**** BEGIN LOGGING AT Sun Jul 03 23:59:56 2005 Jul 04 00:06:11 I've got a question about adding a package to OE. It's a newer version of openntpd than is in the feed. The problem is that it requires a user in /etc/passwd. Is it OK if I add a ntpd user? Jul 04 00:07:47 I'd add it in the install part of the bb, but adduser asks for a passwd. **** BEGIN LOGGING AT Mon Jul 04 00:13:30 2005 Jul 04 00:53:45 good morning Jul 04 01:08:20 What happened to the monotone servers? Jul 04 01:09:34 oh... now vanille.de is back Jul 04 01:44:16 morning Jul 04 01:44:20 ~curse monotone Jul 04 01:44:21 May you be reincarnated as a Windows XP administrator, monotone ! Jul 04 01:44:31 ouch, that's a bad one ;) Jul 04 01:45:13 I miss bitkeeper... Jul 04 01:45:14 hrw|work: what you done to break it :-) Jul 04 01:45:46 XorA: I have 1.4GHz athlon here - "monotone pull" takes 20% cpu and is working since 10:18 Jul 04 01:46:20 hrw|work: You're gmt+2 (ie, 10:46)? Jul 04 01:46:32 NAiL: yes Jul 04 01:46:55 That's... like.. long Jul 04 01:47:02 and it's not exactly going to get faster :( Jul 04 01:47:11 hrw|work: my amd64 it takes about 2 mins Jul 04 01:47:29 hrw|work: working in 32bit mode Jul 04 01:47:47 NAiL: Europe/Poland Jul 04 01:47:53 hi mickeyl Jul 04 01:47:57 hrw|work: Europe/Norway ;) Jul 04 01:47:59 good morning Jul 04 01:48:09 mickeyl: morning Jul 04 01:48:40 monotone should have -verbose switch to give more info during pull - now it looks like zombi Jul 04 01:49:19 hrw|work: => add that to their task tracker Jul 04 01:49:45 There have been some cool patches posted to the mt list last week Jul 04 01:50:44 mickeyl: good morning Jul 04 01:51:00 mickeyl: question... I asked a question almost two hours ago. Who's the right persons to ask? Jul 04 01:56:11 who are, even Jul 04 01:56:24 someone here use monotone on <2GHz machine? Jul 04 01:56:32 32bit one Jul 04 01:57:01 Does XP2000+ count? Jul 04 01:57:37 NAiL: can Jul 04 01:58:23 NAiL: how long 'monotone pull' takes on your machine? Jul 04 01:58:43 Ages. 15-20 minutes I'd say. Jul 04 01:59:21 I haven't timed it. I'm doing a build of openslug right now, so timing it now would skew the results a bit Jul 04 02:03:19 NAiL: me of course Jul 04 02:03:38 mickeyl: ok. Can I add the user? :) Jul 04 02:03:42 NAiL: yep Jul 04 02:03:48 hrw|work: me, at home I have only 1 GHz Jul 04 02:04:10 mickeyl: neat. :) Jul 04 02:04:21 it looks like next time I will checkout whole repo instead of pull/merge Jul 04 02:05:27 uf.. monotone pull finished.. 10:18 -> 11:04 Jul 04 02:05:28 mickeyl: we have more discussion about the lameness of monotone than about OE. I think that's a bad sign. Jul 04 02:05:40 schurig: i don't think so. Jul 04 02:06:03 mickeyl: yestereday evening when I was online people talked about this, now I'm only, people again talk about this Jul 04 02:06:04 schurig: a new tool always has impact Jul 04 02:06:13 months ago we had memoryhog (bitbake), now we have cpuhog (monotone) Jul 04 02:06:15 mickeyl: that's not a broad data basis ... :-) Jul 04 02:06:54 hrw|work: hmm, it wasn't actually THAT slow: "time monotone pull": real 1m19.171s Jul 04 02:07:07 hrw|work: however, I have a tree where I never made any change so far Jul 04 02:07:34 I have to talk with management here - one unused machine has xp2000+ cpu (I have 1700+ here) Jul 04 02:07:41 time monotone update: real 0m11.851s Jul 04 02:08:05 schurig: I have no changes here - monotone update took 3minutes Jul 04 02:08:32 machine has usual tasks running - kde341, oe build Jul 04 02:09:14 the problem is that monotone does things that most people don't like. verifying lots of SHA1 signatures (or maybe even RSA keys, not sure) Jul 04 02:09:22 ah, don't need Jul 04 02:09:29 not don't like Jul 04 02:21:50 to put a last word on to this. It's a new tool and this has quite an impact for us, it's totally clear that this will be the number one topic for a couple of days. I won't participate in a discussion if the decision to use it was good or not good until I have been presented something that counts as an alternative. Jul 04 02:22:06 * mickeyl goes back to work Jul 04 02:23:05 * NAiL was thinking more along the lines "how can this be fixed" anyway. I'm kinda happy with monotone. Jul 04 02:23:18 * florian_kc likes monotone... it is quite some improvement. Jul 04 02:27:31 monotone gives us more freedom - so some more people will use OE Jul 04 02:27:54 I hope so :-) Jul 04 02:28:19 http://lists.gnu.org/archive/html/monotone-devel/2005-07/msg00016.html Jul 04 02:30:00 Looks like a good start to me ;) Jul 04 02:30:44 hrw|work: I agree with this Jul 04 02:31:14 hrw|work: yesterday I went to the webpage of some CMS/Blog utility and saw that they have BK as respository. I instantly went away from the web page ... Jul 04 02:37:20 morning all Jul 04 02:37:38 morning RP Jul 04 02:37:43 hi RP Jul 04 02:37:55 hi RP Jul 04 02:43:45 morning folks Jul 04 02:49:28 hello ade|desk Jul 04 02:50:02 hi ade Jul 04 02:50:45 Is there a way to un-commit my changes before a sync? Jul 04 02:53:25 try disapprove Jul 04 02:57:04 hmm, might have worked. I got two heads though ;) Jul 04 02:57:22 two head better then one ?? Jul 04 02:57:42 yeah Jul 04 02:57:43 ;) Jul 04 02:57:49 mickeyl: Yeah, worked. Thanks :) Jul 04 02:59:53 excellent. would you be so kind to add this question and answer into our MonotonePhraseBook wiki entry? Jul 04 03:00:27 Yes. Where is this monotonephrasebook? Jul 04 03:00:50 http://oe.handhelds.org/cgi-bin/moin.cgi/MonotonePhraseBook Jul 04 03:04:27 mickeyl: BTW, NAiL is our newest nslu2-linux core team member. He is going to be managing the OpenSlug package feed ... Jul 04 03:07:22 mickeyl: wiki updated Jul 04 03:07:33 rwhitby-away: i see. thanks for the info Jul 04 03:07:37 welcome on board, NAiL Jul 04 03:07:48 thanks NAiL Jul 04 03:07:57 np, and thank you :) Jul 04 03:27:37 is there anyway to find out what should go into RDEPENDS ?. i know DEPENDS must contain package name, RDEPENDS ; sometimes what u want to install is same as what u build example:busybox, but sometimes, it is different as in the case of xserver. is there any way to find it. i know .ipks are in ipk directory. but if u have already installed that ipk. then it's hard to find out ?. Jul 04 03:52:06 03florian 07org.openembedded.dev * re60c3e0f... 10/packages/gkdial/gkdial_1.8.1.bb: Fix build problem. Jul 04 03:55:00 mickey|lunch: can you please send me the CIA script you are using? Jul 04 03:55:06 mickey|lunch: looks like pull, merge, sync does the trick. Jul 04 03:58:56 Jumbo: some RDEPENDS are caluclated automatically, e.g. library dependencies Jul 04 03:59:26 Jumbo: for the others, the "way to find out what should go into RDEPENDS" is knowledge of the packages involved, what they do etc Jul 04 04:01:19 what is the output of the package (name of the ipk): is this what u mean ? Jul 04 04:04:14 Jumbo: in RDEPENDS you need to specify the package name of the IPK, not the name of the .bb file Jul 04 04:13:18 what i meant is the name of the ipk file Jul 04 04:21:36 am i right ? Jul 04 04:24:47 Jumbo: file name of the ipkg, e.g. RDEPENDS += "update-modules" if the file name is "update-modules_1.0-r3_all.ipk" Jul 04 04:40:05 mickey|lunch: Are there any problems with monotone.vanille.de? I have web access, but mt tells me "connection refused". Jul 04 04:41:25 hi Jul 04 04:43:17 hi dirk Jul 04 04:43:24 monotone: fatal: std::runtime_error: trapped network exception: recv failure: No route to host Jul 04 04:43:36 our network connection sometimes drops for a couple of seconds Jul 04 04:43:46 i need to write a restart loop around monotone Jul 04 04:43:51 hi mickeyl Jul 04 04:44:21 restarted. Jul 04 04:47:27 mickeyl: thanks Jul 04 04:53:28 03florian 07org.openembedded.dev * r31f722a1... 10/packages/tslib/ (tslib/nokia770/ts.conf tslib_cvs.bb): Add configuration file for Nokia 770. Jul 04 04:53:29 03florian 07org.openembedded.dev * r1e3ae573... 10/packages/maemo/gconf-osso/gconf-daemon-dbus-oe.sh: Add missing startup scripts. Jul 04 04:53:29 03florian 07org.openembedded.dev * r4f22ecbf... 10/packages/maemo/osso-core-config_200515.1.bb: Jul 04 04:53:29 Update postinst task to create default user properly and link Jul 04 04:53:29 to stuff located in initrd. Jul 04 04:53:43 ah excellent. no forks :) Jul 04 04:57:57 does anyone think we need this gcc MULTILIB stuff? on ppc, -msoft-float currently adds a "nof" to all libgcc/libstdc++ paths, screwing up the install/packaging process Jul 04 04:59:48 so my thought was to simply remove the MULTILIB names instead of making the do_install much more complicated. I think nobody will ever want to produce binaries with and without softfloat with the same environment. Jul 04 05:30:39 hey Jul 04 05:32:47 tmbinc: sounds fair enough Jul 04 05:39:10 hey zecke Jul 04 05:40:41 zecke: the thing is, gcc3-build.inc does have an --enable-multilib. is that done on purpose? or does it affect something else? (honestly, i'm a bit confused by --disable-multilib, --without-multilib, --enable-multilib=no) Jul 04 05:47:09 any plans for updating OE gettingstarted with monotone info? Jul 04 05:51:01 tmbinc: For arm we do not use multilib (or it is not working), so building without multilib sounds okay Jul 04 05:51:23 tmbinc: but pb_ would have the right(tm) answer Jul 04 05:57:29 ok i'll ask him when he's back Jul 04 06:17:11 ~lart 2.4.20 Jul 04 06:17:11 * ibot gives 2.4.20 an extra strength ACME sleeping pill, sending 2.4.20 to sleep for 150 years, and awakening to seven strange dwarfs and a large apple Jul 04 06:17:27 BigAl: what's the status on spitz/2.6? i can't stand 2.4.20 :/ Jul 04 06:17:43 do13: How are things progressing with tosa? :) Jul 04 06:19:28 mickeyl: Not much progress over the last couple of weeks unfortunately, had exams, and then got a cold the week after they finished. Jul 04 06:21:05 feelling better yet? Jul 04 06:21:42 Yeah, today I felt pretty good. Jul 04 06:22:58 excellent. health is most important anyway Jul 04 06:23:19 Yup. It's a bit hard to do anything when you're feeling terrible. Jul 04 06:24:41 mickeyl: slowly. At this time I'am looking at the battery charging. Jul 04 06:27:37 mickeyl: During my vacation I used the tosa with 2.6 kernel and a opie image. Some parts are rough! :) But the machine is usable. Jul 04 06:28:47 03mickeyl 07org.openembedded.dev * r2b0c7532... 10/conf/machine/zaurus-clamshell-2.4.conf: Zaurus/2.4 Conf: readd apmd and remove some things for the new spitz bootstrap Jul 04 06:31:44 do13: excellent. i'll give it a try asap. But how about charging? Jul 04 06:32:32 Not yet :) Jul 04 06:33:04 ok, so to let it charge, we can reboot and let 2.4.18 take over. Jul 04 06:34:19 2.4.18 isn't necessary. The bootloader did the same job. Jul 04 06:35:07 Ok. I hope in a few days I have a first charging version. Jul 04 06:35:16 do13: I'm undecided about what to do with the charging code for the c7x0. Its extremely ugly :-/ Jul 04 06:37:05 RP: The tosa is more complicated. There is another battery in the jacket. So there are 2 batteries which needs to be charged. Jul 04 06:37:38 do13: what a mess :-( Jul 04 06:37:49 do13: Is the charging code common or totally different? Jul 04 06:38:04 different Jul 04 06:38:51 Maybe we can extract some common parts and form machine specific hooks. Jul 04 06:39:44 Or maybe we just try and evolve seperate drivers. I guess if we do that we can merge any common parts left later... Jul 04 06:39:58 No easy solution... Jul 04 06:40:07 Indeed. Jul 04 06:41:24 My biggest problem is the ADC. These are inside the AC97 codec. Jul 04 06:42:58 do13: You could make the chip itself a SoC device. I'm doing that for the tsc2101 in the ipaq hx2750 Jul 04 06:43:27 the tsc2101 is a sound chip, touchscreen and adc with a couple of gpios Jul 04 06:44:10 (So far I've not implemented the sound) Jul 04 06:46:39 Did someone here notice any problems with serial console in latest images? Jul 04 06:46:47 don't worry too much about the jacket Jul 04 06:46:53 it has been sold in _extremely_ limited quantities Jul 04 06:46:59 florian_kc: you mean in opie images? Jul 04 06:47:04 er, gpe images Jul 04 06:47:15 hi kergoth Jul 04 06:47:20 RP: The sound interface is implemented (some mixer settings are wrong): ALSA - AC97 - WM9712. But the kernel is too big if I enable the alsa stuff. Jul 04 06:47:38 florian_kc: I found the latest xserver seems to jam /dev/ttyS0 on the c7x0 Jul 04 06:48:07 RP: I built a gpe and a maemo image - both images suffer from a missing serial console. Jul 04 06:48:14 mickeyl: If I have no jacket to test, I won't support this :) Jul 04 06:48:24 do13: that's the spirit :D Jul 04 06:48:43 RP: hrm... i use the same Xserver like i did before. Jul 04 06:49:27 florian_kc: the serial consoles are fine until xserver starts after which I have to kill the getty/shell on the port to get it working again Jul 04 06:49:45 florian_kc: Its only with the June xservers rather than the Feburary ones I was using... Jul 04 06:50:16 do13: I guess we'll face the space problem on c7x0 as well. We'll probably have to make sound modular Jul 04 06:50:58 RP: I'm using the Xomap package which used to work... Jul 04 06:51:35 florian_kc: Your problem is probably something else then... Jul 04 06:53:33 RP: I did this already on the tosa. Jul 04 06:54:03 * mickeyl builds tosa/2.6 Jul 04 06:56:34 hmmm Jul 04 06:56:59 RP: Is that problem maybe related to x11 instead of xserver? Jul 04 06:57:31 florian_kc: Could be. I would have rebuild that at the same time... Jul 04 06:57:34 mickeyl: one hint. You need a new defconfig for tosa. http://www.do13.in-berlin.de/openzaurus/temp/defconfig-tosa Jul 04 07:01:41 do13: ah ok, thanks. should I commit that? Jul 04 07:02:12 Yes please. Jul 04 07:02:19 k Jul 04 07:02:23 2.6.12-mm1 ? Jul 04 07:02:32 yes. Jul 04 07:02:35 *nod* Jul 04 07:02:58 Has anyone tried 2.6.13-rc1-mm1 yet? Jul 04 07:03:54 mickeyl. This defconfig builds pcnet driver as module. Jul 04 07:04:18 do13: righto. that is consistent with the c7x0 defconfig. i already changed that in BOOTSTRAP_EXTRA_RDEPENDS Jul 04 07:05:07 mickeyl: yep. opie-image won't package without this module. Jul 04 07:05:22 I recognized this 30 minutes ago. Jul 04 07:06:35 RP: nope, any major changes? Jul 04 07:06:47 RP: Not yet. The last days I was on vacation. So I have to read some emails first. Jul 04 07:06:50 RP: as far as I can gather 2.6.12-mm1 broke Irda Jul 04 07:07:02 oh no, is that true? Jul 04 07:07:06 XorA: really? Jul 04 07:07:09 so much for the new development model :/ Jul 04 07:07:30 RP: seems to have, I can no longer connect to my phone, but Im clueless when it comes to irda Jul 04 07:07:39 XorA: I'm epecting the backlight class to break soon as well... Jul 04 07:08:10 According to John Lenz's email... Jul 04 07:09:02 03mickeyl 07org.openembedded.dev * rb378da25... 10/packages/linux/linux-openzaurus-2.6.12-mm1/defconfig-tosa: Update defconfig-tosa for linux-openzaurus-2.6.12-mm1 - courtesy Dirk Opfer. Jul 04 07:23:52 re Jul 04 07:24:20 I tried 2.6.13-rc1 on desktop only - get back to 2.6.12.1 as reiserfs was nasty Jul 04 07:26:53 wb Jul 04 07:41:24 ~lart kwallet, kdepim for useless working Jul 04 07:41:24 * ibot nabs the moon and broadsides kwallet, kdepim with the sea of tranquility for useless working Jul 04 07:42:39 let kde devs turns into gnomes... Jul 04 07:43:51 hrw|work: use evolution ;) Jul 04 07:43:59 hrw|work: now I need to only embed kmail into evo Jul 04 07:44:26 zecke: sooner I will use m$ lookout depress (so called OE) then evo Jul 04 07:58:53 re Jul 04 08:03:59 hey, ibot is back! Jul 04 08:04:04 ibot: welcome! Jul 04 08:11:49 03jbowler 07org.openembedded.nslu2-linux * r6edd8711... 10/packages/openslug-init/ (2 files in 2 dirs): Make a raw, unconfigured, NSLU2 default to DHCP boot proto Jul 04 08:11:50 03jbowler 07org.openembedded.nslu2-linux * rbfb64dfb... 10/conf/distro/openslug.conf: Enable 2.6.12[.*] kernel Jul 04 08:11:51 03jbowler 07org.openembedded.nslu2-linux * r6f01d56b... 10/ (23 files in 18 dirs): Jul 04 08:11:51 propagate from branch 'org.openembedded.dev' (head e7deec567712930f830c6a392453b58268463120) Jul 04 08:11:51 to branch 'org.openembedded.nslu2-linux' (head bfb64dfb19eb6055f394a9ee933ebb8c93bcbe75) Jul 04 08:11:52 03nail 07org.openembedded.nslu2-linux * r5dcdc1c2... 10/packages/base-passwd/base-passwd-3.5.9/ (openslug/root-home.patch root-home.patch): Added ntpd user for openntpd. Jul 04 08:11:55 03nail 07org.openembedded.nslu2-linux * r07fb1e97... 10/packages/base-passwd/base-passwd-3.5.9/ (openslug/root-home.patch root-home.patch): disapproval of revision 5dcdc1c208d97529b980412b7bf1e2cabcc78297 Jul 04 08:11:58 03nail 07org.openembedded.nslu2-linux * rbead1190... 10/packages/openntpd/files/init: Added openntpd init-script Jul 04 08:12:01 03nail 07org.openembedded.nslu2-linux * re0d6457b... 10/packages/openntpd/files/makefile-install.patch: Fixed makefile to ignore that the ntpd user may not exist on the build system. Jul 04 08:12:04 03nail 07org.openembedded.nslu2-linux * r4f2dd7b3... 10/packages/openntpd/files/adjtimex-3.7p1.patch: Added patch for openntpd-3.7p1 to use the adjtimex() call. Jul 04 08:12:07 03nail 07org.openembedded.nslu2-linux * r5c1a5813... 10/packages/base-passwd/ (3 files in 3 dirs): Added ntpd user (for openntpd privsep), and bumped revision. Jul 04 08:12:10 03nail 07org.openembedded.nslu2-linux * re04a7f68... 10/packages/openntpd/openntpd_3.7p1.bb: Added new version of openntpd. Jul 04 08:12:13 03rwhitby 07org.openembedded.nslu2-linux * rd6fdb02a... 10/ (21 files in 16 dirs): Jul 04 08:12:15 propagate from branch 'org.openembedded.dev' (head b378da256e147a5c55663f95c0cc3df118a4a63d) Jul 04 08:12:17 to branch 'org.openembedded.nslu2-linux' (head e04a7f684ef5855d717ea0e66d63fd31465c451a) Jul 04 08:12:19 03nail 07org.openembedded.nslu2-linux * rfc7a16d8... 10/packages/meta/openslug-packages.bb: Added working openntpd. Jul 04 08:12:43 wohooo Jul 04 08:15:34 mickeyl: is the above going to make us unpopular? Each time we do a propagate from nslu2-linux to oe and push, that will happen (i.e. all changes since last time will be reported). Whilst it ensures that #oe's CIA is a true record of all openembedded changes, I wonder if anyone will get upset by a whole bunch of them happening at once. Jul 04 08:16:31 mickeyl: my rev r07fb1e97... is what happens when I disapprove. So a commit, disapprove, and a new commit will be three messages. Just so you know :) Jul 04 08:16:50 rwhitby-asleep: we need something like MT_SILENT=1 before pushing - like it was with BK_SILENT Jul 04 08:17:27 hrw|work: since these will always be on org.openembedded.nslu2-linux, you can just ignore that branch in ciabot_monotone.py if it's a problem Jul 04 08:19:25 probably yes Jul 04 08:20:31 hrw|work: in nslu2-linux, we are ignoring the org.openembedded.dev branch, so you should ignore us in return. Jul 04 08:20:52 mickeyl: did you hear that? :) Jul 04 08:21:22 rwhitby-asleep: that only happens when you propagate (push) into our oe tree? Jul 04 08:22:18 zecke: yes. We get them individually reported in #nslu2-linux as they happen on org.openembedded.nslu2-linux, and you get them as a batch when we propagate and push back. Jul 04 08:23:14 so each change is only reported here once, but all in one batch. Jul 04 08:23:56 so it's your choice (in the project_for_branch routine in ciabot_monotone.py) as to whether you report them here or not. Jul 04 08:25:01 I'm actually tempted to turn it back on in #nslu2-linux so when we propagate from org.openembedded.dev to org.openembedded.nslu2-linux, we see all the changes (like we used to with bk pull) roll past ... Jul 04 08:25:08 mickeyl: Is the website kept in monotone too? Jul 04 08:25:45 rwhitby-asleep: sorry our music raid is faulty, and the kernel loves to oops when setting this ole disk faulty... Jul 04 08:26:03 [00:52] rwhitby-asleep: zecke: yes. We get them individually reported in #nslu2-linux as they happen on org.openembedded.nslu2-linux, and you get them as a batch when we propagate and push back. Jul 04 08:26:08 [00:53] rwhitby-asleep: so each change is only reported here once, but all in one batch. Jul 04 08:26:15 [00:53] rwhitby-asleep: so it's your choice (in the project_for_branch routine in ciabot_monotone.py) as to whether you report them here or not. Jul 04 08:26:21 [00:54] rwhitby-asleep: I'm actually tempted to turn it back on in #nslu2-linux so when we propagate from org.openembedded.dev to org.openembedded.nslu2-linux, we see all the changes (like we used to with bk pull) roll past ... Jul 04 08:26:28 rwhitby-asleep: personally I do not feel disturbed and welcome to 'monitor' the progress you make Jul 04 08:26:29 you're now up to date :-) Jul 04 08:27:23 brb Jul 04 08:33:58 *cry* Jul 04 08:36:40 ~lart warwick Jul 04 08:36:41 * ibot gives warwick an extra strength ACME sleeping pill, sending warwick to sleep for 150 years, and awakening to seven strange dwarfs and a large apple Jul 04 08:41:51 Good evening. Jul 04 08:46:27 evening SirFred Jul 04 08:46:37 RP: Hi. Jul 04 08:49:21 rwhitby-asleep: No problem with these changes. They only appear once and will keep us up to date. Jul 04 08:49:26 s/changes/commit messages/ Jul 04 08:50:03 florian_kc: no, the oe website is @ hh.org/home/oe/ Jul 04 08:55:41 andersee: hey Jul 04 08:56:10 andersee: you recently posted some nice patches to monotone-devel. I've tried to apply them against HEAD but all of 'em failed. Would it be possible that I could grab a complete tarball of your mt version? Jul 04 08:56:41 mickeyl: morning Jul 04 08:56:52 mickeyl: monotone? Jul 04 08:56:54 andersee: hey Jul 04 08:56:58 ups Jul 04 08:57:08 oh well time to go brave the riots, cya Jul 04 08:57:14 mickeyl: afraid that must be somebody else... Jul 04 08:57:18 andersee: sorry, never mind. I just mixed you with Eric Anderson Jul 04 08:57:34 similar enough Jul 04 08:57:41 * andersee is 'Erik Andersen' Jul 04 08:57:50 mickeyl: shame on you Jul 04 08:57:51 so close Jul 04 08:58:08 heh, close but no cigar Jul 04 08:58:24 mickeyl: should be possible to apply them to f56500ddd113aa716173f6d94db5d59c88bdc20 (the id in the message). Jul 04 08:58:43 jbowler-away: ah right. i overlooked the id Jul 04 08:58:47 I found that building off the head was problematic... Jul 04 08:59:45 ah cool, my message made it to the monotone-list Jul 04 08:59:54 and the response is satisfactory Jul 04 09:00:05 err, nice. Jul 04 09:00:09 note_netsync_{revision,cert,pubkey}_received. Don't recall if they Jul 04 09:00:09 exist in 0.19, will definitely be in 0.20. Jul 04 09:00:21 that will get us a push interface for ciabot_monotone.py Jul 04 09:25:24 mickeyl: ah okay... i just wondered how to extend the "Success Stories" section :-) Jul 04 09:25:47 Why doesn't CVSDATE_xserver-kdrive=20050630 force using that CVS date? Jul 04 09:27:12 Luke-Jr: maybe xserver-kdrive_cvs isn't the default one. Jul 04 09:28:03 Luke-Jr: We have some kdrive bb with fixed dates now and at least familiar defaults to one of these. Jul 04 09:28:51 florian_kc: I doubt there would be fixed dates for yesterday/broken... Jul 04 09:31:19 Luke-Jr: indeed... Jul 04 09:47:08 ok. opie upgraded to 20050704 can go now Jul 04 09:47:10 cu all Jul 04 09:47:38 monotone is slow :| Jul 04 09:52:16 Luke-Jr: yeah, "verifying revisions" after sync seems to take awhile (as it suggests). Local operations seem to be pretty quick though. Jul 04 10:27:22 03florian 07org.openembedded.dev * r33378305... 10/ (4 files in 4 dirs): Jul 04 10:27:22 Move responsibility for setup links to extra package. Jul 04 10:27:22 Default to osso gtk version in maemo images. Jul 04 11:30:12 Do we have an autoconf guru around? Jul 04 11:52:57 evening pb_ Jul 04 11:53:03 hi rp Jul 04 11:58:57 I'm trying to get minicom to compile. It appears to have an autoconf problem. It fails to compile, basically because config.h contains /* #undef POSIX_TERMIOS */ rather than #define POSIX_TERMIOS Jul 04 11:59:28 configure says the system does have posix TERMIOS though... Jul 04 12:01:17 config.log says ac_cv_sys_posix_termios=yes configure.in tests $am_cv_sys_posix_termios ... Jul 04 12:02:42 I don't know autotools very well :-/ Jul 04 12:14:28 http://www.rpsys.net/openzaurus/temp/configure.patch - the first hunk is what I think needs adding. Can someone who knows what they're doing sanity check this please? :) Jul 04 12:15:47 mickeyl: I like the interactive mode. I can't seem to get some of the commands to work though... Jul 04 12:38:39 ah, hi pb_. I'm thinking of disabling multilib in gcc, what do you think of that? (-msoft-float on ppc screws up the build process here because of the nof/-suffixes, and i generally don't see any reason to have it enabled.) is multilib required on some architecture? Jul 04 12:43:25 eww @ multilib Jul 04 12:48:16 hm? Jul 04 12:54:58 tmbinc: go ahead and disable it for ppc, if you like. I would prefer you not to do that globally though. Jul 04 13:24:05 RP: Which commands in particular are not working for you? Jul 04 13:24:45 mickeyl: I tried something like "configure minicom force" Jul 04 13:24:59 ah Jul 04 13:25:02 even "configure minicom" didn't work... Jul 04 13:25:03 that's not how it's supposed to work Jul 04 13:25:09 force is a command on its own Jul 04 13:25:16 configure minicom should work though Jul 04 13:25:17 let me try Jul 04 13:25:43 Try with minicom as its broken for me - then you can check my fix ;-) Jul 04 13:25:54 ok Jul 04 13:25:56 btw. Jul 04 13:26:06 i got some first hand tests for 3.5.4-rc from an experienced user Jul 04 13:26:10 some kernel problems are mentioned Jul 04 13:26:37 Do yo have a link? Jul 04 13:26:42 no, it's german anyway Jul 04 13:26:45 here it comes: Jul 04 13:26:49 Ah :) Jul 04 13:26:54 - screen looks like 60Hz interlace Jul 04 13:26:58 - infrared broken Jul 04 13:27:03 - sd not recognized Jul 04 13:27:20 - errornous stylus clicks Jul 04 13:27:29 - complete freeze once (not reproducable) Jul 04 13:27:30 that's it Jul 04 13:27:40 the rest is userland Jul 04 13:28:27 Answers are setfastpll clock in sysfs, no idea, sd isn't recognised after booting - remove and replace card and it will, not see this before and seen but also not been reproduced Jul 04 13:28:29 evening Jul 04 13:28:39 Does that make sense? :) Jul 04 13:28:46 a bit Jul 04 13:28:57 We should set fastpllclock by default on the c7x0 images Jul 04 13:29:04 how would we do that? Jul 04 13:29:33 echo 1 > /sys/devices/platform/corgi-ssp/sharp-scoop/w100fb/fastpllclock Jul 04 13:29:39 or something like that Jul 04 13:29:44 ok, I'll take care of it. the proper location would be the devices script. Jul 04 13:30:11 You need to check the path as that above is probably wrong :) Jul 04 13:30:22 i reckon irda has been broken by mainline Jul 04 13:30:30 there were updates in 2.6.12 Jul 04 13:30:37 This was with the 2.6.12 kernel btw? Jul 04 13:30:37 irda, updates, to be detailed. Jul 04 13:30:41 yes, 2.6.12-mm1 Jul 04 13:31:03 Ok, the other known issue is the cf memory cards not umounting nicely Jul 04 13:31:27 I'd be interested to know if SD worked if he reinserted it. Jul 04 13:31:28 unfortunately it's not only cf memory Jul 04 13:31:46 cf cards as well? Jul 04 13:32:04 my BT card responds with a nice oops as well when I don't stop bluetooth and remove the modules :/ Jul 04 13:33:08 oh and something more regarding BT Jul 04 13:33:12 we have a race condition here Jul 04 13:33:28 regarding the uart? Jul 04 13:33:30 apparantly it takes longer to setup the serial port on the card than cardmgr likes Jul 04 13:33:38 which makes hciattach fail Jul 04 13:33:40 ah Jul 04 13:33:46 manually starting it afterwards works Jul 04 13:33:59 I don't know BT at all... Jul 04 13:34:01 mickeyl: I just got your botmail... i guess i managed to do it in the right way today :-) Jul 04 13:34:03 i guess i could "fix" that with a sleep 2 in /etc/bluetooth Jul 04 13:34:19 florian: yeah, you played nice with us today :)) Jul 04 13:34:32 the botmail was from yesterday nite Jul 04 13:34:52 florian: The nokia 770 "defconfig" in oe - is that real or a mock up? Jul 04 13:35:38 mickeyl: I suspect the delay might be easiest for now :-/ Jul 04 13:35:51 RP: That's actually the default defconfig for that CPU, i'll replace it with a real one as soon as i have it. Jul 04 13:35:57 Too many bigger bugs... Jul 04 13:35:59 RP: ok, the good news from the tester: "it's damn fast and responsive. suspend/resume works very solid." Jul 04 13:36:13 florian: Ok, I assume we have no kernel source yet? Jul 04 13:36:22 mickeyl: irda: try modprobe pxaficp_ir. This needs an module_autoload entry. Jul 04 13:36:41 do13: will try. thanks for the suggestion. Jul 04 13:36:48 mickeyl: That's nice to hear :) 2.6.12 is looking good in a lot of ways Jul 04 13:37:01 RP: bbshell: configure doesn't do what it's supposed to do. I'll investigate Jul 04 13:37:15 mickeyl: When is 3.5.4 due to ship? I should try and fix a few bugs before that... Jul 04 13:37:36 RP: yep. could we calm down for 3.5.4 now? I'd like us to ship it mid august. Jul 04 13:37:57 RP: Yes... they seem still to work on it, but the basically the 770 should work with the upstream kernel. Jul 04 13:38:39 florian: Ok, thanks. Just curious. I have a few things in mind to fiddle with when it becomes available ;-) Jul 04 13:38:48 RP: 2.6.12-mm1 looks like a solid base for 3.5.4 + some patches we might come around in the next weeks Jul 04 13:39:19 mickeyl: I'm tempted to make 2.6.12-mm1 the new "stable", remove 2.6.11 and add a new development kernel? Jul 04 13:39:19 RP: The only blocking issue for w100 per default seems to be the suspend/resume freakout Jul 04 13:39:22 RP: heh cool... what are you thinking of? Jul 04 13:39:27 RP: ACK. Jul 04 13:39:43 hmm... w100?! Jul 04 13:39:57 florian: ATI W100 in Zaurus C7x0 Jul 04 13:40:03 mickeyl: That is a userspace issue Jul 04 13:40:07 ahh! Jul 04 13:40:18 RP: ok. i'll continue to pester Sirfred|away then :D Jul 04 13:40:22 RP: please help me ;) Jul 04 13:40:39 mickeyl: Sirfred|away and I have been working on things ;-) Jul 04 13:40:44 Why doesn't CVSDATE_xserver-kdrive=20050630 force using that CVS date? Jul 04 13:40:55 It does. Jul 04 13:40:56 RP: excellent. you two are doing a helluva RE job Jul 04 13:41:01 pb_: not for me :( Jul 04 13:41:14 I think we managed 1000 letter blits in random colours in 32ms today :) Jul 04 13:41:19 RP: WOOOT! Jul 04 13:41:37 RP: cool! Jul 04 13:42:03 mickeyl: It just proves we can use the MonoChar function and interfacing it to qte will be fun but Sharp managed it so I'm hopeful :) Jul 04 13:44:29 yep Jul 04 13:44:59 Luke-Jr: ah, that's very sad Jul 04 13:45:12 mickeyl: re the sus/res problem, I think all Sirfred needs it a call to say the device is suspending and another when it resumes Jul 04 13:45:13 pb_: so no idea why it's misbehaving? Jul 04 13:45:46 nope. nobody else has complained, though, so I guess it must be some local problem at your end. Jul 04 13:46:34 when you say it's "misbehaving", what exactly happens when you build xserver_cvs.bb? Jul 04 13:46:44 xserver-kdrive_cvs.bb, that is Jul 04 14:00:17 pb_: it uses the most recent date Jul 04 14:01:39 meaning 20050704? Jul 04 14:01:43 hmm... nm 'most recent', was reading the date wrong. Jul 04 14:02:03 20050207, which is a frozen snapshot Jul 04 14:02:15 wonder why it fails compiling and why it suddenly stopped using 'cvs' Jul 04 14:02:50 ah, so you aren't actually building xserver-kdrive_cvs.bb? Jul 04 14:03:00 if you are using some other .bb file, that would certainly explain why CVSDATE is ignored. Jul 04 14:05:03 yea, now the question is why BB suddenly decided not to use the _cvs.bb Jul 04 14:05:09 it used to Jul 04 14:05:30 PREFERRED_VERSION should work, I suppose Jul 04 14:27:49 RP: hats off to you. you triggered a bug in BK :) Jul 04 14:27:56 s/BK/BB/ Jul 04 14:28:18 RP: It's not the shell's problem *phew* Jul 04 14:28:24 I've added the problem as #131, and #132 Jul 04 14:28:27 g'night Jul 04 14:29:48 mickey|zzZZzz: 'night Jul 04 14:29:58 mickey|zzZZzz: Thanks for looking into it :) Jul 04 14:33:00 np Jul 04 14:33:17 while you're working with bbshell, use fileconfigure instead. that one should always work Jul 04 14:33:28 all the file* commands should work Jul 04 14:33:33 * mickey|zzZZzz wanders off Jul 04 14:33:39 mickey|zzZZzz: good night Jul 04 14:34:07 mickey|zzZZzz: thanks and g'night :) Jul 04 14:35:14 pb_: Did you look at the configure patch for minicom above? Jul 04 14:35:46 nope Jul 04 14:35:49 what does it do? Jul 04 14:36:38 It changes a couple of AM references to AC which makes minicom detect a POSIX_TERMIOS seeminly correctly Jul 04 14:37:24 oh, right, very good Jul 04 14:37:39 I think we used to work around that in the site file, but fixing it properly would be nice. Jul 04 14:37:41 Is that the correct solution? Jul 04 14:37:46 let me have a look Jul 04 14:37:58 03florian 07org.openembedded.dev * r23bb3a52... 10/conf/machine/nokia770.conf: Update to new udev naming. Jul 04 14:37:58 03florian 07org.openembedded.dev * r2d388161... 10/packages/maemo/osso-esd_0.5.bb: Limit RPROVIDES to the binary package. Jul 04 14:38:00 03florian 07org.openembedded.dev * r041417bf... 10/packages/maemo/ (nokia770-init/fixup-770.sh nokia770-init_1.0.bb): Jul 04 14:38:00 Add package to do some initialisation and fixups like unmounting stuff Jul 04 14:38:00 from initrd and create some useful links. Jul 04 14:38:27 pb_: Keep in mind I don't know much about autotools... ;-) Jul 04 14:38:39 yeah, that looks ok to me Jul 04 14:38:53 AM_SYS_POSIX_TERMIOS is the old school. Jul 04 14:39:29 right. I'll check that in then Jul 04 14:39:34 pb_: did you say anything to my MULTILIB issue? i pinged out after asking that question.. Jul 04 14:40:04 tmbinc: go ahead and disable it for ppc, if you like. I would prefer you not to do that globally though. Jul 04 14:40:07 pb_: just wanted someone who has some knowledge ot what that does to sanity check it, thanks :) Jul 04 14:40:25 RP: you might like to zap the site entries for that at the same time. Jul 04 14:41:00 pb_: but won't the packaging break anyway if any architecture produce a libgcc with a multilib suffix? Jul 04 14:41:19 or is multilib more than this naming scheme? Jul 04 14:42:01 (at least it breaks for ppc, not sure for any other platform which uses that) Jul 04 14:42:36 tmbinc: multilib is a bit more than that. I think there is something a bit funny about the way ppc uses it. Jul 04 14:43:13 right now, I don't think any platform is actually using multilib, but we might want to start doing that for arm at some point. Jul 04 14:44:47 well, it's documented (at least i've understand it that way) that every MULTILIB_OPTION (like msoft-float) corresponds to a MULTILIB_DIRNAME (like nof, in ppc's case). so whenever you specify one of them, it will rename the libgcc, thus breaking the packaging/install Jul 04 14:44:59 so at i think in the current form it can never work Jul 04 14:45:34 that's correct, but you should also get a "normal" libgcc.a in the usual place. Jul 04 14:45:45 for use when no multilib option is specified at runtime Jul 04 14:45:46 just that arm, for example, has no MULTILIB_OPTION in their arm/t-linux Jul 04 14:45:59 pb_: There are no site entries for that other than for sh3 and sparc. That's probably why it was broken for me :) Jul 04 14:46:25 RP: yeah, probably. I don't think anybody has really been using minicom other than on those platforms. Jul 04 14:46:58 (i once used minicom on ppc, after fixing the AM/AC stuff, it worked quite well.) Jul 04 14:46:59 actually, I thought minicom was also used on i586, but maybe I imagined that. Jul 04 14:48:22 pb_: i can't say for sure, but at least libstdc++ is NOT at the normal place, only with a nof/ added in the path. iirc it was the same for libgcc (only libgcc_nof available, no libgcc) Jul 04 14:48:55 yeah, that's what I meant by "there is something a bit funny about the way ppc uses it". Jul 04 14:49:03 That doesn't happen on other platforms. Jul 04 14:49:40 yes, because they don't have any MULTILIB_OPTIONS/MULTILIB_DIRNAMES defined, thus not allowing any multilib at all (or am i wrong here?) Jul 04 14:50:07 but i'll disable it only for ppc then Jul 04 14:50:17 no, I mean it doesn't happen on other platforms even when multilib is enabled. Jul 04 14:50:35 like I say, no platform other than ppc is currently using multilib, but arm has used it in the past and might want to use it again in the future. Jul 04 14:51:17 do you know what to fix in the ppc's config to generate "proper" multilibs then? Jul 04 14:51:27 no Jul 04 14:51:34 ok Jul 04 15:22:10 How can I check the changes monotone is going to send without actually doing it? Jul 04 15:24:57 monotone log will give me the outline I guess... Jul 04 15:28:31 03rpurdie 07org.openembedded.dev * rfdf14f17... 10/packages/sharp-binary-only/sharp-aticore-oss_1.0.1.bb: Update atilib to avoid a segfault Jul 04 15:33:52 03rpurdie 07org.openembedded.dev * r8f8edb0b... 10/packages/minicom/minicom-2.1/configure.patch: minicom: Fix configure to correctly use POSIX_TERMIOS Jul 04 15:38:42 kergoth: question for you... Jul 04 15:38:53 good night Jul 04 15:39:59 andersee: if its about oe, i've been out of it for like months now, so.. :) Jul 04 15:40:15 kergoth: heh Jul 04 15:40:17 kergoth: ok Jul 04 15:40:21 me too Jul 04 15:40:44 * andersee just realized oe is no longer using bk... Jul 04 15:40:44 i think i need a new project, or even more than that, a new hobby entirely Jul 04 15:40:45 heh Jul 04 15:41:08 You could always take up farming.... Jul 04 15:41:31 Just like open source, lots of work with little to show for it. Jul 04 15:41:34 :-) Jul 04 15:41:38 hehe Jul 04 15:43:25 * andersee nukes about 20 GB work of old bk based attempted oe builds Jul 04 16:23:36 'night all Jul 04 22:41:08 hi friends, i would like to know how we can get access to a ram on a remote system ? Jul 04 22:56:20 does that even make sense? Jul 04 22:56:54 emte, was that question for me ? Jul 04 22:57:44 yeah i think so Jul 04 22:58:07 more in general because i was wondering if anyone else understood it Jul 04 22:58:13 nope Jul 04 22:58:24 emte, my issue i have to get the ram image of a system over the network Jul 04 22:58:38 okay that makes more sense Jul 04 22:58:43 now did you get it ? Jul 04 22:58:53 maisheri: ram? or flash? Jul 04 22:58:56 just drop "ram" next time Jul 04 22:59:05 ram Jul 04 22:59:17 and to transfer an image across a network, pick a transport agent of choice Jul 04 22:59:27 i didn't get you Jul 04 22:59:29 :( Jul 04 22:59:41 there are many ways Jul 04 23:00:01 dd and netcat might work Jul 04 23:00:10 my current guess is you want an rboot or bootp system Jul 04 23:00:12 dd ? Jul 04 23:00:25 NAiL, how using dd ? Jul 04 23:00:46 which means it "pulls" in the image off a remote server when it powers up Jul 04 23:01:08 aha... Jul 04 23:01:16 I think emte is better at guessing than me Jul 04 23:01:30 yeah Jul 04 23:01:37 emte somthing like that Jul 04 23:01:39 lol Jul 04 23:01:48 those two are the easiest Jul 04 23:02:02 but how ... Jul 04 23:02:07 to be exact ... Jul 04 23:02:13 since you can burn an eeprom and drop it in a network card to do just that Jul 04 23:02:15 if the two OS are different ? Jul 04 23:02:37 the os really has nothing to do with it Jul 04 23:03:01 you can still view a webpage no matter what platform the webserver is on Jul 04 23:03:01 i would like to clarify a bit more ... Jul 04 23:03:25 i have to get the ram image from a pda over the usbnet Jul 04 23:03:29 all you are doing is telling it WHERE to find it's image Jul 04 23:03:58 why didnt you say so ? Jul 04 23:04:20 did you read the howto pages? there are quite a few on image backup and restore Jul 04 23:05:01 thats far simpler that what i thought you were trying to do Jul 04 23:05:12 emte, i think still things are not clear ... Jul 04 23:05:20 apperently not Jul 04 23:05:23 i will try to elucidate the scenario Jul 04 23:05:29 how about a better description Jul 04 23:05:33 yeah Jul 04 23:07:23 connecting a pda to the linux desktop or a laptop and then getting the ram image of the pda without disturbing the contents of the ram .. Jul 04 23:07:41 this is my objective Jul 04 23:08:03 i hope it explains better than i tried earlier Jul 04 23:08:56 emte, did you get it Jul 04 23:08:57 ? Jul 04 23:09:06 when i run ts_print i get the error : ts_open: No such file or directory Jul 04 23:12:49 emte, waiting for response ... Jul 04 23:29:58 sorry Jul 04 23:30:01 gf was talking to me Jul 04 23:31:00 so you want to backup the data in your ram and store it on the network ... Jul 04 23:31:43 that is what multisync is supposed to do ... but i havent got it to work for "everything" myself Jul 04 23:32:06 just the basic contact, calandar stuff ... Jul 04 23:32:58 multisync, synce, etc Jul 04 23:36:09 Good morning Jul 04 23:36:23 i get the error TslibInit() failed to open tslib when i run Xfbdev. i have ts.conf file and /dev/input/event1 both are present. libraries releated to touch screen are present. dont know what it is looking for ? Jul 04 23:36:48 but i am looking for ram and rom image not for the PIM ... Jul 04 23:36:51 :( Jul 04 23:39:24 yeah i am not sure if there is a full image app Jul 04 23:39:51 you might have to do as NAiL said and dd the rom and transfer it manually Jul 04 23:40:11 bbl Jul 04 23:40:17 bbl ? Jul 04 23:42:06 emte dd is possible only if the pda where a linux system ... if it is some other proprietory system... how comes the question ? Jul 04 23:44:57 NAiL, you there Jul 04 23:45:06 ? Jul 04 23:48:07 NAiL, what is ur suggestion what could be done .. in this scenario Jul 04 23:48:08 ? Jul 04 23:50:16 i am having an error when i run Xfbdev. i get TslibInit ( ) failed to open tslib. how do i resolve my problem ?. Jul 04 23:50:44 i have used oe to compile my xserver. **** ENDING LOGGING AT Mon Jul 04 23:59:56 2005