**** BEGIN LOGGING AT Fri May 04 02:59:56 2007 May 04 03:12:51 is there a specific DISTRO setting I need for angstrom? May 04 04:27:58 polyonymous: ok, grabbing image now :) May 04 05:37:25 OE finaly starting to get handy for me! May 04 05:55:57 polyonymous: Woooot :) It boots, sound kinda works (old problem with poodle sound) Thanks for building the image for me ): May 04 05:56:18 s/):/:) May 04 06:21:39 03lenehan 07org.oe.documentation * r67b9d3a1... 10/ (1 usermanual/Makefile usermanual/README): May 04 06:21:39 usermanual: Modify the Makefile so that "make " works for the May 04 06:21:39 following types: dvi, pdf, ps, rtf, tex, texi and txt. This is in addition May 04 06:21:39 to html and xhtml were already working. Update the readme to give a guide as May 04 06:21:39 to what packages you need on the host for these to work. May 04 07:03:44 good morning May 04 07:07:34 hi likewise May 04 07:07:46 hello May 04 07:20:00 good morning koen May 04 07:20:13 good morning all May 04 07:43:34 morning all May 04 07:44:55 hey RP May 04 07:46:18 OE user manual: http://www.openembedded.org/~koen/usermanual.pdf May 04 07:46:22 morning RP, pgfeller May 04 07:46:36 likewise: good morning May 04 07:47:48 koen: bugsquash weekend for the manual? May 04 07:48:52 that would be a good idea May 04 07:48:59 * koen opts to go to the zoo instead May 04 07:49:37 koen: yeah, you might see more animals there May 04 07:56:47 03lenehan 07org.oe.documentation * r1d471462... 10/ (4 files in 3 dirs): May 04 07:56:47 usermanual: Work some more on the using bitbake/OE chapter. Enable it - May 04 07:56:47 while it's not great but I think there is probably a few useful items in May 04 07:56:47 there for new users now. May 04 08:11:09 bitbake question: As a user I would like to know the reason for skipping a package. If the .bb file parser/handler raises an exception, load_bbfile() will return skip="True". IOW: How can I debug why the exception was raised? May 04 08:19:43 morning May 04 08:20:03 bitbake -DDDD tried? May 04 08:20:21 hrw: yes, even 5 times May 04 08:20:51 then no idea May 04 08:20:59 looks like opie is starting to fly again :-D May 04 08:21:44 hrw: good morning BTW May 04 08:21:48 someone did grub setup without accessible menu? May 04 08:22:00 hrw: long time ago, yes May 04 08:22:18 hrw: you mean password protected, or none at all? May 04 08:22:36 hrw: ah, you mean grub installation without having the grub-menu? May 04 08:22:47 none (option to show on keypress would be fine) May 04 08:23:09 hrw: yes, I did, why? May 04 08:23:26 I am thinking about doing such one on my laptop May 04 08:23:48 now it has pass on power on, then pass to mount rootfs/swap/home lvm May 04 08:25:59 likewise: Does bitbake -v -DDD help? May 04 08:27:08 RP: no. I know what is the problem (glib requiring USE_NLS="yes"), but I want this to be made clear somehow. May 04 08:27:30 RP: because that way, I can show another issue more easily :-) May 04 08:29:16 likewise: It should show up in the debug logs... May 04 08:29:57 likewise: Heh, I had that issue yesterday when trying my uclibc builds with 1.8.2 ;) The error bitbake showed was about bootstrap-image was unavailable (due to task-base pulling in cups which pulled in dbus which pulled in glib-2.0). Took ages to figure it all out... May 04 08:30:28 v8jlene: *exactly* the problem I was about the explain (and then another one) May 04 08:30:37 blaster8: hi May 04 08:30:46 hey there :) May 04 08:31:22 likewise: There is a bitbake bug already open about the logging of failed providers May 04 08:31:39 Its not as simple as it sounds though May 04 08:31:54 v8jlene,rp : OK, the other problem I see is not bitbake related, but task-base related. May 04 08:33:48 task-base will want to satisfy all dependencies of all of task-* (for example task-bluetooth), even if COMBINED_FEATURES does not include bluetooth. As bluez depends on glib, which depends on NLS, my image won't build. May 04 08:34:32 likewise: Thats a nasty one :/ May 04 08:34:33 These are two well-hidden problems that users run into when slightly adjusting the options (USE_NLS, and stripping FEATURES). May 04 08:34:48 likewise: Yeah, that task-base one's a bit annoying. I only had about 400 recipes symlinked in, now I have them all just to make task-base happy. And as you say it's building lots of stuff that you don't use. May 04 08:34:57 RP: koen suggested that splitting up the task-* into different .bb files would solve this. May 04 08:35:02 likewise: Its debian.bbclass at fault - remove that and it will probably work May 04 08:35:21 * RP hates debian.bbclass for this May 04 08:35:35 likewise: BTW, the NLS one I solved like this: include/slugos.inc:USE_NLS_glib-2.0 = "yes" May 04 08:35:42 in my own distro May 04 08:36:06 RP: OK, but suppose that we want to solve this issue (for the dumb user, such like me :-) ), would we need to fix debian.bbclass somehow, or redo task-base? May 04 08:36:33 likewise: Splitting task-base is a rather evil idea but would work. I would prefer to find another way though May 04 08:36:33 v8jlene: isn't that dangerous to do? May 04 08:37:18 likewise: No idea. But since I'm not using glib or anything that's built with glib... May 04 08:37:32 RP: yes. Both issues are show-stoppers for building minimal distro's currently. May 04 08:37:44 v8jlene: does glib get build? May 04 08:38:35 likewise: can you split task-base into separate files and make task-base DEPEND only on needed ones? May 04 08:38:45 likewise: Hmmm... good question. uclibc doesn't build for me with 1.8.2 yet, so I haven't got far enough to test that yet. May 04 08:39:19 hrw: I probably could (with some help), but I want to know first if this is the way to go, within the BB/OE framework. May 04 08:40:07 likewise: I like task-base idea but it really gives too much to build May 04 08:40:24 v8jlene: let me know, because the strange this is, even glib is not build, it must be buildable in the current task-base approach. May 04 08:41:55 hrw: yes, but that's only because it requires dependencies to be buildable (even if you do not want them). You have so satisfy these dependencies (often by enabling stuff that makes the build bigger). May 04 08:42:02 Sorry for repeating myself, just wanted to make sure I was not the only one hitting these. :-) May 04 08:49:31 03ifaistos 07org.oe.dev * r6dd82167... 10/ (1 classes/magicbox-image.bbclass): May 04 08:49:31 classes/magicbox-image.bbclass: Add a class that handles the generation of flash images for Magicbox May 04 08:49:31 It creates "ready-to-burn" squashfs and squashfs-lza images and a single file kernel+jffs2 image May 04 08:49:38 03ifaistos 07org.oe.dev * rf9d679b4... 10/ (1 conf/machine/magicbox.conf): May 04 08:49:38 conf/machine/magicbox.conf: Clean up the image generation parameters , inherit magicbox-image.bbclass in order to May 04 08:49:38 create "ready-to-burn" images May 04 08:49:45 in general the default OE image seems to include a lot of stuff that are not need it for small systems May 04 08:50:49 examples? May 04 08:50:52 on a 4MB flash with jff2 (+kernel) it leaves something like 300K free May 04 08:51:49 ok. so give example what to remove from it May 04 08:53:32 trying to get the system plugged in.... May 04 08:54:13 hrw: In general, the minimal system OE should be able to build is a kernel + one userspace (init) program. May 04 08:54:59 Although a default init does not make sense, init can be a custom multithreaded application provided by the user. May 04 08:55:43 hrw, Ifaistos: We are at a 3.5 MB squashfs-lzma image, and that's still using glibc. May 04 08:56:16 likewise: I know - but to get smaller we need to know what to remove May 04 08:57:34 hrw: I would turn around the question: what do some minimal setups need. Because otherwise the question would become "remove from where or from what?" May 04 08:58:15 likewise: task-boot gives small image and I do not know what to remove from it May 04 08:59:29 hrw: i think something similar to the openwrt just boot to busybox shell May 04 08:59:56 openwrt use uclibc so they win May 04 09:00:21 Ifaistos: task-boot should give getty login May 04 09:00:32 * hrw tested similiar thing in Poky May 04 09:00:53 likewise: yes squashfs-lzma is the best option in terms of compression/sizes May 04 09:02:12 hrw: glibc or uclibc should not make a difference .The image should allow a system just to boot and provide a console... May 04 09:02:37 then try task-boot only image May 04 09:02:39 hrw: let the rest be handled by the distro May 04 09:03:33 hrw: that's used by bootstrap image? May 04 09:03:48 bootstrap-image use task-base May 04 09:04:13 angstrom-minimal-image use task-boot + ssh daemon May 04 09:05:28 we probably need to clean up that all May 04 09:05:35 Can we please try to find a way of dealing with debian.bbclass instead of splitting up task-base. If we slpit it up, it going to make a big mess :-( May 04 09:05:47 I'll use v8jlene's approach to work around the task-base dep issue. May 04 09:05:59 RP: +1 for me May 04 09:06:09 yes that's the idea. May 04 09:20:06 koen|away: did you find your a780 ? May 04 09:20:20 morning all btw May 04 09:21:40 morning ade|desk May 04 09:22:01 v8jlene: what are your uclibc builds? powerpc only? May 04 09:26:56 powerpc uclibc ... brave brave man May 04 09:27:09 next brave is sparc imho May 04 09:27:44 likewise: sh4 May 04 09:45:29 hrw: building angstrom-minimal-image fails because perl 5.8.7 won't build (not that I need it in the image, but this is the task-base issue coming up) May 04 09:45:39 hrw: I'll try bootstrap-image instead. May 04 09:46:28 both use task-base recipe May 04 09:50:19 hrw: yes, I have now removed debian.bbclass to see if the deps disappear May 04 09:50:25 * v8jlene should probably make perl 5.8.8 the default May 04 10:01:42 hrw, v8jlene, Ifaistos: without debian class, after rm -rf build/tmp/cache, I still get all the deps build (such as cups, which is depended upon by bluez-utils, etc). May 04 10:01:56 hrw, v8jlene, Ifaistos: (that's for bootstrap-image, btw) May 04 10:33:14 koen|away: the Copyright notive is incorrectly parsed from the xml http://www.openembedded.org/~koen/usermanual.pdf. You should add ", " between the names. May 04 10:56:37 Copyright notice May 04 11:02:30 How many programmers does it take to change a light bulb? One hundred. One to hold the bulb and 99 to debug the house. May 04 11:02:35 That is, good morning. May 04 11:04:42 polyonymous: or none as this is hardware problem May 04 11:05:13 hrw, yup. And "How many hardware engineers it take? None, we'll fix it in software" May 04 11:11:05 => hrw is an engineer May 04 11:11:09 ;-) May 04 11:11:22 hardware or software or both? :) May 04 11:12:13 my last two lines made no sense May 04 11:12:21 I was making an incorrect inference May 04 11:13:00 who cares as long as you're |gone :) May 04 11:14:48 polyonymous: rather software then hardware May 04 11:15:11 Thought so. Your nick may sound misleading though :) May 04 11:15:12 polyonymous: You want me to be gone May 04 11:15:22 Here you are ;-) May 04 11:15:29 Laibsch, it wasn't me who did your /nick :)) May 04 11:15:36 I postponed shopping May 04 11:16:23 so, having amd64 means I'm out of luck with qemu? May 04 11:16:31 polyonymous: "as long as your gone" could be interpreted as "thank goodness he is gone" which I know you did not mean. Was just picking a pun May 04 11:16:50 polyonymous: I think there was a workaround in the bug, wasn't there` May 04 11:17:00 But yes, AMD64 has lots of issues May 04 11:17:06 well, not the chip May 04 11:17:10 but software support May 04 11:17:37 Laibsch, ;-) It's up to you to interpret, but as you pointed out, I didn't mean that exactly. It was more like "nevermind, we don't have to push this senseless argument" :) May 04 11:17:40 Yeah May 04 11:17:50 I think precompiled 32 bit binaries may help, tho. May 04 11:18:12 I'm more concerned about failing bitbake gcc of all the bugs I can't resolve. May 04 11:29:29 03Laibsch 07org.oe.documentation * r05bae6b6... 10/ (1 usermanual/usermanual.xml): usermanual.xml: denominate for useless use of cat award ;-) May 04 11:31:08 I take it konqueror-embedded current versions are all totally broken and it may make sense to try the latest 2007 one from bugzilla? May 04 11:56:39 victor_rx1950: so, what's exactly your problem? GettingStarted should describe what you should have in local.conf May 04 11:56:40 psokolovsky: i set BBFILES like in wiki May 04 11:57:14 psokolovsky: and TARGET_ARCH would be arm? May 04 11:58:02 polyonymous: I think a clean start might be the best solution indeed. May 04 11:58:03 victor_rx1950: no, generally, besides paths, all you need to set is MACHINE and DISTRO May 04 11:58:14 polyonymous: Got RW access yet? May 04 11:58:15 victor_rx1950: machine should exist of course ;-) May 04 11:58:36 psokolovsky: where i can look list of machines? May 04 11:58:39 Laibsch, Just wanted to make sure, I can't break it (that is - it ain't fixed:)) May 04 11:58:45 victor_rx1950: her's my local.conf: http://pastebin.ca/470618 , but note it contain much more stuff than you need or should use ;-) May 04 11:58:59 thx May 04 11:59:01 Laibsch, not that I'm aware of, but I do not have the commit ready (that is for konq), yet :) May 04 11:59:18 victor_rx1950: under config/ dir. spend some time looking thru OE tree first - both source, and build tmp/ May 04 12:02:33 ~curse mozilla build May 04 12:02:36 May you be reincarnated as a Windows XP administrator, mozilla build ! May 04 12:06:47 hrw: you crazy fool May 04 12:07:14 XorA: ;D May 04 12:10:18 which package provides glib2.0? May 04 12:20:47 glib-2.0 May 04 12:25:57 polyonymous: you sent out that key? May 04 12:26:18 Laibsch, yup. Assuming, monotonephrasebook has correct email addresses :) May 04 12:26:48 Laibsch, I wouldn't be surprised if guys had better things to do and just didn't have time. May 04 12:26:56 XorA: check claws-mail in Poky - Tomas did some UI changes to make it better fit in VGA May 04 12:27:03 See no reason for nagging them. May 04 12:28:04 hrw: sorry. I missed it May 04 12:28:12 hrw:thanks May 04 12:28:49 polyonymous: well, especially since you sent them out within the last 12 hours or so. Nagging would be kinda pushy ;-) May 04 12:28:59 Laibsch, exactly my point. May 04 12:29:29 Laibsch, it's not like I can't sit here and hack around konq, without the key :) May 04 12:29:58 * hrw flash tosa May 04 12:30:03 Laibsch, upon receiving access I'd have to "waste" time on reading more of mtn docs, right now I'd rather hack konq instead :) May 04 12:46:06 shit.. Z lack wifi stuff May 04 12:46:18 hi Thomas May 04 12:47:01 hi Marcin May 04 12:47:38 hey thesing May 04 12:47:53 ~curse usb May 04 12:48:00 hi XorA May 04 12:48:06 May the fleas of a thousand camels infest your most sensitive regions, usb ! May 04 12:49:45 suspend on collie works now. But the battery driver doesn't work after resume. (I think its the ucb driver.) May 04 12:50:00 thesing: cool May 04 12:52:31 nearly every collie specific driver doesn't seem to like suspending... :( May 04 12:52:50 on spitz, my g-ether can't survive suspend. May 04 12:52:57 That is, even unloaded and reloaded. May 04 12:53:14 it never sees the link afterwards. May 04 13:08:16 !oebug 1253 May 04 13:08:17 * * Bug 1253, Status: REOPENED, Created: 2006-07-30 00:06 May 04 13:08:18 * * oe_bugs(AT)matt.kincaid.name: 2.6.17 kernel on Tosa fails to recognize NAND partitions May 04 13:08:19 * * http://bugs.openembedded.org/show_bug.cgi?id=1253 May 04 13:08:36 do13: we do not support resizing flash and do not plan to - right? May 04 13:11:35 !oebug 1253 May 04 13:11:36 * * Bug 1253, Status: RESOLVED (WONTFIX), Created: 2006-07-30 00:06 May 04 13:11:37 * * oe_bugs(AT)matt.kincaid.name: 2.6.17 kernel on Tosa fails to recognize NAND partitions May 04 13:11:38 * * http://bugs.openembedded.org/show_bug.cgi?id=1253 May 04 13:28:42 I got a little confused in this page: http://www.openembedded.org/wiki/UpgradingPackages May 04 13:29:08 because? May 04 13:29:25 xjqian, just the way to preserve history. What's confusing? May 04 13:29:30 you do mean to `cp packages/$pkg/$file_v2.bb packages/$pkg/$file_v1.bb` before making any change to $file_v2.bb do you? May 04 13:30:25 or is it really before the final step? May 04 13:30:57 xjqian, I'd put it as `mtn mv v1 v2`, `cp v1 v2`, `mtn add v1` May 04 13:30:58 i.e. after making changes to $file_v2.bb? May 04 13:31:27 Before making any changes. May 04 13:31:34 In case you want to preserve old package. May 04 13:31:38 yeah, that makes more sense to me May 04 13:31:39 If you don't, just mtn mv -e v1 v2 May 04 13:32:25 Basically, the result is the same, except for the page gets you to move -e and then copy back. From mtn POV it's the same. May 04 13:32:36 but the description " before the final step" really make people feel the other way May 04 13:32:55 xjqian, I hardly read the wording, snippets were enough for me :) May 04 13:33:23 huhu May 04 13:33:28 huhu May 04 13:33:39 :> May 04 13:35:40 btw, the latest mtn 0.3.4 seems not supporting the -e switch anymore. anybody can confirm? May 04 13:36:03 --execute one? May 04 13:36:23 xjqian, 0.34 you mean? I have 0.33... May 04 13:36:34 hrw, I think yes, mtn mv -e May 04 13:36:41 - The commands 'drop', 'rename' and 'pivot_root' default to May 04 13:36:41 always perform the operation in the file system as well. May 04 13:36:41 They do not accept '--execute' any more, but will instead May 04 13:36:41 take '--bookkeep-only' in case the user only wants to affect May 04 13:36:42 bookkeeping files. May 04 13:36:53 yes, I use 0.34 May 04 13:36:55 ah, just reverted defaults. May 04 13:36:59 yes May 04 13:37:26 wow! ssh-agent support in monotone! :DD May 04 13:37:59 I never got to use ssh-agent yet... shame on me. May 04 13:38:34 hrw, do you, by chance, know the answer to my question in #opie? May 04 13:38:34 hrw: thanks for the explanation of this change May 04 13:39:54 * hrw upgrades to 0.34 May 04 13:41:51 db migrate took <1m May 04 13:42:12 is it safe to migrate with dirty working area? May 04 13:42:34 should be May 04 13:42:49 ok. just wondering. Not that I'm going to do it right away. May 04 14:12:40 ERROR: This autoconf log indicates errors, it looked at host includes. Rerun configure task after fixing this. Path was '/home/hrw/devel/build/openmoko/tmp/work/armv4t-angstrom-linux-gnueabi/firefox-1.0.7-r0/mozilla' May 04 14:12:44 ;d May 04 14:15:03 insane is approtriately neamed for firefox May 04 14:19:53 hrw, I bet it picks up some *-config script from host... May 04 14:20:26 anyway I got 2.0.0.3 built. but it will probably does not work (as usual) May 04 14:20:33 snafu :) May 04 14:29:43 polyonymous: hehe... May 04 14:40:33 PR = should or usually start from r0 or r1 May 04 14:40:53 I think omitted PR == r0 May 04 14:41:42 so if set it explicitly, should start from r1 May 04 14:41:44 ? May 04 14:41:56 When I omit PR, it has a tendency to set the PR to "0", not "r0" May 04 14:42:08 bullshit May 04 14:42:12 hmm.. I think not. May 04 14:42:16 check conf/bitbake.conf May 04 14:42:22 PR = "r0" should be there... May 04 14:42:34 zecke: possible May 04 14:42:40 Dunno why that happened May 04 14:42:40 NAiL, you summoned evil bitbake spirit :) May 04 14:42:44 haha May 04 14:43:30 this is not important at all, but just think I'd ask if you guys has a hard line set for this May 04 14:43:41 well, anyway ... May 04 14:44:17 well, you see, opinions are like assholes - everyone has one. May 04 14:44:52 xjqian: if you omit PR it is taken from conf/bitbake.conf, it starts at "r0" (unless you changed it in your local.conf) May 04 14:45:16 xjqian: normally PR = "r0" is omitted from the .bb files May 04 14:48:22 I understand that. I'm asking what's the best/common practice. So what's the first version number people start with their bb file? 0? 1? or everybody to his own? May 04 14:49:04 r1 May 04 14:49:06 this is open source project. so probably the last one! May 04 14:49:31 xjqian: well, for every change effecting the resulting package you increment it by one :) May 04 14:50:46 zecke: yeah, at least I had that bit of common sense8-). I am asking what if you start from fresh. May 04 14:53:34 well, I take it as "everybody to his own". not worth anymore discussion. thanks for all replied. I'm leaving... May 04 14:54:49 xjqian, do not put any PR for the first package ever. May 04 14:56:38 First package == r0. As r0 is inside bitbake.conf omit from the bbfile May 04 14:57:35 my suggestion really amounts to "get back to the problem when you really have it" :) May 04 15:06:58 polyonymous: got your spirit. thanks May 04 15:13:29 do you guys have a list of what can be put in SECTION = " " ? Does SECTION determines where the package goes to in the feed? May 04 15:14:04 if so, the list is the feed skeleton, isn't it? May 04 15:15:12 xjqian: yes, and yes, there's a section list in OE manual May 04 15:22:43 psokolovsky: will read koen's new release of OE manual this weekend. thanks May 04 15:46:20 re May 04 15:48:13 hi koen May 04 15:50:42 hey hrw May 04 15:52:07 koen: how goes rpath changes? May 04 15:52:17 hrw: no idea May 04 15:52:26 Crofton and likewise are more into it May 04 15:52:57 I have been very busy with real work :( May 04 15:53:05 hrw: not very well, I'm lacking libtool/autoconf/configure knowledge to find the sources of evil. May 04 15:53:40 If I get a chance, what package shoud I look at? May 04 15:53:57 I like RP's install patch, but we need to (re)fix do_deploy May 04 15:54:28 priority for me is the issue discussed this morning (task-base and glib-2.0/NLS, and reporting of failed providers in BB) May 04 15:54:53 Crofton: this looks like a nice bug to fix as it's a standalone pkg: http://bugs.openembedded.org/show_bug.cgi?id=2208 May 04 15:54:55 I agree with RP that splitting up task-base will lead to a mess May 04 15:55:24 likewise: we could special-case gcc-cross(-initial) in the rpath check May 04 15:55:33 koen: the suggestions that it was debian class related seems false. I can reproduce without debian as well. May 04 15:56:07 koen: yes, the old test (against ../work) already skipped gcc-cross. Shall I revert to that search string (works here)? May 04 15:56:32 Has the fix to gcc-cross been committed? May 04 15:56:37 I think RPATH shouldn't contain work nor staging May 04 15:57:19 Crofton: which fix?rpath fix?? May 04 15:57:39 gcc-cross was faling on rpath, right? May 04 16:03:10 ok, just updated and am baking madplay May 04 16:03:15 I'll see where that takes me May 04 16:04:15 Crofton: yes, but there was no fix yet (AFAIK) May 04 16:07:25 lots of bb "inherit pkgconfig", but lots among them don't use .pc. Is it because people are lazy (don't bother to check whether there are .pc files in the src)? What's the best practice if I've checked that there are no .pc files: fail safe to always inherit pkgconfig or keep the bb file slim? May 04 16:12:23 its safe to inherit but can be removed if not needed May 04 16:13:15 it could even be inherited by default in base.bbclass safely May 04 16:14:08 koen: that's what I thought. If it's always there and safe, why not default it May 04 16:14:38 and then binconfig? May 04 16:18:02 xjqian, koen: Maybe we should add this to the insane.bbclass? May 04 16:26:16 * Laibsch noticed with some satisfaction that people are reading my UpgradingPackages page May 04 16:28:56 Laibsch: that's me. :) May 04 16:30:12 Laibsch: ...reading that page with a little dissatisfaction :) May 04 16:30:21 it is a wiki May 04 16:30:47 you want me to modify it a little bit May 04 16:30:49 But luckily it is not only you, the main devs seem to keep follow the advice, too May 04 16:31:11 xjqian: Sure. I will revert your changes if I don't like them :-D May 04 16:32:21 xjqian: I don't think that page is really for you May 04 16:32:38 xjqian: It serves only people with RW access to OE repo May 04 16:32:59 Unless you have that you can safely ignore it May 04 16:38:48 I just made it clear that you want to mtn add the original version of the bb file back in. not the revised new version May 04 16:39:42 wiki change committed, please feel free to revert May 04 16:45:21 thesing: Is there a way to gather the needed info for you to add support for non-working cards into the driver? May 04 16:45:37 I though an SD was basically just an SD May 04 16:46:22 Laibsch: the sd and mmc specs have multiple revisions and vague parts May 04 16:46:31 and not every SD cards conforms to the spec iirc May 04 16:46:45 (just like usb) May 04 16:47:52 Laibsch: Yes. use the modules compile with debug-info, do "echo "1" > /sys/bus/locomo-bus/drivers/locomo-spi/verbose" and insert a card. May 04 16:48:56 Laibsch: But you will get a lot output. May 04 16:49:21 thesing: no problem May 04 16:49:29 thesing: I got a ton of lines scrolling by on the screen May 04 16:49:45 thesing: I will redirect them into a file and attach to the BTS. May 04 16:49:50 Laibsch: you need serial cable for it May 04 16:50:04 syslog is too slow May 04 16:50:13 * Laibsch forgot to think about CF card as "sneaker net" while USB is not yet functional May 04 16:50:32 hrw: I don't have a serial cable, but see above May 04 16:50:38 I think that might work May 04 16:50:51 dmesg buffer is too small? May 04 16:50:54 Laibsch: I had that problem ~year ago May 04 16:50:58 koen: much too small May 04 16:51:06 Does anybody here have experience debugging problems in pxafb? I would like to work on a fix for #2120 May 04 16:51:16 hrw: Redirection to a file is not possible??? May 04 16:51:52 Laibsch: can try via syslog May 04 16:51:58 Laibsch: What is the problem with your cards are they recognized? May 04 16:52:26 thesing: I am the one with all the bug reports in the BTS May 04 16:53:27 Laibsch: I know ;) May 04 16:54:26 thesing: You are talking about all the stuff flying by on the console, are you? May 04 16:54:37 So I do not need to activate anything further, right? May 04 16:55:26 Wrong. May 04 16:56:07 OK May 04 16:56:11 Then I will do that May 04 16:56:22 How do I redirect output into a file? May 04 16:56:37 I know how to do that generally May 04 16:56:49 You will see every byte the card sees and writes ;) May 04 16:57:25 How to capture? May 04 16:58:53 I'm not sure. Maybe something like "cat > file 2>1" ? May 04 16:59:06 Well, that I know ;-) May 04 16:59:19 Question is *what* to capture May 04 16:59:29 IOW, where is this normally written to? May 04 17:00:09 I don't know May 04 17:00:24 kernel log buffer -> syslog iirc May 04 17:01:09 ok. collie booted May 04 17:02:52 koen: I already tested /var/log/syslog which is not there May 04 17:03:38 its about 5KB/s May 04 17:03:50 /var/log/messages? May 04 17:04:04 koen: no - minicom logging May 04 17:04:23 hrw: I was talking to Laibsch :) May 04 17:04:40 ;D May 04 17:04:42 koen: Only file in /var/log is wtmp May 04 17:04:50 But hrw already told you ;-) May 04 17:06:46 thesing: http://blog.haerwu.biz/tmp/collie-log.text.bz2 is card insert of my 512M sd card (not recognized) May 04 17:07:09 1.1M uncompressed, 10K compressed May 04 17:07:30 Laibsch: you have to edit /etc/syslog.conf to get a logfile. May 04 17:07:46 hrw: thanks. May 04 17:08:05 thesing: look inside and tell does it is usefull. if yes then I can test other cards May 04 17:08:51 hrw: Would be cool if you attached that to the BTS May 04 17:09:07 I will write up how to capture this info (unless you beat me to it) May 04 17:09:32 which bug? May 04 17:09:43 !oebug 2161 May 04 17:09:44 * * Bug 2161, Status: NEW, Created: 2007-04-29 07:48 May 04 17:09:46 this? May 04 17:09:46 * * bugs.openembedded.org(AT)rolf.leggewie.biz: unrecognized cards with new collie SD card driver May 04 17:09:47 * * http://bugs.openembedded.org/show_bug.cgi?id=2161 May 04 17:09:47 There is a collie meta-bug which links it May 04 17:09:52 Yes, that is the one May 04 17:10:22 I think that is where we should discuss unsupported cards May 04 17:11:10 hrw: What did you add to syslog.conf to make it work? May 04 17:11:26 nothing May 04 17:11:29 Just looking at it I would have thought it already produces a log May 04 17:11:39 hrw: So how did you capture? May 04 17:12:20 Laibsch: kernel sends all to serial May 04 17:12:27 Lucky ouy May 04 17:12:29 you May 04 17:12:34 Laibsch: serial connected to laptop. laptop runs minicom May 04 17:12:35 hrw: The log is nearly useful. The version of the driver you used has a typo, which I already fixed locally I will update the patches. May 04 17:12:57 thesing: Can you help me with syslog.conf? May 04 17:12:58 Laibsch: we need to organize amazon buy of ce-170ts cables? May 04 17:13:30 I usually don't use a serial cable May 04 17:13:35 Laibsch: yes. wait a moment. May 04 17:13:45 thesing: DESTINATION? May 04 17:13:59 file May 04 17:14:53 hrw: Well, I'd have 5$ to spare, though ;-) May 04 17:15:11 for LOGFILE I use a file in /tmp the flash is much too little for the huge logs ;) May 04 17:15:54 Laibsch: but amazon does not sell electronics outside US May 04 17:16:12 hrw: I have contacts in the US May 04 17:16:31 Laibsch: then buy few cables May 04 17:17:22 Maybe it would be even easier to ask one of the guys hanging around here May 04 17:17:34 hvontres|home: Are you listening ;-) ? May 04 17:17:40 Crofton|home: ping? May 04 17:18:05 thesing: I assume that you have serial cable - right? May 04 17:18:19 hrw: Your collie kernel is not current. OE already includes the latest mmc-patches. May 04 17:18:25 hrw: right. May 04 17:18:59 thesing: ok. will build newer May 04 17:19:44 current modules should do. May 04 17:19:51 mmc-spi r4? May 04 17:20:19 yes. but the important one is locomo-spi r3 May 04 17:20:28 ok May 04 17:20:29 thesing: I put in a line "FILE=/tmp/syslog" and restarted syslog May 04 17:21:40 Laibsch: it should be LOGFILE May 04 17:22:10 thesing: It already had a LOGFILE line May 04 17:22:19 set to /var/log/messages which was not there May 04 17:22:26 * Laibsch retries May 04 17:22:39 Laibsch: http://www.openembedded.org/repo/org.openembedded.dev/packages/busybox/busybox-1.2.1/angstrom/syslog.conf May 04 17:23:23 morning May 04 17:24:10 mreimer_: good morning! May 04 17:24:38 hi mreimer_ May 04 17:26:02 Laibsch: you need to comment the ROTATE* lines out. May 04 17:26:54 got syslog working now May 04 17:27:24 hvontres|poodle: ping May 04 17:27:30 Laibsch: pong May 04 17:28:14 Can you scroll back to our serial cable discussion and tell us if you can help us? May 04 17:28:25 We are looking for a buyer/shipper in the US. May 04 17:28:35 hvontres|poodle: which will send stuff to europe May 04 17:28:37 Laibsch: I might be able to :) May 04 17:28:42 Cool May 04 17:29:01 Laibsch: how many do you guys think you need? May 04 17:29:06 no idea May 04 17:29:11 hvontres|poodle: http://www.amazon.com/Sharp-Zaurus-SL-5500-CE-170TS-RS-232C/dp/B0000645V7 is url May 04 17:29:15 I don't really need it May 04 17:29:25 But it is too cheap to pass ;-) May 04 17:29:53 Laibsch: if you get "locomospi: tx: " lines in the log, you need to update the mmc modules too. May 04 17:29:53 hvontres|poodle: I guess we should look around and then come back to you May 04 17:30:36 hvontres|poodle: I think that we can get 10 cables for start May 04 17:30:39 thesing: OK. Do you know if koen's image from April 22nd has the debug info turned on? Otherwise I'd need to build first and reflash May 04 17:31:38 make[2]: *** No rule to make target `drivers/mmc/mmc_spi.c', needed by `drivers/mmc/mmc_spi.o'. Stop. May 04 17:32:20 Laibsch: they had. but I updated the drivers on April 24 . May 04 17:32:51 hrw: I thought I already corrected this ... May 04 17:33:17 I am trying to build x11 image (angstrom distro) but it failes when making prismstumbler. Is this s known bug? : pastebin.co.uk/14016 May 04 17:33:23 thesing: I only have /sys/bus/locomo-bus/drivers/locomo-backlight, not /sys/bus/locomo-bus/drivers/locomo-spi so I guess a recompile is in order, right? May 04 17:33:39 thesing: OK, I will recompile May 04 17:34:00 Laibsch: did you load the modules? May 04 17:34:04 Laibsch: modprobe locomo_spi May 04 17:34:20 goxboxlive: prismstumbler? iirc yes - check bugtracker please May 04 17:34:34 goxboxlive: have you searched bugzilla? May 04 17:34:57 * koen points to rule #5 from http://www.angstrom-distribution.org/ May 04 17:35:21 hrw: Yes I was right: mmc_spi-r4.patch creates this file. May 04 17:35:41 hrw: how do you guys want to handle this? I can give you my paypal account May 04 17:36:07 thesing: 675d93727582e370a9a1b4d4f4c42c57 patches/mmc_spi-r4.patch May 04 17:37:18 hvontres|poodle: I can pay for cables. you will get them from amazon and then send do the ones who needs them. they pay post costs May 04 17:37:48 hrw: ok, sounds good. May 04 17:37:55 Laibsch: write mail to oz-devel/angstrom-devel about that offer - maybe few other non-US guys will want cable too May 04 17:38:21 ce-170ts is not comfortable cable but it is very cheap ;D May 04 17:38:34 hrw: I have a different one. (if this is a md5 sum) May 04 17:38:48 thesing: I see May 04 17:39:48 thesing: my is without mmc-spi.c file.. May 04 17:40:26 reapplied patch, building May 04 17:42:58 BTW: it would be best if you could give both a short a verbose log of nonworking cards. May 04 17:44:02 short and verbose? you mean with locomo debug and without? May 04 17:44:14 reverse :) May 04 17:44:30 I can provide you such one May 04 17:44:34 exactly May 04 17:44:37 but probably not today May 04 17:44:45 * hrw -> food May 04 17:46:51 hrw: Cables Ordered :) May 04 17:51:48 how many? and how much for you? May 04 17:52:38 hrw: "I can pay for cables. you will get them from amazon and then send do the ones who needs them. they pay post costs" -> did you mean shipped individually from the US? May 04 17:52:46 I think that is going to get expensive May 04 17:52:50 hrw: 10@$63.99 what is your e-mail and I'll forward you the Confirmation May 04 17:53:03 hvontres|poodle: openembedded@hrw.one.pl May 04 17:53:19 Laibsch: we can get pack from hvontres|poodle and share them in europe May 04 17:53:24 hrw: sent May 04 17:53:36 hrw: Yes, I think bulk shipping to Europe is better May 04 17:53:51 Laibsch: sure May 04 17:53:57 BTW: I put my current quilt patch stack at www.informatik.hu-berlin.de/~tkunze/zaurus/patches/quilt_stack. Just in case somebody wants to test the nearly ready suspend. May 04 17:54:09 Laibsch: so you will handle European part May 04 17:54:17 If you want me to May 04 17:54:51 Laibsch: I have one such cable already and no need for second May 04 17:55:00 Laibsch: I'll let you know when they show up. Any hints on how to get them through customs cheaply ? May 04 17:55:02 Laibsch: you lack one so you have interest in it May 04 17:55:17 hvontres|poodle: you have only poodle? May 04 17:55:20 hvontres|poodle: I was thinking about that May 04 17:55:26 hvontres|poodle: declare them as gift usually works May 04 17:55:27 * hvontres|poodle feels a bit silly for building one himself May 04 17:55:50 Laibsch: you check cable with spitz - if it will work in both ways then send one to Richard May 04 17:55:56 * hrw goes to shop May 04 17:56:02 hvontres|poodle: First of all, it needs to be declared as a gift. No invoices or such things inside. Then I guess crummy "mum-and-dad" packaging may help May 04 17:56:03 hrw: yup. and right now the sharp cable wouldn't work anyways May 04 17:56:23 Laibsch: I can do that :) May 04 17:56:53 hvontres|poodle: I will send you my address May 04 17:57:00 * hvontres|poodle wonders if he should "gift wrap" the cables to confuse the inspector if he opens the box :) May 04 17:57:23 hvontres|poodle: I think there is also a value threshold, something like 35€ or so. May 04 17:57:34 I will check up on it and let you know May 04 17:57:49 !oebug 2096 May 04 17:57:50 * * Bug 2096, Status: NEW, Created: 2007-04-09 08:20 May 04 17:57:51 * * hvontres(AT)sbcglobal.net: 2.6 Kernel on Poodle disables VCC (Pin11) on IO connector May 04 17:57:52 * * http://bugs.openembedded.org/show_bug.cgi?id=2096 May 04 17:59:23 hvontres|poodle: Don't forget to move that one away from OZ or it is at risk of being autoclosed with WONTFIX ;-) May 04 18:01:02 Laibsch: I'll test it on Angstom tonight. it should do the same thing tho.. May 04 18:01:43 I assume so, yes May 04 18:03:48 wow, important news. "Don't hug the queen" May 04 18:04:04 zecke: ???????????? May 04 18:04:10 germany has a queen? May 04 18:04:30 koen: hehe, the queen of the U.K. Just came in the news :) May 04 18:04:35 "I didn't vote for her" May 04 18:04:36 koen: I think Germany has lots of "Queens" ...:) May 04 18:05:03 * mreimer_ makes a weak allusion to The Holy Grail May 04 18:18:43 hi all May 04 18:18:49 hi mrdata May 04 18:19:22 woglinde: hi, have you test the new patches for kernel 2.6.21? May 04 18:21:54 hvontres|poodle: Got my message in private IRC? May 04 18:22:31 mrdata no net yet May 04 18:22:43 my actual project is porting phoneme to qt4 May 04 18:22:59 woglinde: no problem May 04 18:23:13 mrdata hm I also started to look in the sound driver May 04 18:23:20 to ported it to 2.4 May 04 18:23:25 aeh 2.6 May 04 18:23:48 it seems not to difficult May 04 18:23:51 woglinde: a new one or the old oss driver? May 04 18:24:05 mrdata when port it to alsa May 04 18:24:28 woglinde: alsa (soc) would be the right solution i think May 04 18:24:51 but but first needed is the mcp/ucb sound stuff I guess May 04 18:25:36 woglinde: yes May 04 18:26:13 hi mrdata May 04 18:26:25 woglinde: for kernel 2.4 i had never sound working May 04 18:26:30 thesing: hi ;-) May 04 18:26:42 mrdata hu? May 04 18:26:44 mrdata: Does the mcp driver work on simpad after resume? May 04 18:27:12 mrdata: I only get timeouts May 04 18:27:34 thesing: not realy, i must build it as module, unload it while suspend and reload after resume, then need to start ts_test May 04 18:28:35 thesing: you must unload/load ucb1x00_ts and mcp_sa11x0 May 04 18:29:18 mrdata: the touchscreen works after suspend, but I get no battery voltage May 04 18:29:53 thesing: battery is measured over your locomo right? May 04 18:30:12 thesing: or also over ucb1200 adc May 04 18:31:20 mrdata: over ucb May 04 18:32:25 thesing: battery measurment works after suspend May 04 18:33:52 mrdata: thanks. May 04 18:34:03 thesing: i have a ucb1x00-simpad module and this works also after suspend May 04 18:35:29 mrdata: where can I find the patch for this module? May 04 18:36:41 ~change 64 USD to pln May 04 18:36:48 i got some error trying that May 04 18:36:56 ~change 64 usd to pln May 04 18:37:02 i got some error trying that May 04 18:38:36 thesing: i have only a local one, why we would make a "new" battery-monitor driver, using the new battery-class from hh.org kernel May 04 18:39:04 thesing: but i could send you my version, if you wish May 04 18:40:41 thesing: and this module is for measurment simpad accu-voltage, power-supply-voltage and charging "current" May 04 18:40:59 mrdata: Do you still have my email address? May 04 18:41:23 thesing: i should have it from last week i think May 04 18:42:39 thesing: i have you email address ;-) May 04 18:43:43 mrdata: Ok. Then let the homer fly. May 04 18:43:50 xjqian: You caught an important logical error. Thank you. May 04 18:44:56 re May 04 18:45:01 wb May 04 18:45:06 hvontres|poodle: mail me your paypal account May 04 18:48:49 mrdata: can you tell me where in der hh.org kernel tree the battery-class is? May 04 18:49:26 thesing: battery driver is extracted by bd2 into separate git tree May 04 18:49:52 thesing: http://git.infradead.org/?p=users/cbou/battery2-2.6.git May 04 18:50:11 hrw: thanks. May 04 18:50:34 thesing: it will be used in olpc and it was blessed by dmwm2 May 04 18:55:30 hrw: sent May 04 18:55:48 mail to oz-devel/angstrom-devel sent too May 04 18:56:12 thesing: mail is send, and comes from googlemail May 04 18:56:50 mrdata: I got it. Thanks. May 04 18:57:05 thesing: I attached both a verbose and non-verbose log to bug 2161 May 04 18:57:08 http://bugs.openembedded.org/attachment.cgi?id=1416 May 04 18:57:12 http://bugs.openembedded.org/attachment.cgi?id=1419 May 04 18:57:12 thesing: have fun with it May 04 18:58:34 thesing: I do see the tx stuff, so I guess I will need to recompile May 04 18:58:39 Laibsch: sorry but you didn't use the latest modules May 04 18:59:11 Laibsch: I will give you latest May 04 18:59:37 hrw: you are too fast ;) May 04 19:00:00 http://www.angstrom-distribution.org/unstable/images/collie/20070422/modules-2.6.20.6-collie.tgz ? May 04 19:00:14 no - they are old May 04 19:00:22 OK May 04 19:02:41 Laibsch: http://blog.haerwu.biz/tmp/collie/t.tar.gz May 04 19:02:44 in a moment May 04 19:02:48 now ;) May 04 19:04:37 got it May 04 19:06:21 ~hail thesing for locomo-spi r3 May 04 19:06:26 thesing: rulz May 04 19:06:31 * ibot bows down to thesing for locomo-spi r3 and chants, "I'M NOT WORTHY!!" May 04 19:07:07 http://www.mistralsoftware.com/html/product/omap_products/products-omap2430_starterkit.php May 04 19:07:16 thesing: mmcblk0: mmc0:0000 SD512 500224KiB May 04 19:07:17 hrw: your card works now? May 04 19:07:21 yes May 04 19:08:16 to be exact: this fix was done in mm_spi-r4 ;) May 04 19:08:20 ok May 04 19:08:37 now 1GB one May 04 19:09:50 thesing: mmcblk0: mmc0:0000 SD 976896KiB May 04 19:10:45 now md5sum test May 04 19:10:49 lol May 04 19:11:38 root@collie:~$ for f in `seq 1 3`;do time md5sum /dev/mmcblk0 >>/tmp/md5sums;done May 04 19:11:52 this will take some time... May 04 19:12:32 thesing: after that test I will check rs-mmc and microsd cards May 04 19:13:33 maybe you should build non-debug modules and test with them. the md5sum test will be much faster. May 04 19:15:00 thesing: witch transfer rates for read and write are now for sd/mmc ? May 04 19:15:20 thesing: I know and will build ones May 04 19:15:56 mrdata: AFAIR 500KB/s May 04 19:16:37 * Laibsch added HOWTO to bug 2161 that details steps necessary to collect this debug info May 04 19:17:36 thesing: not bad, for kernel 2.4 on simpad (software-bitbang-spi) i had only 125KB/s May 04 19:17:56 hrw: Will I need the kernel as well or just the .ko modules? May 04 19:18:11 I did both May 04 19:18:11 Laibsch: the modules will do. May 04 19:18:22 OK May 04 19:19:03 thesing: if the rest of kernel works on simpad, i will return to my mmc/sd-card driver and port it to kernel 2.6 May 04 19:20:50 mrdata: with the mmc_spi.patch and the spi-bitbang modules this should be easy. May 04 19:21:51 thesing: that sound's great ;-) May 04 19:22:47 thesing: you mean your mmc_spi-r4.patch and spi-bitbang from kernel? May 04 19:24:26 yes. May 04 19:25:03 http://www.vidvoting.com/xoops/modules/xstreamer/index.php?class=video&id=52 - cocaine production in easy steps May 04 19:26:57 hrw: and if you still wand to take some after learning how to process it...nothing like gasoline and battery acid to make your day :) May 04 19:27:47 mrdata: http://www.uwsg.iu.edu/hypermail/linux/kernel/0701.3/0737.html somewhere in this thread I got the mmc_spi.patch (Just in case I my version is broken on non collies) May 04 19:28:32 thesing: thanks, i will have a look on it May 04 19:28:34 thesing: rs-mmc not recognized May 04 19:29:34 hrw, thesing. First card is recognized by new driver May 04 19:30:43 second one, too May 04 19:31:19 the 1G card is not. May 04 19:32:36 good evening. May 04 19:33:14 Laibsch: Did you try a few time eject an reinsert? sometimes this is needes. (especially if one forgets to unmount the cards.) May 04 19:33:17 thesing: added rs-mmc logs to @2161 May 04 19:33:39 thesing: I am doing that now May 04 19:33:42 so far no luck May 04 19:35:08 With the first two I got an error of "no valid FAT filesystem being present". The partition was automounted and shown with correct size by df -h, though Furthermore I would get an error of "mmc0: error requesting CID: 4" upon removal May 04 19:36:53 thesing: No, this card just won't work. I will attach the logs May 04 19:37:45 Laibsch: This is normal. May 04 19:38:16 OK. May 04 19:40:23 thesing: mmcblk0: mmc0:0000 SU128 123008KiB May 04 19:40:29 thesing: microSD 128M May 04 19:41:27 thesing: so only rs-mmc one does not work here. but I will check 16M mmc May 04 19:41:35 which worked May 04 19:42:46 recognized May 04 19:42:54 koen, we must get one of these omap2430 starter kits :) May 04 19:43:06 thesing: mmcblk0: mmc0:0001 16M 15680KiB May 04 19:45:26 hrw: fine. so nearly all cards work now. I have an idea what your rsmmc might need. May 04 19:46:40 great May 04 19:47:04 Crofton: that's no problem :) May 04 19:47:26 Crofton: Rishi will send both of us a OSK if we ping him :) May 04 19:47:29 the mistral website says delivery is 16 weeks May 04 19:47:39 I just pinged him on something else May 04 19:47:56 you see the link I pasted a while back, they list sdr as an app May 04 19:48:11 I think I saw the link May 04 19:48:34 I also have one of the lyretch expensive SDR boards May 04 19:48:45 * Crofton explodes from to much to do May 04 19:49:08 is teh lyrtech OE-able? May 04 19:49:11 thesing: md5sum not always match - first one is wrong, next ones are same (no mmc_debug) May 04 19:49:15 I hope so May 04 19:49:17 or too much binary crap in the wrong places? May 04 19:49:29 it comes with inegrity, must cure that problem May 04 19:49:55 I suspect some of the people involved will want Angstrom on it May 04 19:50:14 there may be some hairy driver work to speak to the FPGA May 04 19:50:25 Can I load Angstrom on a 'Compaq OHH BootLoader'? May 04 19:51:03 hrw: the fs-layer seems to write sometimes to the card. Maybe that's the reason. May 04 19:51:18 maybe May 04 19:51:50 On a H5550 May 04 19:51:51 vervain: if you can put other linux on it then angstrom too May 04 19:52:56 Oh ok... I have familiar 8.4 from a while back and was willing to give Angsrom a whirl if it's the new sheriff... But don't see any H5xxx docs in the Angstrom Wiki. May 04 19:53:12 vervain: installation procedure is same May 04 19:53:14 iirc May 04 19:53:52 it is May 04 19:54:14 but with the caveat that 2.6 isn't finished yet on the h5xxx, psokolovsky should know the status of that May 04 19:55:29 koen: do you have working oabi builddir? May 04 19:55:47 hrw: no, but I plan to have one tomorrow :) May 04 19:56:04 koen: then please do build for collie May 04 19:56:12 * koen is building armv5teb now May 04 19:56:23 * hrw tries to build oabi again May 04 19:57:21 frigging ti May 04 19:57:29 http://focus.ti.com/dsp/docs/dspsupporttechdocsc.tsp?sectionId=3&tabId=409&familyId=1302&abstractName=sprue20a May 04 19:57:36 the pdf has the text backwards May 04 19:57:52 heh May 04 19:58:14 * koen can't wait till july when the bulk of the ti DSP stuff goes open source May 04 19:58:23 I suppose I need to find acrocrap for Linux May 04 19:58:35 spreading rumours again? May 04 19:58:53 if rumours is "rishi said it", yes :) May 04 19:58:57 heh May 04 19:59:07 * Crofton hopes rishi knows what he is talking about May 04 19:59:22 off to a meeting May 04 19:59:39 * koen puts on an episode of NCIS May 04 20:00:04 * Crofton wonders why people are obsessed with American TV shows May 04 20:03:07 Crofton: because uk ones suck, although I prefer Japanese May 04 20:03:29 i think all tv sucks equally May 04 20:03:52 * Crofton watched BBC America :) May 04 20:04:07 the grass is always greener on the other side of the Atlantic:) May 04 20:04:35 http://blog.haerwu.biz/2007/05/04/sdmmc-for-collie26-progress/ May 04 20:04:52 Crofton: main thing is you guys have such a vast output, a few of them are actually good May 04 20:05:03 Crofton: wherease we make a couple and so they tend to suck ass May 04 20:05:07 real 2m 47.07s May 04 20:05:07 user 0m 21.55s May 04 20:05:07 sys 0m 3.90s May 04 20:05:31 this is time of md5sum /dev/mmcblk0 on collie without mmc debug - 128M microsd card from Neo1973 May 04 20:06:31 hrw: that newet patch in OE? May 04 20:07:12 it is May 04 20:07:23 for over a week May 04 20:07:24 * XorA|gone goes to kick off a build May 04 20:07:33 I must have missed the commit notice May 04 20:07:56 hmm May 04 20:08:02 Timing buffered disk reads: 4 MB in 5.15 seconds = 795.01 kB/sec May 04 20:08:04 I think we are getting over recent events May 04 20:08:26 I am in a meeting where someone is suggesting we attempt to obtain guilt funding May 04 20:09:57 XorA|gone: http://www.burtonini.com/blog/computers/die-libgnome-die-2007-05-04-16-20 May 04 20:10:03 Laibsch: you didn't use the current locomo-spi module. could you get it from http://www2.informatik.hu-berlin.de/~tkunze/zaurus/images/locomo_spi.ko and resubmit your verbose log? May 04 20:10:26 ah scooby snax are ready back later May 04 20:10:46 koen: it is in Poky ready to take May 04 20:11:14 koen: cool May 04 20:11:15 Laibsch: I need to get everything what happends from the moment on where you put the card in. May 04 20:11:49 thesing: hm. without mmc debug it does not like 512M card which it liked before May 04 20:12:02 mmc_spi .0: CMD55 response SPI_R1: INVALID RESPONSE, 82 May 04 20:12:02 mmc_spi .0: CMD55 response SPI_R1: INVALID RESPONSE, 80 May 04 20:12:02 mmc_spi .0: CMD1 response SPI_R1: INVALID RESPONSE, 80 May 04 20:12:04 sounds like timing issues May 04 20:12:10 yep May 04 20:12:21 the aim for the proposal is to go on a pub crawl in Ireland May 04 20:12:41 * hrw reboots collie May 04 20:13:36 Crofton: you need moral support there? May 04 20:13:48 heh May 04 20:13:52 :-) May 04 20:13:58 first we need funding May 04 20:14:19 acroread is spitting out syntax errors on TI pdf file .... May 04 20:14:39 As you said "the grass is always greener on the other side of the Atlantic:)" May 04 20:14:39 hmm, all pdf May 04 20:16:55 thesing: without debug recognized are: 16M mmc, 128M microSD. fails: 512M and 1GB (invalid response to CMD55/58/1), no reaction at all to rs-mmc one May 04 20:17:34 hvontres|poodle: did not get mail from you May 04 20:19:26 hrw: I sent a message from paypal, since I am not sure how the user-id works exactly... May 04 20:20:36 so let me login into paypal May 04 20:23:25 hrw: the email I use on paypal is henry.von.tresckow@orhtodyne.com May 04 20:23:57 ok May 04 20:26:32 hvontres|poodle: sent May 04 20:27:11 thesing: http://bugs.openembedded.org/attachment.cgi?id=1424 May 04 20:27:22 and http://bugs.openembedded.org/attachment.cgi?id=1423 May 04 20:28:10 thesing: I did not use your module but the one that hrw provided earlier May 04 20:29:00 Laibsch: that shouldn't matter. May 04 20:29:04 hvontres|poodle: got mail? May 04 20:30:43 Laibsch: the verbose log seems to be truncated. Did you comment the ROTATE* lines in /etc/syslog.conf? May 04 20:31:41 hrw: not yet... May 04 20:31:59 hvontres|poodle: its paypal so it can take week May 04 20:32:29 hrw: np. I'll let you guys know as soon as the cables arrive at my house May 04 20:33:07 thesing: I had them on but later joined two files with "cat syslog syslog.0 > syslog-long" May 04 20:33:34 hrw, koen: do you guys know at what stage during image creation modules.symbols etc are built? I just tried re-building angstrom-console-image and got nothing May 04 20:34:28 Laibsch: But they only preserve 64 KB. And the log gets much bigger. May 04 20:35:25 OK, I will remove the rotation May 04 20:35:43 Laibsch: set log to atleast 1-2M in size May 04 20:36:05 hvontres|poodle: do_compile of kernel? May 04 20:42:18 likewise: http://lists.freedesktop.org/archives/cairo/2007-May/010540.html May 04 20:42:47 hrw: not sure...but that file is nowhere to be found in tmp....:( May 04 20:43:46 hrw: NOTE: package gcc-cross-4.1.2-r3: task do_compile: started (collie) May 04 20:44:23 hvontres|poodle: so looks like it is not created at all (not found in my builds) May 04 20:45:04 thesing: http://bugs.openembedded.org/attachment.cgi?id=1426 May 04 20:45:59 hrw: yet it shows up in the rootfs tarball for angstrom....root@poodle:/media/cf/angstrom# tar -tzf Angstrom-console-image-test-20070422-poodle.rootfs.tar.gz |grep modules.symbols May 04 20:46:02 ./lib/modules/2.6.20.6/modules.symbols May 04 20:47:14 hvontres|poodle: depmod started during creating of images? May 04 20:47:56 hrw: you guys would know better than me... May 04 20:49:42 koen: yes, saw that thread, but I'm not sure if it matters much for non-X. May 04 20:50:35 koen: I am not sure if we should back port the gcc-cross patch back into 4.1.1, 4.1.0 etc. Maybe we should check with ptxdist. May 04 20:51:01 * koen is happy with 4.1.2 May 04 20:51:14 koen: although angstrom and slug might be the only two active users of arm5teb eabi, which can use 4.1.2 May 04 20:52:19 * likewise is building a GCC 4.1.2 toolchain with newlib for compiling the redboot bootloader May 04 20:59:01 Laibsch: I think I can fix your card too. But not today ;) May 04 20:59:47 I'll go now. CU all. May 04 21:09:33 hrw: With the advent of ?=, is the use of auto.conf still advantageous? I'd like to move it to local.conf and rm auto.conf. I share site.conf as a hard link between the distros that I compile for. May 04 21:10:11 Laibsch: for you? I don't know, for me yes :) May 04 21:10:36 what is the advantage? May 04 21:10:55 for my autobuilder I can echo "" > without deleting anything May 04 21:10:58 Can I remove auto.conf or will that trigger some kind of error or problems? May 04 21:11:17 Laibsch: you can remove it (likely) May 04 21:11:22 zecke: It would be easier to just "MACHINE=foo bitbake bar" May 04 21:11:24 ;-) May 04 21:11:55 hrw: it looks like the problem is happening in do_rootfs... May 04 21:12:17 zecke: I have "MACHINE ?= collie" in auto.conf so I can easily override that. May 04 21:12:33 * hvontres|poodle needs to dig into this more when he gets home May 04 21:12:34 likely :) May 04 21:12:50 Laibsch: for machine in akita spitz hx4700 h2200 hx2700 ixp4xxle;do echo "MACHINE=\"$machine\"" >conf/auto.conf;bitbake angstrom-console-image;done May 04 21:12:50 this is advantage May 04 21:12:50 very simple way of spending cpu time with building for multiple machines without interactivity May 04 21:12:50 with one line you can build angstrom for all supported machines May 04 21:13:14 cu May 04 21:13:46 my laptop will build angstrom-oabi during weekend (I hope) May 04 21:18:01 hrw|gone: Yes, I understand that was the initial reason (I think ?= was not yet there back then) May 04 21:18:12 You could do the same today like this May 04 21:18:43 for machine in akita spitz hx4700 h2200 hx2700 ixp4xxle;do MACHINE=$machine bitbake angstrom-console-image;done May 04 21:20:11 Crazy! 2G SD card for 10€ May 04 21:20:14 http://www.mediaonline.de/shop/ProductDisplay?storeId=5000&jspStoreDir=mediamarkt&productId=305049&partnerId=psm&catalogId=5000&langId=-3&ddkey=MMShopByPartner May 04 21:20:29 It looks like I might need more. May 04 21:20:51 the wikipedia recompressed to 1,7G and I want at least .de and .jp May 04 21:23:13 Laibsch: wow, how do you compress? May 04 21:23:47 Are you impressed by how little or how much that is? May 04 21:27:06 yes May 04 21:27:17 1.7 of mostly text is much May 04 21:27:36 I was about to write a mail to the user list of angstrom. I found the numbers very disappointing. The original .7z file was 760MB. squashfs recompressed that to 2.0G. squashfs-lzma which AFAIK uses the same algo as 7-zip came out at 1.7G. Uncompressed size is 13G. I wonder how for example the collie would keep up with this amount of data :-D May 04 21:27:37 wikipedia is a lot of mostly text. May 04 21:28:03 but 1.7G vs 700M is disappointing May 04 21:28:43 Laibsch, how well does resulting fs compress with, say, gzip? May 04 21:32:13 polyonymous: I can test later May 04 21:32:24 Laibsch, just wondering how dense it is. May 04 21:32:30 Laibsch: where to get the original file from? May 04 21:33:21 static.wikipedia.org May 04 21:33:35 I will dig up the url that I used, just a minute May 04 21:34:22 thank you May 04 21:34:37 Laibsch: you could try to convert the html to plucker format May 04 21:34:59 zecke: Did you follow the original discussion with Ifaistos, koen, et al.? May 04 21:36:08 http://static.wikipedia.org/wikipedia/ja/%E8%AA%BF/%E3%81%B9/%E3%82%82/Wikipedia%7E%E8%AA%BF%E3%81%B9%E3%82%82%E3%81%AE%E6%A1%88%E5%86%85.html May 04 21:36:20 wrong one May 04 21:36:30 no :) May 04 21:36:33 that is a sample article May 04 21:37:06 seems this was in angstrom May 04 21:37:10 * Laibsch digs out those logs. May 04 21:37:19 interesting article :) May 04 21:37:28 Laibsch: well, don't do it now. I will go to bed :) May 04 21:38:00 http://static.wikipedia.org/downloads/November_2006/ja/ is it May 04 21:38:20 zecke: yes, that is the one May 04 21:38:57 ~logs May 04 21:39:01 logs is, like, apt/ibot/infobot/jbot/purl all log daily to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz May 04 21:40:28 zecke: Too bad, it seems there are no logs for the angstrom channel. May 04 21:40:38 But if you are interested I can send you a copy May 04 21:40:46 tomorrow ;-) May 04 21:41:20 let us see. I plan to take a deep look into textbooks :) May 04 21:41:42 why, Laibsch, I do have ~/.xchat/xchatlogs :) May 04 21:43:13 zecke: Good for you. Take the studies serious (although they will certainly suck at times) May 04 21:43:16 polyonymous: ? May 04 21:43:26 I have mine readily available too May 04 21:43:29 Laibsch, the log for angstrom channel is there :) May 04 21:43:32 But zecke was leaving May 04 21:43:51 I keep personal logs, too May 04 21:43:55 Laibsch: I still have the possibility to finish in time... even if it is a bit unlikely May 04 21:44:06 ahh, you wanted a link to log. I thought you wanted to find link IN the log. May 04 21:44:11 zecke: Well, in time is faster than most May 04 21:44:24 polyonymous: The link is up above ;-) May 04 21:44:24 :} May 04 21:44:31 Laibsch, nevermind :) May 04 21:44:44 polyonymous: Thanks for the implicit offer :-) May 04 21:45:17 Yes, I even typed the offer of sending 1.8M uncompressed over dcc, but erased, thinking that you have it, anyway :) May 04 21:56:21 lrg_: better get a new passport quick :-D May 04 21:57:56 XorA|gone: i know, very worrying :/ May 04 21:58:25 lrg_: I guess a Lab Con LDP coolition on the cards then May 04 21:59:27 XorA|gone: enjoy the talk and beer ? May 04 21:59:45 talk was pish, apparently meritocracy is now called eXceed May 04 21:59:59 buzzwords May 04 22:00:01 I guess Tom Laird is a Vista eXceed user May 04 22:00:27 humming bird eXceed I mean May 04 22:00:57 we had best eXceed (tm) then May 04 22:01:05 lrg_: yup :-D May 04 23:29:00 my laptop died in the middle of a 'bitbake nano', is there a way to resume from where I left off? May 04 23:29:17 bitbake nano May 04 23:29:26 zecke_: ok, thanks May 04 23:29:38 I'm used to gentoo's 'emerge --resume' so I wasn't sure May 04 23:29:41 ;-) May 04 23:29:48 most operations should be interruptible May 04 23:29:54 Laibsch: I'm in bed :) May 04 23:30:09 thanks, I'll be back in a bit May 04 23:30:16 except as summatusmentis May 04 23:30:21 That is where I am heading now, too May 04 23:30:26 good night y'all May 04 23:31:06 night May 04 23:32:39 sleep time, cyas May 05 00:48:01 does anyone have issues with beryl going insane on the CPU while running bitbake? May 05 00:54:21 summatusmentis: no (of course I don't run beryl so my answer is not very useful ;) May 05 00:55:08 v8jlene: yeah, that's helpful :-P for some reason when I was running bitbake nano, beryl jumped to 86% CPU usage May 05 00:56:11 That doesn't really make sense... does it always happen or just now and then? May 05 00:57:44 I don't know.. it's done it a couple of times now... I'm working on setting up OE May 05 01:56:13 03lenehan 07org.oe.documentation * rfdc83f33... 10/ (1 usermanual/chapters/usage.xml): usermanual: Add some details in the usage section on how to bring up a devshell. **** ENDING LOGGING AT Sat May 05 02:59:56 2007