**** BEGIN LOGGING AT Fri Jul 14 02:59:57 2006 Jul 14 05:11:28 http://pastebin.ca/87774 during a fresh openzaurus-unstable build Jul 14 05:11:34 any ideas? Jul 14 05:16:43 Luke-Jr: i was able to build glibmm here earlier Jul 14 05:23:40 anybody know how to re-index your ipkg directory? Jul 14 05:26:31 morning Jul 14 05:27:45 * zwelch waves to TEMic Jul 14 06:00:28 ~ping hvontres Jul 14 06:00:49 pong hvontres Jul 14 06:31:02 Greetings ! Jul 14 06:33:57 morning all Jul 14 06:36:33 hi do13 ! Jul 14 06:36:58 hey Ifaistos Jul 14 06:37:54 Ifaistos: How is the weather in athens? Jul 14 06:38:20 hot and humid. its time to go the beach ;) Jul 14 06:38:43 ;) Jul 14 06:39:13 Ifaistos: In which part of athens do you live? Jul 14 06:39:44 south part,close to the Marina Jul 14 06:48:33 have to reboot. be back shortly Jul 14 06:49:59 good morning all Jul 14 06:53:49 hey koen Jul 14 06:56:15 morning all Jul 14 06:56:42 hi RP Jul 14 06:57:21 morning RP ! Jul 14 06:58:49 RP: what are the ep changes you made to the pxa27x_udc driver about? with your version I have to unplug/reconnect the usb cable after suspend/resume to get usb networking back Jul 14 06:59:20 hey Richard, koen Jul 14 07:01:08 pH5: Once the chip is initted, you can't change the endpoint configuration. Those ep changes allowed the ep config to be specified so the chip knows how to setup Jul 14 07:01:29 pH5: The unplug/reconnect sounds like a bug Jul 14 07:03:30 pH5: The problem comes when you need to switch configuration at the hosts request - if you didn't setup that configuration, you can't switch to it so all configurations need to be known at chip init time Jul 14 07:08:46 RP: How would you proceed to find that bug? The functions I provide in udc_mach_info are never called. Could that be related? Just after booting I have to unplug/replug the cable even with the hh.org cvs version. Jul 14 07:10:22 pH5: The machine specific setup parts of that driver needs a lot of TLC. There is a bit hardcoded for then zaurus in there which could be an issue Jul 14 07:11:28 pH5: Those bits aren't in the suspend/resume paths though... Jul 14 07:12:20 pH5: It could be you need to setup the machine specific regs in udc_enable Jul 14 07:12:51 pH5: It looks like I hardocoded that for the Zaurus and it is on the resume path Jul 14 07:13:13 RP: ah right, after your FIXME the udc_command invocation is commented out. Thanks! Jul 14 07:14:54 pH5: It needs fixing properly :) Jul 14 07:18:35 RP: is there anybody working on it? doesn't look like this driver is a very popular playground.. Jul 14 07:19:43 pH5: Nobody I know of it. I needs some time to fix RNDIS, tidy up the machine support , maybe implment autoswitching between ohci and udc and then try and get it upstream... Jul 14 07:24:10 RP: that would be really great. pxa27x_udc, gpio-keys and the pxa_ll_pm wince bootloader suspend/resume stuff are the only things that separate me from the upstream kernel. Jul 14 07:24:49 pH5: Which device is that? Jul 14 07:27:45 RP: htc magician. uses the pxa272 for most things, a nice target to start learning with. Jul 14 07:28:50 pH5: Is the core machine support in mainline? Jul 14 07:32:42 RP: no. all I have so far is the patch at http://userpage.fu-berlin.de/~zabel//magician/magician-2.6.18-rc1-20060710.tar.bz2, and there is an initial version in the hh.org cvs. Jul 14 07:49:05 Can somebody please push the stuff from bug 1117 before OE goes read-only? Jul 14 07:51:42 Laibsch: I compile with -j4 and xproto seems to work fine, what -j setting do you use? Jul 14 07:51:56 -j 3 Jul 14 07:52:03 Maybe it was that icecc stuff. Jul 14 07:52:13 I will try again today. Jul 14 08:00:49 After updating my db, task-boostrap says 'Nothing provides runtime dependency locale-base-en-us'. Why? Jul 14 08:01:02 koen: I also had it that packages would only compile on the second attempt. Whatever the reason. Jul 14 08:01:29 Laibsch: that usually idicates that they are not parallelmake safe Jul 14 08:02:26 I see. I will keep an eye on that next time. Is there a way to override in OE parallel compilation for individual packages? Jul 14 08:03:29 PARALLEL_MAKE = "" inside the .bb Jul 14 08:03:31 morning Jul 14 08:03:53 hey tkp Jul 14 08:04:33 morning Jul 14 08:05:31 koen: Thanks. Jul 14 08:05:38 morning XorA Jul 14 08:06:41 hey XorA Jul 14 08:07:27 wow Jul 14 08:07:39 mb-wm trunk has the dialog resizing feature Jul 14 08:08:24 koen: cool Jul 14 08:10:07 After updating my db, task-boostrap says 'Nothing provides runtime dependency locale-base-en-us'. Why? Jul 14 08:11:40 is there a class that creates a bootable image (besides the bootimg one - which I can't seem to get working properly)? Jul 14 08:11:56 TEMic: we dont like the Americans :-) Jul 14 08:13:09 :-) It wasn't my decision Jul 14 08:14:25 How do I integrate this package? Jul 14 08:15:14 (bootstrap..(r)depends) Jul 14 08:15:14 03arfankai 07org.oe.dev * r0ca021ad... 10/packages/linux/linux-ezx_2.6.16.13.bb: linux-ezx: enable extra 16MB of ram via the cmdline Jul 14 08:15:22 03koen 07org.oe.dev * r77ac9087... 10/packages/linux/linux-ezx_2.6.16.13.bb: linux-ezx: and bump PR.... Jul 14 08:15:47 ah, --author works Jul 14 08:21:04 <|BundaBRG|> Quick question, and I hope its a stupid one ;-). I've just installed bitbake, monotone and the latest OE (org.openembedded.dev). I've copied org.openembedded.dev/conf/local.conf to my build/conf/local.conf. Tried to edit it for a c7x0 (For my c860). It looks similar to http://de.pastebin.ca/45932. Tried 'bitbake nano', and it gets to glibc, faling with an error 'symbol __bind already defined'. What am I doing wrong? Jul 14 08:22:23 have you updated your db? Jul 14 08:22:32 'monotone pull && monotone update' Jul 14 08:22:35 <|BundaBRG|> Yep Jul 14 08:22:41 How do I set the locale thing in my conf? Jul 14 08:22:43 I should have fixed that error yesterday Jul 14 08:23:40 this one: http://www.openembedded.org/viewmtn/revision.psp?id=76c8df086a972bf977e792a3dc937701548faf7a Jul 14 08:23:44 <|BundaBRG|> -=> koen <=- I'll just do a pull and update once more to be sure. Jul 14 08:24:48 <|BundaBRG|> 'monotone --db=OE.db pull monotone.openembedded.org org.openembedded.dev' <-- thats the right syntax Jul 14 08:25:14 that looks OK Jul 14 08:29:19 Hmm. The w100 accel patches look to mean the processor gets to spend a load of time in a busy loop :-( Jul 14 08:29:46 RP: slightly counter productive then Jul 14 08:30:13 XorA: Well, I suspect we don't lose or gain anything except perhaps increased power usage Jul 14 08:30:36 Thankfully only the console is affected Jul 14 08:30:58 <|BundaBRG|> -=> koen <=- I'm having trouble remembering the syntax for monotone update. Error I'm getting is (with just monotone update) is 'monotone: misuse: working copy directory required but not found Jul 14 08:30:59 <|BundaBRG|> ' Jul 14 08:32:26 * TEMic is ignored from all :-( Jul 14 08:33:03 RP: which accell patches? Jul 14 08:33:34 |BundaBRG|: cd org.openembedded.dev ; monotone update Jul 14 08:34:00 <|BundaBRG|> Merci Jul 14 08:42:29 |BundaBRG|: Check with "monotone log|grep 76c8" after doing your pull and update Jul 14 08:42:40 You could also grep for glibc. Jul 14 08:43:07 and rebuild binutils Jul 14 08:43:49 see http://www.openembedded.org/wiki/OeFaq section 2.5 Jul 14 08:44:37 <|BundaBRG|> Thanks. I've rm -Rf'ed my tmp dir, and doing a bitbake nano build again just to be sure I've got a pristine setup. If it errors again I'll follow those leads. Jul 14 08:45:32 Ruby sources for 1.8.2 are no longer available -> bug 1171 Jul 14 08:45:49 koen: indeed, the "culprit" is "inherit icecc" Jul 14 08:45:57 <|BundaBRG|> I have noted I get this error each time I run bitbake, during the scanning process: NOTE: Handling BitBake files: | (1591/3549) [44 %]ERROR: Unable to generate local paths for SRC_URI due to malformed uri: cvs while parsing /usr/src/OE/org.openembedded.dev/packages/linux/handhelds-un-2.6_2.6.16.bb Jul 14 08:48:19 yeah, the handhelds-un bb is broken Jul 14 08:49:57 wtf? Jul 14 08:49:59 SRC_URI = "cvs -d :pserver:anoncvs@anoncvs.handhelds.org:/cvs checkout -d linux-2.6.16-hh2 linux/kernel26" Jul 14 08:50:15 that is soooo wrong in too many ways Jul 14 08:55:07 pH5: could have have a look at the hh-pxa-2.6_cvs.bb? I suspect the versions are wrong Jul 14 08:56:01 koen: The accel patch Mardy added to the kernel Jul 14 09:00:41 koen: HHV needs updating. what is it at now? 4? Jul 14 09:00:56 pH5: no idea Jul 14 09:01:05 * pH5 waits for hh.org cvs... Jul 14 09:02:29 yes, hh4 it is. Jul 14 09:07:36 03koen 07org.oe.dev * re6747840... 10/packages/linux/handhelds-un-2.6_2.6.16.bb: handhelds-un: fix parse-error in SRC_URI, tag might be wrong, leaving that up the to MAINTAINER Jul 14 09:12:08 <|BundaBRG|> Ok same glibc error ;-) Following the FAQ entry now. Jul 14 09:12:46 03pH5 07org.oe.dev * r1fe9f8d4... 10/packages/linux/handhelds-pxa-2.6_cvs.bb: handhelds-pxa-2.6_cvs: bump version to 2.6.16-hh4 Jul 14 09:16:27 kone r u there? Jul 14 09:16:30 koen Jul 14 09:30:48 hey mikearthur Jul 14 09:30:57 hey Jul 14 09:31:08 I've just been given a Sharp PDA Jul 14 09:31:14 mikearthur: which one Jul 14 09:31:25 Tosa Jul 14 09:31:30 looks like a tricorder Jul 14 09:31:41 lucky bugger Jul 14 09:31:54 it's the project one, apparently Jul 14 09:33:12 morning all Jul 14 09:33:17 wheeee Jul 14 09:33:18 http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=a9da396a37fb3da5d0f212c41c6e96bcbf8fe590 Jul 14 09:33:37 koen: did you send me you address ? Jul 14 09:34:05 hey Liam Jul 14 09:34:07 lrg: will do now Jul 14 09:34:53 koen: I have NBP and NBP pcb for you. You can keep the working pcb, but can long term borow the NBP. iirc both have 256MB RAM Jul 14 09:34:54 * koen runs to the station Jul 14 09:35:07 lrg: ah, cool Jul 14 09:35:16 * koen really gone now, back on saterday Jul 14 09:35:26 cu Jul 14 09:36:10 do13: I hear that you have a Tosa? Jul 14 09:37:29 lrg: I added that extra debug, and I either screwed up or those if statements are not entered Jul 14 09:39:30 XorA: interesting. I'll add some extra debug and release pre4 today. Been swamped with support all week :( Jul 14 09:39:59 lrg: is cool, I had serious bitbake problems monday/tuesday then swamped at work rest of week so Ive been slow on testing Jul 14 09:47:11 hi lrg Jul 14 09:47:19 hey Richard Jul 14 09:47:44 lrg: I have a patch for a new machine for you :) Jul 14 09:48:03 RP: excellent :) what machine ? Jul 14 09:48:12 lrg: http://www.rpsys.net/openzaurus/patches/poodle_audio-r0.patch Jul 14 09:48:34 lrg: I presume you saw the cc'd email? Jul 14 09:49:47 RP: I'm not sure, I've been firefighting a lot of customer issues this week. It's all been a daze :( Jul 14 09:50:32 lrg: http://openzaurus.org/wordpress/2006/07/12/sound-support-for-poodle26/ Jul 14 09:51:08 lrg: Feel free to merge the ASoC parts of that patch - I'll take care of fixing up locomo Jul 14 09:51:28 The locomo function call names will change but the arguments will not Jul 14 09:51:39 RP: Is there already a device using uda1380 with ASoc support? Jul 14 09:51:53 pH5: No, someone needs to test it Jul 14 09:52:30 RP: wiil do, many thanks :) btw, we now almost have AT91RM9200 platfrom support Jul 14 09:52:45 lrg: excellent Jul 14 09:53:00 lrg: poodle shows the same external speaker bug as corgi btw (I think anyway) Jul 14 09:54:42 RP: ok, will send XorA a lot of debug in the pre4. Hopefully, we will crack it Jul 14 09:55:42 lrg: I was really pleased with the way poodle basically "just worked" :) Jul 14 09:57:03 :) Jul 14 10:00:08 RP: my udc suspend/resume problem really was just the wrong UP20CR setting Jul 14 10:00:23 pH5: Great :) Jul 14 10:01:13 hey liam Jul 14 10:01:32 mikearthur: yes I have a tosa Jul 14 10:01:53 can I flash through USB? Jul 14 10:02:16 mikearthur: I haven't used this method Jul 14 10:02:26 what do you use, CompactFlash? Jul 14 10:02:30 yep Jul 14 10:02:58 03mickeyl 07org.oe.dev * r0b32c326... 10/packages/tslib/tslib/htcuniversal/tslib.sh: tslib.sh(htcuniversal) remove opie specifics Jul 14 10:03:57 mikearthur: for kernel testing kexec is very convenient Jul 14 10:04:08 morning folks Jul 14 10:04:18 anyone know offhand how to handle discontiguous memory in machine fixup? Jul 14 10:04:27 hey mickey|office Jul 14 10:04:34 yeh, I need to flash the non-sharp image to it though, and don't, currently, have a CompactFlash or wireless :s Jul 14 10:05:25 that's bad. Do you have a sd card? Jul 14 10:09:50 I don't Jul 14 10:09:52 Liam might :) Jul 14 10:10:02 I have one now Jul 14 11:19:42 03florian 07org.oe.dev * r41ea97e1... 10/packages/linux/ (handhelds-un-2.6_2.6.16.bb handhelds-un-2.6_2.6.16-hh4.bb): handhelds-un-2.6: Rename to include extra version. Some major fixes. Jul 14 11:24:15 can the bitbake svn fetcher only fetch annomously? Jul 14 11:29:31 http://www.vanille.de/sources/config_cvs.savannah.gnu.org_20050701.tar.gz (missing) Jul 14 11:30:07 There are a couple of patches in the bug tracker that I would like to see pushed before OE goes read-only: 1114 (does not compile, though), 1117 (my last two patches) and 1171. Can somebody please do it? Thank you. Jul 14 11:36:51 Laibsch: I'll commit 1114 and fix 1171. What about poboxserver? Didn't it compile once? Jul 14 11:37:30 pH5: I never got it to compile. But I have always been struggling with just getting OE to run ;-) Jul 14 11:37:44 pH5: Thanks for fixing it. Jul 14 11:37:54 it=1114 and 1171 Jul 14 11:38:47 pH5: re 1114, nunome is of course already in OE. That patch has been incorporated. Jul 14 11:40:08 pH5: What about 1117? That one does compile fine. Jul 14 11:43:32 Laibsch: I meant 1117 when I wrote 1114, actually. Do the packages work? Jul 14 11:45:09 I never figured out how to activate them. Jul 14 11:45:14 They did install fine. Jul 14 11:47:59 03kristoffer 07org.oe.dev * r78497884... 10/ (45 files in 8 dirs): Jul 14 11:47:59 Moving linux-jlime-${ARCH}-* -> linux-jlime-${machine}-*: Jul 14 11:47:59 * Changing from arch->machine due to the addition of 56x support. This Jul 14 11:47:59 requires different kernel but since both 72x and 56x are arm we need to Jul 14 11:47:59 use MACHINE to define instead. Jul 14 11:48:00 conf/local.conf : Changing ARCH -> MACHINE. Jul 14 12:00:28 03pH5 07org.oe.dev * r9f029b90... 10/packages/ruby/ (4 files): Jul 14 12:00:28 ruby: add 1.8.4, remove 1.8.2 Jul 14 12:00:28 - 1.8.2 is gone upstream, closes bug 1171 Jul 14 12:17:10 any ideas of http://pastebin.ca/87774 during a fresh openzaurus-unstable build? Jul 14 12:23:23 03bugs.openembedded.org 07org.oe.dev * rbf485b02... 10/packages/uim/ (4 files in 3 dirs): uim: packaging updates by Tomoaki Takebayashi, closes bug 1117 Jul 14 12:29:44 Luke-Jr: what does file say? Jul 14 12:29:46 bugs.openembedded.org ??? Jul 14 12:30:56 zecke: which file? the x86_64 lib that shouldn't be used? Jul 14 12:37:26 who own the commit address bugs.openembedded.org? Jul 14 12:37:44 mickey|office: I had to try --author. I guess this email address is a bit misleading if you don't see the domain part.. Jul 14 12:38:12 ah Jul 14 12:38:21 hmm k Jul 14 12:38:31 btw., you left a file Jul 14 12:38:33 i'm adding it Jul 14 12:38:43 please don't commit the anthy stuff, it has issues Jul 14 12:38:56 i added a new comment to the bt Jul 14 12:39:23 or nunome Jul 14 12:39:37 * mickey|office mixes up all those japanese input stuff names Jul 14 12:41:18 mickey|office With the updated qte-common.inc opie-bluepin refuse to compile. When i go back a revision it builds. Can you try it and see if that also happends to you? Jul 14 12:41:38 mickey|office: thanks for fixing it up Jul 14 12:41:42 03mickeyl 07org.oe.dev * r90d48d1b... 10/packages/uim/uim.inc: uim add .inc file that has been left out in the uim commit Jul 14 12:41:55 goxboxlive_: k, let me try Jul 14 12:42:31 my htc universal won't come any time soon btw. Jul 14 12:42:44 i guess i have sent my money to a scammer Jul 14 12:43:13 i am in contact with him now and he promised to send the money back asap :/ Jul 14 12:43:21 i won't bid on another one until i got it back Jul 14 12:44:12 goxboxlive_: opie-bluepin 1.2.2 compiles fine here. what's your error? Jul 14 12:46:09 mickey|office I dont havce the error message anymore since i went back to the last revison . But after i have build some images i can try to build the opie-bluepin again and then let you know. Jul 14 12:55:03 ok Jul 14 12:55:42 note that after all the RTTI changes, it's best to rebuild everything depending on Qt/E from scratch Jul 14 12:55:47 zecke: ...? Jul 14 12:57:58 hm! Jul 14 13:08:43 mickey|office I have rebuilt everything, i started with a fresh .dev toady. I got the fail message again compiling rdesktop. But i am going to collect them and sending you it to you on mail i a txt file. It's easier since i am building over vnc. Jul 14 13:09:51 goxboxlive_: no problems with glibmm? Jul 14 13:11:55 no Jul 14 13:15:52 hey zecke Jul 14 13:15:55 man my english is good, my english teacher would have been real prud of me: "them and sending you it to you on mail" Jul 14 13:16:02 goxboxlive_: ok, do that. => mickeyl@handhelds.org Jul 14 13:16:12 thx sir Jul 14 13:16:21 mickey|office: re bug 1114: Are you sure you ever touched anything to do with pobox? As I wrote in my comment in the BTS, the nunome stuff shall not be committed (and I should not have included it in that report in the first place) Jul 14 13:16:45 zecke: mikearthur has the project tosa now. he has more time than me atm. he will complete touch and battery Jul 14 13:17:07 zecke: wiki has been updated Jul 14 13:17:17 lrg: I'm fine with that, but who is mikearthur? Jul 14 13:17:37 zecke: he's my Padawan learner Jul 14 13:17:51 zecke: student at wolfson ;) Jul 14 13:18:26 ah :) Jul 14 13:20:08 Laibsch: dunno offhand, i touch way too many files. Facts are standing that it contains bogus stuff, at least the patch -p1 Jul 14 13:20:43 Laibsch: also I don't like Jul 14 13:20:45 -DESCRIPTION = "A japanese inputmethod for Qt/Embedded based palmtop environments." Jul 14 13:20:52 +DESCRIPTION = "Japanese input method plugin" Jul 14 13:21:01 that's what i mean with not incorporating my improvements Jul 14 13:21:15 Well, but I think you are talking about the nunome patch which you had committed in 1095 and improved. The pobox stuff was so far never touched, I believe. Jul 14 13:21:17 yes, this may be nitpicking but there's is more than just a description Jul 14 13:21:44 no, I'm talking about http://bugs.openembedded.org/attachment.cgi?id=797&action=view Jul 14 13:21:51 which is in 1114 Jul 14 13:21:53 mickey|office: I try to safe-guard your improvements, but it is not always easy for me. Jul 14 13:22:04 does the bootsplash package work? Jul 14 13:22:18 I thought the kernel needed to be patched to use bootsplash Jul 14 13:22:32 and I can see that it does that in the bootsplash bb file Jul 14 13:22:51 of course not Jul 14 13:22:56 OE is flexible Jul 14 13:23:06 when you chose to include bootsplash, then you should use a patched kernel Jul 14 13:23:15 it's not bootsplash.bb task to patch a kernel Jul 14 13:23:19 it wouldn't even know which one Jul 14 13:23:20 ah, ok Jul 14 13:23:20 heh Jul 14 13:25:02 Laibsch: I'll try to apply 797 and remove the bogus portions Jul 14 13:25:13 Wow, that would be great. Jul 14 13:25:27 ~lart web developers Jul 14 13:25:27 Sorry to make it more cumbersome than it should have been. Jul 14 13:25:27 * ibot throws web developers's poor little doggy off a cliff Jul 14 13:25:50 I was going to throw out stuff that I could obviously spot as deteriorations. Jul 14 13:26:15 I will look more carefully at the stuff in future. Jul 14 13:26:20 are you sure nkf-native is no longer needed? Jul 14 13:26:26 no Jul 14 13:26:35 That was one thing I was going to revert also. Jul 14 13:26:51 I think tota is working from his own tree and not using patches. Jul 14 13:27:10 *nod* that's what I don't like Jul 14 13:27:17 it should be a two way Jul 14 13:27:18 not a fork Jul 14 13:27:21 That is why it would be good to get his stuff in quickly and in small increments rather than the big tar balls he sends. Jul 14 13:27:34 * Laibsch nods Jul 14 13:27:37 certainly. Jul 14 13:27:52 ok, now we have a cross compile badness because of /usr/include Jul 14 13:27:53 nothing serious Jul 14 13:28:19 I will try to get them accustomed to OE style. I think they are shy and too afraid to make mistakes. Jul 14 13:28:27 So please don't yell at 'em ,-) Jul 14 13:28:54 Dealing with Japanese takes time, but once it is rolling I expect many things to come in. Jul 14 13:29:17 I will also try to learn as much as I can as a non-programmer. Jul 14 13:30:29 I will try to teach tota about quilt so he can submit small patches frequently. Jul 14 13:30:38 you won't get me to yell easily... you got to do a lot to make that happen Jul 14 13:30:41 :D Jul 14 13:30:46 uff Jul 14 13:30:55 I can yell :) Jul 14 13:30:58 zecke: he will *hopefully* complete touch and battery :D Jul 14 13:31:03 zecke: hurray! Jul 14 13:32:02 mikearthur: failure would be fatal, so I'm confident you will get it done :) Jul 14 13:32:31 zecke: which file? the x86_64 lib that shouldn't be used? Jul 14 13:32:47 Luke-Jr: could you point me to the pastebin again? Jul 14 13:34:24 mickey|office: do you remember QPF? Was t15.qpf 270 degree rotated? Jul 14 13:35:38 yes Jul 14 13:35:44 5 = 90 degree Jul 14 13:35:49 and so on Jul 14 13:36:13 Packaged contents of poboxserver into /local/pkg/oe/c7x0/tmp/deploy/ipk/poboxserver_1.2.5-r2_armv5te.ipk Jul 14 13:36:20 piece of cake Jul 14 13:36:21 :D Jul 14 13:36:26 mickey|office: I think 10 is 270 degree? Jul 14 13:36:40 10? then i can't follow the TT logic Jul 14 13:36:43 but that wouldn't be something new Jul 14 13:36:45 *cough* Jul 14 13:36:52 isn't that documented on doc.trolltech.com? Jul 14 13:37:10 I had issues with QColor yesterday... but I enjoy hacking QtopiaCore Jul 14 13:37:23 mickey|office: it can be something really promising Jul 14 13:37:37 he zecke Jul 14 13:37:45 damn mac-pc konversion Jul 14 13:37:46 zecke: i believe that. Jul 14 13:37:55 zecke: you remember the project I've talked to you? Jul 14 13:37:56 woglinde: I'm sooo tired Jul 14 13:38:05 zecke: they chose kdrive + GTK+ :/ Jul 14 13:38:58 mickey|office: not directFB+GTK+? Jul 14 13:39:01 zecke: http://pastebin.ca/87774 Jul 14 13:40:11 Luke-Jr: oh okay Jul 14 13:40:31 -L/home/luke-jr/src/oe2005/build.alt/tmp/staging/x86_64-linux/lib Jul 14 13:40:42 * Luke-Jr nods Jul 14 13:40:43 -L/home/luke-jr/src/oe2005/build.alt/tmp/work/armv5te-linux/gcc-cross-4.1.1-r5/gcc-4.1.1/build.x86_64-linux.arm-linux/ Jul 14 13:40:58 Luke-Jr: now let us find the evil .pc or .la file Jul 14 13:41:20 Luke-Jr: so take a look at config.log and search x86_64 Jul 14 13:41:50 staging/arm-linux/lib/libsigc-2.0.la:dependency_libs='... Jul 14 13:41:59 * Luke-Jr used grep o.o Jul 14 13:42:49 Laibsch: the qpobox changes from toto are mostly bogus Jul 14 13:43:15 I can apply the patch to the code though Jul 14 13:43:28 03mickeyl 07org.oe.dev * r08ecc189... 10/packages/poboxserver/ (3 files in 2 dirs): poboxserver 1.2.5 update and make it compile. patch based on Tota's and Laibsch's work as detailed in #1114 Jul 14 13:45:57 mickey|office: I see. Jul 14 13:46:47 mickey|office: yes t10 is 270 degree Jul 14 13:47:03 i.e. _never_ ever point to an unversioned tarball in a versioned bb Jul 14 13:47:11 that's why i mirrored it under a new name Jul 14 13:47:16 zecke: ok. odd anyway Jul 14 13:47:46 also, don't DEPEND on something when it's ok to RDEPEND Jul 14 13:47:57 Packaged contents of qpobox into /local/pkg/oe/c7x0/tmp/deploy/ipk/qpobox_0.5.4-r0_armv5te.ipk Jul 14 13:48:40 I thought RDEPEND was "stricter" than DEPEND? Jul 14 13:48:57 RDEPEND: compile it and install it at runtime Jul 14 13:49:03 zecke: so any ideas? Jul 14 13:49:04 DEPEND: compile it Jul 14 13:49:31 RRECOMMENDS? Jul 14 13:49:39 mickey|office: Am I mistaken? Jul 14 13:49:58 Laibsch: RDEPEND is actually just "install it at runtime" Jul 14 13:50:08 it doesn't imply compile unless you set a flow Jul 14 13:50:09 Luke-Jr: more than looking into config.log, *.pc and *.la - no sadly not Jul 14 13:50:10 flag* Jul 14 13:50:25 zecke: ? Jul 14 13:50:29 Hm, I remember seeing an RDEPENDS being compiled which is why I thought it was stricter. Jul 14 13:50:42 Laibsch: task-* set the flag Jul 14 13:51:19 RDEPENDS being compiled is just a matter of having them present when you make an image Jul 14 13:51:35 that doesn't mean they're a compile time necessity Jul 14 13:52:07 * mickey|office closes 1114 Jul 14 13:52:13 and that's about all i can do for OE today Jul 14 13:52:14 l8er Jul 14 13:52:29 later Jul 14 13:52:33 mickey_away Jul 14 13:52:36 later and thank you mickey_away. Jul 14 13:52:46 y*sum* Jul 14 13:53:12 Luke-Jr: I have not understood it 100%. But I think I somehow get the idea. Jul 14 14:01:47 !!! Jul 14 14:01:48 Luke-Jr: I have no frelling idea! Jul 14 14:04:39 so bitbake was building glibmm when it errored... now I tell it to build the same providee, and it doesn't even try glibmm... Jul 14 14:04:47 it doesn't try any dependencies, in fact Jul 14 14:05:09 is this going to work? :\ Jul 14 14:06:08 I don't get what your saying Jul 14 14:06:26 I don't understand your refusal of using grep as well.. Jul 14 14:06:36 *g* Jul 14 14:06:49 zecke: after I tell bitbake to build gpe-image again, it goes straight to gpe-image and skips gtkmm and whatever else it was initially trying Jul 14 14:06:53 what refusal to use grep? Jul 14 14:07:49 see above. You have a failing package with a bogus -L and you don't want to use grep to find the sucker Jul 14 14:07:54 anyway need to work Jul 14 14:07:59 ... Jul 14 14:08:24 I found the sucker using grep... and never complained about using grep... wtf? Jul 14 14:11:21 bye Jul 14 14:15:00 sigh... guess I'll just need to manually build my dependencies Jul 14 14:21:11 Luke-Jr, the stamps exist signalling to BB that the package exists Jul 14 14:21:35 you need to remove the stamps and work dirs manually or use the clean command Jul 14 14:23:06 emte: then why does BB build the package from scratch if I request it specifically? Jul 14 14:23:20 this behaviour i belive is a side-effect of world, where you want to continue and just skip/ignore the errored package Jul 14 14:23:32 because your overriding it Jul 14 14:23:56 luke-jr@tsurukikun tmp $ ls stamps/*/gaim* Jul 14 14:23:56 ls: stamps/*/gaim*: No such file or directory Jul 14 14:24:18 try ls -R stamp |grep gaim Jul 14 14:24:43 what is the purpose of this directory: tmp/staging/x86_64-linux/share/pkgconfig/ ? Jul 14 14:24:46 it uses long names for stamps Jul 14 14:25:08 luke-jr@tsurukikun tmp $ ls -R stamps/ | grep gaim Jul 14 14:25:09 luke-jr@tsurukikun tmp $ Jul 14 14:25:16 build-time library Jul 14 14:25:49 well Luke-Jr it looks like gaim was never even started Jul 14 14:25:57 I know it wasn't Jul 14 14:25:59 that's the problem Jul 14 14:26:02 which makes me thing a requ is missing Jul 14 14:26:14 it's not just gaim either Jul 14 14:26:17 think* Jul 14 14:26:29 it's gpe-bootsplash, bluez-utils-dbus, matchbox, and a whole list of maybe 50+ more Jul 14 14:26:49 sounds like you didnt actually build your image Jul 14 14:26:51 after the first build errored out, when I told it to build the image again it skipped everything Jul 14 14:27:15 how big is your work dir? Jul 14 14:27:36 du -sh tmp Jul 14 14:27:44 tmp or tmp/work? Jul 14 14:27:50 either Jul 14 14:28:24 * Luke-Jr waits for it to finish counting Jul 14 14:28:32 you didnt use -s? Jul 14 14:28:38 yes Jul 14 14:28:41 it still takes time Jul 14 14:28:53 not usually Jul 14 14:29:03 maybe you have a slow drive Jul 14 14:29:10 'build gpe-image-ljr' 'build gpe-image-ljr' Jul 14 14:42:33 done yet? Jul 14 14:48:39 Hello Jul 14 14:48:51 Hello, sirfred Jul 14 14:55:00 03mickeyl 07org.oe.dev * r64e3a9ae... 10/packages/qpobox/ (4 files in 2 dirs): qpobox: update Tota's code patches, but keep my .bb changes intact. closes #1114 Jul 14 14:55:04 03mickeyl 07org.oe.dev * re94e511b... 10/conf/distro/ (familiar-unstable.conf openzaurus-unstable.conf): familiar-unstable.conf: include fixed cvs dates Jul 14 14:55:09 03mickeyl 07org.oe.dev * re466ce38... 10/conf/distro/ (65 files in 2 dirs): Jul 14 14:55:09 conf/distro cleanup. conf/distro should only contain files that are ok to be directly included Jul 14 14:55:09 by means of setting DISTRO. conf/distro/include should contain those which doesn't. Jul 14 14:57:58 03mickeyl 07org.oe.dev * rb2503585... 10/conf/distro/ (debianslug.conf openslug.conf ucslugc.conf): slugos distro configurations: fix includes Jul 14 15:00:36 would someone mind giving me a hand with some linking issues please? Jul 14 15:01:23 damn CFLAGS are screwed again Jul 14 15:01:48 and Im'm getting CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/directfb Jul 14 15:02:00 I have no idea why it's trying to look in /usr/include Jul 14 15:02:13 and I don't know how to make it stop Jul 14 15:02:26 look at the makefile and see if its hardcoded Jul 14 15:02:50 if it is it will need ot be patched Jul 14 15:02:54 to* Jul 14 15:03:01 is that quite common then? Jul 14 15:03:10 very Jul 14 15:04:06 remember diff -Nurd is your friend Jul 14 15:04:27 includedir = /usr/include Jul 14 15:04:44 but the Makefile only exists after I have run do_configure Jul 14 15:04:50 (oe_runconf) Jul 14 15:05:06 look at your confinure or your Makefile.in/am Jul 14 15:05:15 configure* Jul 14 15:05:26 not or, and Jul 14 15:06:09 in configure, it is set to: includedir='${prefix}/include' Jul 14 15:06:22 and ther is also: oldincludedir='/usr/include' Jul 14 15:07:36 no, it's not hard coded in either of those files Jul 14 15:07:58 would I need to pass ssome CFLAGS? I thought oe_runconf would do it all for me Jul 14 15:08:35 http://rafb.net/paste/results/ZMhyQq91.html Jul 14 15:09:10 somehow -I /usr/include/directfb is being passed to ccache Jul 14 15:12:47 most common in Makefile is -I${includedir} being hardcoded Jul 14 15:13:53 emte: oops, heh yeah Jul 14 15:14:00 emte: 3.0 GB Jul 14 15:14:01 in Makefile.in, it's set to includedir = @includedir@ Jul 14 15:14:19 tkp: unset CPATH? Jul 14 15:15:22 nope :( Jul 14 15:15:25 this is what I have: http://rafb.net/paste/results/sMRSuP52.html Jul 14 15:17:49 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Jul 14 15:21:22 when oe_runconf runs, it seems to set --includedir=/usr/include. Is that normal? Jul 14 15:22:10 actually, thats todo with the target system I think Jul 14 15:31:06 guess I'll just go ahead and build dependencies manually then! Jul 14 15:36:03 grr. damn CFLAGS :( Jul 14 15:37:32 http://pastebin.ca/88141 Can someone tell me what might be going on? This is what I get when do_rootfs fails. Jul 14 15:38:03 Wow - thats a new one Jul 14 15:39:20 cinix: thats a corrupt ipk Jul 14 15:39:29 anyway time to go Jul 14 15:39:33 thanks Jul 14 15:48:10 I extract the ipk with dpkg -x and it seems to be intact. It's also the very first package that just happens to fail, seems too coincidental. Jul 14 15:57:43 hi Jul 14 15:59:34 Hi CoreDump|home Jul 14 15:59:40 is orinoco-modules_0.15rc supposed to build? It borks my poodle images Jul 14 15:59:45 hi hvontres|poodle Jul 14 16:00:49 CoreDump|home: Not sure..I think I took the wlan stuff out of my poodle.conf file, since I don't have a card... Jul 14 16:01:05 hvontres|poodle: heh Jul 14 16:01:51 CoreDump|home: Do you know where the 2.4 kernels hide their keymap? I am trying to figure out what the "OK" key needs to map to to work right in opie Jul 14 16:02:08 Apparently it is not "return" :( Jul 14 16:02:57 well IIRC it is mapped to KBD_ENTER or something like that. w/ kernel 2.4 the keymap is in the kernel sources somewhere Jul 14 16:03:47 CoreDump|home: I think the orinoco_modules should come from the kernel, so you should be able to get rid of them as a sepearate dependancy... you might want to chek with RP or hrw on that tho Jul 14 16:04:26 well it is not built for akita, so I just removed the dependency Jul 14 16:05:17 CoreDump|home: Is there some special trick to the the vt switching to work in opie? If I try to use arrow form the console window it works, but when I am at the opie vt it does not. :( Jul 14 16:05:53 opie problem, works fine w/ GPE Jul 14 16:06:01 ~lart opie Jul 14 16:06:02 * ibot executes killall -TERM opie Jul 14 16:06:37 I guess it is time to build a gpe image this weekend...:) Jul 14 16:07:08 yep, you might wanna fix up the hotkey mappings in GPE Jul 14 16:12:59 ~seen hrw Jul 14 16:13:11 hrw was last seen on IRC in channel #openzaurus, 3d 17h 53m 53s ago, saying: 'http://www.hrw.one.pl/2006/07/10/tosa-and-26-kernel/'. Jul 14 16:18:39 ~seen hrw|gone Jul 14 16:18:43 hrw|gone was last seen on IRC in channel #oe, 4d 8h 3m 55s ago, saying: 'http://www.linuxdevices.com/news/NS6437147703.html'. Jul 14 16:19:07 * Kerwood frowns Jul 14 16:20:09 well, I'm glad Marcin finally got a break Jul 14 16:28:08 Kerwood: I think his ISP is on the fritz :( Jul 14 16:33:24 * JustinP wonders what the problem is Jul 14 16:33:50 Kerwood, hvontres|poodle: if you need a newer image I can build one...I just built e-images for all of the machines so it woudln't take long Jul 14 16:36:56 JustinP: Even Poodle??? Jul 14 16:36:58 hvontres|poodle: no, hrw will be on holiday through the end of July Jul 14 16:37:08 hvontres|poodle: yeah Jul 14 16:37:19 Kerwood: Ahh, good for him :) Jul 14 16:37:21 hvontres|poodle: would you like to try it? Jul 14 16:37:38 JustinP: With 2.6? Jul 14 16:38:04 ~angstrom Jul 14 16:38:05 [angstrom] a unit of length (0.1 nm), or a Swedish physicist (Anders Jonas Angstrom) Jul 14 16:38:15 hvontres|poodle: I just thought you were checking to see if he visited in the morning (as he said he might) Jul 14 16:38:50 ibot: no, angstrom is [angstrom] a unit of length (0.1 nm), or a Swedish physicist (Anders Jonas Angstrom), or a new distribution for handhelds utilizing EABI Jul 14 16:38:51 okay, JustinP Jul 14 16:39:04 ~angstrom Jul 14 16:39:07 i guess angstrom is [angstrom] a unit of length (0.1 nm), or a Swedish physicist (Anders Jonas Angstrom), or a new distribution for handhelds utilizing EABI Jul 14 16:39:10 ibot: botsnack Jul 14 16:39:11 JustinP: aw, gee Jul 14 16:39:12 Anyway, I'm outta here Jul 14 16:39:26 Kerwood: Actually, I was just gonig to ask him to add irssi to the feeds Jul 14 16:39:44 ah Jul 14 16:39:47 ~eabi Jul 14 16:39:59 hvontres|poodle: looks like my build was for the 2.4 kernel. I can try building 2.6 if you like. Will pretty much only take rebuilding the kernel Jul 14 16:41:09 JustinP: That would be cool. I have 2.6 running ritht now...:) Trying to see where I want to go now, stick wiht opie, try gpe or maybe e ..:) Jul 14 16:41:20 ^_^ Jul 14 16:41:36 ~hail RP Jul 14 16:41:37 * ibot bows down to RP and chants, "I'M NOT WORTHY!!" Jul 14 16:42:50 hvontres|poodle: is it KERNEL_VERSION = "2.6"? Jul 14 16:43:51 I've used that before, just want to make sure it works for poodle Jul 14 16:44:07 hvontres|poodle: you looking for .dev or .oz354x? Jul 14 16:44:12 JustinP: I believe so. I would need to check my configs at home to be sure... Jul 14 16:44:21 will binaries created for OZ and other systems run in Angstrom? Jul 14 16:44:28 or only EABI binaries? Jul 14 16:44:44 raduga: no Jul 14 16:44:48 raduga: OZ is not using EABI Jul 14 16:44:55 raduga: Angstrom will be the first Jul 14 16:44:55 justinp: i understand that. Jul 14 16:45:12 justin: question was, "will non-eabi bins run in angstrom?" Jul 14 16:45:21 I don't think so Jul 14 16:45:25 ok. Jul 14 16:45:42 JustinP: I am currently running .oz354x, but the .dev should also work. I was using the .dev 2.6.17 stuff until it moved to .oz354x Jul 14 16:46:05 hvontres|poodle: is the 2.6 stuff for poodle in oz354x? (I built oz354x, not dev) Jul 14 16:46:15 hvontres|poodle: it's ZKERNEL_VERSION = "2.6" for poodle it seems Jul 14 16:48:49 is there a wiki, or site, or mailinglist i can ping, to learn more about Angstrom? Jul 14 16:48:59 or just browse oesf :/ Jul 14 16:49:11 * raduga attempting to avoid asking repeated annoying questions in #oe Jul 14 16:52:58 JustinP: for poodle the default is still 2.4 Jul 14 16:53:13 thanks (i think i found what i needed) Jul 14 16:54:40 CoreDump|home: I assumed it was, which was why I needed the KERNEL_VERSION(s) Jul 14 16:55:15 CoreDump|home: but is the newest poodle 2.6 in .oz354x? Jul 14 16:55:36 JustinP: yeah, (2.6.17) Jul 14 16:56:02 2.6.17 in .oz is default_pref -1 for all but poodle Jul 14 16:56:21 ok Jul 14 16:56:25 will try building, then Jul 14 17:03:25 Can anyone tell me what the procecure is if I want to do my own distribuiton that is similar too, but not exactly like one of the existing distributions? i.e I need to modify the kernel defconfig as well as including additional modules... Jul 14 17:16:14 n7ipb: you can just edit a defconfig in OE to change it and make a copy of whatever image bb you want Jul 14 17:16:34 n7ipb: you could also make a copy of the kernel bb and defconfigs/patches so that you're not editing OE stuff Jul 14 17:16:54 n7ipb: or make a new kernel bb which includes the one you want to use Jul 14 17:18:49 JustinP: The problem is that I want to distribute this to others as well. I've already done a modified version of the defconfig for my testing, but I'm a little reluctant to be handing out patches to existing files to people and saying "Here, copy this to this location etc." Jul 14 17:20:03 jo zecke Jul 14 17:21:00 n7ipb: you could edit locally and commit to a local branch, then open up your monotone db for people to pull from Jul 14 17:21:35 n7ipb: I suggest committing to a local branch either way, but there's another way without editing files... Jul 14 17:22:05 JustinP: hey Jul 14 17:22:31 JustinP: you're right, a local branch is still a good idea, but I'm looking for a way that might become part of the main. Jul 14 17:22:56 n7ipb: make a copy of the image you want to modify (or include the bbfile and make adjustments to vars in yours) then make a bbfile for the kernel which includes the kernel you want to modify and add an _append to the appropriate task which replaces/patches the defconfig and sets the patch dir to the old kernel's Jul 14 17:23:07 n7ipb: well....what are your defconfig changes? Jul 14 17:23:18 n7ipb: perhaps you could define a new machine and make your changes for that machine instead? Jul 14 17:23:32 zecke: hey Jul 14 17:23:59 JustinP: I hope you didn't feel commanded or such :} Jul 14 17:24:30 zecke: commanded? Jul 14 17:24:55 JustinP: My stuff is almost identical to the openslug distribution, but with the kernel ham radio support turned on in defconfig, the reiserfs stuff left out of the image and the support tools, apps and libraries for ham radio included in the default image. Jul 14 17:24:56 JustinP: oh well, just another language issue on my side Jul 14 17:25:37 n7ipb: could you make the ham stuff modules? perhaps that could make it into the normal defconfig... Jul 14 17:25:46 zecke: I'm not sure what you're referring to Jul 14 17:26:28 JustinP: Most of it I've been able to do with local.con, and the creation of some bb files for the ham applications and tools (which I plan on submitting to oe soon) Jul 14 17:27:02 JustinP: most of the ham stuff is configured as modules, with the exception of the core network support... Jul 14 17:27:20 JustinP: I referred to the bugreport on bitbake, and the default setting to make it private Jul 14 17:27:53 n7ipb: well, you could ask the kernel maintainer if your changes are ok...or perhaps an option could be set at add a patch that enables it..? Jul 14 17:28:06 zecke: ah, that. Sorry about that, I just left the default Jul 14 17:29:19 JustinP: For my own use and partly as a learning tool I created my own distribution (hamslug) but still didn't know what to do a bout the kernel... I'll ask the kernel maintainer... Jul 14 17:32:37 a var which when set sets en extra defconfig patch or something would be my best advise Jul 14 17:32:52 but if the core support is small enough the kernel maintainer may just do it Jul 14 17:33:26 * mrz80 is back (gone 02:15:38) Jul 14 17:33:50 JustinP: personaly I don't think the core support is very large at all, but then again I would... :-) Jul 14 17:37:18 JustinP: Thanks for the help... Jul 14 17:38:49 omg Jul 14 17:39:01 it's been months and sqlite's bb still isn't fixed >_< Jul 14 17:39:35 luke-jr Jul 14 17:39:36 hm Jul 14 17:40:06 Luke-Jr: I would ask my support contractor ;) Jul 14 17:40:46 zecke maybee I should ask for a oe account Jul 14 17:40:50 an Jul 14 17:41:18 are those handed out? maybe I should too Jul 14 17:41:24 then I can just fix this stuff myself =p Jul 14 17:41:48 luke-jr thats sounds familiar in opensource projekts Jul 14 17:41:55 hi katossi Jul 14 17:41:59 Luke-Jr: what is that sqlite bug? Jul 14 17:42:12 zecke: it uses the host's pointer size instead of the target's Jul 14 17:42:13 Luke-Jr: I feel sort of responsible for sqlite3 Jul 14 17:42:38 main.mk implies cross-compilers are supposed to create config.h on their own with the correct value Jul 14 17:42:55 #define SQLITE_PTR_SZ 4 Jul 14 17:43:00 4 for Zaurii anyhow Jul 14 17:43:08 Luke-Jr: and other platforms? Jul 14 17:43:14 no clue, sizeof(char *) Jul 14 17:43:17 Luke-Jr: if not autoconf, how should one know? Jul 14 17:43:30 how should one know the ptr size of the target Jul 14 17:43:34 no idea Jul 14 17:43:41 I assume OE must have some means Jul 14 17:44:03 maybe the cross-GCC can return it Jul 14 17:46:29 I just know x86_64_sizeof(void*) != armv5te_sizeof(void*) Jul 14 17:48:50 zecke: BTW... I found the 'sucker' using grep... and never complained about using grep... wtf? Jul 14 17:49:02 Luke-Jr: yes? Jul 14 17:49:17 Luke-Jr: who was the sucker? Jul 14 17:49:42 Luke-Jr: BTW: my arm site has a ac_cv_sizeof_void_p so sqlite should use that Jul 14 17:50:10 Jul 14 14:01:23 staging/arm-linux/lib/libsigc-2.0.la:dependency_libs='... Jul 14 17:50:49 Luke-Jr: awesome, now one just need to fix it Jul 14 17:51:08 curious that sqlite uses autoconf, but not for pointer size... Jul 14 17:51:21 I wonder why Jul 14 17:51:45 I hacked the .la manually-- not sure how libsigc's bb would need fixing Jul 14 17:52:15 I don't know much of build scripting stuff-- certainly not whatever makes those dumb .la files Jul 14 17:52:26 Luke-Jr: libtool Jul 14 17:52:51 eh, well libtool then seems from my experience to be more trouble than it solves ;p Jul 14 17:53:16 right Jul 14 18:09:34 hvontres|poodle: what are you using to set the kernel version for your poodle? Jul 14 18:20:40 * JustinP hacks the 2.6.17 bbfile Jul 14 18:37:23 JustinP: it should beused automatically as this version is hard-wired in poodle-2.6.conf Jul 14 18:45:08 CoreDump|home: it's not... Jul 14 18:45:32 CoreDump|home: I don't see a PREFERRRED_VERSION in there... Jul 14 18:46:00 PREFERRED_VERSION_linux-openzaurus = "2.6.17" Jul 14 18:46:05 CoreDump|home: just a PREFERRED_PROVIDER. I tried setting the PREFERRED_VERSION in my local.conf but is didn't take Jul 14 18:46:23 ah...I had it backwards Jul 14 18:46:29 it's in poodle-2.6.conf Jul 14 18:46:31 but it's not in the poodle-2.6.conf.... Jul 14 18:46:39 hmm Jul 14 18:50:04 CoreDump|home: I am definately up to date.... Jul 14 18:51:14 well, now that I think of it,it may just be something I added heh Jul 14 18:52:04 yep, sorry Jul 14 18:52:12 lol Jul 14 18:52:53 ;) Jul 14 18:55:33 why does the curl BB have IPv6 and SSL disabled? Jul 14 18:55:41 kinda restricts the use... Jul 14 19:04:30 Luke-Jr: probably onlylinks with openssl or something... Jul 14 19:04:41 CoreDump|home: how about orinoco-modules? it failed on me Jul 14 19:05:06 JustinP: yeah, well it needs it for gpe-mini-browser to do https Jul 14 19:05:22 JustinP: same here, I removed them Jul 14 19:05:25 ah....I didn't realize that was what was stopping it... Jul 14 19:06:55 same with IPv6, I assume Jul 14 20:04:24 AS woglinde pointed me to: "http://www-128.ibm.com/developerworks/library/l-dist2.html" Jul 14 20:04:31 we should try harder to push patches Jul 14 20:05:50 zecke: so what of sqlite and that c++ lib? =p Jul 14 20:07:22 Luke-Jr: still busy working Jul 14 20:09:07 Luke-Jr: I'm tempted to say "Use the source, Luke" Jul 14 20:09:10 but I don't Jul 14 20:09:29 =p Jul 14 20:09:46 well, I'd add a 'echo "..." >config.h' for sqlite =p Jul 14 20:09:57 but something tells me you don't want that method Jul 14 20:10:04 Luke-Jr: right Jul 14 20:10:20 Luke-Jr: open an autoconf manual, see how one checks for sizes of types Jul 14 20:10:43 and then update sqlite to use that instead of custom code Jul 14 20:12:34 hvontres|poodle: NOTE: package gpe-image-1.0-r22: task do_rootfs: started ^_^ Jul 14 20:12:52 hvontres|poodle: no orinoco modules, the build failed so I removed it Jul 14 20:20:31 so build them Jul 14 20:20:49 Luke-Jr: wh? Jul 14 20:20:54 Luke-Jr: err....eh? Jul 14 20:22:58 build orinoco-modules? Jul 14 20:23:00 O.o Jul 14 20:26:48 Luke-Jr: it fails....as I said... Jul 14 20:29:21 hvontres|poodle: NOTE: package gpe-image-1.0: completed Jul 14 20:31:29 hvontres|poodle: privmsg ok? you registered? Jul 14 20:38:27 Kerwood_: fyi, I have some poodle 2.6.17 images brewing if you'd like them Jul 14 20:45:59 is there a regression test procedure/policy for bumping packages? Jul 14 20:50:03 zwelch: rule of thumb at best Jul 14 20:50:16 zwelch: you mean upgrading a package to a newer release? Jul 14 20:50:39 zecke: yeah. i want to bump openssl, ffmpeg, and a bunch of gtk related stuff Jul 14 20:50:41 anyone experienced with va_args? Jul 14 20:50:54 zwelch: check with the maintainers of these files Jul 14 20:51:13 zwelch: Currently it is: Make sure it works Jul 14 20:51:41 well, koen seems to have tried to delegate gtkmm/glademm maintenence to me, though i don't think he actually "authorized" me in any real capacity Jul 14 20:51:48 zwelch: look for anything that DEPENDS or RDEPENDS on them, built it, test it Jul 14 20:52:17 JustinP: i can't afford the time to use such a brute force process Jul 14 20:52:44 zwelch: :-P then push it as a new bbfile with DEFAULT_PREFERENCE = "-1" Jul 14 20:52:54 zwelch: then be glad this version is present Jul 14 20:52:57 and test as you have time (or get others to help) Jul 14 20:53:13 this is why i was hoping there is some kind of remedial automation that i can run it through to at least give me a basic idea that i haven't done something stupid Jul 14 20:53:37 i.e. something that automatically scans for its dependenants and builds them for you Jul 14 20:53:41 grep -r & build....that's about it AFAIK Jul 14 20:53:49 zecke: va_args is pretty simple... o.O Jul 14 20:53:49 I don't think we have anything like that, no Jul 14 20:53:51 zwelch: well building is one thing Jul 14 20:54:05 sure it is Jul 14 20:54:05 Luke-Jr: can you call va_start and va_end multiple times? Jul 14 20:54:09 and, yes, that's only half of the solution... but it is the first half Jul 14 20:54:27 (maybe even more since you have 1) identify, 2) build, and 3) test) Jul 14 20:54:41 so call it 52% ;) Jul 14 20:54:42 zecke: sure Jul 14 20:54:55 zecke: doesn't an example do that? o.o Jul 14 20:55:48 urm, re: va_args... in what context would you want to use start//end multiple times? Jul 14 20:56:24 multiple vprintfs? Jul 14 20:56:27 do you mean within a single function, or calling a varargs based function within another varargs based function Jul 14 20:56:32 method(foo,...) {va_list l; for_each(oberser : Observers) observer.method(fmt,va_list) Jul 14 20:56:44 oh, yeah, that works fine Jul 14 20:56:47 multiple vprints using the same arguments Jul 14 20:57:13 i am pretty sure that i've written stuff like that in the past Jul 14 20:57:17 zwelch: Building is one thing. Testing it works is more tricky :( Jul 14 20:57:28 zwelch: e.g. there is a Unified Linux testing project, but we do not make use of it Jul 14 20:57:29 well, testing is something that you need everyone helping too Jul 14 20:57:48 regression testing automation would be sweet, if the qemu stuff really gets solid Jul 14 20:57:59 zwelch: e.g. OpenSSL. You need to make sure it is not again binary incompatible Jul 14 20:58:08 zwelch: or Gtk+, there might be users not wanting cairo Jul 14 20:58:16 heh Jul 14 20:58:23 zwelch: so it is more a social question of properly upgrading packages :) Jul 14 20:58:41 re binary compatibility, i am happy to see these deployed with the next major release Jul 14 20:58:50 i.e. dist-upgrade (or equivalent) Jul 14 20:59:21 so glibmm will be fixed? =p Jul 14 20:59:28 and yeah, i did not expect the policy/process for bumping to be a purely techinical one; users have non-technical requirements of their packages ;) Jul 14 20:59:41 Luke-Jr: yes, that's my plan... but as i said... i built it here recently Jul 14 20:59:59 i just nuked my whole tmp and have been building up from scratch again Jul 14 21:00:01 zwelch: well its trying to link with x86_64 libs for me :( Jul 14 21:00:13 Luke-Jr: oh, are you not actually on x86_64? Jul 14 21:00:31 and you're trying to build zaurus, right? Jul 14 21:00:40 * zwelch supposes he should look at it more closely Jul 14 21:00:49 however, i suspect you have something else in your tree that's bonkers Jul 14 21:00:57 zwelch: I'm on x86_64 trying to build zaurus Jul 14 21:01:05 Luke-Jr: find out if sigcmm is using our libtool or another libtool Jul 14 21:01:10 after nuking my tmp, i was able to build things that were previously giving me problems Jul 14 21:01:20 wtf is sigcmm? Jul 14 21:01:21 o.O Jul 14 21:01:28 and this is after I nuked tmp thrice Jul 14 21:01:28 whatever Jul 14 21:01:35 ah, sorry Jul 14 21:02:10 Luke-Jr: well, don't ask the project what it can do for you, ask the developers what you can do for them Jul 14 21:02:21 * Kerwood_ appreciates the currency of the perfectly good word "thrice" Jul 14 21:02:28 zecke: so what is libcmm and how can I tell the difference? Jul 14 21:02:40 Kerwood: O.o;; Jul 14 21:02:52 Luke-Jr: I suck at names, you need to find out what I mean Jul 14 21:03:08 Luke-Jr: so for 8 hours you know that your libtool file is broken Jul 14 21:03:28 Luke-Jr: now the next logical step is to find how it is broken and then why it is broken Jul 14 21:04:12 Luke-Jr: my build should be back to glibmm within a few hours; i'll let you know if i have problems with it Jul 14 21:04:48 Luke-Jr: one source of evil could be. This tool is not using the OE version of libtool Jul 14 21:05:05 zecke: how it's broken: it references the wrong staging lib dir; why: no idea how to determine Jul 14 21:05:32 JustinP: Sorry for the delay. I appreciate the offer, but as a spitz owner I'm in "fat city" right now Jul 14 21:05:50 Luke-Jr: well, don't tell me. Try to fix it yourself, it will be fun Jul 14 21:06:06 Luke-Jr: I will need to work the whole week, so I can not spare too much time Jul 14 21:06:23 zecke: I have been trying Jul 14 21:06:32 I can't even get it to work editing the files manually Jul 14 21:08:05 Kerwood_: ah....I thought you were one of the 2 looking for poodle earlier Jul 14 21:09:15 JustinP: nope, not me, but thanks for thinking of me... Jul 14 21:12:58 zecke: my guess is that libsigc++ does NOT use OE libtool Jul 14 21:13:12 most other work/*/*/libtool look like this Jul 14 21:13:14 -# Generated automatically by (GNU gpgme 1.0.3) Jul 14 21:13:25 libsigc++'s, however... Jul 14 21:13:27 +# Generated automatically by config.status (GNU libsigc++) 2.0.6 Jul 14 21:14:09 Luke-Jr: a) is autoreconf called on it? Jul 14 21:14:16 Luke-Jr: b) does it say something about libtool? Jul 14 21:14:41 Luke-Jr: c) if you freshly unpack, does it have a libtool file, or even an m4 for libtool? Jul 14 21:16:26 in the bb? Jul 14 21:16:40 Luke-Jr: would that make sense? Jul 14 21:17:03 Luke-Jr: how many bb files have you seen carrying a verbatim copy of m4 files or shell scripts Jul 14 21:17:17 4 Jul 14 21:17:27 tell me the names Jul 14 21:17:32 actually, I was referring to (a) Jul 14 21:17:55 Luke-Jr: 4? how is that related to a Jul 14 21:18:17 is autoreconf called in the bb? <-- question? Jul 14 21:18:28 no, on it Jul 14 21:18:37 on the package, we are talking about source code Jul 14 21:19:09 so where am I supposed to see if it's called? Jul 14 21:19:46 Luke-Jr: like with any other 'task' that bitbake is executing Jul 14 21:19:54 Luke-Jr: if it is a shell task, it will create a shell script Jul 14 21:20:07 Luke-Jr: in any case a logfile will be created Jul 14 21:20:09 so look in temp? Jul 14 21:20:21 Luke-Jr: in this log file you can read the output of the task Jul 14 21:20:37 yes, I see autoreconf Jul 14 21:20:40 after automake Jul 14 21:22:33 Luke-Jr: is it saying something about libtool Jul 14 21:22:46 autoreconf: configure.ac: not using Libtool Jul 14 21:22:51 checking if libtool supports shared libraries... yes Jul 14 21:22:56 config.status: executing libtool commands Jul 14 21:24:21 iirc, sqlite is an example where we remove a custom copy of libtool Jul 14 21:24:31 in favor of our copy Jul 14 21:24:46 you could take a look and see if that applies to this library as well Jul 14 21:24:52 ... Jul 14 21:27:23 JustinP: sorry, I was away for a while :) Jul 14 21:34:41 hvontres|poodle: understoos Jul 14 21:40:16 hvontres|poodle: you seeing my privmsgs? Jul 14 21:42:58 JustinP:no Jul 14 21:43:24 does anyone know how to identify what sources to use to put otgether the latest arm kernel for linux? Jul 14 21:43:49 hvontres|poodle: :-P Jul 14 21:44:05 hvontres|poodle: e-mail/IM? I have a slow uplink and would rather not post my URLs here for now Jul 14 21:45:15 JustinP:send me a note at hvontres (at) sbcglobal (dot) net :) Jul 14 21:45:22 I know this isn't info directly on topic for OE (you folks use bitbake more) but I want to see if I can assemble the sources Jul 14 21:45:55 any help to be had? Jul 14 21:46:19 chuckr: "asssemble the sources"? Jul 14 21:46:32 chuckr: have you tried something like 'bitbake -i fetch gpe-image' or something? Jul 14 21:46:45 * zwelch wonders if that will do what he intuits it should Jul 14 21:46:56 except s/-i/-c/ Jul 14 21:47:24 chuckr: hm? just grab a current 2.6 kernel, that supports arm out of the box. whether it supports _your machine_ is somethign else entirely Jul 14 21:47:54 if this was gentoo, the answer to his question would be 'emerge --fetch-only world' Jul 14 21:48:01 (if i am reading him correctly) Jul 14 21:48:20 zwelch: -c Jul 14 21:48:31 bitbake has 'world' too. last time i did a bitbake -k world in oe it took up 25+ gigs of hd space and took days ;) Jul 14 21:48:34 zwelch: but there is a bitbake 'bug' with RDEPENDS Jul 14 21:48:42 I am trying to see what' son the site www.linux.kernel.org, and the arm site from tere, and get what's advetised as the latest version (2.6.14-rc3) Jul 14 21:48:48 i.e. gather all of the tarballs first, for building later (e.g. offline) Jul 14 21:49:03 heh, okay, i was wrong :) Jul 14 21:49:09 but there isn't any tarball with naming that includes rc-anything Jul 14 21:49:19 eh? Jul 14 21:49:21 chuckr: www.kernel.org Jul 14 21:49:25 the naming is unlike anything I read there Jul 14 21:49:28 download a damn kernel and call it a day Jul 14 21:49:32 -rc isnt arm Jul 14 21:49:37 its sa release candidate of a linux kernel Jul 14 21:49:53 you dont need a -rmk patch anymore. just grab a kernel.org kernel. Jul 14 21:50:07 kergoth: would it be useful to have a (hopefully fast, multiproc/core) server doing that on a period basis? Jul 14 21:50:53 I am not trying to get the oldest kernel they have on the site, Look, if you don['t want to anser, fine, but please don;t block me from getting an answer. All I want is hte name o fthe current kernel that is serving as the patching base, and the name of the latest patch file Jul 14 21:51:21 chuck, put down the crack pipe Jul 14 21:51:29 if you want an arm kernel, go download the latest linux kernel on kernel.org Jul 14 21:51:32 it supports arm out of the box Jul 14 21:51:34 chuckr: finger www.kernel.org Jul 14 21:51:40 kergoth *g* Jul 14 21:51:49 why, precisely, is what I ask either crazy or even unfreindly? Jul 14 21:51:58 because i've answered 3 times now? Jul 14 21:52:03 and you still havent stopped asking? Jul 14 21:52:04 jesus Jul 14 21:52:05 chuckr: it's the wrong place, to start Jul 14 21:52:14 * zwelch tries to hug kergoth Jul 14 21:52:20 sorry it is finger kernel@www.kernel.org Jul 14 21:52:21 I went o that darn site, and it had bad info Jul 14 21:52:32 ok Jul 14 21:52:36 and i told you where to go Jul 14 21:52:37 kernel.org Jul 14 21:52:47 if the arm site is out of date, that isnt our problem Jul 14 21:53:54 I never saw that before, and that's looking like what I was after. I was polite, I asked for something reasonable, and you ascted like a wild man. Please give me a chance. Jul 14 21:54:11 15:13 < kergoth> chuckr: hm? just grab a current 2.6 kernel, that supports arm out of the box. whether it supports _your machine_ is somethign else entirely Jul 14 21:54:26 that was not impolite, or unclear, and you continued this nonsense after that Jul 14 21:54:37 if you'd actually read what i wrote, it wouldnt have gotten to this point, now would it? Jul 14 21:54:57 * zwelch sighs and starts doing a dance to distract everyone's attention Jul 14 21:55:27 irc sucks for communication :) Jul 14 21:55:37 tempers are too easy to flare up Jul 14 21:55:47 * zwelch got himself severely flamed in #ffmpeg this morning Jul 14 21:55:55 i won't be going back there anytime soon Jul 14 21:56:26 all i asked was "why don't you use autotools?" and "why don't you produce regular versioned releases?" Jul 14 21:56:40 and tried to have what i thought was a rational conversation Jul 14 21:56:45 ouch, no wonder :) people take such things personally Jul 14 21:56:47 big friggin mistake :) Jul 14 21:57:05 there's often a nearly evangelical opinion about autotools, in one direction or the other Jul 14 21:57:10 you didn't mention "vi" or "emacs" did you? Jul 14 21:57:29 hahah Jul 14 21:57:30 that's true. The data at www.kernel.org is for linux in general, and the other site I was at, www.arm.linux.org, says that 2.6.14 is the latest for the arm, and that contradists this data 9but makes sense if you limit it to the arm) but I don't know what files to use as base or as patch file Jul 14 21:57:33 of course, i am somewhat antagonistic :) Jul 14 21:57:40 i did use the work fork and mention guft.org Jul 14 21:57:43 * zwelch shrugs Jul 14 21:58:21 chuckr: the point kergoth was trying to make is that arm/linux == linux as far as the kernel goes Jul 14 21:58:21 chuckr: again, as i've explained before, you just need to download the linux kernel sources. there is no patch file involved. go download a kernel. if you cant figure out how to find a kernel on kernel.org, theres nothing more anyone can do for you Jul 14 21:58:57 anything older is just obsolete Jul 14 21:59:07 stuff that people haven't taken down from the web Jul 14 21:59:22 i know it's confusing, because it seems like the arm stuff is out of date Jul 14 21:59:39 but the fact is, it's a good thing - it means there is no longer a separate branch/patch series to handle Jul 14 21:59:58 chuckr: what specifically are you looking for, again? Jul 14 22:00:02 * zwelch hated dealing with the dual -rmk-np SA1100 series Jul 14 22:00:04 zwelch: right, only patches that missed the merge window of the kernel Jul 14 22:00:29 right, and you generally don't need/want stuff that's in that queue Jul 14 22:00:37 you want to see it get into the next release Jul 14 22:01:07 (because if it doesn't ... umm... there's probably a reason for that) Jul 14 22:01:33 raduga, I am trying to assemble the latest sources for the ARM, the latest kernel sources Jul 14 22:02:06 so, what have you learned from this discussion? please re-cap for us :) Jul 14 22:03:42 I've been pointed At data for the general Linux kernl, which is at 2.6.17.something, but when I check the site www.arm.kernel.org (pointed to from the www.kernel.org site) it tells me that the lates arm kernel is 2.6.14-rc3. I don't have any idea what files to use for this Jul 14 22:03:52 * zwelch sighs Jul 14 22:04:02 okay, let me be explicit: Jul 14 22:04:13 chuckr: arm.linux.org.uk's information is IRRELEVENT. go download 2.6.17. Jul 14 22:04:14 chuckr: you recognize that website can get outdated? Jul 14 22:04:16 if I was going for, say, i386, I would be gone by now Jul 14 22:04:22 its the same kernel as i386 Jul 14 22:04:23 ignore www.arm.kernel.org Jul 14 22:04:27 use kernel.org Jul 14 22:04:30 identical kernel Jul 14 22:04:32 no seperate kernel for arm Jul 14 22:04:36 * zwelch sighs Jul 14 22:04:42 I a, using www.arm.kernel.org Jul 14 22:04:44 this is what, the 4th or 5th time we've explained this? Jul 14 22:04:53 kergoth: ignore him Jul 14 22:04:53 go to www.kernel.org Jul 14 22:04:58 jesus christ, fucking idiots Jul 14 22:05:00 chuckr: well, stop it :) Jul 14 22:05:00 which is what www.kernel.org points me to Jul 14 22:05:01 chuckr: If you know better, don't ask Jul 14 22:05:15 * zwelch cackles maniacally Jul 14 22:05:53 know better than what? I *think* I'm doing just precisely what you're telling me to do Jul 14 22:06:39 if I donwload the files for the general distribution, I violate the info from the www.arm.kernel.org site Jul 14 22:07:08 oh, forget, it, if you'd rather cackle than explain, forget it Jul 14 22:07:09 chuckr what? Jul 14 22:07:13 for the last time: ignore www.arm.kernel.org Jul 14 22:07:36 ignore it? Jul 14 22:07:43 * thejapa ignores it Jul 14 22:07:44 ok, I can try that Jul 14 22:07:47 chuckr whats your problem? Jul 14 22:07:59 so far, you are Jul 14 22:07:59 * zwelch hugs chuckr Jul 14 22:08:07 thank you for finally listening Jul 14 22:08:17 chuckr cant you decide which is your right und which is your left hand? Jul 14 22:08:28 I am trying to just be striaght and not put any anger or anything into my posts, Jul 14 22:08:40 i can understand your frustration Jul 14 22:08:48 I think that most flames are counter-productive Jul 14 22:08:56 thanks Jul 14 22:09:04 kergoth: you might just kick him Jul 14 22:09:08 "all i did was ask a simple question and *wham* suddenly i'm covered in napalm" Jul 14 22:09:14 kergoth: if he can't evaluate information Jul 14 22:09:23 zecke na we he have fishis nice little fishis Jul 14 22:09:24 some folks seem to think I misunderstand on purpose, to try to bother folks, and geeze, I'm not Jul 14 22:09:39 * zecke hugs ignore lists Jul 14 22:09:56 churck: i can understand his frustration too; he is a founder of this project, and your question really is not apropos for this venue Jul 14 22:10:01 * raduga hugs fishes Jul 14 22:10:31 chuckr: by being in #oe you should be asking about OE; you have a generic Linux kernel question that could have been answered in #linux or something Jul 14 22:10:33 maybee its a test from computerbild on how long oe-people can take some people which really trolls Jul 14 22:10:47 woglinde: hehe, this is what I have though as well Jul 14 22:10:55 was the original question, "Where can I find the latest arm linux kernel?" Jul 14 22:11:16 chuckr: ftp.arm.linux.org.uk Jul 14 22:11:29 chuckr: ftp.arm.linux.org.uk Jul 14 22:11:30 chuckr: ftp.arm.linux.org.uk Jul 14 22:11:50 03mickeyl 07org.oe.dev * rcede3a0f... 10/conf/distro/ (41 files in 2 dirs): distro configurations: rename include files to *.inc and adjust those which include it accordingly Jul 14 22:12:02 chuckr: it is FTP so you can download, and it has 'arm' and 'linux' in the domain name Jul 14 22:12:19 If I am where in bad mood i would suggest the next questions, why linux-2.x.x.tar.bz dont boot Jul 14 22:12:26 Luke-Jr: my build is starting on glibmm... i should know if it's "just you" soon Jul 14 22:12:40 woglinde: no. Why isn't Linux 5.3 working Jul 14 22:12:43 when it says oops, what does that mean? Jul 14 22:12:57 woglinde: i printed out the kernel sources, stuffed them into my toaster, but the toaster doesn't boot. Jul 14 22:13:04 and it won't even toast bread anymore. Jul 14 22:13:06 CosmicPenguin: my KDE isn't working, it says oops Jul 14 22:13:07 and its getting really warm Jul 14 22:14:50 chuckr: are you more interested in fixing the apparent problem with tech support on OE, or in findin an answer to your question? Jul 14 22:15:09 chuckr: if the latter, msg me. if the former, enjoy flames ;) Jul 14 22:15:14 zwelch: honestly, the fact that the question was inappropriate wasn't as much of an issue as his inability to listen to the answers to his questions. there's nothing worse than contacting tech support and assuming you know better than they do. Jul 14 22:15:58 kergoth: sometimes its difficult to listen to answers, or to recognise them as answers Jul 14 22:16:05 kergoth: you know, it's pretty often that tech support is dumb in general ... =p Jul 14 22:16:13 g'night Jul 14 22:16:19 nite mickeyl Jul 14 22:16:41 and if the answer isn't presented in a way that makes sense to you, where you are at that moment, there's a tendency to ignore it, and try again. Jul 14 22:16:50 Luke-Jr: and it's even more often that users are. Jul 14 22:17:35 its one way that users deal with information overload. Jul 14 22:17:46 Which is a pretty serious problem :( Jul 14 22:17:47 ok, I never understood, you were telling me that the data I read at www.arm.kernel.org is all completely incorrect? Jul 14 22:18:04 not all, but mostly out-dated Jul 14 22:18:07 chuckr: ftp.arm.linux.org.uk Jul 14 22:18:10 chuckr: some of it is correct, some of it is incorrect. some is old, some is useful, some is just not maintained. Jul 14 22:18:10 hvontres|poodle: see the e-mail? Jul 14 22:18:26 thanks zecke, I'm looking now Jul 14 22:18:48 chuckr: what people are suggesting, is that you don't rely on this resource, because its no longer well maintained, or a place where people are putting much effort into support Jul 14 22:19:12 chuckr: but www.kernel.org got source code that works perfectly for ARM Jul 14 22:19:28 chuckr: similarly, calling Sharp-USA and asking for zaurus help, is probably not going to get you much help Jul 14 22:19:57 heh Jul 14 22:20:16 chuckr: ftp.arm.linux.org.uk was once useful, was once well-maintained, and was once current. Jul 14 22:20:34 chuckr: finger kernel@www.kernel.org with 99% probaility will show the latest version of Linux kernel able to run on ARM Jul 14 22:20:54 chuckr: and the appendix hanging off your stomach once was used to digest food, by your remote ancestors. Jul 14 22:21:54 chuckr: everything has to be taken with a grain of salt on the internet; things get published and fall out of date. that said, there is probably still a lot of useful information on those sites, but it has to be tempered with the added knowledge that it may not all still apply Jul 14 22:22:03 chuckr: but, strangely, you look at your digestive tract, and guess what!! Jul 14 22:22:11 the appendix is still there. its still attached. Jul 14 22:22:31 you could complain to God, that he should update his web page. Or you could just ignore it. Jul 14 22:22:32 raduga: great metaphor :) Jul 14 22:22:38 it boggles the mind that someone cannot grasp something as simple as a site being out of date Jul 14 22:22:41 it happens every day Jul 14 22:23:17 kergoth there are enough people who think all what Bush says is 100% the truth Jul 14 22:23:17 one could argue it's symptom of insufficient skepticism Jul 14 22:23:23 hahah Jul 14 22:23:26 woglinde: ditto Jul 14 22:23:30 woglinde: it's not 100% lies? Jul 14 22:23:35 now... if you stuff food into the appendix (as some people's stomachs do) eventually it may get infected, and have to be removed Jul 14 22:24:13 people believe politicians, because they tell people nice things, that make them feel good. Jul 14 22:24:29 they validate people's hopes and fears. Jul 14 22:24:55 * zwelch stands by his last statement and starts jumping up and down while pointing at it Jul 14 22:25:04 chuckr: if you want, send email to the webmaster at kernel.org Jul 14 22:25:29 and ask them to update the link on their home page (or at least to make a comment tag that the site is no longer current) Jul 14 22:26:16 some things that kernel.org says are not lies. Jul 14 22:26:25 some things that Bush says are not lies either. Jul 14 22:27:00 even chuck norris looks at kernel.org Jul 14 22:27:25 the 'r' stands for 'norris'? Jul 14 22:27:59 later Jul 14 22:28:02 woglinde: maybe we should be a little more respectful, or he will kick our butts Jul 14 22:28:10 actually, zwelch, why are you proud aboout jumping up and down like a maniac? Jul 14 22:28:17 args Jul 14 22:28:19 thanks, oe devs Jul 14 22:28:23 thanks, oe trolls Jul 14 22:28:29 lol Jul 14 22:28:30 chuckr: we are derived from apes Jul 14 22:28:43 chuckr: it must be in our nature Jul 14 22:28:45 didnt complaided about the cuck in chuckr Jul 14 22:28:56 * zwelch cackles madly Jul 14 22:29:02 well, now that I separated the screaming from the data and got my stuff, I;'m pleased Jul 14 22:29:09 chuckr: read the feedback you got, evaluate it Jul 14 22:29:09 jesus, somebody kick him already, there's no need to tolerate this off topic trolling Jul 14 22:29:13 chuckr: consider your options Jul 14 22:29:23 I have what I want now Jul 14 22:29:32 thats fine Jul 14 22:29:36 good nite Jul 14 22:29:39 exactly Jul 14 22:32:45 Luke-Jr: glibmm built for me fine; gtkmm is building now. conclusion: it's all you :) Jul 14 22:33:15 good ole WFM Jul 14 22:33:17 ~emulate zautrix Jul 14 22:33:18 this is exactly what the GPL is about. grab the stuff and run away Jul 14 22:33:43 kergoth: it's nice to be on the WFM side of it too Jul 14 22:34:06 it's never fun to be on the receiving end Jul 14 22:34:23 hehe Jul 14 22:36:03 WFM? Jul 14 22:36:08 ~WFM Jul 14 22:36:11 extra, extra, read all about it, wfm is (Wired For Management Baseline) This is an Intel hardware specification that is designed to allow for compliance with easier management of desktop PCs in a networked environment. The specification calls for computers to be compatible with a pre-boot protocol that can be used to update the system or perform other management ... Jul 14 22:36:23 worksforme. its one of the states a bug can be set to in bugzilla Jul 14 22:36:33 hehe Jul 14 22:36:52 lol Jul 14 22:46:18 so, i have essentially finished the new metadata files for minisplat to build with OE/bb Jul 14 22:50:26 so nite Jul 14 22:50:51 * raduga has a probably inappropriate OZ kernel question, but haven't seen anyone respond in other channels or oesf :/ Jul 14 22:51:18 since running 2.6, I'm frequently getting: "Error: AC check failed" "Charging error" Jul 14 22:51:34 battery otherwise seems ok, and charger seems working. Jul 14 22:52:01 has anyone seen this? is it a sign that my AC, or battery, or charger actually are failing? Jul 14 23:19:21 JustinP: Thanks for the images.. I will try flashing when I get home... yhou probably saved me ~15-20 hrs of building a gpe image :) Jul 14 23:20:17 hvontres|poodle: :-) Jul 14 23:21:53 JustinP: Those poor PII's can only go so fast .. It actually took me 15 hrs to build opie, and that was with the toolchain already built..:( Jul 14 23:24:07 well , almost time to go home..:) see y'all later Jul 14 23:26:04 any idea what this "socket: Address family not supported by protocol" means? GPE/OZ 3.5.4 Jul 14 23:27:25 RickSeymour: in what context are you getting it? Jul 14 23:27:52 when i run NSQLD (the daemon needed for sqlite connections for gpesyncd) Jul 14 23:28:31 * zwelch shrugs Jul 14 23:28:42 when i do a "netstat -an" i have a couple of interesting lines.....netstat: no support for `AF INET6 (tcp)' on this system. Jul 14 23:28:54 which i assume is that there is no IPv6 support Jul 14 23:29:45 * kergoth attempts to get caught up on the bitbake source Jul 14 23:30:21 (PS if anyone knows where i can get my mitts on a 2.6 kernel for Poodle .. i would be a happy bunny) Jul 14 23:35:08 IPv6 should not be a requirement though Jul 14 23:35:13 well, goddammit, I spend a moment with my family and miss a spectacular slapping match between kergoth and some n00b Jul 14 23:35:33 * Kerwood_ thinks kergoth is the shizzle Jul 14 23:35:35 Kerwood_: do you like to feed trolls? :) Jul 14 23:35:41 you've no idea Jul 14 23:36:16 First time I've ever seen kergoth and shizzle in the same sentence Jul 14 23:36:21 heh Jul 14 23:36:31 at times, i find irc can offer a fascinating study in human behavior :) Jul 14 23:36:33 Not that he isn't, but its really hard to type 'shizzle' and sound intellegent Jul 14 23:38:16 * zwelch waits for a joke about GWB using that word at a press conference Jul 14 23:42:33 ~lart statistics! Jul 14 23:42:33 * ibot executes killall -TERM statistics! Jul 14 23:43:32 zwelch: or x86_64 Jul 14 23:43:56 Luke-Jr: yes, that is still a free variable Jul 14 23:44:20 it wouldn't be the first time, either Jul 14 23:44:29 i believe you Jul 14 23:44:44 the sqlite bug I was talking about earlier doesn't exist with x86_32, since its pointer size matches arm Jul 14 23:45:01 thats not x86_64s fault Jul 14 23:45:07 nope Jul 14 23:45:12 if developers are supid enough to assume the size of unsigned long, then fuck em Jul 14 23:45:23 lol Jul 14 23:45:34 they #define pointer size based on sizeof(char *) during compile Jul 14 23:46:11 which isn't a problem for x86_32 because it has the same size as arm Jul 14 23:46:34 if arm was 64-bit, it'd be the reverse Jul 14 23:47:22 sqlite comments *do* say cross-compilers should define the pointer size manually Jul 14 23:47:28 but zecke wants to use autoconf Jul 14 23:47:31 so use it Jul 14 23:47:35 we use autoconf all the time Jul 14 23:47:46 why do you think an x86_64 site file exists? Jul 14 23:48:00 heh, I can fix sqlite's problem manually Jul 14 23:48:09 so glibmm is more of a priority to fix for me Jul 14 23:48:39 I can't even tweak it to work manually Jul 14 23:53:25 anyway, I'm not too surprised it works on x86_32 Jul 14 23:53:38 that would explain why it might be using a non-portable libtool Jul 14 23:53:44 (or non-cross-compile) Jul 14 23:59:39 hmmm, it appears that a recent update broke my configuration: http://rafb.net/paste/results/63u0EE54.html Jul 15 00:00:25 ... and my DISTRO="openzaurus-unstable", which still appears in the dev branch conf/distro/ directory. i'm not sure what's up. :/ Jul 15 00:07:19 so how do I make libsigc++ and glibmm use OE's libtool? Jul 15 00:08:13 you know who I haven't seen in a while? pb_ Jul 15 00:08:17 is he hiding from OE these days? Jul 15 00:08:31 * zwelch acts as pb_ repellant Jul 15 00:08:40 but i don't know why :/ Jul 15 00:08:43 no ideas? :( Jul 15 00:08:49 zwelch: well, I know one of mickeyl's last acts today was to rearrange the conf files in org.embedded.dev/conf Jul 15 00:09:07 ~seen pb_ Jul 15 00:09:13 Luke-Jr: first, you might figure out how libtool is being used Jul 15 00:09:21 pb_ was last seen on IRC in channel #oe, 4d 3h 52m 56s ago, saying: 'hi kergoth, zecke'. Jul 15 00:09:25 zwelch: ? Jul 15 00:09:30 dayyum Jul 15 00:09:31 dunno, what does libtool try to do? Jul 15 00:09:35 from there, it's probably a matter of running libtoolize or fixing the usage Jul 15 00:09:47 Luke-Jr: run 'info libtool' and read up Jul 15 00:09:53 I don't have info Jul 15 00:10:12 * zwelch points at the world wide web and shrugs Jul 15 00:10:15 =p Jul 15 00:10:37 what, you want me to just say "RTFM" and leave it at that? ;) Jul 15 00:10:46 * zwelch was just trying to be nice ;) Jul 15 00:11:25 sure =p Jul 15 00:11:35 run libtoolize where? O.o Jul 15 00:11:40 or should I use an eclass Jul 15 00:11:41 ? Jul 15 00:11:42 to be more constructive, i would bet it's a problem with the glibmm makefiles Jul 15 00:11:50 which might be fixed in a newer version Jul 15 00:11:53 I'm trying it on libsigc++ first Jul 15 00:11:59 oh Jul 15 00:11:59 ? Jul 15 00:12:22 well, koen asked me to look at those packages for that reason: to bump their versions to newer releases Jul 15 00:12:33 i will need to in order to get all of minisplat working Jul 15 00:12:56 but why is it likely fixed in a new version? Jul 15 00:13:09 either way, the .la from libsigc++ refer x86_64/lib Jul 15 00:14:03 i said "might" not "likely" Jul 15 00:14:31 but yeah, if it's a dep, then it may not show up until later either Jul 15 00:14:32 true, you only bet on it being glibmm's makefiles not that it was fixed ;) Jul 15 00:15:02 actually, the problem might be a missing --with-stdlibc++ option in glibmm or something Jul 15 00:15:41 i had that problem with speex, where it was producing a .la with links to the native libstdc++, and it only showed up when trying to link against my audio codec package Jul 15 00:16:15 well, how can I test if adding libtoolize helps? =p Jul 15 00:16:21 the fix was to tell the configure script where it really should be looking for the libraries (in ${STAGING_LIBDIR} and ${STAGING_INCDIR} Jul 15 00:17:11 well, you need to figure out exactly a) where the problem is originating, and b) what the problem actually is. libtoolize is a hypothetical solution for one possible problem Jul 15 00:17:50 Kerwood_: btw, i am aware of mickey's changes; this is why i brought the pastebin straight here ;) Jul 15 00:17:59 * zwelch wants to here WFM or otherwise Jul 15 00:18:04 s/here/hear/ Jul 15 00:18:16 zwelch: trial and error is probably the best I can do... :\ Jul 15 00:18:41 and from what I can see, libtoolize on sigc++ seems most likely Jul 15 00:18:42 right, but have a strategy for isolating the fault Jul 15 00:18:51 well, that just replace the libtool files Jul 15 00:19:01 grep shows x86_64/lib in sigc++'s .la's within staging/arm-linux Jul 15 00:19:12 if libstdc++ is somehow producing an .la that includes your native libraries, you need to tell it to knock it off Jul 15 00:19:16 libtoolize won't do that Jul 15 00:19:20 necessarily Jul 15 00:19:39 i.e. heck, it might, but there's no way to be sure at this point Jul 15 00:20:20 an easy way to hack the .bb? i dunno offhand; i'm still debugging some stuff like this :) Jul 15 00:23:32 Luke-Jr: incidentally, what are you trying to build that needs glibmm? Jul 15 00:25:31 zwelch: GPE? Jul 15 00:25:45 Luke-Jr: oh, nothing big then... ;) Jul 15 00:25:48 ... Jul 15 00:25:51 hehehe Jul 15 00:26:13 this is looking to be only slightly easier to fix than my KDE BBs Jul 15 00:26:29 i really think we need to look at the latest versions Jul 15 00:26:38 but that takes time and patience Jul 15 00:26:39 =p Jul 15 00:27:00 i have both... in varying quantities, depending on the phase of the moon Jul 15 00:27:09 I find my primary motivator to be my impatience sometimes Jul 15 00:27:26 if I weren't impatient, I'd probably have just ignored the problem and figured it'll work in a month Jul 15 00:27:37 instead of trying to figure it out Jul 15 00:28:20 * zwelch wonders if mickey's changes that "break" openzaurus are actually meant to force people to think about using the configuration and manually assert their sanity Jul 15 00:28:35 ... Jul 15 00:28:42 in which case, the bug is not really a bug.... except the output trace is sort of annoying Jul 15 00:28:43 mickey broke it? Jul 15 00:28:58 by my new found reasoning, no Jul 15 00:29:16 and sorry, i mistyped, i meant to say openzaurus-unstable Jul 15 00:29:28 that's what I'm trying to build Jul 15 00:29:52 my local.conf had that set as its DISTRO, and my most recent mt pull gave me the pastebin above Jul 15 00:30:29 now, reading the full error and considering things, i can see the rational to force users to manually invervene before building an "unstable" tree Jul 15 00:30:55 most folks probably should be using the stable branch Jul 15 00:31:40 since i need to bump a lot of packages, i don't think that's my best path and have selected dev and unstable Jul 15 00:32:39 so, you can see this as a bug, if you think anyone should be able to build the unstable branch without having to think about it first :) Jul 15 00:33:39 * zwelch wishes he had received that revelation from mickey's emails to the list, but it's all good now Jul 15 00:33:54 ... Jul 15 00:37:30 so, here's a silly feature request: how about versioning the tarballs from svn repos by their revision, rather than by date? then, instead of getting a new tarball each day, we only get one when the repo actually changed that day Jul 15 00:37:39 is that silly, or what? Jul 15 00:38:12 * zwelch can hardly take himself serious at times Jul 15 00:38:20 hmmm. +ly? Jul 15 00:38:42 isn't that a cvs feature called tags? hehe Jul 15 00:38:44 zwelch: here are the comments from the relevant patch: Jul 15 00:38:47 (11:12:26 AM) CIA-9: conf/distro cleanup. conf/distro should only contain files that are ok to be directly included Jul 15 00:38:47 (11:12:26 AM) CIA-9: by means of setting DISTRO. conf/distro/include should contain those which doesn't. Jul 15 00:39:37 yes, but openzaurus-unstable is still in conf/distro/ directly Jul 15 00:40:06 zwelch: hey, I just report the news... Jul 15 00:40:13 the difference appears to be that it doesn't set its DISTRO, letting it be set by the include/openzaurus.inc Jul 15 00:40:23 since they don't match, the configuration is presumed insane Jul 15 00:40:39 Kerwood_: likewise, i'm just calling it like i see it :) Jul 15 00:41:46 after following its directions, everything started working again, but it reflects the fact that you need to be prepared to fix everything yourself ;) Jul 15 00:42:32 yup Jul 15 00:43:07 overall, he probably did exactly the right thing, given the circumstances Jul 15 00:50:28 thejapa: btw, afaik, cvs doesn't have the equivalent revision support; timestamps are the only way to get each "changeset" in the same manner Jul 15 00:51:35 in effect, the svn repo revision is another monotonically increasing package version, and while bb has support for specifying a specific revision this way, i don't see a way to say "the latest HEAD" Jul 15 00:51:58 or, maybe it has it and i just need to try it ;) Jul 15 00:52:25 zwelch: sorry, i'm actually very ignorant on cvs and other source control apps :) Jul 15 00:52:29 effectively, it's just a little extra logic to say "if asking for HEAD, get it, then convert that to an actual rev" Jul 15 00:52:57 i was with cvs on my head because i'm just trying my first steps in it. :) Jul 15 00:53:17 my problem is that i am ignorant of bitbake, so i'm not sure exactly how easy this would really be Jul 15 00:53:37 ....because it involves very late binding of the package revision... i.e. after fetch Jul 15 00:54:18 now after thinking about it, i can see why it's not trivial and probably not worth implementing :) Jul 15 00:54:40 see, i told you it was a silly request Jul 15 00:55:15 yeah, i dumbily(sp?) agree Jul 15 00:55:32 that's double dumb I guess Jul 15 00:55:51 * zwelch stands on his head and tries to make sense of that Jul 15 00:57:19 dumb because i don't have the bitbake knowledge to understand the difficulty and double dumb because i don't know how to write dumbly/dumbilly/dumbily :) Jul 15 00:57:36 triple dumb now Jul 15 00:58:08 * thejapa thinks me is losing communication skills Jul 15 01:07:48 nifty. i just fixed a bug in pkgconfig.bbclass Jul 15 01:13:02 fwiw: http://bugs.openembedded.org/show_bug.cgi?id=1174 Jul 15 01:22:03 so, i have a question about the notes in the manual... when it says "this only works with .bb and .bbclass files", one can presume the same is true for files "include"d or "require"d by such files? Jul 15 01:22:28 e.g. in .inc files required by .bbfiles Jul 15 01:22:49 * zwelch is using that model and is getting bugs Jul 15 01:25:35 libtoolize doesn't work on libsigc++ :( Jul 15 01:25:42 its configure script appears to undo it Jul 15 01:29:57 the autotools.bbclass runs autoreconf during do_configure Jul 15 01:37:26 Luke-Jr: good luck. i'll be back in several hours Jul 15 01:40:25 CoreDump|home: morning :) Jul 15 02:07:44 * kergoth considers possible major conceptual changes in the way bitbake works Jul 15 02:10:30 damnit, why do all the other core bitbake devs have to be in europe :P Jul 15 02:13:58 because we only let minimum wage employees come to the USA Jul 15 02:14:24 hehe Jul 15 02:16:41 zwelch: indeed, but autoreconf doesn't help it Jul 15 02:17:16 zwelch: I moved libtoolize to post-configure, so next time it tries building we'll see Jul 15 02:17:31 unfortunately, my bitbake has once again decided it doesn't need glibmm =p Jul 15 02:17:49 or fortunately for me if it works, I guess =p Jul 15 02:18:20 it should be libtoolizing automatically, autoreconf calls it out Jul 15 02:18:27 kergoth: the configure script undoes it Jul 15 02:18:47 or seems to, from my limited knowledge of this stuff Jul 15 02:19:10 the 'libtool' program in the work directory is different from the normal OE 'libtool' programs Jul 15 02:20:55 kergoth: only about 4 more hours till daylight over there :) Jul 15 02:21:09 checking if libtool supports shared libraries... yes Jul 15 02:21:09 config.status: executing libtool commands Jul 15 02:21:16 that part is probably related Jul 15 02:25:22 Luke-Jr: libtoolize should be overwriting the ltmain.sh in the source tree, which configure uses to generate the local libtool. Jul 15 02:25:34 hvontres|home: :) Jul 15 02:28:18 kergoth: well, the .la is ending up with staging/x86_64-linux (host) paths... Jul 15 02:28:27 kergoth: you'd know better than I Jul 15 02:29:02 dont use the term 'host' Jul 15 02:29:04 its ambiguous Jul 15 02:29:10 forcing libtoolize *after* configure doesn't leave a libtool prog Jul 15 02:29:13 build Jul 15 02:29:20 okay, weird Jul 15 02:29:51 well, it makes sense if configure creates it from ltmain Jul 15 02:30:00 I don't know about getting a good .la tho Jul 15 02:31:18 ~seen JustinP Jul 15 02:32:06 justinp was last seen on IRC in channel #oe, 2h 49m 34s ago, saying: '~lart statistics!'. **** ENDING LOGGING AT Sat Jul 15 02:59:56 2006