limitless, we could check what the last good build was to help bisect May 19 19:34:59 yes, and changes since, and changes of deps since etc May 19 19:35:36 timtimred: ah, deps, that's a nice one May 19 19:35:58 jeremy_laine: http://amethyst.openembedded.net/oe/viewmtn/viewmtn.py/revision/info/ May 19 19:36:11 hrw: heh.. May 19 19:36:44 ;) May 19 19:36:48 having the option to archive the package build dir for some kinds of errors for download might also be useful May 19 19:36:55 never trust user data May 19 19:37:19 obviously that takes up potentially a lot of space, but a useful option :) May 19 19:37:26 timtimred: I do not see how I can push those gigabytes thrugh my 512kbps link May 19 19:37:58 ok but for others with build servers on larger pipes ;) May 19 19:38:21 build failures are sometimes not repeatable on other hardware May 19 19:38:31 or even s/hardware/systems May 19 19:38:47 indeed, but are you going to compile my distros for me? :) May 19 19:38:51 ok, git daemon is working May 19 19:39:54 jeremy_laine: Would it be a problem if we always returned a BuildCompleted event with a number of failures? May 19 19:40:32 RP: sure, would you export a var? May 19 19:40:36 RP, Laibsch: thanks :) May 19 19:40:46 RP: if you have a patch for me I'll give it a spin May 19 19:40:57 jeremy_laine: Where is the code you're using this? May 19 19:41:01 the new calss? May 19 19:42:06 RP: classes/oestats-client.bbclass May 19 19:46:41 jeremy_laine: BuildBase.getFailures() already exists in 1.8 so its just a case of connecting the event correctly May 19 19:53:21 jeremy_laine: http://www.rpsys.net/openzaurus/temp/fire.patch May 19 19:56:29 RP: identation is off by one in cooker.py:549 May 19 19:57:40 jeremy_laine: oops, sorry about that. Other than that does it solve the problem? May 19 19:57:48 RP: AttributeError: TaskFailure instance has no attribute '__len__' May 19 19:57:54 RP: same line May 19 19:58:47 RP: looks like fnids is not an array May 19 19:59:27 jeremy_laine: That will teach me to test code. I started with a counter, then thought I could take a shortcut ;-) May 19 20:00:19 jeremy_laine: I've just changed the patch at the above address May 19 20:04:27 Tartarus: Hi... my configure virt/kernel finish now I tried the devshell. it didnt work. bitbake said there's no action called devhell May 19 20:04:47 hernan: devshell or devhell? May 19 20:04:56 * mwester is often in "devhell" with monotone lately... ;) May 19 20:04:57 looking in the kernel.bbclass file i found do_menuconfig.. so doing ""bitbake -c menuconfig virtual/kernel"" works? May 19 20:05:06 hernan: Try INHERIT += "devshell" in local.conf May 19 20:05:09 devshell :) May 19 20:05:15 ok May 19 20:05:16 trying May 19 20:06:44 interesting.. when i tried with -c menuconfig i've got another terminal window with menuconfig ready for me to edit May 19 20:08:01 Yep, that's how it should work. Rather cool, really... May 19 20:08:14 indeed May 19 20:08:31 * RP is pleased someone likes it :) May 19 20:08:36 omg!!! May 19 20:08:48 it works.. inherit += devshell May 19 20:09:02 it opened a shell for me in the kernel src dir May 19 20:09:07 thanks a lot May 19 20:10:29 ;) May 19 20:10:44 RP: brilliant, works for me! May 19 20:10:53 see: http://oestats.jerryweb.org/builds/ May 19 20:11:19 RP: from now on there are 3 states: Started, Succeeded, Failed May 19 20:11:32 jeremy_laine: great :) May 19 20:12:48 RP: let me know when you get a chance to commit, will update oestats-client.bbclass accordingly May 19 20:13:26 * hrw waits for 4th one May 19 20:13:28 Completed May 19 20:14:27 03rpurdie 07bitbake-1.8 * r1071 10/ (ChangeLog lib/bb/cooker.py): cooker.py: Make sure Build Completed events are generated even when tasks fail May 19 20:16:08 03rpurdie * r1072 10/ (4 files in 4 dirs): BBHandler.py: Correctly handle '-' characters in class names (#2958) May 19 20:17:09 03jeremy_laine 07org.oe.dev * rf92eb79f... 10/ (1 classes/oestats-client.bbclass): oestats-clients.bbclass: discriminate between Succeeded and Failed builds May 19 20:21:28 jeremy_laine: do you plan to add 'Completed' status? May 19 20:22:14 hrw: no, Completed is mapped to: Succeeded when failures=0 or Failed if failures>0 May 19 20:23:20 phew. i've finally configured my kernel .config May 19 20:23:23 jeremy_laine: but when you have set of pacakges in progress and look at build page there are lot of 'succeeded' which are 'in progress' ones May 19 20:23:42 what should i do now? make bzImage or go back to build and "bitbake bootstrap-image" ? May 19 20:23:45 hrw: ah, initially a build is in the "Started" status May 19 20:23:55 hernan: back to May 19 20:24:04 ok May 19 20:24:06 hernan: first copy your config somewhere May 19 20:24:29 .config you mean? May 19 20:25:00 jeremy_laine: build is in 'started', then tasks run. currently is any task finished you list package as 'succeeded'. if build of package fails you show 'failed'. May 19 20:25:03 hernan: yes May 19 20:25:08 ok May 19 20:25:17 done, i am bitbaking now May 19 20:25:43 its going to take a big while .. thanks a lot May 19 20:25:48 i will be back tomorrow May 19 20:25:48 jeremy_laine: if package has 9 tasks to run you show that package 'succeeded' after 1st task end properly. same when 2,3,4,5,6,7,8 and 9th one. May 19 20:25:52 hrw: correct, there is no "package" item in the database (only builds and tasks), so I can't store the fact that it's "in progress" at the package level May 19 20:26:17 jeremy_laine: export BB_DEFAULT_TASK and check for it? May 19 20:26:30 hrw: right, "Succeeded" means "no errors so far" for packages May 19 20:27:01 hrw: I could introduce a "package" db entry but it's a pain, you have to decide what makes up a "package" May 19 20:27:21 hrw: so far it's something like (builder, package, version, revision) May 19 20:27:28 bitbake has no knowledge of what a 'package' consists of... May 19 20:28:03 RP: right, the "package" view in oestats is just a convenience May 19 20:29:14 anyhow, it's much better now, thanks for the change to bitbake May 19 20:29:52 jeremy_laine: np - you're pushing some barriers and the event infrastructure is far from perfect - let me know if you find any further glitches May 19 20:30:18 That was an unusual case of a fix being in trunk but not in 1.8... May 19 20:31:10 RP: can we add this for tomorrow run on buildbot? could be nice set of test data May 19 20:31:53 hrw: Is it useful from buildbot? May 19 20:33:02 bigger amount of tasks then any current test data for use May 19 20:33:07 s/use/sure May 19 20:34:08 I've no objection to trying it I guess... May 19 20:34:17 Does it matter that its poky as against OE? May 19 20:34:48 RP: I already submitted Poky build into it and oestats handled it fine May 19 20:35:21 except METADATA_REVISION = "" for which probably hack^Wexception in code exists May 19 20:35:37 hrw: ok, feel free add it if you want... May 19 20:36:15 I will do some runs first on my machine - should generate some amount for jeremy_laine to test May 19 20:36:44 kind of 'for machine in LONG_LIST;do ...' May 19 20:37:59 RP: does that bitbake change landed also in Poky? May 19 20:39:51 hrw: not yet May 19 20:39:55 03rpurdie * r1073 10/ (4 files in 4 dirs): manuals: Add missing options documentation (thanks to zecke for noticing and the patch) May 19 20:42:13 hrw: it has now May 19 20:44:34 for mach in ../meta/conf/machine/*.conf;do M=`basename $mach .conf`;MACHINE=$M bitbake poky-image-sato;MACHINE=$M bitbake poky-image-sdk;bitbake MACHINE=$M bitbake meta-toolchain done May 19 20:44:56 jeremy_laine: this should give your service lot of test data ;D May 19 20:45:17 hrw: woah :) May 19 20:46:05 18 machines, 3 targets for each one May 19 20:46:32 I just hope that this will end in 12h May 19 20:46:56 because if it not then will have to kill it rather as this can really occupy this machine May 19 20:47:41 renice May 19 20:48:12 kario: and then renice it everytime when new bitbake starts? May 19 20:48:23 :) May 19 20:57:00 jeremy_laine: I wonder is there a way to get information which target(s) are to be build (poky-image-sato for 'bitbake poky-image-sato', 'mc nano' for 'bitbake mc nano') May 19 21:00:32 will also show me how much time it takes May 19 21:01:02 jeremy_laine: please add one column to builds list: amount of time already spent on building May 19 21:06:24 jeremy_laine: btw - why build 12 is completed and 23 not? May 19 21:16:23 So, hmm May 19 21:17:41 This is a loaded question I suppose, but do _some_ people who commit to .stable watch/commit bugs/fixes against the stable branch? If so (and so I know who to prod once I start filing / fixing things) do I prod about it? May 19 21:18:06 {watch bugs,commit fixes} May 19 21:18:17 .stable policy is: patches/fixes should go to ML May 19 21:19:04 openembedded-devel@openembedded.org or ? May 19 21:19:14 we all have good intentions, but mailing list stuff gets seen first May 19 21:19:40 somehow, I always forget to look at bugzilla unless I am really bored :( May 19 21:20:33 I don't mind starting out emails saying This is bug # (http://...), just need to know where to send 'em to :) May 19 21:21:10 openembedded-stable@ May 19 21:21:31 ok May 19 21:22:40 openembedded-stablebranch@ ? May 19 21:24:19 iirc May 19 21:26:14 see the article at www.openembedded.org May 19 21:52:23 bye May 19 21:55:04 * * OE Bug 4280 has been created by trini(AT)embeddedalley.com May 19 21:55:06 * * The /usr/include/scsi headers are missing in some cases May 19 21:55:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4280 May 19 22:14:35 Tartarus, can you explain a little more about how you use your customizations to OE as an 'overlay' to the OE tree? Are you simple adding files to the tree, or adding/modifying/deleting? May 19 22:16:27 03tmbinc 07org.oe.dreambox * rf7b51c2c... 10/ (3 files in 3 dirs): initscripts-opendreambox: set videomode on bootup May 19 22:28:19 what are the chances of getting simple programs from .dev to .stable, if they're not in .stable? May 19 22:28:20 (thinking dt) May 19 22:59:28 trying to understand the following 'NOTE: preferred version 2.2.5 of glibc not available (for item libsegfault)' how can that not be available... its supposed to download the sources and build it right? May 19 23:02:49 tharvey what are you building? May 19 23:02:59 glibc-2.2.5 is really old May 19 23:03:05 * * OE Bug 4281 has been created by trini(AT)embeddedalley.com May 19 23:03:07 * * mtd-utils-dev is missing headers May 19 23:03:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4281 May 19 23:07:25 woglinde, thought I was doing something very basic and simple - MACHINE=nslue2be DISTRO=unslung May 19 23:09:09 somehow I think my local.conf isn't getting read? May 19 23:12:56 I'm trying real hard to understand how OE is setup - I'm just trying to build a firmware image for something simple May 19 23:14:11 tharvey: Unfortunately, Unslung is not a very representative distro for OE. We're happy that we can build it at anymore. May 19 23:14:32 tharvey: You can ignore that "NOTE:", IIRC. May 19 23:15:03 ok... so can anyone give me a simple configuration to build? May 19 23:15:23 Use the makefile: http://www.nslu2-linux.org/Makefile May 19 23:15:38 It will set up the entire configuration for you. May 19 23:16:00 Or do you mean that you wish to try a simple DISTRO other than Unslung? May 19 23:16:01 my goal is not necessarily to build firmware for NSLU2 but to understand how OE works... May 19 23:16:29 Ah, ok. Angstrom is probably what you want; it is well-maintained and is the most representative of best practices. May 19 23:16:52 (Unfortunately, I maintain Unslung, and don't have a running Angstrom environment right now) May 19 23:16:56 yes... a simple distro that has a fairly high probability of actually producing a kernel and rootfs image May 19 23:17:53 http://www.angstrom-distribution.org/building-angstrom May 19 23:18:17 That should do it for you! May 19 23:18:56 Don't forget to delete the "tmp" directory with the partial build stuff for Unslung; it may conflict with angstrom. May 19 23:25:12 ok, thanks I'll give that a try. I guess I should consider openembedded.org.dev unstable. May 19 23:25:46 Yep. Pick the stable branch if you dislike surprises. May 19 23:26:14 what about bitbake... do I really need to force bitbake-1.8.8 vs whats installed on my system (1.8.10) May 19 23:28:33 so Angstrom is a general purpose embedded distro using ipkg? Is that probably the best example to play around with and follow? May 19 23:28:36 I really don't know for sure. I have switched around on other distros, and it has been compatible, but that may not be universally true. To be safe, I'd go back to 1.8.8 May 19 23:29:58 Yes, Angstrom is the best to play with to learn. I'm not sure if the stable branch is still using ipkg; the dev branch has replaced ipkg with opkg (upwards compatible, and maintained). May 19 23:35:18 does local.conf have to exist in a build/conf subdir? seems strange that you need to create an empty skeleton to put a project conf file in May 19 23:40:24 needs to be in BBPATH, yes May 19 23:47:49 Tartarus, could you explain what you were talking about the other day regarding you overlaying your customizations on top of the OE tree - are you saying that you SVN a skeleton of files that get added/changed from OE? May 19 23:48:11 tharvey, kinda May 19 23:48:21 We have 3 dirs May 19 23:48:32 1 is oe.dev at a specific revision, unchanged from that May 19 23:48:48 2 is a custom overlay dir which contains our machine conf files and package additions/changes May 19 23:49:10 3 is the build output dir and local.conf May 19 23:49:18 2 is in SVN expanded May 19 23:49:27 1 we get via script kept in svn that just checks out the correct rev May 19 23:49:43 3 doesn't live in SVN at all as its the build dir May 19 23:51:45 so 2, you call an 'overlay' but it really does not overlay OE at all. I imagine that you have 2 in your BBPATH before 1 so that anything there takes priority over 1? May 19 23:52:56 yes May 19 23:53:05 iirc the overlay term comes from OE docs May 19 23:53:22 same dir structure as the main OE checkout May 19 23:53:30 files in there override files found in the others May 19 23:54:13 right... makes completely sense. I think I was thrown off by the 'overlay' term. I thought you were putting some sort of sparse svn checkout over the top of an OE checkout May 19 23:54:24 ah May 19 23:54:43 trying to build anstrom I get an error right away: /proc/sys/vm/mmap_min_addr is not 0. This will cause problems with qemu so please fix the value (as root) May 19 23:55:13 and no such file? May 19 23:56:22 nm :) May 19 23:57:31 exists and is 65536 - what is this all about? I didn't get this previously when trying to build unslung. Perhaps this is something due to bitbake 1.8.8 vs 1.8.10? May 19 23:57:48 nope May 19 23:57:58 angstrom uses qemu to generate locale files May 19 23:58:02 Tartarus, so do you ever run into a situation where a file exists in oe that you want 'removed' instead of modified? May 19 23:58:11 qemu wants that file to have a specific value (0) May 19 23:58:49 tharvey, sometimes external-toolchain.bb makes me pull my hair out, but you can DEFAULT_PREFERENCE=-1 stuff May 19 23:58:55 and that's as good as removing May 19 23:59:49 ahh... good to know May 20 00:01:13 so you can either change that file to read 0 or in local.conf set ENABLE_BINARY_LOCALE_GENERATIHN = 0 May 20 00:01:16 or so May 20 00:01:20 see angstrom-2008.1.conf May 20 00:01:40 thanks May 20 00:06:27 not really seeing anything like that in angstrom-2008.1.conf? May 20 00:10:06 still getting 'preferred version x of glibc not available' - for angstrom-2008.1 "NOTE: preferred version 2.6.1 of glibc not available" - this is not very intuitive May 20 00:12:21 tharvey, I hate that message also May 20 00:12:33 It builds anyway, right? May 20 00:13:02 dunno... I'm not sure what messages are errors and what are not :) **** ENDING LOGGING AT Tue May 20 02:59:56 2008