**** BEGIN LOGGING AT Fri Dec 08 02:59:59 2006 Dec 08 06:19:40 03awilcox 07org.oe.dev * rf14146ee... 10/ (1 packages/apache2/apache2-native_2.2.3.bb): apache2-native: replace i686 references with proper ${HOST_SYS} and ${STAGING_...} Dec 08 07:31:37 03likewise 07org.oe.dev * rdd905d1f... 10/ (6 files in 4 dirs): lighttpd: add 1.4.13, closes #1662 Dec 08 07:35:45 good morning all Dec 08 07:37:17 morning Dec 08 07:37:24 multiple heads in .dev again :( Dec 08 07:38:26 so merge them Dec 08 07:38:31 it's not rocket science Dec 08 07:42:18 never did that... wouldn't even know how Dec 08 07:43:54 'mtn merge', like it says in the mtn manual Dec 08 07:44:00 like I said, it isn't rocket science Dec 08 07:46:41 hello Dec 08 07:46:55 hey Genesis Dec 08 07:48:29 koen: okay, that was indeed easy :) Dec 08 07:48:40 :) Dec 08 07:48:53 packaging some stuff , i saw that lot of them was shipping with some m4 files that define some AM_PATH_THING Dec 08 07:49:12 is this way better that a pkgconfig .pc metadata file ? Dec 08 07:49:41 that could be check with a PKG_CHECK_MODULES() Dec 08 07:50:38 i find that work very well with the pkgconfig bbclass Dec 08 07:52:56 * Genesis needs packager skills for advise :) Dec 08 08:18:30 morning Dec 08 08:18:37 morning all Dec 08 08:20:02 hey XorA & RP Dec 08 08:42:59 03koen 07org.oe.dev * ree3f1c72... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.3: use glibc-initial 2.3.2 for powerpc, should close #1637 Dec 08 08:54:45 03koen 07org.oe.dev * r26c5fc34... 10/ (1 site/arm-linux): arm site: add tcpnodelay + cork test Dec 08 09:22:02 Good morning! Dec 08 09:22:36 What input methods are available for gpe? http://ewi546.ewi.utwente.nl/tmp/hrw/fb/?action=section§ion=gpe%2Finputmethods does not exist Dec 08 09:22:47 Looking for examples. Dec 08 09:25:46 Good morning Dec 08 09:25:51 Laibsch: what does the openzaurus feed have to do with inputmethods? Dec 08 09:26:05 Laibsch: or for that matter #oe with gpe input methods? Dec 08 09:26:10 hey sirfred Dec 08 09:26:13 * koen hints at the gpe mailinglist Dec 08 09:26:19 XorA: Hello. Dec 08 09:26:24 sirfred: I finally got round to trying your mplayer patch Dec 08 09:26:29 XorA: Great. Dec 08 09:26:38 sirfred: is there a later one available than one on sf.net? Dec 08 09:27:23 XorA: No, I'm working in a new version of the library, and have updated the patch consecuently, but it's still not released. Anyway, there's no improvements in the overlay section, but a simpler way to manage surfaces. Dec 08 09:27:41 XorA: The patch is now shorter Dec 08 09:27:50 sirfred: cool Dec 08 09:28:14 XorA: The bad thing is that it doesn't like Xw100 Dec 08 09:28:40 And I still don't know what's the best way to fix that Dec 08 09:29:23 koen: The feed browser is a source of quick reference. #oe is about dev work. As you know I am interested in getting a Japanese input method into OE. Dec 08 09:30:14 Laibsch: yes, but your assumption that input methods are always under SECTION="gpe/inputmethods" AND that they are in the oz feed is wrong Dec 08 09:30:51 I did not assume that. I know that OZ are not complete and that OE has more. Dec 08 09:30:55 and you probably want gtk inputmethods Dec 08 09:31:23 First of all, I want to understand the differences between gpe and gtk input method, I guess. ;-) Dec 08 09:31:33 I did not know that there was a difference. Dec 08 09:31:57 What I am trying to understand is how to get input method X working under GPE. Dec 08 09:36:37 sirfred: for the level of usage of w100, I would share the device via libw100 rather than get Xv working in kdrive, which I suspect is a longer job Dec 08 09:37:12 sirfred: and if as a by product you get xranrd support for 320x240 working that would be cool as well Dec 08 09:37:48 XorA: Yes, but I should need to inject w100 knowledge in kdrive Dec 08 09:38:04 sirfred: thats fine Dec 08 09:38:16 XorA: The problem is the offscreen handling of kdrive Dec 08 09:38:43 XorA: It supposes that the offscreen memory if the memory after the visible framebuffer and that's all. You have to pass it the end of your visible screen. Dec 08 09:39:07 And libw100 have a more complex memory management than that, other than the imageon have two memory areas, in different places. Dec 08 09:39:11 I have to leave now, cu Dec 08 09:40:57 sirfred: cool Dec 08 09:57:14 03koen 07org.oe.dev * r08d21ad9... 10/ (1 site/powerpc-linux): powerpc site: add entries needed for glibc-initial 2.5 Dec 08 10:03:21 morning Dec 08 10:03:36 hey likewise Dec 08 10:04:06 koen: FedEx just arrived. Yippie! Dec 08 10:07:31 03koen 07org.oe.dev * r04e2b1d3... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.3: powerpc can now use glibc-initial 2.5 due to the recent changes to the site file Dec 08 10:13:23 morning all Dec 08 10:15:31 hey ade|desk Dec 08 10:31:25 * koen gets annoyed with glibc Dec 08 10:32:26 * Paperface is annoyed with uclibc Dec 08 10:43:08 hi all Dec 08 10:43:48 mhhh seem there is some trouble when a download timeout and restart ... the download is resumed but the output file change with have a '.1' addedd every time Dec 08 10:43:53 mhhh seem there is some trouble when a download timeout and restart ... the download is resumed but the output file change with a '.1' addedd every time Dec 08 10:45:23 hi all! i'm having troubles configuring and building stuff with OE... can someone please help me..?? Dec 08 10:45:46 my machine is dm320 based (e.g., arm926 + dsp, but i guess arm926 is good enough) Dec 08 10:46:27 I receive "file_move: ERROR: failed to rename //usr/lib/ipkg/status.tmp to //usr/lib/ipkg/status: No such file or directory" when installing files with my 0.99.163, self-compiled ipkg. Is this likely to be a packaging or upstream problem? Dec 08 10:52:46 Laibsch: on the target, or on the host? It seems some prefix is missing there. Dec 08 10:53:08 shirour: is that a DaVinci/ Dec 08 10:53:51 gremlin[it]: is that with wget? Dec 08 10:53:55 likewise: I am sorry, if I was easy to misunderstand. This is for the compiled package on the target. I install files with "ipkg install $package" and receive that error. Dec 08 10:54:58 Laibsch: Then I am clueless. What exactly *is* in your /usr/lib/ipkg ? Dec 08 10:56:13 why does bitbake execute its scripts with #!/bin/sh instead of #!/bin/bash Dec 08 10:56:17 imo that's a bug Dec 08 10:56:50 giel: no, bashisms inside those script are a bug Dec 08 10:56:57 koen: okay, fair enough ;) Dec 08 10:57:03 i'll remove my brace expansion ;) Dec 08 10:57:04 "When running Wget without -N, -nc, or -r, downloading the same file in the same directory will result in the original copy of Dec 08 10:57:06 file being preserved and the second copy being named file.1." Dec 08 10:57:17 looks like a file is in the wrong location for the linux kernel package I got this error cp: cannot stat `/OE/org.openembedded.dev/packages/linux/linux/defconfig': No such file or directory I copied the file from /OE/org.openembedded.dev/packages/linux/linux/progear/defconfig/progear/defconfig to the expected location and it seems to have passed that spot Dec 08 10:57:32 where do I submit bug reports? Dec 08 10:57:46 zezom: bugs.openembedded.org Dec 08 10:58:00 thanks Dec 08 10:58:11 likewise: much like. dm320 (neuros) Dec 08 10:59:00 zezom: http://wiki.openzaurus.org/Bugtrackers Dec 08 10:59:11 it really depends. Dec 08 10:59:27 * Laibsch expects a slap from koen Dec 08 10:59:28 thanks Laibsch Dec 08 10:59:48 "When running Wget with -r, but without -N or -nc, re-downloading a file will result in the new copy simply overwriting the old." Dec 08 11:00:30 gremlin[it]: maybe OE should run wget with -r by default? Dec 08 11:00:40 likewise: /usr/lib/ipkg contains status information for installed packages. Dec 08 11:00:53 likewise: I have some interesting photo's for you Dec 08 11:00:58 Laibsch: is your fs writable? Dec 08 11:01:11 yes, that is not the problem. Dec 08 11:01:14 koen: hit me. Dec 08 11:01:29 likewise: http://dominion.kabel.utwente.nl/koen/cms/index.php/v/Koen/Nederland/enschede/Hardware/F2FA7317.jpg.html?q=gallery Dec 08 11:01:41 likewise: and http://dominion.kabel.utwente.nl/koen/cms/index.php/v/Koen/Nederland/enschede/Hardware/F2FA7318.jpg.html?q=gallery Dec 08 11:01:45 Laibsch: well, I am clueless, I would try to see which file exactly is missing, and when it should be created. Dec 08 11:02:37 * likewise is big-smiling Dec 08 11:02:54 * likewise curses that he left the 9250 at his work and is at home today. Dec 08 11:03:02 likewise: radeon 9250 over dvi Dec 08 11:03:13 minus faceplate Dec 08 11:03:17 * likewise ...and because his car broke down yesterday... Dec 08 11:03:41 koen: yes, in the way with the Eth/USB tower :-) Dec 08 11:04:20 likewise: I folded an A4 so the card doesn't make contact with the disk Dec 08 11:04:39 koen: proper thinking :-) Dec 08 11:08:20 likewise: btw, you can fit the agp card in *both* ways in the riser Dec 08 11:09:12 koen: OK. I am just thinking my 9250 is a PCI card. I do have a AGP 9700 though at home. Dec 08 11:09:32 time to make some room on my desk (I'll be back in three hours :-) ) Dec 08 11:09:39 the faceplate has to obstruct the eth/usb tower Dec 08 11:17:29 ~seen b0ti Dec 08 11:17:42 b0ti was last seen on IRC in channel #openzaurus, 323d 15h 54m 55s ago, saying: 'any news when 3.5.4 will be released for c7x0?'. Dec 08 11:21:37 I didn;t think I'd seen him for a long while Dec 08 11:42:25 * koen loads the debian installer over tftp Dec 08 11:46:13 koen: nice work on angstrom/ppc. Tested it yet? :) Dec 08 11:46:40 I'd like to test with my ts101 eventually Dec 08 11:50:59 I have to test one patch for glibc Dec 08 11:51:04 (after the weekends) Dec 08 11:51:10 * koen heads to germany Dec 08 11:51:12 later all Dec 08 12:09:45 Sourceforge seems to be picking mirrors automatically now. Maybe we should all change our sf.net mirror from heanet to http://downloads.sourceforge.net. I think this would be fair. Dec 08 12:10:14 Maybe we can get rid of ${SOURCEFORGE_MIRROR} altogether. Dec 08 12:11:53 What do you guys think? Dec 08 12:16:09 Laibsch: I'd not get rid of it in case it changes in the future Dec 08 12:20:05 morning guys Dec 08 12:20:07 RP: you got mail Dec 08 12:21:02 RP: OK, makes sense not to remove the code, I agree. But I guess it does make sense to promote the use of downloads.sf.net. Are you OK with me changing local.conf.sample in such a way? Dec 08 12:21:14 mickeyl: hi! Dec 08 12:22:05 hey schurig. how are things? Dec 08 12:22:15 mickeyl: lot's of stress :-( Dec 08 12:22:24 mickeyl: this week i was more on the road than in the office Dec 08 12:22:25 bummer. business wise? Dec 08 12:22:32 mickeyl: yes Dec 08 12:22:53 mickeyl: Dave quit his job and I am one of the persons who has to do part of his tasks ... Dec 08 12:23:07 RP: I meant bitbake.conf Dec 08 12:23:13 you guys should hire more people then!? Dec 08 12:24:02 Laibsch: As about it on the mailing list. sf has causes more than a few problems in the past and I'm not the best person to ask Dec 08 12:24:12 OK Dec 08 12:24:23 mickeyl: Thanks, looks good. Its already corrected one misunderstanding I had :) Dec 08 12:24:32 good :) Dec 08 12:25:03 RP: A change in bitbake.conf should be easy to revert in case of problems. I would assume downloads.sf.net is more robust than just heanet. Dec 08 12:25:51 Laibsch: I'm not sure that is a good assumption - please ask on the mailing list or ask hrw|gone as I think he knows more about sf.net than I do Dec 08 12:26:34 More robust because it will not fail when heanet goes down. I will sent a mail to the list. Dec 08 12:27:19 Laibsch: I think the problem is that not all sf mirrors have all the files. If it suddenly resolves to a mirror without the right files, build failures results Dec 08 12:27:37 Laibsch: We've been there before... Dec 08 12:28:58 I am sure that sf.net put some logic into downloads.sf.net. Just like before where you had that intermediate page which only listed the mirrors having the file. Of course that intermediate page was not scriptable, but downloads.sf.net now is. Again, easy to revert to heanet in case of problems. But I think we should play fair and use the mirrors. Dec 08 12:35:08 mickeyl: One question - Can xmlrpc be bidirectional or is both the UI and the backend going to need to act as a server? Dec 08 12:35:24 the latter Dec 08 12:35:31 if you want free communitcation you need two servers Dec 08 12:35:36 however Dec 08 12:35:56 why should the frontend need to communicate self-originated? Dec 08 12:36:38 mickeyl: A user tells the UI to build package foobar, it needs to tell the backend this Dec 08 12:37:00 mhhh if i can say my opinion ... (i have a slow internet connection) i don't like auto mirror cause my local proxy became unusable :( ... i prefear have only one dite to download from Dec 08 12:37:21 s/dite/site/ Dec 08 12:37:39 RP: hmm... is the backend doing other things at the same time? if not, why shouldn't it just block in get_next_job? Dec 08 12:37:55 free flow communication is very complicated Dec 08 12:38:04 we can workaround that with a clever flow logic Dec 08 12:38:06 imo Dec 08 12:38:49 but maybe my sight is too limited here Dec 08 12:39:06 and given that a xmlrpc server is just a couple of lines, it probably doesn't matter Dec 08 12:40:27 mickeyl: I don't think simply blocking would work. Its the UI which has to initiate the event and the backend then has multiple events to report back in no fixed order/sequence Dec 08 12:41:03 fine with me as long as we don't get deadlocks Dec 08 12:41:40 mickeyl: No, the model I have in mind shouldn't do that Dec 08 12:41:46 ok Dec 08 12:42:11 btw., i have some more clever tricks for the xmlrpc (I worked 3 years with xmlrpc on python), but I'll add these when the basic model works Dec 08 12:43:00 mickeyl: I just needed to confirm the communication isn't bidirectional, thanks. I'll mail you an example event flow so you can see what I mean and hopefully you'll see it shouldn't cause lock problems Dec 08 12:43:09 excellent Dec 08 12:43:22 off topic question : the filesystem where i work with OE is xfs ... when i have to wipe out tmp dir (or other very crawled dir) it take minutes ? you know where i can find an how-to for xfs tuning ? ... thanks Dec 08 12:43:57 RP: nearby... that was one of the reasons I would've loved to use PyLinda, which is a shared tuple space abstraction which is by nature bidi... but it's an extra package. Dec 08 12:44:41 mickeyl: xmlrpc seems quite simple and I think it can work. Lets see whether my example event flow will work with it... Dec 08 12:44:55 yep Dec 08 13:07:39 mickeyl: You have mail :) Dec 08 13:07:39 * RP -> food Dec 08 13:10:11 ~bon appetit Dec 08 13:10:28 well, bon appetit is smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito. Buen apetito Bom Apetite. buen apetito Dec 08 14:28:12 morning Dec 08 15:39:33 * gerwinin has been fired today Dec 08 15:43:52 gerwinin: as in booted out of the door? Dec 08 15:47:35 Xora yes I came in this morning and they told me to give the key and stuff Dec 08 15:47:49 They closed down my computer Dec 08 15:47:57 They deleted my stuff from the netwok Dec 08 15:48:00 They deleted my stuff from the network Dec 08 15:48:09 gerwinin: ouch Dec 08 15:48:20 XorA: yes unbelievable Dec 08 15:48:40 XorA : And that only because I wanted to put sources online of something I made Dec 08 15:49:03 gerwinin: sounds like a case for a lawyer to me Dec 08 15:49:18 or petrol bomb Dec 08 15:49:32 XorA: yeah I know , but I am going to report them to bsa :) Dec 08 15:49:47 XorA : for all that illegal windows software they are running there Dec 08 15:49:56 gerwinin: :-D heh heh, always the best way, that reqard is quite tempting Dec 08 15:50:10 s/reqard/reward/ Dec 08 15:50:24 XorA: What comes arround goes arround even for these aseholes Dec 08 15:51:15 XorA: And they still need to pay me my salary over last month Dec 08 15:51:59 gerwinin: Hi. That's quite shitty to hear. ok, so your plan B is going into action earlier. Dec 08 15:52:00 gerwinin: :-( Dec 08 15:54:27 Likewise: yes :( Dec 08 15:54:52 Likewise: if you want to work never work for a company called i-products Dec 08 15:55:03 gerwinin: What's your plan now? Start for yourself, or look somewhere else? Dec 08 15:55:14 * likewise is setting a reminder. Dec 08 15:55:44 Likewise: First I will need to see how far I financially can keep my head above the watter Dec 08 15:55:44 gerwinin: what's your background btw? I forgot to ask you during tdose. Dec 08 15:55:56 Likewise : company information science Dec 08 15:56:30 Likewise: do you have a job currently ? Dec 08 15:56:49 gerwinin: What would you like to do? Dec 08 15:57:01 hi Dec 08 15:57:03 gerwinin: Yes, but it's a 45 min drive from Ehv. Dec 08 15:57:27 Likewise: some easy job like systemadministration or something easy so I can spent more time on my opensource projects Dec 08 15:57:45 Likewise: at this company I was an engineer Dec 08 15:58:06 Likewise: what is your background ? Dec 08 15:58:26 RP: so many bitbake idea's, I feel overflowed Dec 08 15:58:53 zecke: me too :-/ Dec 08 15:59:09 RP: what makes you ill with TmpHandler? Dec 08 15:59:40 zecke: I had difficultly easily seeing what it does. The final death blow was the embedded tab Dec 08 15:59:48 adding new events luckily trivial Dec 08 16:00:06 RP: the parser should warn about tabs! Dec 08 16:02:54 gerwinin: I did MSc Electrical Engineering & Information Technology Sciences - I started as an embedded s/w eng 6 years ago at axon.tv -- I currently do systems architecting and am lead engineer for a few projects. Dec 08 16:03:17 RP: more critical question Dec 08 16:03:27 RP: do we want events to have a return value? Dec 08 16:03:43 zecke: I missed that off the list :} I propose not Dec 08 16:03:54 RP: in this case we could put code in OE which makes decisions :) Dec 08 16:03:55 We want all handlers to see all events? Dec 08 16:04:33 RP: I think we should have two different event handlers Dec 08 16:04:42 RP: the ones we currently load with INHERIT += Dec 08 16:04:55 RP: we should make them HANDLER += "sanity foobar" Dec 08 16:05:11 and these things see all events, and then we have per bb class event handlers Dec 08 16:05:56 sanity is a bbclass event handler though? Dec 08 16:06:02 hmm I still feel pissed, back to bed :) Dec 08 16:06:20 RP: I think they can be bbclasses Dec 08 16:06:33 RP: but we should not abuse INHERIT to load these handlers Dec 08 16:07:46 zecke: I don't see how HANDLER += is different to addhandler Dec 08 16:08:18 these are two different things Dec 08 16:08:18 addhandler will mark a python function as handler Dec 08 16:08:26 INHERIT is 'abused' to load the bbclass to make sure the handler is added Dec 08 16:08:53 You can't have python code in a conf file though? Dec 08 16:09:23 pb,kergoth,schurig (one of them,some...) decided against that Dec 08 16:09:52 I see no problem with having things like sanity as a bbclass file... Dec 08 16:10:09 RP: I don't have a problem with that either Dec 08 16:10:35 RP: check bin/bitbake, currently it loads all INHERITS to make sure the handlers are loaded Dec 08 16:10:42 RP: and this is the thing that makes me feel it is abusive Dec 08 16:11:03 you mean lib/bb/cooker.py? ;-) Dec 08 16:11:19 hmm ;) Dec 08 16:12:49 zecke: I'd say we should INHERIT as instructed for configuration Dec 08 16:13:04 zecke: We also want any variables declared by those classes present... Dec 08 16:13:30 RP: for INHERIT e.g. with package_ipk, multifoo I agree Dec 08 16:13:49 RP: for handlers I do not agree. They just waste time and resources Dec 08 16:14:09 RP: e.g. you don't want sanity.bbclass be parsed for each bbfile but only once :) Dec 08 16:14:37 zecke: yes, I actually want sanity to only run once... Dec 08 16:15:38 hence my comment about a configuration loaded event to have sanity to hook in too Dec 08 16:18:28 zecke: I do think we could do something better with the handler registration. Why doesn't the parser register handlers directly? Dec 08 16:19:04 iit does Dec 08 16:19:45 then why the handler code in parseConfigurationFile? Dec 08 16:20:17 andersee: i'm getting the following error when trying to play a file: "ERROR: src/plugins/nms/nms.c:147: Error opening output" Dec 08 16:20:26 inc, .bbclass and .bb honor INHERIT Dec 08 16:20:46 e.g. the nslu2 ninjas don't load a .inc file -> no handlers registrated Dec 08 16:20:58 when no bb file was loaded (e.g. on a cached run) Dec 08 16:21:35 -> adding HANDLER variable Dec 08 16:21:44 zecke: doesn't configuration always include base.bbclass? Dec 08 16:22:07 before write a bug , has anyone an idea about this error : Dec 08 16:22:08 | Collected errors: Dec 08 16:22:08 | Cannot find package libltdl. Dec 08 16:22:11 | Check the spelling or perhaps run 'ipkg update' Dec 08 16:22:20 an error in the libltdl recipe ? Dec 08 16:22:48 no only .inc, .bb, .bbclass inherit classes Dec 08 16:23:36 zecke: Should we not just change that so .conf files do too? Dec 08 16:24:38 If I put INHERIT = "multimachine" in a .conf file and it doesn't, I'd class that as a bug... Dec 08 16:25:40 no Dec 08 16:28:13 zecke: I get the feeling I'm missing something... Dec 08 16:28:52 hi...anyone good with regular expressions? I'd like to use it on a number (which is i.e. "3.3") and make it match all values greater-than 3.0 Dec 08 16:30:20 RP: I will be back in 20 minutes Dec 08 16:30:40 zecke: ok Dec 08 16:30:42 but this is a chicken egg problem one should avoid :) Dec 08 16:31:06 zecke: I just don't understand what the problem is... :) Dec 08 16:31:43 you would have to parse base.bbclass before the conf file :) Dec 08 16:31:55 and i don't want to do it Dec 08 16:34:26 zecke: I agree with that. How about parse conf files, batch up the inherits then process after the implicit INHERIT base ? Dec 08 16:40:43 is there a kernel tun package anywhere for OZ 3.5.4 ? Dec 08 16:48:54 follow up question, has the suspend problem been fixed in CVS for poodle for OZ 3.4.5.2? Dec 08 16:49:10 which cvs? Dec 08 16:49:51 which suspend problem, which kernel? Dec 08 16:53:42 zecke does bonsia work with svn? Dec 08 16:53:46 RP: okay now to the trinity I see :) Dec 08 16:53:49 Crofton|oe: yupp Dec 08 16:53:59 from ozilla cvs? Dec 08 16:54:05 no Dec 08 16:54:11 * Crofton|oe hates tinderbox setup ... Dec 08 16:54:11 Crofton|oe: ah the 'true' bonsai Dec 08 16:54:22 No it does not work with svn, and it is ugly perl Dec 08 16:54:55 RP: We have configuration which starty by parsing conf/bitbake.conf Dec 08 16:55:09 RP: we have handlers which should be global and are implemented as bbclasses Dec 08 16:55:38 RP: we have bitbake files which inherit base.bbclass, all INHERITS and what ever the author thought is necessary Dec 08 16:55:38 bye Dec 08 16:55:53 RP: I would really like to keep configuration and BitBake file separate Dec 08 16:56:30 RP: and avoid loading global handlers into each parsed BitBake file (even if the new parser handles this quickly) Dec 08 16:57:31 RP: my solution for globally loading handlers would be Dec 08 16:57:41 RP: remove them from INHERIT, place them in HANDLER Dec 08 16:58:22 RP, with the latest release candidate for OZ, suspend does not work on my 5600 Dec 08 16:58:35 I'm just wondering if its been fixed Dec 08 16:59:44 so I guess its the latest kernel Dec 08 17:00:34 heath_: I can't remember what kernel version was in the last OZ release for the 5600. Was it a 2.4 or a 2.6 kernel? Dec 08 17:01:13 zecke: Calling them HANDLER isn't right as they behave more like bbclass files than just HANDLERs Dec 08 17:01:36 RP, its 2.6 Dec 08 17:01:56 zecke: The thing is if every .bb file is going to include the things thing INHERIT, why not do this immediately after parsing the configuration? Dec 08 17:02:01 RP: but not the same way as bbclasses and bitbake files :) Dec 08 17:02:13 RP: The new parser is doing this :) Dec 08 17:02:41 zecke: I'm fine with that - we should change the old parser to do this too ;-) Dec 08 17:02:53 RP: I do not want to touch the old parser :) Dec 08 17:02:57 heath_: Was there a bug report filed? Dec 08 17:03:16 RP: but still. I do not want tinderclient,sanity, base_handler (does not exist yet) Dec 08 17:03:27 I am trying to analyze what program is calling a library that I had installed from my personal feed but that was creating problems. Is strace what I am looking for? Dec 08 17:03:28 RP: pollute variables Dec 08 17:03:49 zecke: Those handlers might rely on the variables? Dec 08 17:04:00 Laibsch: calling as in loading the library? Dec 08 17:04:10 Laibsch: either strace or ltrace (which used to be broken on ARM) Dec 08 17:04:17 zecke: yes Dec 08 17:04:41 RP: hmm, my hangover might be too bad to express myself here Dec 08 17:04:54 Laibsch: strace is your friend then, also ldd might be handy? Dec 08 17:05:06 RP: INHERIT and HANDLER are two different concepts Dec 08 17:05:20 zecke: I think I see the difference, I don't like it though :-/ Dec 08 17:05:36 RP: INHERIT is used to automatically inherit a set of BitBake classes for every parsed BitBake file Dec 08 17:05:40 zecke: Your approach suggest we need a new file format - .bbhandler Dec 08 17:06:11 RP: a HANDLER is loaded once and operates globally Dec 08 17:06:14 RP, this weekend I think I'm going to put together a build env. and have a look at it Dec 08 17:06:32 RP: Now to mix the two things again :) Dec 08 17:06:50 RP: We have global handlers, and packages could decide to install one as well (for whatever reason) Dec 08 17:08:02 RP: and I think it is fine to keep .bbclass for handlers as well Dec 08 17:08:02 RP: They will pollute the global namespace anyway Dec 08 17:08:02 zecke: I fail to see why we need to make this distinction Dec 08 17:08:02 RP: because I want to is probably the best one :) Dec 08 17:08:02 RP: What this INHERIT mess forced me to do Dec 08 17:08:22 zecke: Personally, I think it adds complication for no added benefit Dec 08 17:08:24 RP: is to parse the configuration, look at INHERIT, load all classes, throw the result away Dec 08 17:08:44 zecke: Why throw the result away? Dec 08 17:08:49 RP: parsing 2*4000 .bbclasses less Dec 08 17:09:00 RP: because we don't need it. We have the compiled handler code Dec 08 17:09:17 RP: and (which is another issue) we pass them the data dict of the current bbfile Dec 08 17:09:29 heath_: FWIW I wrote the poodle 2.6 suspend/resume code. I can't remember any poodle 2.6 suspend/resume bug being reported so I can't comment whether it was fixed or not Dec 08 17:09:31 RP: so handlers currently have no persistence (I would like to change that as well) Dec 08 17:10:08 heath_: http://www.angstrom-distribution.org/ rule number five Dec 08 17:11:48 zecke: I think there is a simpler way to do this. INHERIT is only set in config files. We parse config data to create the configuration. We then inehit base and anything in INHERITS. This adds in the handlers. When we come to parse other .bb files, we already have those classes inherited, no need to look at them again Dec 08 17:12:14 RP: that is dirty and cheating Dec 08 17:12:25 RP: again these are two different concepts! Dec 08 17:12:32 zecke: Its efficent. Why would we need to do anything else? Dec 08 17:12:48 RP: and now I remembered persistence they are even more different Dec 08 17:12:58 RP: INHERIT -> applies to EACH BitBake file Dec 08 17:13:10 RP: a handler applies to the global scope! Dec 08 17:13:23 and this is a fundamental difference Dec 08 17:13:28 zecke: so we make INHERIT apply to the global scope Dec 08 17:13:41 basicly the code in cooker.py should look at HANDLER Dec 08 17:14:00 RP: no I have a bad feeling about that Dec 08 17:14:17 zecke: I think its a much simpler solution to the problem Dec 08 17:14:51 It is not a solution at all Dec 08 17:15:28 all the time I have written a eventhandler I wanted to keep a state/the state Dec 08 17:15:47 I don't think you want to execute the anonfuncs of the INHERITED classes when the configuration was parsed Dec 08 17:15:58 yes one can cheat as well and keep them and execute them delayed Dec 08 17:16:14 and you can even cheat more the next time this will introduce an error :) Dec 08 17:16:33 Having event handlers store state is a really nasty problem... Dec 08 17:16:56 I don't like the idea at all :-/ Dec 08 17:17:32 anonfuncs would just be queued, executed once the final .bb was parsed Dec 08 17:17:33 i see but your cheating won't pay off Dec 08 17:18:23 :) Dec 08 17:18:46 I think I should write some stuff down Dec 08 17:19:04 We both need to think about this... Writing it down might make it clearer Dec 08 17:19:14 Its why I wrote that email last night... Dec 08 17:19:52 RP: For me the fundamental difference is: We apply INHERITS to n bitbake files, Our current handlers should only be loaded once though Dec 08 17:20:27 RP: I have written three eventhandlers and I always felt bad to use INHERITS just to get my handler installed Dec 08 17:20:58 RP: and when writing my handler, I always wanted to keep the global state Dec 08 17:21:21 RP: and this is a clear indication that handlers are global Dec 08 17:21:37 zecke: Marking a class file as just containing a handler is equally bad though... Dec 08 17:21:50 The syntax just feels all wrong... Dec 08 17:22:05 RP: adding .bbhandler to the old parser is easy :) Dec 08 17:22:21 I don't think we need another identical file format :-/ Dec 08 17:22:40 anyhow, I need to go, sorry :-( Dec 08 17:23:14 RP: don't worry, I will change event.py while you are away :) Dec 08 17:23:36 "You have five seconds to veto or I will commit" 54321 "okay no veto" Dec 08 17:25:16 hi, all! Dec 08 17:26:30 for some package I need to have 3 config files for machine1 2 config files for machine2 and all ${D}${sysconfdir} for all others. how could I implement it? Dec 08 17:28:18 slapin_nb: SRC_URI_append_machine1 += " the specific files" Dec 08 17:28:33 slapin_nb: do_install_append_machine1 () { install the specific files } Dec 08 17:29:13 zecke: I should be online later ;-) Dec 08 17:29:42 RP: you better should, I think I feel so bad I will just exchange license headers... Dec 08 17:30:52 ok, I've got an idea, thanks! Dec 08 17:33:34 slapin_nb: if the filenames are the same but the content is different Dec 08 17:33:55 slapin_nb: you can create subdirs like machine1/ machine2/ (create them at the 'right' place) Dec 08 17:35:00 zecke, subdirs in ${D}? Dec 08 17:35:47 slapin_nb: no Dec 08 17:36:04 slapin_nb: imagine SRC_URI += "file://fooo.bar" Dec 08 17:36:22 this is a relative path and OVERRIDES are used to find the best fooo.bar Dec 08 17:36:39 slapin_nb: e.g. we do that in busybox or basefiles to install a different fstab depending on distro/machine Dec 08 17:38:37 thanks! Dec 08 17:52:31 sirfred: do you have a few minutes to help me with a libw100 problem? Dec 08 18:09:37 has anyone used probe_irq_on() before? Dec 08 18:10:36 Good Morning Dec 08 18:12:32 I was wondering, does anyone know how to simulate a right-click when pressing Alt? Dec 08 18:12:40 Ala Win CE Dec 08 18:12:53 Because xmonobut is flaky, at best Dec 08 18:13:07 At least in my current setup Dec 08 18:13:55 Maybe some option in X? Dec 08 18:14:12 Or some program to run in the backround Dec 08 18:14:45 Looking in the Open Embedded Software Foundation didn't help much Dec 08 18:30:31 koen|away: http://lists.freedesktop.org/archives/cairo/2006-December/008837.html Dec 08 18:32:50 heath_: What suspend problem are you refering too? I can suspend from the menu/apm --suspen just fine. The only thing that does not work right now is suspending with the "Cancel" button Dec 08 18:35:29 zecke: I am not making much progress. How come ldd says qpe is not a dynamic executable but that when calling it there is an error about some lib missing? I'd really appreciate your help. strace output is at http://rafb.net/paste/results/M0HJyX28.html Dec 08 18:35:46 ldd at http://feeds.openzaurus.org/3.5.4/feed/x11/ Dec 08 18:35:57 ldd at http://rafb.net/paste/results/g2AHSy42.html Dec 08 18:55:10 hvontres|poodle, is that in the latest release candidate? Dec 08 18:56:20 heath_: Actually, I am running the latest kernel built from the .354x branch with a hentges 1.0 rc Dec 08 19:04:24 heath_: BTW, the latest kernel also makes the volume control work in opie :) Dec 08 19:04:34 hvontres|poodle, thats nice Dec 08 19:04:41 I haven't even played with sound Dec 08 19:05:04 hvontres|poodle, do you have your initrd and zImage hosted somewhere? Dec 08 19:05:15 * hvontres|poodle has been for a while, since I pushed RP into making it work in the first place :) Dec 08 19:05:41 hi Dec 08 19:06:40 heath_: nope. but if you get OE setup it's pretty straight forward. The latest kernel recipie will build a kernel that boots the rootfs from SD. Dec 08 19:07:29 * heath_ needs to go out an buy an SD card Dec 08 19:07:36 or can you boot from CF ? Dec 08 19:07:46 heath_: nope. Dec 08 19:08:14 heath_: Of course Dec 08 19:08:14 heath_: the rootfs tarballs are here : http://www.hentges.net/tmp/snapshots/Hentges-ROM/1.0.0/20061108/poodle/tarballs/ Dec 08 19:08:57 heath_: plus I have found you can kexec the kernel from SD as well. Dec 08 19:09:57 heath_: so, if you untar the rootfs to an SD card, copoy the zImage over and use kexec -l ;kexec -e, you should be able to test WITHOUT flashing anything :) Dec 08 19:12:24 nice Dec 08 19:13:53 hvontres|poodle: I am working on adding a script to altboot the help with the whole kexec thing. I think eventually, the hentges rom will have some sort of rescue Image on internal flash and the normal image on SD. Dec 08 19:14:07 I'll have to fiddle with that this weekend, its nice that SD cards are really cheap now Dec 08 19:15:14 hvontres|poodle: plus I think 2.6 can support cards bigger than 1GB. I haven't tried on poodle yet tho. (Just have a 1GB card right now) Dec 08 19:15:34 yea, 1 GB for like $14 on newegg Dec 08 19:16:13 heath_: do you have a fry's neearby? Dec 08 19:17:14 no I'm on the east coast Dec 08 19:17:22 I miss frys Dec 08 19:31:07 Laibsch: still there? Dec 08 19:31:13 yes Dec 08 19:31:16 Laibsch: because you would need arm-linux-ldd Dec 08 19:31:25 Laibsch: or use ldd on your device Dec 08 19:31:48 Laibsch: http://www.busybox.net/lists/busybox/2001-December/005270.html Dec 08 19:32:00 Laibsch: if you have debian install binutils-multiarch Dec 08 19:32:19 I did use ldd on the device, see http://rafb.net/paste/results/g2AHSy42.html Dec 08 19:32:20 Laibsch: and then use objdump -x app | grep NEEDED Dec 08 19:32:44 Laibsch: eeek, use objdump on your host then Dec 08 19:32:49 OK Dec 08 19:33:05 Laibsch: this is so scary I don't want to think about it now Dec 08 19:33:29 scary? Come on. These are just computers ;-) Dec 08 19:33:42 Laibsch: yes, scary computers Dec 08 19:36:10 Can I have objdump on the host as well? It would make things easier. Dec 08 19:36:17 target! Dec 08 19:36:27 Laibsch: compile binutils Dec 08 19:36:32 OK Dec 08 19:36:37 bitbake binutils Dec 08 19:37:07 Or just install it: http://ewi546.ewi.utwente.nl/tmp/hrw/fb/?action=details&pnm=binutils ? :-D Dec 08 19:37:29 Laibsch: -x extracts the headers of the elf object and shared libraries are marked as SO_NEEDED Dec 08 19:37:53 Laibsch: eek a GPL violation :( Dec 08 19:38:23 where? Dec 08 19:38:35 Laibsch: on your link :) Dec 08 19:38:44 Laibsch: no offer to get the source code Dec 08 19:39:43 hey holger Dec 08 19:40:07 zecke: bug koen. but http://ewi546.ewi.utwente.nl/mirror/www.openzaurus.org/official/unstable/3.5.4/sources/ is right there. Dec 08 19:40:59 Laibsch: I know we aim to comply with the GPL...:) Dec 08 19:41:42 Wollen wir mal nicht päpstlicher sein als der Papst, gelle? Dec 08 19:41:47 :-D Dec 08 19:41:49 hi Dec 08 19:41:51 Laibsch: http://packages.debian.org/stable/games/bygfoot Dec 08 19:42:00 Laibsch: has a link to the source+patch Dec 08 19:42:18 Laibsch: I don't assume FSF will revoked the binutils license but we should try harder :) Dec 08 19:42:20 I don't like soccer Dec 08 19:42:32 Laibsch: I searched for a package called 'foo' Dec 08 19:42:46 Laibsch: and wanted to check if I have download source option Dec 08 19:43:32 Laibsch: and I do not want to have more boyscouts than the pope Dec 08 19:43:33 zecke: honestly, I do not think that a compilation of links is what FSF would distribution. http://ewi546.ewi.utwente.nl/mirror/www.openzaurus.org/ and below is where the fun is happening and see it being compliant. Dec 08 19:43:41 zecke: want to see GPL violations? Check the output of amr-uclibc-ldd on the dlink media servers (DSM320*) media player apps... Dec 08 19:44:05 Man I need to learn how to type. Dec 08 19:44:11 Bernardo: I know I'm strange :) Dec 08 19:44:16 zecke: under the GPL no offer of source needs to be made if the source is on the same media (and media can mean ftp site) Dec 08 19:44:30 that clause has an "or" in it Dec 08 19:44:34 XorA|gone: I would define website as media in this case Dec 08 19:44:42 zecke: I am trying to get the fsf interested on those, so that dlink will have to release their source Dec 08 19:45:02 Bernardo: gpl-violations.org even if harald is busy Dec 08 19:45:03 zecke: luckilly FSF forgot to define media in the GPL :-) Dec 08 19:45:20 XorA|gone: we don't even state the license of the software Dec 08 19:45:33 XorA|gone: and just allow to download :) Dec 08 19:46:36 zecke: that's a idea... I need to document my findings on the firmware images a little more, but it might be a path to get the source Dec 08 19:46:38 zecke: Remember that GPL is trying to preserve the rights of the user. I do not seem them endangered in any way here. Really. Dec 08 19:46:44 XorA|gone: so same media == internet :) Dec 08 19:47:40 zecke: Even media == internet might be fine. But the problem is, how do you make you some webserver does not pull the plug on some sources. that is IMHO the only reason you need to host yourself. Dec 08 19:48:07 s/make/make sure/ Dec 08 19:49:05 admitingly I'm strange in this regard. Dec 08 19:49:33 * zecke searches for the section demanding that the tools to build the software are published as well Dec 08 19:49:56 * Bernardo just wants to use his media player the way he wants... :) Dec 08 19:50:27 Bernardo: DLink got sued and lost several times against harald welte... Dec 08 19:50:51 Bernardo: harald needs to spread his copyright in the kernel a bit more, otherwise dlink builds stuff without netfilter :} Dec 08 19:50:59 Laibsch: did objdump help you? Dec 08 19:51:04 hi all ... Dec 08 19:51:29 does in last few weeks was done some work on GPE to reduce library size or dependencies ? Dec 08 19:52:04 zecke: the DSMs are a pain... They don't provide the redsonic bootsector, they don't provide the application sources... All you get is a 2.4.18 kernel plus some heavy patches plus busybox... Dec 08 19:52:08 gremlin[it]: cairo got optimized Dec 08 19:52:32 gremlin[it]: library sizes memory footprint was not changed Dec 08 19:52:47 on my h3600 i have more than 1M more of free space ! Dec 08 19:53:19 zecke: Yes, thank you. It seems like exactly what I was looking for. Dec 08 19:53:35 I think I can manage now together with strace Dec 08 19:54:59 gremlin: 1mb is beautiful :) Dec 08 19:55:09 sure ... now it's usable ... Dec 08 19:55:15 Jin|away: I'm the jbbrs at sourceforge's mediatomb forum... :) Dec 08 19:56:10 meybe i'll test better but i use a my 'gpe-image-small' ... so i thing the only way to reduce size is smaller packeges rathen than fewer packeges ... i'll check btw Dec 08 19:56:40 zecke: gplviolations.org redirects to http://chaosradio.gnumonks.org:2342/, is this correct? If so, I need to open another port at the firewall... Dec 08 19:56:58 hmmm Dec 08 19:57:19 not here Dec 08 19:57:22 hi zecke :) ... Dec 08 19:57:37 Bernardo: maybe your dlink router blocks gpl-violations.org? Dec 08 19:57:54 Bernardo: http://www.jbb.de/judgment_dc_frankfurt_gpl.pdf Dec 08 19:58:11 zecke: I don't have a dlink router, just a media center... :) Dec 08 19:59:56 I do have a sagem router between my firewall and the telephone cable, but it is in transparent mode... probably just my ISP acting up Dec 08 20:01:10 ok, I had written gplviolations without the dash... :( Dec 08 20:01:14 bbl, dinner time Dec 08 20:01:56 Wireless G Dec 08 20:01:57 Network media Storage DSM-G600 Dec 08 20:02:11 so it wasn't a router this time Dec 08 20:06:03 zecke: Found the package. Will try to downgrade it to get back to working condition. Thanks a lot! Dec 08 20:08:12 hi zecke Dec 08 20:14:43 pb_: master Dec 08 20:15:07 pb_: would you mind BitBake being GPLv2 only and to reevaluate once v3 is released? Dec 08 20:15:38 zecke: I don't know if there is still code from me in BitBake, but I wouldn't mind Dec 08 20:15:42 zecke: no, that sounds reasonable to me Dec 08 20:16:09 schurig: hehe, I'm playing code monkey and almost every file has Dec 08 20:16:16 #Based on functions from the base bb module, Copyright 2003 Holger Schurig Dec 08 20:16:54 * zecke adds proper GPL headers to BitBake... Dec 08 20:18:13 03zecke123 * r717 10bitbake/lib/bb/parse/parse_c/: Dec 08 20:18:13 Remove the first attempt to integrate Marc's flex/lemon Dec 08 20:18:13 Remove the first C implementation. I'm too lazy to Dec 08 20:18:13 create proper license headers for what will be replaced Dec 08 20:18:13 with the bitbake-parser code soon(tm). Dec 08 20:21:16 03zecke123 * r718 10bitbake/doc/COPYING.GPL: Dec 08 20:21:16 Update the GPL to the current version of the FSF Dec 08 20:21:16 Update the address of the FSF Dec 08 20:26:03 what a monkey task... Dec 08 20:26:25 zecke: Worked wonders. Everything is in order now and back to where I can continue testing. Thanks a bunch. Dec 08 20:26:41 03zecke123 * r719 10bitbake/ (36 files in 6 dirs): (log message trimmed) Dec 08 20:26:41 Add proper GPLv2 headers to all BitBake files Dec 08 20:26:41 BitBake trunk is now GPLv2 only, no mix of Dec 08 20:26:41 MIT,FreeBSD License is left. Dec 08 20:26:41 Update GPL headers to point to the correct Dec 08 20:26:41 address of the FSF Dec 08 20:26:42 Update the list of authors. Uli Luckas, Dec 08 20:27:06 zecke: It's nice to have the right tools and some nice people to help out. Dec 08 20:31:39 * zecke pets his zeckeATselfish.org email address Dec 08 21:14:35 RP: for your consideration: http://frodo.vpop.net/~mreimer/w100fb.patch Dec 08 21:14:53 RP: the first part just changes memcpy_{from|to}io() to memcpy() Dec 08 21:15:08 RP: the second part turns off external memory to save power Dec 08 21:26:26 hi mreimer Dec 08 21:26:33 hi pb_ Dec 08 21:49:12 nite guys Dec 08 22:22:26 mreimer: I can't change memcpy - it was a condition of making mainline Dec 08 22:22:46 RP: even for memory-mapped framebuffer? Dec 08 22:26:16 mreimer: Yes, it was explicitly asked for - see past mails on linux-fbdev-devel Dec 08 22:26:28 RP: ok. What about the extmem part of the patch Dec 08 22:26:29 ? Dec 08 22:26:54 mreimer: What does it do? Dec 08 22:29:14 RP: it sets mem_invert_mck, mem_tristate_mcke, mem_tristate_ma, mem_tristate_mck, and resets mem_ext_enable. Apparently it just turns off ext mem. Dec 08 22:30:59 mreimer: that sounds harmless enough. I'll need to test on a c7x0... Dec 08 22:31:07 RP: thanks Dec 08 22:31:50 RP: are you seeing "w100fb: FIFO Timeout!" on suspend or resume? I am, with the imgblit patch Dec 08 22:32:46 mreimer: I've not tried it yet but that doesn't sound good :-/ Dec 08 22:32:54 sirfred would be the person to ask Dec 08 22:32:58 RP: otherwise it works fine. Dec 08 22:33:01 ok Dec 08 22:33:16 I'm lacking time atm to look after these details :-( Dec 08 22:34:29 RP: I understand. Is there something I can do, like put a bug in a db somewhere, to help you track it? Dec 08 22:36:31 mreimer: Personally, email works best. I have emails going back over a year which are marked as pending action. I just need some time to deal with them. That doesn't help others track them though :-/ Dec 08 22:36:44 RP: ok, I'll send you patches Dec 08 22:37:13 mreimer: Did you ever understand rmk's reply on lakml about memcpy? Dec 08 22:37:44 RP: no. let me re-read Dec 08 22:38:02 RP: doesn't make sense to me to assume that byte accesses are ok Dec 08 22:38:48 I understand mappings might not be 1-1. I don't understand how you could ever use memcpy over such an address range though Dec 08 22:42:21 RP: what do you think is the fix for this? Dec 08 22:43:25 mreimer: I don't know. Before I can suggest solutions, I need to understand the problem Dec 08 22:44:05 RP: I don't see how the current situation is any better, because some memory might be configured for 16-bit accesses and abort on 8-bit accesses, right? Dec 08 22:47:40 Presumably that is a less common problem Dec 08 22:54:45 mreimer: I'm tempted to rip that code out entirely, save the problem... Dec 08 22:54:56 RP: which code? io.c? Dec 08 22:55:08 mreimer: save/restore video memory Dec 08 22:55:14 RP: Spyro agrees Dec 08 22:55:43 RP: by my measurements, even with memcpy that would save us about .6 seconds suspending Dec 08 22:59:00 mreimer: Reading back the other channel, Spyro was talking about registers, not memory Dec 08 22:59:12 We do setup registers from scratch Dec 08 23:00:55 I read scotch Dec 08 23:00:57 I must need booze Dec 08 23:06:53 good nite all Dec 08 23:10:16 CosmicPenguin: We have a server called 'pug'. You know your subconcious is trying to tell you something when you ssh into pub :) Dec 08 23:11:21 heh Dec 08 23:47:00 RP: still around? Dec 08 23:47:14 RP: I have another patch for w100fb to make the FIFO timeout errors go away Dec 08 23:57:21 mreimer: I'm still here but not for much longer Dec 08 23:58:07 RP: the imageblit function was being called after suspend. Calling fb_set_suspend() in w100fb_suspend() gets the proper state set in fb_info so the imageblit function can know not to try to do anything Dec 08 23:59:14 mreimer: sounds good Dec 08 23:59:32 RP: I'll send a patch I worked up after consulting the neighboring drivers Dec 09 00:00:47 mreimer: ok, thanks Dec 09 00:00:52 'night all Dec 09 00:01:21 good night RP Dec 09 00:17:16 I'm getting a patch reject while trying to build uboot. I've fixed the patch error do I just exit sh to continue? Dec 09 00:53:37 Hey, just a heads up Dec 09 00:54:16 bugs.oe.org will be going down for a few hours in a week or two for a UPS upgrade. Dec 09 00:55:34 When we get closer, I should have more precise date and timeframe Dec 09 01:32:38 how do I conitinue after a patch fails and I'm droped into a shell? I've fixed the file that needed patching but if I type exit to exit the shell the compile stops Dec 09 01:33:58 the interesting thing is that the file was actually patched correctly but it claims that it failed... Dec 09 01:34:59 do I have to delete the config.mk.rej file for it to continue Dec 09 01:48:18 do I just type make? Dec 09 02:05:56 is it safe to delete file in build/tmp/work after the compiling is finished? **** ENDING LOGGING AT Sat Dec 09 02:59:58 2006