**** BEGIN LOGGING AT Tue Sep 13 02:59:56 2005 Sep 13 03:00:20 koen: iirc its pxa255 only Sep 13 03:00:36 my h2210 is a pxa255 :) Sep 13 03:00:42 It is - we'll need something else for pxa270 Sep 13 03:01:23 255 or 25x? Sep 13 03:01:40 hmm.. hx4700 is 2500 pln (~600 eur) Sep 13 03:01:52 depends on CPU_FREQ && PXA25x Sep 13 03:10:58 .... why do I feel cache today is completely useless Sep 13 03:11:30 Hmm. Do we switch to 2.6.13-mm3 or 2.6.14-rc1 Sep 13 03:16:46 any new code have to wait until 2.6.14 anyway Sep 13 03:18:17 RP: lots of your code in -rc1 Sep 13 03:19:14 hrw|work: and a bit more yet to come (spitz patches) Sep 13 03:19:37 RP: yeah Sep 13 03:20:31 hrw|work: why did I get that problem earlier? Sep 13 03:20:48 AvengerMoJo: ipkg .152 have a problem with provides Sep 13 03:20:56 .153 has a problem with depends Sep 13 03:20:58 hrw|work: thanks Sep 13 03:21:00 oh Sep 13 03:21:02 .... Sep 13 03:21:08 that's nice ... Sep 13 03:21:19 what should I do next :) Sep 13 03:22:43 AvengerMoJo: do what you want Sep 13 03:22:55 :) Sep 13 03:23:10 hrw|work: but you said it get depends problem Sep 13 03:25:06 AvengerMoJo: OE has 'fix' for it Sep 13 03:25:19 AvengerMoJo: fix works or not works - it depends Sep 13 03:29:47 NOTE: package sylpheed-2.1.1-r0: task do_build: completed Sep 13 03:29:57 scary Sep 13 03:30:00 latest development Sep 13 03:30:59 anyway its still patched to gnutls - I'm curios does someone sent those patches up Sep 13 04:06:07 hrw|work: yeah!!! Sep 13 04:06:13 hrw|work: thanks alot ..2.4 done Sep 13 04:06:42 hrw|work: I work a script to pin down the CVSDATE in the bb file .... see if you guys want to use it Sep 13 04:09:01 hi pb Sep 13 04:09:11 AvengerMoJo: what that script has to do? Sep 13 04:09:23 hi hrw|work Sep 13 04:19:36 NOTE: package gpe-image-1.0: completed Sep 13 04:19:38 finally Sep 13 04:20:23 hello again.. Sep 13 04:20:51 hm, are the xfonts-xorg the ones to use with xserver-kdrive? or is there a special font package for it? Sep 13 04:22:01 hrw|work: I pin the CVSDATE to a specific date or remove in the bb file Sep 13 04:22:19 hrw|work: so you don't have to download tar ball like everyday Sep 13 04:22:56 AvengerMoJo: CVSDATE=20050912 in local.conf is not enough? Sep 13 04:26:06 03hrw 07org.oe.dev * r917ed884... 10/packages/opie-pcmciaapplet/ (opie-pcmciaapplet_1.2.1.bb opie-pcmciaapplet_cvs.bb): opie-pcmciaapplet_1.2.1.bb: set PV to 1.2.1 as it is now available in CVS Sep 13 04:27:58 hrw|work: oh I didn't know you can do that :) hehe Sep 13 04:33:37 now you know Sep 13 04:33:54 anything which is in bitbake.conf can be overwritten by local.conf Sep 13 04:40:17 hrw|work: ok :) thanks Sep 13 05:30:01 hrw|work: get some ARM question Sep 13 05:30:20 if we compile for PXA270 .. will 255 able to use it? Sep 13 05:30:33 I mean the kernel Sep 13 05:43:24 AvengerMoJo: no Sep 13 05:48:12 thanks Sep 13 05:50:19 ~seen pigi Sep 13 05:50:24 pigi was last seen on IRC in channel #oe, 4d 14h 35m 3s ago, saying: 'nite all'. Sep 13 05:50:32 4.5d :( Sep 13 05:53:42 ..... Sep 13 05:53:55 hrw|work: what's wrong? Sep 13 05:54:35 ~seen AvengerMoJo Sep 13 05:54:37 avengermojo is currently on #oe (8h 16m 3s). Has said a total of 68 messages. Is idling for 2s Sep 13 05:55:11 * AvengerMoJo is going to test the 2.4.20 kernel Sep 13 05:59:53 03koen 07org.oe.dev * r7e21aa6e... 10/packages/gpe-bluetooth/ (files/bluez-pin-gpe.glade gpe-bluetooth_0.45.bb): packages/gpe-bluetooth/files/bluez-pin-gpe.glade: install gladefile for pin requester Sep 13 06:07:32 :o Sep 13 06:08:05 hmm.. 10 applets changed Sep 13 06:08:29 oh Sep 13 06:09:24 hrw|work: have you ever use and sandisk reader (internal to laptop like C3000) Sep 13 06:10:50 AvengerMoJo: I have Hama x-in-1 usb 2.0 card reader Sep 13 06:16:59 usb work :( Sep 13 06:17:10 but I get one in my laptop that didn't work Sep 13 06:17:22 usb one is pretty simple Sep 13 06:17:46 but I don't know what internal one become that difficult Sep 13 06:17:59 I can't even find any howto or whatever on it either Sep 13 06:18:43 Is it a PCI device then? Sep 13 06:18:52 yes Sep 13 06:18:57 AvengerMoJo: it can be hooked anywhere.. is can be pci device, isa device, umbaumba device.. Sep 13 06:18:57 My laptop has an internal TI based SD reader. No specs available though. :( Sep 13 06:19:07 0000:02:03.1 CardBus bridge: Ricoh Co Ltd RL5c476 II (rev ac) Sep 13 06:19:29 0000:00:1f.3 SMBus: Intel Corp. 82801DB/DBM (ICH4) SMBus Controller (rev 01) Sep 13 06:20:03 when I plug it in the pcmcia get some memory probe Sep 13 06:21:53 [alex@QThemAll root]$ cat /sys/devices/pci0000:00/0000:00:1e.0/0000:02:03.0/0.0/modalias Sep 13 06:21:53 pcmcia:m0000c0000fFEfn00pfn00paD9F522EDpbC3901202pc00000000pd00000000 Sep 13 06:22:24 product info: "RICOH", "Bay1Controller" Sep 13 06:22:24 manfid: 0x0000, 0x0000 Sep 13 06:22:24 function: 254 ((null)) Sep 13 06:22:44 then I'm lost .... don't know where to get into the device Sep 13 06:23:03 or even host what kind of driver into the pcmcia Sep 13 06:23:45 There probably isn't a driver. Sep 13 06:24:01 The TI stuff is the same as the PCMCIA controller too. Sep 13 06:24:29 hum Sep 13 06:26:00 Noodles: any idea where I can look into Sep 13 06:26:14 Datasheets from RICOH? Sep 13 06:28:20 hum ... Sep 13 06:34:03 I think I know what I should do ... ide-cs may be? Sep 13 06:44:04 zecke: ping Sep 13 06:44:09 mickeyl: ping Sep 13 06:44:10 pong Sep 13 06:44:22 zecke: opie.bbclass will need some love I think Sep 13 06:44:44 zecke: now I have 10 applets to push with small change to their postinst() - 'do not use pidof on host' Sep 13 06:44:46 hrw|work: in what way Sep 13 06:44:53 hrw|work: ah Sep 13 06:45:16 03nail 07org.oe.dev * r4c6d0f0d... 10/packages/syslog-ng/syslog-ng_1.6.8.bb: syslog-ng: Add flex dependency Sep 13 06:45:32 zecke: let opie.bbclass generate postinst()/postrm() for applets, inputmethods Sep 13 06:46:14 zecke: http://home.hrw.one.pl/oe/diffs/0913_applets.diff Sep 13 06:46:33 can you comment? Sep 13 06:54:01 :( 0.0: GetNextTuple: No more items Sep 13 06:54:32 hrw|work: leave PR alone ;) Sep 13 06:55:02 zecke: bumped PR will force rebuild Sep 13 06:56:59 * hrw|work want ipkg which will handle provides: and depends: in good way... Sep 13 06:58:09 my opie-image fails because of opie-mediaplayer2 problem with skin Sep 13 06:59:26 zecke: http://www.rpsys.net/openzaurus/temp/add_git.patch - It seems to mostly work ;-) Sep 13 06:59:35 http://pastebin.ca/22808 is not normal for me Sep 13 07:00:26 For a native .bb, if I include the none-native version and then inherit native, does that update the dependencies to native ones? Sep 13 07:00:27 RP: why GITDIR at all? svn fetcher use CVS_DIR Sep 13 07:00:58 hrw|work: CVS and SVN are very similar. GIT it totally different and deserves to be separate Sep 13 07:01:04 ok Sep 13 07:01:15 Thats a stupid distinction Sep 13 07:01:43 I hope that I dont have to install/have git in system to be able to build images for Z Sep 13 07:01:58 hrw|work: I have .bb s to add for it Sep 13 07:02:04 CP|Laptop: IMO, the svn should be separate as well. Sep 13 07:02:34 or, the cvs directory gets renamed to workarea or something which probably better describes what it does Sep 13 07:03:17 RP: I agree - if we distinguish some RCSes, we should distinquish them all Sep 13 07:03:18 RP: its your choice to have CVS_DIR=/somewhere/far/far/away/workspace_for_cvs/ Sep 13 07:03:43 I have ~/devel/sources/cvs_dir/ and ~/devel/sources/dl_dir/ Sep 13 07:03:45 The point being the git fetcher code generates a totally different structure to the cvs/svn code and I don't see the point in mixing and having a total mess Sep 13 07:04:01 This means hrw can set git_dir somewhere else :) Sep 13 07:04:33 RP: what will be fetched from git? Sep 13 07:07:24 hrw|work: mainly kernels I expect Sep 13 07:09:00 hrw|work: I have ideas about the development zaurus kernels using a git:// url in their .bb Sep 13 07:10:32 RP: how much space does git repo take? Sep 13 07:10:43 currently 85MB Sep 13 07:10:56 checked out 45MB (tarred) Sep 13 07:11:32 zecke: so should I submit it as it is? or you really want all those PR stripped? Sep 13 07:12:05 hrw|work: now you have done that, keep it Sep 13 07:16:26 zecke: monotone: committed revision 58b2ce8acc4750a52b3dc8dc9d0aeb8c319d5c46 Sep 13 07:17:45 mickeyl: I will - once at home - mail oz and oe that I've one of the Zaurii Sep 13 07:17:59 03hrw 07org.oe.dev * r58b2ce8a... 10/packages/ (30 files in 10 dirs): Sep 13 07:17:59 OPIE applets: fix postinst() to not be used during image create Sep 13 07:17:59 - usage of "pidof -s qpe" is useless on build machine and make dead processes here Sep 13 07:17:59 - close bug #309 Sep 13 07:17:59 - creation of postinst()/postrm() should be moved to opie.bbclass Sep 13 07:18:42 only 5 files changed in repo - nice Sep 13 07:25:24 wow http://www.sdcard.org/sdio/index.html Sep 13 07:26:50 did they open up the spec yet? Sep 13 07:27:18 Heh Sep 13 07:27:20 not enough Sep 13 07:27:44 http://www.sdcard.org/sdio/Simplified%20SDIO%20Card%20Specification.pdf doesn't seem to be very usefull Sep 13 07:28:18 Most of it is useful - its missing the most important thing - the description of the card registers Sep 13 07:31:03 that sdio digital tv tuner is pretty sick Sep 13 07:31:16 http://tuxmobil.org/Mobile-Guide/mobile-guide-p2c2s1-card-reader.html <--- this really kill me Sep 13 07:31:28 telling my SD basically useless Sep 13 07:31:56 SD is so no useles Sep 13 07:32:58 :( my SD slot is :( Sep 13 07:33:07 who use udev on pda? Sep 13 07:33:23 this look cool though http://www.c-guys.net/products/sd-link11b.html Sep 13 07:34:31 * AvengerMoJo one day will have his only hardware line that only support Linux OS .. Sep 13 07:34:38 hrw|work: I do Sep 13 07:37:53 03rpurdie 07org.oe.dev * ra9d2b39a... 10/packages/git/ (files/Makefile.patch git-native.bb git.bb): git: Add git and git-native packages. git-native may need some addition dependencies (curl-native, openssl-native) - I'm not sure what the policy is on this. Adding them should be straight forward enough though. Sep 13 07:38:28 RP: you could add dummy packages for openssl-native and curl-native Sep 13 07:38:49 koen: I could also add proper ones :) Sep 13 07:38:50 RP: and put a "HAVE THESE ON YOUR HOST" warning inside Sep 13 07:39:47 I've left it open ended for now until we get forced into something one way or another Sep 13 07:39:59 koen: I have udev 068 working but I use it with debian init script - will add it hacked a bit - resulting package should work but who knows.. will push into bugtracker when end Sep 13 07:40:17 hrw|work: ok Sep 13 07:40:28 hrw|work: I heard people had success with the gentoo script too Sep 13 07:41:56 maybe Sep 13 07:42:03 I get one from my working system Sep 13 07:42:38 I'm usually not to fond of gentoo stuff Sep 13 07:46:21 http://oe.handhelds.org/cgi-bin/moin.cgi/ProjectDevices?action=show Sep 13 07:48:13 so this kind of stuff will never work under linux? http://shop.brando.com.hk/sandisk256mbwifisdcard.php Sep 13 07:49:31 AvengerMoJo: the memory part will work under linux today Sep 13 07:49:40 the wifi part is still under NDA Sep 13 07:49:56 So somebody might write a binary driver Sep 13 07:50:18 CP|Laptop: memory... I still can't seem to ready those memory from my laptop Sep 13 07:50:37 CP|Laptop: I'm talking about non-USB type Sep 13 07:51:03 AvengerMoJo: well, I can't speak to your laptop personally, but you can read SD memory on a handful of different products Sep 13 07:51:33 Just because your laptop doesn't support it doesn't mean Linux doesn't in general Sep 13 07:52:01 And thats straight SD I'm talking about - if you go down the usb-storage route, then many different SD readers are supported with LInux Sep 13 07:52:22 CP|Laptop: right I still don't understand why the USB reader able to read it ...but internal reader can't Sep 13 07:53:17 WEll, the USB reader has an ASIC in it that translates the SD protocol Sep 13 07:53:29 and of course the usb part is open .. and SD is not..... Sep 13 07:53:40 SD is open Sep 13 07:53:40 oh ... Sep 13 07:53:47 SD is ? Sep 13 07:54:07 The simplified SD spec tells you everything you need to know about SD host controllers Sep 13 07:54:23 The code is even going to be in linux 2.6.14 Sep 13 07:54:55 ... Sep 13 07:55:16 Its 2.6.14-rc1, as we speak Sep 13 07:55:18 why all I havn't been reading telling me the other way ... Sep 13 07:55:21 really Sep 13 07:55:36 pfffff... http://gentooforzaurus.opensistemas.com/modules.php?op=modload&name=phpWiki&file=index has been hacked... i hate those stupid guys who do this... Sep 13 07:55:46 AvengerMoJo: The problem is your controller has no open documenation so nobody can write a driver for it Sep 13 07:55:51 AvengerMoJo: because most of those people aren't on the cutting edge Sep 13 07:56:06 Unless you hang out in hh.o circles, you would have never heard of it Sep 13 07:56:20 alan|home: running gentoo on a zaurus is useless anyway Sep 13 07:56:24 RP: right .... but I keep thinking the SD reader is like USB ... some kind of standard Sep 13 07:56:42 AvengerMoJo: SD is a standard, but the host controller interface isn't Sep 13 07:56:42 koen: hehe ^^ Sep 13 07:56:55 AvengerMoJo: Its not. SD controllers are all different. Some we have specs for like the PXA on the Zaurus. Others, we don't... Sep 13 07:57:24 RP: I see Sep 13 07:58:13 CP|Laptop: 2.6.14 is out? Sep 13 07:58:31 AvengerMoJo: no, -rc1 Sep 13 07:58:38 koen: if you want to see something really useless for zaurus, have a look at QuantumSTEP... Sep 13 07:58:43 ok :) Sep 13 07:59:38 CP|Laptop: you try it already? Sep 13 07:59:40 koen: http://bugs.openembedded.org/show_bug.cgi?id=325 Sep 13 08:00:51 AvengerMoJo: Nope - most of my stuff lags behind head Sep 13 08:00:58 bleh Sep 13 08:01:17 hi kergoth Sep 13 08:02:05 morning kergoth Sep 13 08:02:13 hey kergoth Sep 13 08:02:31 koen: I hope that udev 068 will work for you too Sep 13 08:24:01 also it is funny... the day before I was build image ... and in deploy it generate bz2 image Sep 13 08:24:25 today ... it generate tar.gz which is better .... did something get updated? Sep 13 08:24:45 AvengerMoJo: Read the oe-commits list. Yes, something changed... Sep 13 08:25:39 heheh Sep 13 08:25:42 and so it begins Sep 13 08:26:11 CP|Laptop: I haven't actually changed "tar" yet either... Sep 13 08:26:24 This was just for the spitz/borzoi images... Sep 13 08:26:49 hmm, i should finish fixing tailor Sep 13 08:27:42 mickeyl: when you will be online and have some spare time - look into bug #307 for updated meta-opie.bb Sep 13 08:29:02 RP: where is the list? Sep 13 08:34:11 oe@handhelds.org and oe-commits@handhelds.org Sep 13 08:34:39 ugh Sep 13 08:34:49 ~/code/oetailor/wc$ monotone update --branch org.openembedded.dev --revision 7accabfc66965fa53356b781f06244e951957a1e Sep 13 08:34:50 monotone: warning: missing packages/linux/linux-omap1-2.6.12-rc2/nokia770/defconfig Sep 13 08:34:50 monotone: misuse: 1 missing files Sep 13 08:34:50 monotone: misuse: to restore consistency, on each missing file run either Sep 13 08:35:07 that file was added in the revision before the one its trying to apply, yet the file doesnt exist.. Sep 13 08:35:08 hmm Sep 13 08:39:55 hrw|work: my image still get something wrong Sep 13 08:40:09 hrw|work: it miss used the 2.6 kernel module :) Sep 13 08:41:29 god ! lyx is so slooow to compile... Sep 13 08:41:40 alan|home: try giac Sep 13 08:41:48 hrw|work: what is that ? Sep 13 08:43:19 hrw|work: Global Information Assurance Certification ? Sep 13 08:43:38 alan|home: try;) Sep 13 08:44:24 hrw|work: any idea why that happen? Sep 13 08:45:32 AvengerMoJo: sorry - my glassball which I use to predict is in service Sep 13 08:46:52 hrw|work: hehe :P Sep 13 08:46:57 But your sarcasm generator seems to be working well Sep 13 08:47:17 ^^ Sep 13 08:47:25 how to build ipkg repository? Sep 13 08:47:41 AvengerMoJo: rebuild any image or use ipkg-make-index Sep 13 08:47:51 AvengerMoJo: then use tmp/deploy/ipk/ as feed Sep 13 08:48:06 thanks Sep 13 08:48:27 ipkg-make-index.... where is that from? Sep 13 08:48:39 AvengerMoJo: tmp/staging/HOST/bin/ Sep 13 08:48:45 thanks Sep 13 08:51:59 np Sep 13 08:55:15 ~emulate gir Sep 13 08:55:20 i need tacos! i need tacos or i'll explode. that happens to me sometimes. Sep 13 08:56:42 loool Sep 13 09:02:59 03rpurdie 07org.oe.dev * rd0807215... 10/packages/zaurus-updater/spitz/updater.sh: zaurus-updater: Fix some paths. Bad RP. Thanks to gridzero for pointing it out. Also remove code that looks like it would wipe the memory card upon failre?\! Sep 13 09:06:42 Good evening. Sep 13 09:08:40 hi sirfred Sep 13 09:08:47 RP: Hi Sep 13 09:11:44 c Sep 13 09:31:12 hello all Sep 13 09:31:15 For reference, the corgi backlight sysfs bug is not present in 2.6.14-rc1 so its almost certainly a -mm problem Sep 13 09:31:43 and rmk's i2c driver now in mainline is broken - won't even compile Sep 13 09:32:03 I propose renaming conf/machines/h3900.conf to handhelds-pxa.conf. Thoughts, pros, cons? Sep 13 09:32:09 ...2.6 is 30% faster then 2.4 in system branchmark test Sep 13 09:32:24 at my C3000 anyway Sep 13 09:37:19 another question about h3900.conf: I see it contains KERNEL ?= "kernel24", but I don't see this anywhere else in classes or conf. Should this really be KERNEL_VERSION? Sep 13 09:37:51 any package provide gcc for zaurus? Sep 13 09:38:07 yes, 'gcc' :P Sep 13 09:38:09 mreimer: unfortunately the ipaq and zaurus configurations suffer a bit from divergence Sep 13 09:38:25 mickeyl: ok, so I should change it to KERNEL_VERSION? Sep 13 09:38:45 mickeyl: btw, thanks for committing those bb changes, and for your suggestions (very helpful) Sep 13 09:38:59 mreimer: i don't know if that works. some ipaq handhelds are using KERNEL as an override Sep 13 09:39:32 s/handhelds/configurations/ Sep 13 09:40:44 mickeyl: do you mean that for these configurations KERNEL overrides KERNEL_VERSION? Sep 13 09:41:19 mreimer: KERNEL_VERSION is a zaurusism atm. although I'd like it to be used in the other configurations as well Sep 13 09:41:25 kergoth: ...which package bb? I'm sorry I can't figure out .. too many Sep 13 09:41:37 .. Sep 13 09:41:38 KERNEL being added as OVERRIDE means you can do statements like FOO_kernel24 Sep 13 09:41:42 and FOO_kernel26 Sep 13 09:41:47 AvengerMoJo: gcc. Sep 13 09:41:49 or have directories kernel26/someconf.foo Sep 13 09:41:56 not gcc-cross, not gcc-cross-initial, but _gcc_. thats it. Sep 13 09:42:04 heh Sep 13 09:42:13 mickeyl: ok. Just wanted to be sure it wasn't a typo. I'll leave it for now. Sep 13 09:42:15 kergoth: but my deploy doesn't seem to get one ..build Sep 13 09:42:20 ? Sep 13 09:42:23 you did a bitbake gcc? Sep 13 09:42:49 mreimer: after the next set of releases, we should do a palaver on the mailing list to restructure the device configurations files and the bootstrap meta packages. Sep 13 09:43:01 mreimer: there has been a big of wild growth Sep 13 09:43:14 kergoth: I'm trying bitbake -b now .. just trying to figure out which one :P Sep 13 09:43:15 mickeyl: yes, that's good :-) Sep 13 09:43:22 AvengerMoJo: ? Sep 13 09:43:39 kergoth: this look good kergoth thanks gcc_3.4.4.bb Sep 13 09:43:39 you should never use bitbake -b. Sep 13 09:44:17 kergoth: even just building one package? Sep 13 09:44:28 -b bypasses _all_ dependencies Sep 13 09:44:38 unless you can guarantee that everything gcc needs is built, you shouldnt be using -b Sep 13 09:44:52 kergoth: I see Sep 13 09:44:58 this is all documented. Sep 13 09:45:12 oh ..... in bitbake menu? Sep 13 09:45:34 i dont know what "bitbake menu" is. its documented in the bitbake users manual. Sep 13 09:45:51 right manual .. Sep 13 09:45:52 :P Sep 13 09:46:14 So I'm told h3900.conf is for PXA-based devices running the hh.o kernel. Right now I'm trying to clarify what is iPAQ h3900-specific and what is applicable to the other PXA-based devices like h2200. Here's one: IPKG_ARCHS includes 'h3900'. Is that intended to refer to only the h3900 model, excluding h5400 etc., or is it intended to refer to the whole class of PXA-based devices running hh.o kernels? Sep 13 09:49:32 kergoth: I don't really recall I have seen that ..... :( sorry ... I just recall building a single package with -c clean -b bbfile Sep 13 09:50:13 kergoth: and it is much faster then not using -b hehe :P that's why Sep 13 09:50:20 kergoth: I will look into that ... Sep 13 09:51:01 mreimer: it has been used to identify the whole series, i.e. also the h5400 Sep 13 09:51:08 mickeyl: I did think we should covert the zaurus files to use KERNEL_VERSION as an override and convert the ipaq files to use KERNEL_VERSION. Would you agree? Sep 13 09:51:12 mreimer: perhaps renaming it to ipaq-pxa25x would be appropriate Sep 13 09:51:47 mickeyl: yes, I'm very confused about when 'h3900' refers to the whole series, and when just the particular model Sep 13 09:51:59 mickeyl: do you mean h3900.conf should be renamed? Sep 13 09:52:14 RP: I'm not sure if i like adding KERNEL as override. Especially in the light of us adding clamshell as an override after the next set of releases Sep 13 09:52:23 mreimer: yes, i think this would be more clear. Sep 13 09:52:37 mickeyl: I just mean a local override in the files we need it Sep 13 09:52:39 mickeyl: I'll do that Sep 13 09:52:48 hi Sep 13 09:53:16 hi hrw Sep 13 09:53:22 RP: ah right. yes, i would agree to that then. Sep 13 09:53:44 re Sep 13 09:53:49 mickeyl: It means we can lose all the -2.4 and -2.6 files and merge into a more logical file structure Sep 13 09:54:14 mickeyl: I presume you'd want to do this after 3.5.4? Sep 13 09:54:18 what did I lose? Sep 13 09:55:11 mreimer: I suspect the h3900 is used for package architecture purposes. You can probably safely change it all to ipaq-pxa25x Sep 13 09:56:41 RP: we can do that as soon as we have branched Sep 13 09:57:02 RP: i want the next set of releases out asap, since we have some interesting and necessary refactoring to do Sep 13 09:57:38 wb hrw Sep 13 09:58:04 mickeyl: Ok. I'm happy to do that work as and when we branch Sep 13 09:58:35 excellent Sep 13 09:58:50 what change you both plan? Sep 13 09:59:11 hrw: Using KERNEL_VERSION as a local OVERRIDE in the machine.conf files Sep 13 09:59:21 aha Sep 13 09:59:24 It will clean up the existing structure a lot Sep 13 09:59:32 thx Sep 13 10:04:04 mickeyl: do you have an overview what is missing for tosa? Sep 13 10:05:04 zecke_: http://www.do13.in-berlin.de/openzaurus/ Sep 13 10:05:15 http://www.do13.in-berlin.de/openzaurus/ Sep 13 10:05:22 mreimer: yeah, "h3900" means all the devices in the h3900 family. OE does not draw any distinction between the h3900 and the h5400. Sep 13 10:05:23 aargh, too slow again Sep 13 10:05:53 h3900 is now 'any pxa25x pda in hh.org kernel' Sep 13 10:06:14 RP: found already, looks like most stuff is done Sep 13 10:06:24 RP: and OE integration of tosa is 100% done? Sep 13 10:06:35 koen: h3900, h5400, h2200, mypal, axim etc? Sep 13 10:06:43 koen, pb: does it always mean that, wherever I see it in h3900.conf? Sep 13 10:06:45 koen: I get my new ibook soon :) Sep 13 10:06:47 very confusing Sep 13 10:06:48 zecke: i think most important would be charging now Sep 13 10:06:54 zecke_: touchscreen is the critical one and charging Sep 13 10:07:01 zecke: if possible, reusing stuff from Richard Sep 13 10:07:04 hrw: yes Sep 13 10:07:17 zecke_: OE intergration should be ok as most of the other 2.6 Zaurii work Sep 13 10:07:21 mickeyl: but he is on crack ;) Sep 13 10:07:26 RP: sorry Sep 13 10:07:28 mreimer: always Sep 13 10:07:28 heh Sep 13 10:07:39 zecke_: ;-) Sep 13 10:07:50 ~lart ipkg Sep 13 10:07:50 * ibot whips out his power stapler and staples ipkg's foot to the floor Sep 13 10:07:57 mickeyl: I've started a wiki site for the tosa traveling Sep 13 10:08:01 excellent Sep 13 10:08:09 mickeyl: I'm now heading torwards Al... Sep 13 10:08:14 mreimer: like I say, oe does not know any difference between the specific models in that family Sep 13 10:08:24 pb_: so H3900_MODULES_kernel26 should include a list of all the modules that all of the various flavors of pxa 25x-based models? Sep 13 10:08:28 mickeyl: http://bugs.openembedded.org/show_bug.cgi?id=326 Sep 13 10:09:04 pb_: h3900 is an unfortunate name, though I imagine there are historical reasons for it Sep 13 10:09:32 koen: very good. I'll add a bug based on our discussions re. bootstrap reorganization asap Sep 13 10:09:47 mickeyl: it would be nice if KERNELVERSION is a global override so we can use it in the imagename Sep 13 10:10:27 and if it possible then include PR in 'uname -a' output? Sep 13 10:10:46 koen: It doesn't need to be a global override - doing it localling the the files that need it is fine Sep 13 10:11:16 hrw: if you want, go ahead and do that Sep 13 10:11:32 hrw: just change one line in all the zaurus kernel bb Sep 13 10:11:52 mickeyl: it does not break sdmmc module support? Sep 13 10:12:01 hrw: no Sep 13 10:12:08 change EXTRA_OEMAKE = "OPENZAURUS_RELEASE=-${DISTRO_VERSION}" Sep 13 10:12:10 nice Sep 13 10:12:12 hrw: You're only changing 2.6? Sep 13 10:12:13 okay, I'm leaving now... any questions left? Sep 13 10:12:22 that goes only for the 2.4 kernels Sep 13 10:12:28 for the 2.6 kernels we rather use LOCALVERSION Sep 13 10:12:28 RP: 2.6 and 2.4 would be best Sep 13 10:12:56 hrw: True Sep 13 10:12:57 mreimer: right Sep 13 10:13:06 RP: 2.6.11-r16 was in upgrades, then -r19, -r20 on your website so it was hard to get info from user 'which kernel you have?' Sep 13 10:13:42 hrw: Agreed. Sep 13 10:13:56 pb_: thanks. one more question: I see KERNEL ?= "kernel24" in h3900.conf. Does that mean I should set KERNEL = "kernel26" in my build/conf/local.conf, in addition to KERNEL_VERSION = "2.6"? Sep 13 10:14:10 only KERNEL = "kernel26" Sep 13 10:14:19 KERNEL_VERSION is a zaurusism Sep 13 10:14:20 ~lart opie-mediaplayer2 dependency chain Sep 13 10:14:20 * ibot throws a AN/M-8 smoke grenade at opie-mediaplayer2 dependency chain Sep 13 10:14:22 mreimer: yeah, or even instead of. afaik, setting KERNEL_VERSION will not have any effet. Sep 13 10:14:33 nor, indeed, any effect Sep 13 10:14:48 mreimer: Or convert the ipaqs to use KERNEL_VERSION. Why they didn't in the first place is beyond me Sep 13 10:15:03 RP: i'm not sure which came first Sep 13 10:15:10 RP: what changes need to be done? Sep 13 10:15:10 koen: I am ;-) Sep 13 10:15:27 is there agreement that KERNEL should be dropped in favor of KERNEL_VERSION? Sep 13 10:15:47 RP: using overrides is nicer than spreading a zillion extra files in conf/machine IMO Sep 13 10:16:06 mreimer: This really needs the agreement of the ipaq people. KERNEL_VERSION is documented in local.conf.sample and is a more obvious variable Sep 13 10:16:19 yeah, I used KERNEL_VERSION because it was there in local.conf Sep 13 10:16:22 koen: There is nothing stopping us using KERNEL_VERSION as a local override Sep 13 10:16:46 koen: I've agreed with mickeyl that we're switching the zaurus machine files to that format Sep 13 10:16:51 koen: but after 3.5.4 Sep 13 10:17:11 RP: http://bugs.openembedded.org/show_bug.cgi?id=326 Sep 13 10:17:13 bbl, sports time Sep 13 10:17:17 RP: why not now? Sep 13 10:17:18 cu mickey|sports Sep 13 10:17:50 koen: I don't think mickey|sports wants to break things Sep 13 10:18:22 I'll do it whenever - doesn't make any difference to me Sep 13 10:20:00 koen: Are you ok with a standardisation to KERNEL_VERSION? Sep 13 10:20:09 koen: If so, when do you want to do it? Sep 13 10:20:19 I'm OK Sep 13 10:20:33 but I thought it was kind of strange to touch one part, but not the other Sep 13 10:21:30 Making life potentially difficult for the dsitro maintainer just before release is something you need to ask the distro maintainer about... Sep 13 10:21:42 There are a few changes pending the next releases Sep 13 10:22:13 As mickey|sports said, we make them as soon as the branch is made Sep 13 10:22:46 what changes would I need to do to convert h3900.conf to KERNEL_VERSION? Sep 13 10:23:56 sed -i s/KERNEL/KERNEL_VERSION/g h3*.conf Sep 13 10:24:08 mreimer: change the OVERRIDES line to reference KERNEL_VERSION and change the _kernel26 and kernel24 references to _2.6 and _2.4 Sep 13 10:24:27 RP: that's easy Sep 13 10:24:46 mreimer: Indeed. The Zaurus changes are a lot nastier... Sep 13 10:25:34 I'm not especially comfortable with having global "_2.6" and "_2.4" overrides. That seems like a step in the wrong direction. Sep 13 10:26:59 namespace pollution, anyone? Sep 13 10:27:22 kergoth: lovely, give me some Sep 13 10:27:30 hehe Sep 13 10:27:31 re Sep 13 10:27:35 hey hrw Sep 13 10:27:47 pb_: KERNEL_VERSION = "2.6" makes most sense in the user visible config file. KERNEL="kernel24" is far less logical. I agree the namespace isn't ideal though Sep 13 10:27:47 kergoth: right, this would be the Exxon Valdez of namespaces Sep 13 10:28:36 RP: I'm not bothered how you dress it up for the user. You can have whatever you like in the config files, so long as the contents of OVERRIDES is kept under control. Sep 13 10:29:13 pb_: How about linux-2.6? Sep 13 10:29:27 Its entirely possible we may have none linux kernels some day Sep 13 10:30:20 heh.. to get my opie-image done I have to hack ipkg again ;( Sep 13 10:30:26 ~lart ipkg Sep 13 10:30:26 * ibot steals ipkg's mojo Sep 13 10:31:10 4 packages provide umbaumba but if tymtym depends on umbaumba then ipkg find all providers and does not select any Sep 13 10:31:46 I've also been seeing weird ipkg dependency bugs recently :-( Sep 13 10:31:59 That would be OK, I guess. I feel a little bit queasy about the prospect of someone deciding to create linux-2.6_2.6.13.bb and losing, but I gues one can't legislate for every eventuality. Sep 13 10:32:08 RP: there were problems with 152; make sure you're using 153 Sep 13 10:32:21 heh Sep 13 10:32:22 mreimer: 153 has problem too Sep 13 10:32:55 hrw: the specific behaviour you mentioned is by design; ipkg has always done that. Sep 13 10:33:15 So is the consensus KERNEL_VERSION = "linux-2.6" ? Sep 13 10:33:43 pb_: so there is no possibility to install 'tymtym' then? I consider it a bug rather then feature Sep 13 10:34:02 mreimer: Except we can't do that without converting the Zaurus. I suspect we should just go ahead and do that... Sep 13 10:34:02 hrw: if "tymtym" is a virtual, yes. You must install one of the concrete providers. Sep 13 10:34:30 hrw: ipkg has no idea which one you want. if it was to install one without user intervention, _that_ would be a bug. Sep 13 10:34:30 pb_: ok - so the bug is in recommends: now Sep 13 10:34:41 I need to go - back later Sep 13 10:34:43 right, what kergoth said. Sep 13 10:34:51 ok Sep 13 10:34:56 ipkg doesn't have any way to choose between four equally attractive (to it) possibilities. Sep 13 10:35:05 it would be unacceptable for it to just pick one at random. Sep 13 10:35:20 is there any way to add priorities? Sep 13 10:35:44 no, but that is an interesting idea Sep 13 10:36:18 i'd rather just have an easier way for the user to indicate his preferences Sep 13 10:36:28 "look, i like X better than Y. remember that." Sep 13 10:37:06 like an enhanced "ipkg flag prefer"? that might be cute. Sep 13 10:37:11 yeah, exactly Sep 13 10:37:35 i'm almost thinking something vaguely like dpkg --get-selections, really... does ipkg have that yet? :) Sep 13 10:37:44 erm, --set-selections, i mean Sep 13 10:37:53 * kergoth gets something caffinated Sep 13 10:38:06 has anyone got the prism2/3 wireless CF card working with aircrack(aireplay) on the OZ?? Sep 13 10:38:18 abm_y4k: read the topic of this channel. Sep 13 10:38:19 abm_y4k: topic, 3 item Sep 13 10:38:26 oh Sep 13 10:38:33 * pb_ go home now Sep 13 10:38:34 later all Sep 13 10:38:37 later Sep 13 10:38:49 sorry Sep 13 10:38:51 I'm recently unable to get my prism2 card working reliable Sep 13 10:41:40 ARGH.... again usb hub died... I have to use wifi ;( Sep 13 10:44:59 do we have telnetd in repo? Sep 13 10:45:13 ew, telnet Sep 13 10:45:27 i think busybox has a telnetd.. :) dunno beyond that Sep 13 10:45:59 logging into my Z can be nightmare Sep 13 10:46:50 is our ssh still hokey? Sep 13 10:46:56 last time i tried it, doing an ssh localhost segfaulted :P Sep 13 10:47:46 works for me Sep 13 10:47:58 that was an oe build, from about a week ago Sep 13 10:48:01 on i386 Sep 13 10:48:48 I tested on my Z Sep 13 10:49:05 maybe its just broken on x86 or something Sep 13 10:49:06 * kergoth shrugs Sep 13 10:49:28 * kergoth does a monotone pull for the first time in days.. Sep 13 10:49:31 * kergoth taps foot Sep 13 10:49:41 "uhci_hcd 0000:00:10.3: host controller halted, very bad!" - I love that... Sep 13 10:49:47 hehe Sep 13 10:50:21 3 (of 5) hubs works Sep 13 10:50:40 but the most used one (on top-front of case) does not Sep 13 10:50:54 ~curse ipkg once more Sep 13 10:50:57 May the fleas of a thousand camels infest your most sensitive regions, ipkg once more ! Sep 13 10:51:35 on Z: 'ipkg install opie-mediaplayer2' fetch all what depend/recommend. but when doing rootfs it does not Sep 13 10:53:14 sometimes I hate ipkg more then usual Sep 13 11:04:12 <_alwin_> hi Sep 13 11:20:42 bitbake -f is supposed to compile the bb even if it is already found in stamp, right? Sep 13 11:23:31 because when coupled with the -b option (to compile a specific .bb without parsing the tree) it is not working that way Sep 13 11:23:50 i am having to manually remove the corresponding stamps directory to get it to compile... Sep 13 11:36:18 LoDown: -f forces the execution of a task regardless of stamp, however, its important to keep in mind that it _only_ runs that task, compile only, in that case, not configure, etc, and does not re-run the tasks that depend on compile, like install, package, etc. it _forces_ a task to execute, but thats all. Sep 13 11:36:30 (iirc) Sep 13 11:36:45 though if the compile stamp changes itd probably rerun the ones after that.. been too long since i looked at hte code Sep 13 11:38:13 hmmm...thanks kergoth`food Sep 13 11:38:42 i guess for what I'm trying to do, it will be easier to just remove the whole stamp directory... Sep 13 11:39:10 are you modifying the code in its ${S} and trying to recompile/repackage with those modifications? Sep 13 11:39:16 hmm.. looks like I have to test prism2 firmwares to get good working one Sep 13 11:40:11 kergoth`food, yes Sep 13 12:15:46 hello again.. Sep 13 12:17:04 does anyone of you maybe know how to set pxafb display parameters in 2.6 kernel? Sep 13 12:18:02 Delta_: fbset. or probably more what you want - the panel definition in the machine support code Sep 13 12:20:40 is that in include/asm-arm/arch-pxa/machine.h ? Sep 13 12:20:48 I know it was there in 2.4 kernel.. Sep 13 12:21:29 but I defined like that, in 2.6 doesn't work, I can give the parameters as a kernel boot parameter though.. Sep 13 12:23:25 Why is h3900 mentioned twice in /etc/ipkg/arch.conf? I have "arch h3900 36" and "arch h3900 41" Sep 13 12:23:55 $MACHINE gets passed to IPKG_ARCHS by defaults Sep 13 12:24:03 it seems h3900.conf is adding it twice Sep 13 12:24:12 default* Sep 13 12:24:36 IPKG_ARCHS = "all arm armv4 armv4t armv5e armv5te ipaqpxa h3900" Sep 13 12:25:41 jesus, my monotone pull is still going Sep 13 12:25:42 * kergoth taps foot Sep 13 12:48:58 Ciao all Sep 13 12:51:00 hey Pigi Sep 13 12:51:11 hey koen ! Sep 13 12:51:33 did someone reverted ipkg to .152 in oe until I fix the nasty depends bug ? Sep 13 12:53:05 yes Sep 13 12:53:11 cool. Sep 13 12:53:26 this time the bug is really hard to find :) Sep 13 12:53:32 see http://pastebin.ca/22845 for what was said earlier today Sep 13 12:56:05 koen, so the idea is to revert the provides: enhancement if I have understood correctly .... Sep 13 12:56:17 ~seen pb_? Sep 13 12:56:24 pb_ was last seen on IRC in channel #oe, 2h 17m 49s ago, saying: 'later all'. Sep 13 13:13:26 http://oe.handhelds.org/cgi-bin/moin.cgi/KergothsRandomBits Sep 13 13:13:44 that has a proposed .inc for building packages using locally available sources, for development Sep 13 13:18:13 kergoth: you have been hanging out with the maemo people Sep 13 13:18:21 "aiiiieeee" Sep 13 13:18:24 hehe Sep 13 13:18:32 nah, thats from the kernel Sep 13 13:18:32 :) Sep 13 13:18:37 aiiiiieeee, not syncing .. Sep 13 13:18:52 hey france_ Sep 13 13:19:12 koen: does that look sane though? Sep 13 13:19:16 that way is on a per package basis Sep 13 13:19:19 very explicit Sep 13 13:19:29 i have _these_ sources, i want it to build them Sep 13 13:20:26 koen: hi Sep 13 13:20:34 kergoth: looks pretty nice Sep 13 13:24:26 all this fiddling with linux on the pxa made me start drinking coffee.. Sep 13 13:24:35 that's no good.. Sep 13 13:24:40 hmm, maybe it should be more specific than package name, since the local sources are likely more specific than that, a certain PN-PV at a minimum Sep 13 13:24:59 oh well, thisll do for now Sep 13 13:38:59 re Sep 13 13:39:45 hey hrw Sep 13 13:48:17 argh... I dont know what is fscked.. my Z or my router.. yesterday friend managed to connect via wifi to router so router looks ok. but zaurus cannot connect via wifi to it ;( wep configured correctly. will have to try w/o wep probably Sep 13 13:48:43 tv time again.. werbungs ended Sep 13 13:55:50 03ph5 07org.oe.dev * r011329bc... 10/packages/dbus/dbus-native_0.50.bb: dbus: add a dbus-native package for dbus 0.50 Sep 13 14:03:55 hrw...what kinda wireless chip does your Z have? Sep 13 14:05:44 cu Sep 13 14:06:30 LoDown: Z's don't have wireless or bluetooth Sep 13 14:06:47 well...what is hrw using to connect to his router? Sep 13 14:06:57 magic :) Sep 13 14:07:27 I'm guessing a prism2 cf card if the magic runs out Sep 13 14:08:23 LoDown: dlick dcf-660w - prism2 based Sep 13 14:16:28 hmm.. works without wep Sep 13 14:16:40 what would be the nicest way to put the kernel PV into IMAGE_NAME ? Sep 13 14:17:08 look how it is in linux-openzaurus.bb Sep 13 14:17:21 I said nice, not hack Sep 13 14:17:35 ah Sep 13 14:18:27 03ph5 07org.oe.dev * rfc456cb4... 10/packages/module-init-tools/module-init-tools_3.2-pre7.bb: Sep 13 14:18:27 module-init-tools-3.2-pre7: install lsmod symlinks to both /bin and /sbin Sep 13 14:18:27 - fixes Bug #232. additionally changed update-alternatives priority Sep 13 14:18:27 to 60 to prefer module-init-tools binaries over busybox. Sep 13 14:18:56 hrw: and that only names the zImage Sep 13 14:18:58 pH5: look lower in buglist - m-i-t 3.2-pre9 Sep 13 14:19:07 koen: yep - I misunderstood you Sep 13 14:19:32 hrw: I saw this. Should I commit 3.2-pre9? Sep 13 14:19:39 I'm still wondering what the kernel-image packages are good for. Do some machines boot the kernel from out of the rootfs? Sep 13 14:19:47 yes Sep 13 14:19:49 pH5: rather test Sep 13 14:19:57 pH5: most ipaqs Sep 13 14:20:02 oh Sep 13 14:20:09 pH5: zaurus machines have separate partition for kernel ;( Sep 13 14:20:16 hrw: hx4700 too Sep 13 14:20:24 pH5: the hx4700 doesn't do that to support a nonGPL bootloader :( Sep 13 14:20:54 How could we get rid of that kernel-image in the rootfs without disturbing anyone else? Sep 13 14:21:17 leave the kernel out of RDEPENDS Sep 13 14:21:34 pH5: you want to remove that empty package just because its empty? Sep 13 14:21:45 koen: but every kernel module depends on the kernel image... Sep 13 14:21:46 both the hx4700 and hx2750 don't have the kernel inside the image Sep 13 14:22:13 pH5: all the zaurus machines seem to do it Sep 13 14:22:13 Actually, I have the kernel inside the image. It just sits there without use. Sep 13 14:22:29 so i'll have look at the zaurus build Sep 13 14:22:37 c7x0.conf Sep 13 14:22:41 koen: thanks Sep 13 14:23:00 which includes a zillion files... Sep 13 14:23:57 ARGHHH Sep 13 14:24:29 usbnet uber alles Sep 13 14:25:22 bt suxx (pairing problems, sdp not working, slow card), wifi suxx (cannot connect with wep, unstable connection), usbnet works always but is on cable..... Sep 13 14:25:35 bu Sep 13 14:26:15 that's why intel has wireless usb Sep 13 14:27:13 usb and bt are both ugly technologies Sep 13 14:27:39 bt is too ugly Sep 13 14:29:04 * pH5 loves his bt mouse, no matter what Sep 13 14:30:16 pH5: my mx900 is pretty nice Sep 13 14:30:29 except when I use wifi to the max Sep 13 14:31:51 koen: heh, same here. Sep 13 14:50:07 whee, the local_s.inc works great :) Sep 13 14:50:21 now modified it to set SRC_URI as well, for tracking purposese Sep 13 14:50:25 purposes, even Sep 13 14:56:29 ok. time to sleep Sep 13 14:57:01 'night hrw Sep 13 14:57:05 in next few days router will get newer openwrt release and I'll try again to pair it with zaurus Sep 13 14:58:01 kergoth: any additions to http://handhelds.org/hypermail/oe/33/3366.html ? Sep 13 14:58:58 * pH5 already would be happy with the first point alone Sep 13 15:01:01 it would probably involve 10 lines of code to base.bbclass, testing and modifying almost every .bb Sep 13 15:01:13 looks like a good list Sep 13 15:01:20 split staging has been on the todo as long as oe has existed Sep 13 15:01:24 heh Sep 13 15:02:57 the initial thread was annoying Sep 13 15:03:20 especially when someone said that bitbake uses too much memory... Sep 13 15:03:31 * kergoth hasnt really been paying attention to his mailing lists since his mailserver went poof, and had to switch to gmail Sep 13 15:03:38 hah Sep 13 15:04:20 some people at trolltech need to svn up before mailing nonsense.... Sep 13 15:04:48 * kergoth nods Sep 13 15:05:16 unsurprisingly the flamy nature seems to have dampened the motivation of the CELF people Sep 13 15:05:21 :( Sep 13 15:05:33 or they are just being slow as any standards organization Sep 13 15:05:42 hehe Sep 13 15:05:45 wouldnt surprise me Sep 13 15:05:52 though they're a bit better than most Sep 13 15:07:23 yeah, the ISO people haven't waved funding in front of us Sep 13 15:07:29 nite all Sep 13 15:07:39 OpenEmbedded, now ISO 9001 certified Sep 13 15:07:49 lol Sep 13 15:07:52 *cough* Sep 13 15:07:53 scary Sep 13 15:08:26 iso is all about documenting your processes, and following your documented processes. we document jack shit :P Sep 13 15:08:44 did you catch the quote of the philips people? Sep 13 15:08:55 nope Sep 13 15:09:11 "montavista tools suck, that's why we use OE now" Sep 13 15:09:19 haha Sep 13 15:09:21 so true Sep 13 15:09:22 thats great Sep 13 15:09:43 too bad the project itself is under NDA Sep 13 15:09:52 yeah. as usual Sep 13 15:10:53 which reminds me that I need to prod my uni buddies to sponsor a trip to philips Sep 13 15:13:09 lamikr: check http://familiar.handhelds.org/snapshots/v0.8.3/20050913/ in a few minutes Sep 13 15:14:19 koen: So you are building an image h6300 to there? Sep 13 15:14:51 http://familiar.handhelds.org/snapshots/v0.8.3/20050913/gpe-image-h6300-20050913202440.rootfs.ext2.bz2 Sep 13 15:15:03 koen: Nice :-) Sep 13 15:15:11 as well as jffs2 and tar.bz2 versions Sep 13 15:16:27 * ljp thinks some people need to take criticism better Sep 13 15:16:29 koen: We actually need only the tar.bz2 + work/linux-h6300/arch/arm/boot/Image as we can not yet flash. Sep 13 15:17:50 koen: The Image is required for the wince readable partition. Sep 13 15:18:14 zImage uploaded Sep 13 15:18:37 ljp: if you're referring to comments about bitbake's memory usage, it helps if the criticisms are actually grounded in current facts, not speculation. Sep 13 15:18:47 koen: I can test it tomorrow! Sep 13 15:19:21 lamikr: and if they work, please spread the word to the mailing list :) Sep 13 15:19:35 koen: I will :-) Sep 13 15:20:16 I still have some issues with opie-image and multimachine builds, so you only have a bootstrap and gpe image to test Sep 13 15:21:33 right Sep 13 15:21:41 koen: what issues? Sep 13 15:21:51 koen: I tested yesterday an udev based Image and had one problem. I needed to modify /etc/init.d/udev to start also the udevd as nothing started udevd automatically. Do you know is this a all platform bug? (It only started udev not udevd) Sep 13 15:22:29 lamikr: known udev bugs in OE Sep 13 15:22:49 kergoth: .bb that spawn .ipk that not have the same arch Sep 13 15:23:24 kergoth: sysvinit, tslib, base-passwd and opie-button-settings Sep 13 15:24:01 the build failed because I cleaned the _cvs version of opie-button-settings instead of the brand new 1.2.1 Sep 13 15:24:48 kergoth: see http://pastebin.ca/22862 for the script I use Sep 13 15:28:20 koen: Ok. For me the fix was to add "start-stop-daemon -S -b -n /sbin/udevd" to start method and "start-stop-daemon -K -x /sbin/udevd" to stop method in /etc/init.d/udev. Sep 13 15:29:00 lamikr: I used one of the init scripts in the extras directory. You can also just run udevsend to start udevd Sep 13 15:29:45 'night all Sep 13 15:30:07 koen: Good night. Sep 13 15:30:43 mreimer: From where do you start this "udevsend"? Sep 13 15:30:53 night koen Sep 13 15:31:39 lamikr: I used the extras/start_udev script, which I installed as /etc/init.d/udev Sep 13 15:32:02 I did have to add the call to udevsend to get udevd going Sep 13 15:32:09 but I don't know if that's the right way or not Sep 13 15:33:12 mreimer: I do not know either. Well both of the methods seemed to work. :-) Sep 13 15:33:23 I need also go to sleep now. Goodnight. Sep 13 15:46:32 kergoth: we no longer need linear-h2200.c in tslib. Should I send you a patch, or enter a bug into OE's bug db, or ... ? Sep 13 15:49:46 mreimer: just pester me periodically until i get off my ass and remove it Sep 13 15:50:01 kergoth: consider yourself pestered :-) Sep 13 15:50:29 kergoth: rm plugins/linear-h2200.c, edit plugins/Makefile.am, edit configure.ac Sep 13 15:50:34 * kergoth nods Sep 13 15:50:59 kergoth: also, we no longer need a h2200-specific ts.conf in OE. Should I send a patch for that too? Sep 13 17:45:02 what's the fix for "No package 'randrproto' found" when building xserver-kdrive cvs? Sep 13 20:27:02 hum Sep 13 22:07:09 my 2.4 root image still get some problem Sep 13 22:07:24 it won't use the 2.4 module but 2.6's Sep 13 22:33:04 will bootsplash put into kernel? Sep 13 22:45:07 hello, can I run a USB wifi adapter on a sl-5500 or does that require host mode usb? Sep 13 22:47:03 that requires host mode USB Sep 13 22:47:54 bummer...I thought I saw some wifi/USB drivers... Sep 13 22:50:02 arent there some gpio pins that are free in the zaurus? Sep 13 22:50:31 I'm sure there are, but don't expect them to be routed anywhere :) Sep 13 22:51:46 joshua_: but doesn't C3000 get host usb? Sep 13 22:51:52 there's got to be a way outa this... Sep 13 22:51:59 I know the SL-6000 has host Sep 13 22:52:39 greyL: I'm not sure .. Sep 13 22:53:44 I thought I just read it on the net...I'm looking to get usb/wifi on my 5500... Sep 13 22:53:54 handhelds.org has atmel-usb modules for kernel 2.4.19-rmk6-pxa1-hh37atmel wlan drivers Sep 13 22:54:05 I guess that must be for the luck host/usb people Sep 13 22:55:20 hehe Sep 13 22:55:54 I see someone get the C-GUY SDIO Wireless also ... don't know that's actually working Sep 13 22:55:59 but I don't think driver is opensource Sep 13 22:57:26 well my prob is I got the RTM-8000 gprs modem on CF, 1gig flash root on SD, that only leaves the USB... Sep 13 22:57:29 for wifi Sep 13 22:57:55 $132 for USB1.1 Host CompactFlash CF adapter...ouch Sep 13 22:58:02 man Sep 13 22:58:13 gprs modem ... Sep 13 22:58:21 can you swap that Sep 13 22:58:30 you don't need to use both at the same time ... do you? Sep 13 22:59:24 no you're right...(sarcasm)...there's no reason to be talking on the phone while surfing the net... Sep 13 22:59:26 :) Sep 13 23:00:08 greyL: :) Sep 13 23:00:26 greyL: what do you use for phone app? Sep 13 23:00:44 I hacked something together in tcl/tk...it's a hack though... Sep 13 23:00:53 before that I was using...guess... Sep 13 23:01:02 ./term /dev/ttyS3 Sep 13 23:01:12 ha ha ha atdt; Sep 13 23:01:18 of if I was lucky Sep 13 23:01:30 ata to answer incoming calls when I saw the ring... Sep 13 23:01:37 oh Sep 13 23:01:40 man ... Sep 13 23:01:52 greyL: but how much is your card? Sep 13 23:02:06 the biggest problem is that when I suspend my Z...the CF gets powered down...ie no incoming calls Sep 13 23:02:09 gprs... I'm thinking using a phone directly as my modem Sep 13 23:02:19 I cant really remember i think it was $250 Sep 13 23:02:25 greyL: right also ... the calling time is bad Sep 13 23:02:35 greyL: how long is your battery going to last? Sep 13 23:02:41 look up audiovox rtm-8000 Sep 13 23:03:05 battery... well there's that russian guy that sells the 8billion mah batteries... Sep 13 23:03:16 (see openzaurus forums) Sep 13 23:04:08 the other reason I need the USB wifi is that I'm setting up a peer to peer neighborhood network and need a portable signal strength meter Sep 13 23:04:32 anyone know anything about usb2go? Sep 13 23:06:10 ok my solution for host usb on a sl-5500...Compact Flash to PCMCIA Adapter Converter $19.95 + NEC USB 2.0 2-Port PCMCIA PC Card CardBus $7.95 Sep 13 23:06:15 anyone know if it'll work? Sep 13 23:06:37 if it's a 32bit card, then no Sep 13 23:06:43 compact flash is only 16 bit Sep 13 23:07:02 if it's a 16bit card then chances are good Sep 13 23:08:26 hum .. Sep 13 23:08:35 hmm hadnt thought of that...thanks... Sep 13 23:09:26 greyL: how long does it last for you right now? Sep 13 23:09:32 greyL: I mean the battery Sep 13 23:10:22 sorry I don't have the answer...I don't use it enough to tell you...I got a Treo to satiate my internet addiction... Sep 13 23:10:48 let me tell you how lame a treo is! Palm OS does not multitask!!!! or at least the apps dont Sep 13 23:11:21 I wanted VNC on my Z...ran gpe...downloaded tightvnc to my Z... ./configure; make ; make install Sep 13 23:11:25 works! Sep 13 23:11:27 heh palm users can't multitask, it's why they buy palm ;) Sep 13 23:11:33 jk Sep 13 23:11:46 ..... Sep 13 23:11:55 palm OS? no stinking way...pay someone for it...and it's not vnc...so there's no linux clients Sep 13 23:12:24 but when I suspend my treo, it still rings when someone calls... Palm succeeds because of polish Sep 13 23:18:04 anomaly_ - is 32bit = cardbus or can 16bit also = cardbus? Sep 13 23:18:27 pretty sure cardbus = 32bit, but you know what some manufacturers are like with buzzwords... Sep 13 23:18:31 best to find out directly Sep 13 23:26:54 are you sure the CF is only 16bits? Sep 13 23:33:59 to the best of my knowledge, yes Sep 14 01:01:26 morning Sep 14 01:02:10 hi hrw|work Sep 14 01:05:03 morning all Sep 14 01:12:17 morning all Sep 14 01:15:59 morning RP Sep 14 01:17:53 hi RP Sep 14 01:25:36 morning Sep 14 01:26:37 hi XorA Sep 14 01:27:00 start check battery ! <---- what does it mean ... Sep 14 01:27:13 I just see that in my dmesg Sep 14 01:31:01 hum .. I can't do mount bind in Sharp .. so I can only use soft link right? Sep 14 01:31:54 AvengerMoJo: you can - you just lack mount command with bind support Sep 14 01:32:24 some sharp roms allow bind Sep 14 01:32:51 oh ... right Sep 14 01:33:11 hrw|work: meaning replacing the busybox mount with a standard mount Sep 14 01:34:09 yes Sep 14 01:35:21 Since when does busybox not support bind mount? Sep 14 01:35:44 good morning all Sep 14 01:36:31 CP|Laptop: does it? Sep 14 01:36:43 Last time I checked Sep 14 01:36:49 CP|Laptop: I don't know what's why I ask Sep 14 01:36:50 :) Sep 14 01:37:03 best method to chack is mount -o bind /blah /bleh and see if it works Sep 14 01:37:06 what will be the flag? I'm sure it is not --bind Sep 14 01:37:11 No, you said you couldn't bind mount Sep 14 01:37:14 oh Sep 14 01:37:20 thats different then asking if you can bind mount Sep 14 01:37:25 -t bind sorry Sep 14 01:39:01 XorA: no, its -o bind Sep 14 01:42:45 CP|Laptop: thanks alot Sep 14 01:44:42 CP|Laptop: -o bind work with busy box Sep 14 01:45:00 so I can put that into fstab also then Sep 14 01:51:07 RP: how big is your C3000 first partition? Sep 14 01:51:43 AvengerMoJo: Its as standard, whatever that is Sep 14 01:52:16 hum 100M Sep 14 01:52:23 thanks Sep 14 01:52:54 RP: do you use other way to put more application? Sep 14 02:00:20 AvengerMoJo: ipkg -d destination install package Sep 14 02:04:07 hrw|work: thanks ... so I need to update my ld library path by hand do I? Sep 14 02:05:31 use "ipkg-link mount destination" command to create symlinks Sep 14 02:05:41 or resize partitions on microdrive Sep 14 02:05:53 or boot from other partition then hda1 Sep 14 02:06:08 at work I have / on /dev/hda5 Sep 14 02:07:27 hrw|work: thanks ...I'm just worry if I resize the hda1 may affect my image update in the future Sep 14 02:07:33 that's why Sep 14 02:07:48 AvengerMoJo: iirc image is unpacked from tarball? Sep 14 02:08:35 anyway I would like to have that problem ;) now my / on zaurus have ~60M total and only half ot it occupied Sep 14 02:09:03 :) Sep 14 02:09:27 03mickeyl 07org.oe.dev * ref08017d... 10/packages/wlan-ng/ (4 files in 2 dirs): wlan-ng: move two patches that are used by more than one .bb to the files directory Sep 14 02:10:09 mickeyl waked Sep 14 02:13:58 RP: last time you too me about a new meta .... opie-meta-all .. something like that .. can I use that? Sep 14 02:14:15 AvengerMoJo: it was me and it is not in repo Sep 14 02:14:20 AvengerMoJo: build meta-opie Sep 14 02:14:36 2.6 kernel running the branchmark a lot faster then 2.4 about 30% Sep 14 02:14:42 hrw|work: oh :) hehe ok Sep 14 02:14:55 AvengerMoJo: now its meta-opie, soon it will be meta-opie + meta-opie-all (I split 1 into 2) Sep 14 02:15:14 hrw|work: how to make a new meta any link for the doc Sep 14 02:15:38 AvengerMoJo: look into any existing and edit a copy Sep 14 02:15:42 hrw|work: I will try that .... what's the different Sep 14 02:16:07 hrw|work: I mean the meta-opie and -all Sep 14 02:17:16 meta-opie have all opie stuff and is needed for new opie-image Sep 14 02:17:26 meta-opie-all have rest of qpe stuff Sep 14 02:17:35 AvengerMoJo: look into bugtracker for more info Sep 14 02:18:01 ok Sep 14 02:18:13 where is the meta file locate? Sep 14 02:18:31 ? Sep 14 02:18:35 get it Sep 14 02:18:37 :O) Sep 14 02:18:49 it is packages/meta/ .... Sep 14 02:18:55 I never touch that before Sep 14 02:18:56 hehe Sep 14 02:19:47 hrw|work: also ... I still find something wired.. after I build 2.6 then back to 2.4 ... Sep 14 02:20:04 some package still using 2.6 version .. include kernel-module :) Sep 14 02:20:25 what do you think may cause this problem Sep 14 02:20:58 AvengerMoJo: thats simple - machine conf want kernel-module-sco which is provided by kernel-module-sco_2.6 and kernel-module-sco_2.4 - 2.6 is newer Sep 14 02:21:46 ....I see Sep 14 02:22:22 hrw|work: so I need to some how make two seperate build for 2.4. and 2.6 if I want to test both kernel from time to time Sep 14 02:22:33 * AvengerMoJo need to buy a new server Sep 14 02:22:47 my laptop is not going to make it :) Sep 14 02:23:34 AvengerMoJo: or build 2.6 and copy/move all modules outside of ipk/ Sep 14 02:23:39 or other way Sep 14 02:24:20 I see .. moving it out of ipk will do that trick? Sep 14 02:24:24 hum .. that's cool Sep 14 02:24:34 hehe thanks Sep 14 02:24:49 not I know what pcmica and all other stuff are mess up Sep 14 02:25:42 anyway there are more things which are kernel related Sep 14 02:25:56 e.g.? Sep 14 02:26:16 my typo is horrible .... Sep 14 02:26:39 now I know why pcmcia and other stuff are mess up Sep 14 02:26:57 * AvengerMoJo will not understand himself either Sep 14 02:27:32 AvengerMoJo: maybe this iis a silly question, but why do you want to use a 2.4 kernel ? Sep 14 02:28:12 well 2.6 didn't get sound Sep 14 02:28:40 2.6 keymap need to fix and fuctionkey is not there yet Sep 14 02:29:00 AvengerMoJo: so fix it Sep 14 02:29:07 and 2.6 pcmcia didn't get a driver for my wireless card Sep 14 02:29:12 hrw|work: yes I'm going to :) Sep 14 02:29:17 hrw|work: don't worry Sep 14 02:29:25 AvengerMoJo: under 2.4 which driver it use? Sep 14 02:29:31 hrw|work: but I like to feel what is there under 2.4 Sep 14 02:29:40 hrw|work: this is what I'm trying to find out Sep 14 02:29:46 hrw|work: it is still myth to me Sep 14 02:30:04 AvengerMoJo: boot 2.4, insert card, do 'lsmod' Sep 14 02:30:05 this guy /tmp/wl_lkm_722_abg.tar.gz Sep 14 02:30:17 AvengerMoJo: what card is it? Sep 14 02:30:22 hrw|work: I need to compile it myself ... it is not in there Sep 14 02:30:39 hrw|work: eagletec CWB1000 Sep 14 02:30:48 AvengerMoJo: cf? usb? Sep 14 02:30:48 heh - http://www.sourcemage.org/ Sep 14 02:30:56 CF Sep 14 02:30:56 That sumbitch uses a bash based buildsystem Sep 14 02:31:03 re Sep 14 02:31:08 And you thought python was a pita.. :) Sep 14 02:31:20 lol Sep 14 02:31:51 CP|Laptop: iirc bernardo is one of the devs for sourcemage Sep 14 02:32:02 hrw|work: what i"m doing is patching the 2.4 pcmcia with this guy ... and the I try it on my 2.4 Sep 14 02:32:12 hrw|work: if that work then I can port the driver to 2.6 Sep 14 02:32:27 then I can start working .. or else it is a pain without wireless Sep 14 02:32:30 AvengerMoJo: it has own driver? nice challenge it can be Sep 14 02:32:33 I mean network Sep 14 02:32:52 hrw|work: well.. it should be hermes Sep 14 02:33:00 hrw|work: wavelan2 Sep 14 02:33:25 hrw|work: but some variation and the latest hostap driver get something ... not correct though Sep 14 02:33:39 so I'm try to dig into the 2.4 and see the myth of this driver really work Sep 14 02:34:09 the sol reason I'm trying 2.4 actually Sep 14 02:34:34 or else Iwill be have with out the sound and start fixing the keymap stuff already hehe Sep 14 02:34:47 but I get priority hehe :P Sep 14 02:35:09 * AvengerMoJo need to get his wireless card working first Sep 14 02:35:36 also ... this eagle card exist a lot more in China / HongKong .... if I can actually port this card .. may help alot of people Sep 14 02:35:47 so .... Sep 14 02:36:21 koen: what do you think about meta-sdl.bb which will have all sdl based apps/games? Sep 14 02:36:28 the card itself is cheap ... but I don't really want to just get a working card and be done with it Sep 14 02:36:28 doesn't 2.6 have wvlan_cs? Sep 14 02:36:36 hrw|work: sounds good Sep 14 02:36:44 koen that card didn't work :( I don't know why Sep 14 02:36:47 koen: it will make meta-opie-all smaller ;) Sep 14 02:37:02 hrw|work: but someone still needs to remove all the pointless opie stuff from sdl games Sep 14 02:37:07 koen that's why I think this driver claim to be working is a myth Sep 14 02:37:09 koen: but first I have to build meta-opie-all to get full list of it Sep 14 02:37:12 SECTION = opie/games Sep 14 02:37:17 and more of that nonsense Sep 14 02:37:20 koen: yep Sep 14 02:41:38 if I want to extract a tar.bz as a package patch ...where should I put it in? file:// or do_configuer? Sep 14 02:42:39 patches in tarball? Sep 14 02:42:42 You probably want to put it in SRC_URI, and provide some kind of do_munge() method to shuffle files in the working directory. Sep 14 02:44:05 koen: pointless? if they install a .desktop in /opt/Qtopia... they're basicly a Qtopia/Opie app ;) Sep 14 02:44:15 heh Sep 14 02:44:22 koen: I did not know matchbox could read these files Sep 14 02:44:38 matchbox can read .desktop files just file Sep 14 02:44:44 fine* Sep 14 02:44:55 koen: hehe but broken QPE files as well? Sep 14 02:45:07 no, it follows the fdo spec Sep 14 02:45:08 broken as in broken by design Sep 14 02:45:26 yeah Sep 14 02:45:39 qtopia itself is broken by design Sep 14 02:45:54 if someone told me microsoft designed it, I wouldn't be surprised Sep 14 02:46:01 koen: if they would be freepda.o or freesmartphone.o... but freedesktop.org sounds wrong for a mobile computer Sep 14 02:46:05 incompatible junk Sep 14 02:46:17 koen: Microsoft does not design, they buy Sep 14 02:47:14 koen: I need to start a Bitch Qtopia blog and hope TT listens... Sep 14 02:47:45 "we don't need standards because we are doing embedded work" Sep 14 02:47:46 zecke_: add yourself to Qt Blog? Sep 14 02:48:15 koen: no we don't need standards because we have a momentum on embedded linux Sep 14 02:48:26 koen: Montavista and Lineo acknowledge that ;) Sep 14 02:50:55 koen: http://www.linux.org/news/2005/09/14/0003.html Sep 14 02:52:23 zecke_: good morning Sep 14 02:52:39 zecke_: that doesn't mention 'filthy rich' Sep 14 02:52:47 zecke_: MimeType::appsFolderName() return /opt/QtPalmtop/apps/ dir? Sep 14 02:53:47 hrw|work: yes (I wonder if they're translated) Sep 14 02:54:20 zecke_: thx Sep 14 02:54:34 not translated Sep 14 02:54:43 it simply returns the path **** ENDING LOGGING AT Wed Sep 14 02:59:56 2005