**** BEGIN LOGGING AT Sun Feb 11 02:59:57 2007 Feb 11 05:27:58 just built 2.6.20 + modules and installed, runs fine but cant locate modules, neither can insmod unless specify full path to .ko Feb 11 05:38:05 can anyone give me advice? I'm trying to build LyX for a zaurus akita (the recipe from the mtn database), but when it builds LyX, it tries to link against the X libraries in my /usr/src directory, so naturally it gets an incompatible symbol error. All the libraries it needs to link to exist in the staging directory, how can I make it look there? Feb 11 05:38:15 sorry that was /usr/lib Feb 11 07:40:12 03rwhitby 07org.oe.dev * rcc34d79a... 10/ (1 packages/mediatomb/mediatomb_svn.bb): mediatomb: Use id3lib instead of taglib. Update from upstream author. Feb 11 08:14:25 koen: is it possible to override MACHINE in a machine.conf? e.g. in nslu2le.conf, just set TARGET_ARCH=arm, and then set MACHINE=nslu2, so that all the _nslu2 overrides everywhere still work? Feb 11 08:15:30 good morning all Feb 11 08:15:40 morning koen Feb 11 08:16:10 rwhitby: it's possible, yes, but confusing as well Feb 11 08:16:25 and I suspect bitbake wouldn't like $MACHINE to change at that point Feb 11 08:16:52 hmm. since nslu2 and TARGET_ARCH uniquely define whether it's nslu2le or nslu2be, it seems a shame to need two machines Feb 11 08:17:27 (I'm trying to reduce openslug and debianslug into a single slugos distro, which can be used with nslu2le or nslu2be machines) Feb 11 08:18:04 (but don't want to have to go through all of OE and replace _nslu2 overrides with both _nslu2le and _nslu2be overrides. Feb 11 08:18:32 rwhitby: you can just add "nslu" to OVERRIDES by hand. Feb 11 08:19:02 pb_: is there a good way to insert something in the middle of OVERRIDES? Feb 11 08:19:26 oh, you need it in the middle? that's a bit harder, you'd need to use some embedded python. Feb 11 08:20:00 I guess I don't necessarily *need* it in the middle. I doubt there are any places where the position would really matter. Feb 11 08:20:54 OVERRIDES := ${@bb.data.getVar('OVERRIDES',d).split(':')[0:5] + ":nslu2:" + bb.data.getVar('OVERRIDES',d).split(':')[6:]) Feb 11 08:20:58 or, er, something. Feb 11 08:21:11 possibly with more }s on the end Feb 11 08:21:20 sweet. Feb 11 08:22:17 I can use that for the existing _slugos override (except that one will hopefully be done by there just being a single _slugos distro, instead of the multiple slugos distros that exist today) Feb 11 08:23:03 BTW, I finally got arm-kernel-shim to work on both BE and LE, so that will remove the need for any cmdline hacks in ixp4xx-kernel. Feb 11 08:24:23 Is it reasonable to require someone to set TARGET_ARCH correctly in auto.conf, alongside DISTRO and MACHINE? Feb 11 08:24:40 (then I could have a single nslu2 MACHINE, and require the user to set TARGET_ARCH to arm or armeb) Feb 11 08:26:52 seems reasonable enough to me. you can make machine.conf blow up with some suitable error message if it hasn't been set. Feb 11 08:27:21 ok, I'll do that for nslu2 and ixp4xx then. Feb 11 08:29:28 I guess I'll still need to keep MACHINE_ARCH as nslu2le/nslu2be, so that packages are kept separate. Feb 11 08:29:45 but I can set that based on SITEINFO_ENDIANESS Feb 11 08:29:53 bbiab Feb 11 09:09:59 * koen compares last years lab assigments with the current ones Feb 11 09:10:24 * koen changes 2006 into 2007 Feb 11 09:10:39 tada, saved myself 4 hours of matlab + latex :) Feb 11 09:13:56 :)) Feb 11 09:27:04 hey pH5 Feb 11 09:28:18 hey koen Feb 11 09:32:36 awww, too bad WorldWind uses directx Feb 11 09:32:50 it would have been sweeeet to run it other platforms with Mono :D Feb 11 09:33:00 Eblis: there was some talk about making it use opengl Feb 11 09:33:10 that would be nice Feb 11 09:33:20 it uses .Net 1.1 i see which Mono fully supports Feb 11 09:33:31 maybe they've added some directx->opengl goodness too Feb 11 09:35:25 mono i mean ... i haven't looked into it yet Feb 11 09:59:17 pb_: if TARGET_ARCH was not defined, would I raise a bb.parse.SkipPackage in machine.conf, or something else? Feb 11 09:59:45 rwhitby: bb.fatal(), probably. Feb 11 10:00:23 ah, thanks - found usages of that now. Feb 11 10:22:24 pb_: what does the @ at the start of OVERRIDES := ${@bb.data.getVar do, and why isn't there one on the second call to bb.data.getVar? Feb 11 10:22:59 rwhitby: ${@...} tells bitbake to interpret the expression as python, rather than as a straightforward variable expansion Feb 11 10:24:15 hmm - I'm getting a syntax error in the fragment you suggested, even after putting } on the end. Feb 11 10:24:50 mm, right, I probably got something wrong Feb 11 10:25:04 in fact, there's definitely a join() missing there for one thing. Feb 11 10:26:16 oh, so I need to do something like ":".join(...) Feb 11 10:26:29 probably, yeah Feb 11 10:26:41 might be better to write that stuff as a function rather than trying to put it all in one line. Feb 11 10:26:54 def rwhitby_overrides_munging(): Feb 11 10:26:55 ... Feb 11 10:27:08 OVERRIDES := ${@rwhitby_overrides_munging()} Feb 11 10:42:34 * koen stabs taxes Feb 11 10:54:42 koen: any objections to me removing ixp4xxle.conf and ixp4xxbe.conf in favour of requiring the user set TARGET_ARCH, as pb_ recommended above? Feb 11 10:55:23 (ixp4xx.conf has a check to make sure it is set, and calls bb.fatal if it is not) Feb 11 10:56:07 this finally allows me to have a single slugos.conf distro and just set MACHINE and TARGET_ARCH appropriately. Feb 11 10:56:41 next step is multimachine support ... Feb 11 11:39:28 rwhitby: I'm not sure that's a good thing Feb 11 11:39:59 rwhitby: it would mean that for 1 machine OE would blow up when I only specify MACHINE and DISTRO Feb 11 11:40:14 2 machines - nslu2 and ixp4xx :-) Feb 11 11:40:21 and that MACHINE=foo bitbake bar wouldn't work anymore Feb 11 11:40:26 so I don't see the benefits Feb 11 11:40:54 except going back to the confusing state where a single machine can be le or be Feb 11 11:41:05 which isn't a benefit at all Feb 11 11:43:44 morning Feb 11 11:43:48 hey hrw Feb 11 11:44:01 the thing is, that looking through all of OE now (including the unified slugos confs), there is not a single place where an _ixp4xx override needs to be different for LE or BE. Feb 11 11:44:13 (same for nslu2) Feb 11 11:44:32 that doesn't mean there should be only one machine Feb 11 11:44:42 since everything to do with endianness is now done by ${SITEINFO_ENDIANESS} (which comes from TARGET_ARCH) Feb 11 11:45:20 it is entirely valid to not set target_arch when you specify a machine Feb 11 11:45:22 so the other side of the argument is that I can get an inconsistent state if I set MACHINE=nslu2le and set TARGET_ARCH=armeb Feb 11 11:45:32 I don't see why the slug wants be an exception Feb 11 11:45:53 hi greentux Feb 11 11:45:53 Actually, TARGET_ARCH *must* be set somewhere for siteinfo to work. Feb 11 11:45:59 greentux: how was in Austria? Feb 11 11:46:02 hi hrw Feb 11 11:46:16 rwhitby: right, the machine sets target arch Feb 11 11:46:31 hrw: very nice. weather top, some snow for boarding, and: no network access! :) Feb 11 11:46:44 (btw, it's not just the slug, its any processor that can operate in both le and be equally well as far as OE is concerned) Feb 11 11:47:26 cbrake: nice summary post on linuxdevices Feb 11 11:47:48 rwhitby: no, where the machine is usefull when you switch endiannes Feb 11 11:48:07 rwhitby: my ipaq can run in BE as well, but all the interfaces won't work Feb 11 11:48:13 hrw: you got the order? Feb 11 11:48:23 greentux: yes Feb 11 11:49:06 koen: that's what I said - the machine can operate in both modes equally well. if the interfaces won't work, then the machine cannot operate equally well in both modes. Feb 11 11:50:09 hrw: fine, so you can send me the invoice now :) per fax Feb 11 11:50:19 rwhitby: you said "processor" :) Feb 11 11:50:24 greentux: fax? can not it be PDF in mail? Feb 11 11:51:16 koen: I stand corrected. I meant to say machine (as in processor and board combined) Feb 11 11:52:10 actually, the ipaq could be made to work perfectly well in big endian if you wanted to. it's just a pile of kernel work that nobody has bothered to do. Feb 11 11:52:34 hrw: yes, when you sign it with a trusted signing certificate, that would be ok. Feb 11 11:53:30 and, of course, those of us who know that little endian is the only true endianness are quite happy with that situation. Feb 11 11:53:32 pb_: in which case, would you expect to have to go through all of OE, changing _ipaq to _ipaqle and _ipaqbe ? Or just set TARGET_ARCH in your auto.conf .... Feb 11 11:53:33 :-} Feb 11 11:53:36 greentux: cacert one will be ok? Feb 11 11:54:57 rwhitby: indeed, having to change all those files would clearly suck. I think it's pretty clear that this would be a dumb idea. The only interesting debate really is between your original idea of two MACHINEs which share the same overrides, and the alternative idea of a single MACHINE that requires you to also specify the endianness. Feb 11 11:54:59 hrw: cacert is not certified (in EU) Feb 11 11:56:00 greentux: so it looks like I will have to find a fax somewhere. Feb 11 11:56:06 I don't have a very strong opinion either way. I guess I can see the argument that it's convenient for MACHINE+DISTRO to completely specify the output. Feb 11 11:56:45 hrw: yes, looks like. or we use snail mail Feb 11 11:56:50 of course, if you take the view that any given distro is only going to ship for one endianness then you can just set TARGET_ARCH_nslu2 in your DISTRO.conf. Feb 11 11:57:17 if any given distro is likely to want to ship both endiannesses, that'd be a reasonably compelling argument for having two MACHINEs. Feb 11 11:57:38 greentux: ok. will snail it then Feb 11 11:58:05 hrw: fine. Feb 11 11:59:33 pb_: the situation here is that we had independent openslug (be) and debianslug (le) distros, which were identical except for the endianess. For historical reasons, openslug came first and the ethernet wouldn't work le, but then we got it to work le, and could then be compatible with Debian arm packages. This effort is to subsume both of them into a single slugos distro. Feb 11 12:00:20 rwhitby: so you'll have a single distro shipping both le and be binaries? Feb 11 12:00:42 in that case you probably do want the MACHINEs to be different so that you can tell the resulting packages apart. Feb 11 12:00:57 yes. both be and le images, and both be and le packages (in different feeds, to make the Packages file smaller in the jffs2 rootfs) Feb 11 12:01:29 ah, the packages for ixp4xx are already disambiguated by armv5te and armvt5eb Feb 11 12:02:45 oh right, yeah, there'd be no point having the MACHINE as the packaging architecture if your distro is only targetting a single machine. Feb 11 12:02:52 in that case I think it's pretty much a wash between the two options. Feb 11 12:03:34 anyway, there are good arguments on all sides of this discussion, so I'll take another look at it and see whether separate endianess machines (with the common machine override) causes any more work in maintaining the metadata. Feb 11 12:04:46 okay, cool Feb 11 12:06:13 I guess deprecating the nslu2 machine completely (and giving a bb.fatal if someone uses it) is a good way to force people to change their configs to the new structure ... Feb 11 12:06:39 (if I do the nslu2->nslu2le/be and openslug/debianslug->slugos at the same time) Feb 11 12:22:40 Hmm - any idea what this means: "RuntimeError: dictionary changed size during iteration" ? Feb 11 12:27:03 pb_: the worst thing now is all the places that say "if test "${MACHINE}" = nslu2" ... Feb 11 12:48:53 cheers Feb 11 12:51:24 hi mickeyl Feb 11 12:52:54 hi mickeyl Feb 11 12:55:05 hey mickeyl Feb 11 12:57:00 mickeyl: hi, have you got some time for me? Feb 11 12:57:13 heh - the expectancy in the air could be cut with a knife ... Feb 11 12:57:34 * koen wonders at what time mickeyl will do 'mtn sync ; mtn merge ; mtn sync' Feb 11 12:58:12 koen: I hope that OM team first remove AUTH from wiki/bugtracker Feb 11 12:58:18 it takes less time ;D Feb 11 12:58:25 mr_nice: sure, give me 10 minutes for the bathroom, then i'm back Feb 11 12:58:38 mickeyl: :) Feb 11 12:58:39 Geez, I better get these slugos distro changes in before *that* merge ... Feb 11 12:59:16 hrw: I was a bit surprised to get a mail telling me to file bugs in the bugzilla without getting a login :) Feb 11 13:00:55 koen: and I'm tired of seeing links to wiki which all ends in http auth requester Feb 11 13:01:48 'how to switch connection from serial to usb and vice versa using one part of switch' Feb 11 13:06:26 koen, pb_: It looks like multiple machines with a common override is doable. Feb 11 13:06:37 cool Feb 11 13:07:38 I will forcibly deprecate nslu2.conf and ixp4xx.conf at the same time (i.e. they will give a bb.fatal saying you need to choose one of the endian-specific machines) Feb 11 13:22:59 koen, ping Feb 11 13:27:49 pong Feb 11 13:29:32 koen, please do me a favor and invalidate my OE key, I kinda *cough* "lost" *cough* it Feb 11 13:30:00 I can only remove commit access for it Feb 11 13:30:19 that is fine. i'll send you a new one once I get OE back up Feb 11 13:30:34 CoreDump|home: let new one differ from old Feb 11 13:30:50 hrw, I don't have the old one =\ Feb 11 13:30:59 * koen adds if (string.find(identity, "coredump@openembedded.org", 1, true)) then return false end to monotonerc Feb 11 13:31:39 thanks Feb 11 13:32:04 so I better use another ID / email address for the new key huh? Feb 11 13:32:29 @openembedded.org Feb 11 13:32:38 righto Feb 11 13:37:35 mickeyl: did you see the patches from Lars Munch for evas/fb on arm? Feb 11 13:38:10 (on e-devel) Feb 11 13:40:09 re Feb 11 13:40:17 koen: not yet, I'm reading e-devel only once a week nowadays Feb 11 13:40:24 (it's still delivered to the uni account) Feb 11 13:40:28 mr_nice: ok, here i am Feb 11 13:41:19 morning Feb 11 13:42:54 mickeyl: that is fine :). I have completed the switches driver and made a patch against 2.6.20. I will send it to rmk but I am not sure if the driver is good enough for mainline. Also I want to know the needed tasks to integrate it to angstrom. Atm I am on trying to write a propper .bb file but I am a bit puzzeled. Feb 11 13:43:16 can't help you with the driver, but where is your problem with the bb ? Feb 11 13:46:12 mickeyl: Zecke told me to look at linux-rp files. I copied all the files and allready changed the linux-simpad.inc file and the linux-simpad-2.6.20 directory. But the linux-simpad_2.6.20.bb file, which is a copy of linux-rp_2.6.19.bb is quite compley Feb 11 13:46:24 don't bother with complex solutions Feb 11 13:46:29 choose a simple kernel bb Feb 11 13:46:35 then add your patches to SRC_URI Feb 11 13:46:37 then you're done? Feb 11 13:46:40 s/?/. Feb 11 13:47:03 looks at linux kernel .bb for x86 or ezx Feb 11 13:47:34 mickeyl: ok I will try it. thx Feb 11 13:47:36 hi mickeyl Feb 11 13:48:03 rwhitby: okay, very good Feb 11 13:48:04 hi zecke Feb 11 13:48:07 mr_nice: np. if you have a concrete problem, just pastebin here and we'll help. but you really should start simple, if you want to learn :) Feb 11 13:48:08 hey Feb 11 13:48:09 cheers zecke Feb 11 13:48:11 hey pb_ Feb 11 13:51:34 mr_nice: linux_2.6.20.bb is simple recipe for kernel Feb 11 13:53:28 hrw: thx Feb 11 13:54:32 zecke: hi, do you have some time for me? Feb 11 13:56:31 mr_nice: no. I'm sitting at 3gsm the fair will open tomorrow and we need to get stuff done :) Feb 11 13:57:33 zecke: hehe, no problem :) Feb 11 14:08:51 mickeyl: are you ready to commit the openmoko OE stuff? Feb 11 14:09:31 and open wiki? Feb 11 14:09:34 no Feb 11 14:09:42 otherwise i did it already Feb 11 14:09:57 we're having some important discussions atm Feb 11 14:09:58 ok Feb 11 14:10:13 besides, the stuff won't build against .dev atm. Feb 11 14:10:27 i dunno yet how to handle that Feb 11 14:10:37 commit it and let hrw and me fix it :) Feb 11 14:10:47 hmm Feb 11 14:11:08 hehe Feb 11 14:11:23 i don't want .dev and svn.openmoko.org/oe to divergate Feb 11 14:11:29 but need to pinpoint one rev atm. Feb 11 14:11:42 * mickeyl needs to think more about it Feb 11 14:11:47 koen: BTW: Do you feel lonely when going to presentations? Feb 11 14:11:59 hi koen zecke hrw mickeyl Feb 11 14:12:05 zecke: not really :) Feb 11 14:12:12 zecke: want to join? Feb 11 14:12:19 hey chouimat Feb 11 14:12:42 hi chouimat Feb 11 14:12:42 koen: considering it Feb 11 14:13:49 mickeyl: ideally you'll just 'svn delete /oe' after doing 'mtn commit' Feb 11 14:14:06 yes, but like i said that doesn't work yet Feb 11 14:14:20 can't afford tracking a floating .dev at this point Feb 11 14:14:22 right Feb 11 14:14:39 but having 2 repos at the same time also gives out a wrong message Feb 11 14:14:44 yes Feb 11 14:14:54 that's why I dunno atm. Feb 11 14:14:58 rock | mickeyl | hard place Feb 11 14:15:10 hehe Feb 11 14:15:11 yeah Feb 11 14:16:25 mickeyl: is it possible to restrict access to /oe? Feb 11 14:16:44 for svn in general, yes Feb 11 14:17:04 if transport is http Feb 11 14:17:21 Crofton: even with ssh and https and svn protocol it will work Feb 11 14:17:29 ok Feb 11 14:17:41 wasn't sure at what level svn access file was read Feb 11 14:17:54 I though it was mod_auth_svn Feb 11 14:18:01 mickeyl: you could keep /oe closed till you/we/fic/whoever merge it into OE Feb 11 14:18:03 maybe only for per path acl's Feb 11 14:18:10 Crofton: I need to maintain two ACL's though :} Feb 11 14:18:11 and then add a note 'obsolete, use OE' Feb 11 14:18:24 Crofton: one for everyone (ssh) and one prefilter through the apache stuff Feb 11 14:20:58 * koen starts updating the angstrom armv4t feeds Feb 11 14:22:29 hmm mtn 0.31 doesn't like the OE snappshot DB from today. Do I need mtn 0.32? Feb 11 14:22:50 see GettingStarted Feb 11 14:23:24 That's what I'm doing Feb 11 14:24:03 ya ya, didn't read step one *sigh* Feb 11 14:31:39 anybody got a working mtn 0.32 deb for debian sid? Feb 11 14:32:00 CoreDump|afk: x86 one is on monotone homepage Feb 11 14:32:42 hrw, thanks! Feb 11 14:41:04 * chouimat decides to install qemu 0.9 Feb 11 14:44:03 cu Feb 11 15:25:10 sqlite 3.3.7 and 3.3.9 upstream tarballs are gone, instead there's 3.3.12 now. do we just follow and remove old sqlite3 recipes? Feb 11 15:44:24 generally, we try to keep at least one known good recipe working Feb 11 15:44:37 so rather fix 3.3.9 SRC_URI, remove 3.3.7, and add 3.3.12 Feb 11 15:46:29 does anybody know where they move the old sources or if they are available at all? Feb 11 15:48:32 doesn't angstrom have it mirrorred? Feb 11 15:50:43 hmmm Feb 11 15:50:58 I wonders how much trouble migrating from 2.6.8 to 2.6.20 is going to give me Feb 11 15:55:33 cute, ivtv is being merged into the kernel Feb 11 15:56:36 re Feb 11 15:57:16 Jin^eLD: wb Feb 11 15:57:24 hi koen Feb 11 16:20:25 re Feb 11 16:20:58 yo sirfred Feb 11 16:21:06 XorA: Hi Feb 11 16:29:05 hey sirfred & XorA Feb 11 16:29:29 koen: Hi Feb 11 16:54:22 re Feb 11 16:54:35 hrw|tv: wb Feb 11 16:55:24 It seems that the filechooser used in GPE is GtkFileSelection, I thought it was not used in new gtk apps Feb 11 16:55:47 But GtkFileChooserWidget needs really a lot of space Feb 11 16:56:37 'morning' all Feb 11 16:56:53 morning Feb 11 16:56:57 hey RP Feb 11 16:56:57 RP: Morning Feb 11 16:57:05 * RP mutters about an alpha state filechooser in poky Feb 11 16:57:26 tell me folks, how to get *absolute* path of *arbitrary* path. in bash. Feb 11 16:57:30 RP: r1266 & r1267? Feb 11 16:57:46 RP: Is it implementing GtkFileChooserDialog ? Feb 11 16:58:01 RP: That way it could be used as dialog for GtkFileChooserButton, that is what I'm behind. Feb 11 16:58:02 koen: right :) Feb 11 16:58:33 sirfred: I'm not sure, its not my work. I just heard OH people talking about it ;-) Feb 11 16:58:49 is it the one Ross started or a different one? Feb 11 16:59:05 koen: tf wrote it Feb 11 16:59:11 RP: too bad it's for gtk 2.6.8 Feb 11 16:59:41 It should be straightfoward enough to forward port... Feb 11 16:59:53 If only the GtkFileChooserWidget launched from a GtkFileChooserButton could be tuned in some way... Feb 11 17:00:07 question: I'm trying to build LyX for a zaurus akita (the recipe from the mtn database), but when it builds LyX, it tries to link against the X libraries in my /usr/src directory, so naturally it gets an incompatible symbol error. All the libraries it needs to link to exist in the staging directory, how can I make it look there? Feb 11 17:00:19 sirfred: I think some of our gtk hackers were talking about making that possible... Feb 11 17:00:40 RP: That would be interesting. Feb 11 17:01:00 RP: My GtkButton + GtkFileSelection solution stinks, ;-) Feb 11 17:01:42 RP: Did you have time to go into the w100fb changes (rotation and extern memory off) ? Feb 11 17:02:23 sirfred: I've not had time to get around to it yet. I have cleared 30 backlight class related mails from my inbox. The w100 ones are in there too ;-) Feb 11 17:03:02 RP: You work too much, I'm afraid. Feb 11 17:03:04 sirfred: http://svn.o-hand.com/view/poky/trunk/meta/packages/gtk+/gtk+-2.6.8/file* Feb 11 17:03:26 sirfred: Not enough time to do everything I'd like... Feb 11 17:04:05 * RP spent the afternoon taking the MG to bits to fix the battery terminals :/ Feb 11 17:04:37 RP: Thanks for the link. I suppose this is not into gtk+ in OE, isn't it? Feb 11 17:04:48 RP: Only in poky version? Feb 11 17:05:01 sirfred: Its only recently made poky and is a work in progress Feb 11 17:06:22 Hmm, I would like this program to work in current OE, so, I will try a temporary solution, Thanks anyway Feb 11 17:06:43 Time for basket now. Feb 11 17:08:31 sirfred: There is nothing to stop you testing and adding to OE Feb 11 17:09:13 I wouldn't know what I was testing so I won't merge it as it would cause me problems I don't have time for if it went wrong Feb 11 17:37:08 the boost bb file is amusing .... Feb 11 18:08:46 re Feb 11 18:08:59 hrw|tv: wb Feb 11 18:09:16 openmoko wiki still closed ;( Feb 11 18:59:24 osb-nrcit use curl? Feb 11 19:01:46 yes Feb 11 19:02:13 that's why ssl site with a proper cert now 'work' in webcore based stuff Feb 11 19:02:23 * koen still wants webkit Feb 11 19:03:11 * hrw rebuilds curl Feb 11 19:04:32 koen: does that mean you have gtk-webcore working again? Feb 11 19:04:35 NOTE: package osb-nrcit-0.5.0+cvs20060212-r0: task do_configure: completed Feb 11 19:04:37 uf. Feb 11 19:04:39 is it possible to add -dbg packages to image automatically? Feb 11 19:04:47 philippe: no Feb 11 19:05:59 * philippe really needs to get his h3800 to work again... Feb 11 19:07:28 hey guy a question ... ir your opinion what's more important first 1) have usb gadget work on about all StrongArm PDA or 2) have h3800 work (except usb) ??? Feb 11 19:07:53 the latter Feb 11 19:08:19 ok i'll reschedule all ... Feb 11 19:08:57 sincerelly i collect lot of info about usb but i write really few code ... Feb 11 19:09:25 gremlin[it]: I submitted a patch for h3800 a while ago. This already boots the device. Just no X. Feb 11 19:10:01 i see ... what is missing (i think) is asic support ... probably is asic to control power of LCD ... ;) Feb 11 19:18:06 if that's all it is, it should be pretty easy to get it working. the lcd power control is just a handful of gpios. Feb 11 19:18:15 the sequencing is a little bit tricky but you can use the 2.4 source as an example. Feb 11 19:19:26 gremlin[it]: did you tested strongarm usb gadget already? Feb 11 19:19:41 i really thought USB was more important (and i thoug was easy) ... Feb 11 19:20:00 hrw .. yes some test ... but not functional ... Feb 11 19:20:12 well, I guess usb is probably more important for people who don't have h3800s. Feb 11 19:20:28 presumably it can't be all _that_ important though, or someone would have done it already Feb 11 19:20:38 uf.. we have xdm in OE Feb 11 19:20:45 I will be able to remove gpe-dm Feb 11 19:21:00 pb_ yes true :) Feb 11 19:22:16 ok i#ll put my h3800 in devel state :) ... Feb 11 19:23:06 didn't nick bane already write a 2.6 udc driver for sa? Feb 11 19:23:38 koen where's the code ??? on hh.org there is nothing ... Feb 11 19:23:44 gremlin[it]: if you have some hints on debugging it without having to flash a 19MB image over serial I don't mind helping. Feb 11 19:23:51 hh.org *is* nothing Feb 11 19:23:56 koen: psst Feb 11 19:24:02 hmm I just returned in time? Feb 11 19:24:05 gremlin[it]: I sent you link few times Feb 11 19:24:22 gremlin[it]: check linux-rp kernel recipe - this patch is added for collie Feb 11 19:24:57 hrw ... yes but such code don't compile, after some change it compile but hang ... Feb 11 19:25:27 what i write since now compile and react to the plug of the cable ... Feb 11 19:26:36 philippe, is one of teh reasons cause i start working on usb to accelerate dead time in transfer images to ipaq :( Feb 11 19:28:22 why the hell there isn't a unique arm kernel cvs :( ... grrr Feb 11 19:28:47 gremlin[it]: because upstream has arm support nowadays Feb 11 19:28:56 gremlin[it]: there is one: kernel.org is his name Feb 11 19:28:59 having a unique repo would only fragment the developers Feb 11 19:30:07 i have to go let's talk tomorrow ... Feb 11 19:30:31 gremlin[it], btw, isn't there udc driver for SA somewhere? is it that noone hacked one for all this time? Feb 11 19:30:43 at least try t merge the already working usb for sa1110 if really exist ... (damn i spent a couple of month for nothing) ... Feb 11 19:31:11 have to go sorry ... let's see tomorrow Feb 11 19:34:57 philippe: still the same crashes with webcore svn Feb 11 19:35:17 webcore? did I miss anything? Feb 11 19:36:21 philippe: http://rafb.net/p/wJipD449.html Feb 11 19:36:24 koen: grr... I hope I can get something running soon to verify it Feb 11 19:36:53 * koen wants webkit Feb 11 19:37:42 * philippe too Feb 11 19:38:01 well definitely webcore. munmap_chunk() is not a function I use in gpe-mini-browser... Feb 11 19:38:25 koen: have you tried the svn too? Feb 11 19:38:42 which svn? Feb 11 19:38:49 gtk-webcore svn? Feb 11 19:39:00 scroll up 4 minutes Feb 11 19:39:17 [21:07] koen: philippe: still the same crashes with webcore svn Feb 11 19:39:35 root@hx4700:~# ipkg list | grep svn | grep lib Feb 11 19:39:36 libfakekey0 - 0.0+svn20051101-r2 - Feb 11 19:39:36 libjscore0 - 0.5.0+svn20070211-r0 - Feb 11 19:39:36 libnrcit0 - 0.5.0+svn20070211-r0 - Feb 11 19:39:36 libnrcore0 - 0.5.0+svn20070211-r0 - Feb 11 19:39:40 ah ok, sorry. I thought you referred to this osb-nrcit-0.5.0+cvs20060212-r0 Feb 11 19:40:23 * philippe apologizes Feb 11 19:40:56 Good morning! Feb 11 19:41:36 hey Laibsch Feb 11 19:45:31 RP: I'm going to try the filechooser-respect-style.patch against gtk+-2.8.9, it applied cleanly at least. Feb 11 19:46:10 hi all Feb 11 20:07:55 cu Feb 11 20:44:25 I now have created a new simpad 2.6 .bb (http://pastebin.ca/350948) file and it compiles well :). But I am missing a /dev/console entry so it won't boot. how to tell oe that it should create it? Feb 11 20:59:44 mr_nice: sure that it mounts the rootfs properly? did you check the rootfs/ directory to make sure it's really the problem? Feb 11 21:04:30 tmbinc: I get a Warning: unable to open an initial console. And there is no /dev/console entry in ..rootfs.tar.bz2 Feb 11 21:13:57 If someone is interested, I've ported Gsnes9x, a gtk+1.2 snes9x frontend to work on GPE Feb 11 21:14:10 I have made available a first version at: Feb 11 21:14:24 http://www.telefonica.net/web2/mteirap/gpe-snes-0.1.tar.gz Feb 11 21:15:23 http://www.telefonica.net/web2/mteirap/gpe-snes_0.1.bb Feb 11 21:27:20 i'm looking for documentation, the wiki is broken ? Feb 11 21:58:21 counter Feb 11 21:58:41 sorry, wrong channel Feb 11 22:12:28 Night all Feb 11 22:28:50 trem: What do you mean the wiki is broken? Feb 11 22:29:11 http://www.openembedded.org/wiki/MainPage Feb 11 22:29:18 The site got rearranged a bit and not everyone has updated their links, so you're now seeing our fancy 404 page. Feb 11 22:29:19 :( Feb 11 22:35:02 koen: Can you fix the wiki? Feb 11 22:35:15 I need it, too ;-) Feb 11 22:37:00 trem: In the meantime use google. It has a lot of the pages cached. Feb 11 22:37:25 good idea Feb 11 22:37:35 but later, now, time to sleep ;) Feb 11 22:38:29 For example: http://72.14.235.104/search?q=cache:www.openembedded.org/documentation Feb 11 22:39:09 i use org.openembedded.documentation Feb 11 22:39:24 but i can't generate pdf, so I use html Feb 11 22:45:07 trem: so many choices ;-) Feb 11 22:45:28 yes Feb 11 23:13:43 nite all, sweet dreams Feb 11 23:59:24 03Laibsch 07org.oe.dev * r141cd340... 10/ (1 packages/nasm/nasm_0.98.38.bb): nasm: fix SECTION Feb 12 00:21:29 03pfalcon 07org.oe.dev * rc27a1a68... 10/ (4 files in 3 dirs): blueprobe 0.18: Add initial support for h4000. Feb 12 00:21:33 03pfalcon 07org.oe.dev * re32f0276... 10/ (1 classes/opie.bbclass): opie.bbclass: OPIE_CVS_PV: Convert to standard CVS version format. Feb 12 00:27:09 03Laibsch 07org.oe.documentation * r67f3be0b... 10/ (1 usermanual/reference/var_section.xml): var_section.xml: fixed and removed dev-lang sections Feb 12 02:59:56 03Laibsch 07org.oe.documentation * rd9d72bb5... 10/ (1 usermanual/reference/var_section.xml): var_section.xml: fix and remove "x11/apps" section **** ENDING LOGGING AT Mon Feb 12 02:59:57 2007