**** BEGIN LOGGING AT Sun Sep 03 02:59:57 2006 Sep 03 05:42:56 hmm Sep 03 05:43:19 yeees? Sep 03 05:44:47 wondering what to call something. wrote a python function in package.bbclass that makes the -dev and -dbg packages Recommends the corresponding -dev and -dbg packages of that which the main package depends Sep 03 05:44:55 so if your package uses libc6, then your -dev will recommend libc6-dev Sep 03 05:45:10 * emte falls over Sep 03 05:45:19 just wondering how to name that, like the: POSTFIXES = "-dev -dbg" variable, what exactly should i name this? Sep 03 05:45:22 heh Sep 03 05:45:24 kergoth, your working on bitbake again? Sep 03 05:45:28 yep Sep 03 05:45:37 learning monotone Sep 03 05:45:38 heh Sep 03 05:45:50 you forget to take your sanity perscription? Sep 03 05:46:06 mostly pushed class changes and such up, not much package maintaining yet, cause my old build box sucked ass Sep 03 05:46:14 your supposed top work on v2/ng Sep 03 05:46:14 but my new athlon 64 dual core is spiffy Sep 03 05:46:36 ahhj now you get to suffer all the 64 problems? Sep 03 05:47:05 unless it was a 32/64 Sep 03 05:47:31 altho all the dual cores are full 64's arnt they? Sep 03 05:47:55 it can run 32 bit binaries, just still a pain in the ass distro wise Sep 03 05:48:05 had to set up an ia32 chroot for flash and shit Sep 03 05:48:33 what distro? Sep 03 05:48:40 you figured out how to use the new chroot env? Sep 03 05:48:46 yeah works well Sep 03 05:48:57 i tried for a few days and couldnt get it Sep 03 05:49:01 :( Sep 03 05:49:04 firefox32 in my /usr/local/bin runs my 32 bit firefox from the chroot Sep 03 05:49:06 dchroot rocks Sep 03 05:49:11 I was seriously considering getting a dual core Sep 03 05:49:16 * kergoth using debian unstable with a couple packages from experimental Sep 03 05:50:27 back to punchin some more of my own code Sep 03 07:07:14 somebody forget to merge org.openembedded.dev ? "multiple heads" Sep 03 07:07:55 koen: morning Sep 03 07:10:16 hey T0mW Sep 03 07:10:19 good morning all Sep 03 07:10:48 kergoth: I was contemplating writing such a method yesterday after I installed Xaw :) Sep 03 07:32:35 :) Sep 03 07:35:33 I ended up doing this: ipkg list | awk '{print $1}' | grep -- -dev | grep libx | grep -v xine | xargs ipkg install Sep 03 07:53:51 koen: any ideas on what to call it? Sep 03 07:54:47 header-hunter? Sep 03 08:00:39 add_developer_deps? Sep 03 09:29:01 moin Sep 03 09:29:10 hey nap-master zecke Sep 03 09:30:08 zecke: in what part of berlin should I look for a hotel? Sep 03 09:30:54 morning zecke, koen Sep 03 09:30:59 hey pH5 Sep 03 09:31:09 * koen <- berlin n00b Sep 03 09:35:39 koen: hehe, slept all day Sep 03 09:35:56 koen: you could crash at my place Sep 03 10:07:39 zecke: nynke and I will be in berlin from wednesday through monday as a kind of mini-holiday, so we're looking for a ho(s)tel Sep 03 10:07:53 * koen is using his SoC money for the trip Sep 03 10:09:05 koen: okay :), so you need more than crashing Sep 03 10:09:12 koen: I will have to ask woglinde Sep 03 10:09:38 koen: you would want something near the 'U6' metro station Sep 03 10:09:42 zecke: a general direction like 'north berlin' or 'near thanks Sep 03 10:10:31 koen: central, south Sep 03 10:10:37 koen: near a metro station :) Sep 03 10:18:06 koen: the host will rent a new office for october and it is undecided if we will be in the new or old office Sep 03 10:18:12 koen: the old office is here http://www.stadtplandienst.de/map.asp?sid=65005bf5a18ee52de1c83e907728ac02 Sep 03 10:19:41 koen: the new office is likely to be http://www.stadtplandienst.de/map.asp?sid=bc17fa5f813442ac9cf04063568e20e6 Sep 03 10:20:34 so U6 is the best bet, or anything near a metro station in neukoelln, kreuzberg or mitte Sep 03 10:22:32 mickeyl Thinking about buying a Zaurus what is recommended? And what is the kernel status? Sep 03 10:22:40 Battery time and so on Sep 03 10:22:48 hi Sep 03 10:23:20 Kristoffer2: c1000 or c3200 Sep 03 10:23:31 Kristoffer2: second has microdrive inside Sep 03 10:23:44 oki, does it support microdrive (in the CF slot)? Sep 03 10:23:48 yes Sep 03 10:23:59 And how good is kernel support? Sep 03 10:24:08 power/suspend for example.. and battery time Sep 03 10:24:12 Kristoffer2: c3x00 has two cf slots. one internal with microdrive (4 or 6G) and one external Sep 03 10:24:46 Kristoffer2: even if there are some problems those models are the only one worth buying Sep 03 10:24:59 Oki, what kind of problems currently? And is support good in vanilla kernels? Sep 03 10:25:13 Does SD wifi card work well enough? Sep 03 10:25:25 I basicly need both harddrive + wifi to make it usefull Sep 03 10:26:48 sd i/o cards does not work in linux Sep 03 10:27:01 there is no opensource SDIO stack Sep 03 10:27:01 And is it true that it uses same power requirements as playstation? (1 or 2?) Sep 03 10:27:23 Kristoffer2: no - PSP charger fit Z iirc Sep 03 10:27:31 Kristoffer2: same as Ipaq chargers Sep 03 10:27:35 koen oki, so this basicly means that if I want working wifi & harddrive I have to go with 3100? Sep 03 10:27:48 hrw oki so ipaq chargers will work? Sep 03 10:27:58 Kristoffer2: yes - I use one now Sep 03 10:27:58 I will buy this from england so hopefully it will work nicely Sep 03 10:28:00 Kristoffer2: or c3200 Sep 03 10:28:09 How good is battery time? Sep 03 10:28:52 dont know - do not have cxx00 model. and I did not even measured batterylife for Z which I have Sep 03 10:29:40 Do one flash it or does it run directly from CF? Sep 03 10:30:00 c3x00 boot from microdrive (but kernel is in flash) Sep 03 10:30:38 any chance to brick one? Sep 03 10:30:43 no Sep 03 10:30:46 great Sep 03 10:31:03 Best price? Ebay is selling for $560 Sep 03 10:31:11 check pricejapan.com Sep 03 10:31:58 oki thx Sep 03 10:32:08 So its basicly worth buying 3200 instead of 1000? Sep 03 10:32:35 6G microdrive builtin for system and stuff or no 6G microdrive Sep 03 10:33:20 Yeah, personally I would need a free CF slot to get wifi going Sep 03 10:33:36 yep. Sep 03 10:33:53 it has usb host but very low power so usb wifi need extra powered hub Sep 03 10:37:03 you could consider a c3100, same as c3200, but with 4G microdrive Sep 03 10:37:45 or even c3000 which is similar to c3100 but has only 16M flash instead of 128M (does not make difference under OZ) Sep 03 10:37:45 aha thx Sep 03 10:38:01 How come one cannot brick it? Sep 03 10:38:06 does it got a reset of some sort? Sep 03 10:38:29 all pxa zauruses has a diagnostic menu in rom Sep 03 10:38:48 you cannot erase that part and you can use it to restore nand content Sep 03 10:39:01 ah really nice Sep 03 10:40:23 If I buy from Japan, what disadvantages does that have? Sep 03 10:40:36 maybe charger Sep 03 10:41:13 Oki, but nothing else "fixed" stuff that cant be changed Sep 03 10:41:25 all zauruses are for japanese market Sep 03 10:41:58 Oki, goodie. Thx for the info. Sep 03 10:43:33 np Sep 03 10:46:05 Lol, didnt think of one thing. The sl-c1000 also got a USB connection Sep 03 10:46:12 So basicly using wifi through that shouldnt be an issue? Sep 03 10:46:31 you need a powered hub for that Sep 03 10:46:38 Aha, oki doh Sep 03 10:46:41 or a frugal wifi adaptor Sep 03 10:46:47 frugal? Sep 03 10:46:58 <150mA power budget Sep 03 10:47:10 Does it work well then? Sep 03 10:49:38 linux will simply refuse to use it without a hub Sep 03 10:51:24 hmm, looks like the kdrive evdev driver only supports a single keyboard input device. the hx4700 currently has three (asic3 keys, pxa keys and the navpoint) Sep 03 10:53:40 would the input-hotplug branch help with that? Sep 03 10:57:34 or this: http://gitweb.freedesktop.org/?p=xorg/driver/x-input-evdev.git;a=summary ? Sep 03 10:58:41 koen: I have no idea, but thanks for the link. I'll have a look. Sep 03 11:01:40 scummvm fails to build ;( Sep 03 11:01:45 ~lart ICE Sep 03 11:01:46 * ibot DoSes ICE Sep 03 11:07:26 hi people Sep 03 11:07:26 libICE, or gcc ICE? Sep 03 11:07:36 he pavlix Sep 03 11:09:29 koen: ICE train? Sep 03 11:11:02 Any recommended Compact Flash Wifi cards? Sep 03 11:11:16 gcc ice Sep 03 11:11:22 Kristoffer2: any prism2/3 Sep 03 11:11:40 | engines/kyra/screen.cpp: In member function `int16 Kyra::Screen::encodeShapeAndCalculateSize(uint8*, uint8*, int)': Sep 03 11:11:43 | engines/kyra/screen.cpp:1834: internal compiler error: in verify_local_live_at_start, at flow.c:572 Sep 03 11:12:02 hrw thnx, got any example brands? Sep 03 11:12:05 hrw: you can disable the kyra engine to get past that Sep 03 11:12:38 or switch to gcc 4 :p Sep 03 11:12:40 Kristoffer2: linksys, asus, dlink, pretec Sep 03 11:13:34 oki, thx Sep 03 11:14:53 Kristoffer2: if can then avoid 0x0156, 0x0002 cards (pcmcia id) - they work without problem but are handled by orinoco and hostap... Sep 03 11:15:12 Oki, I've ordered my C3200 now btw. Sep 03 11:15:16 Should arrive withing 1-2 weeks Sep 03 11:15:31 cya laterz Sep 03 11:24:17 how do I make bitbake checkout the input-hotplug branch of xorg/xserver? Sep 03 11:25:40 ;head=input-hotplug or something like that? Sep 03 11:26:42 hrm Sep 03 11:26:43 (type, host, path, user, pswd, parm) = bb.decodeurl(data.expand(url, d)) Sep 03 11:28:55 pH5: try ;tag=input-hotplug Sep 03 11:30:49 hm. "git read-tree input-hotplug failed" Sep 03 11:31:21 adding support for heads to the git fetcher doesn't look too hard Sep 03 11:31:46 * pH5 still lacks basic understanding of git Sep 03 11:32:08 git is pretty easy Sep 03 11:32:14 at what place in all those git commands should the head be specified? Sep 03 11:32:23 it just does 'cat /dev/urandom > /dev/console' to make it look 1337 Sep 03 11:32:31 :) Sep 03 11:34:46 * koen declares the git manpage useless Sep 03 11:36:16 * pH5 doesn't understand read-tree Sep 03 11:36:36 bitbake uses read-tree + checkout-index instead of just checkout, right? Sep 03 11:37:07 checkout has a branch parameter, but how do I tell it which branch to use in the read-tree+checkout-index combo? Sep 03 11:37:49 pH5: good question :) Sep 03 11:37:55 ~seen pb_ Sep 03 11:37:56 ~seen pb__ Sep 03 11:37:59 bit a '"git branch %s", branch in between? Sep 03 11:38:02 pb_ was last seen on IRC in channel #oe, 16d 12m 39s ago, saying: 'how is sbrshd different from regular ssh?'. Sep 03 11:38:04 pb__ was last seen on IRC in channel #oe, 14d 23h 25m 1s ago, saying: 'hi philippe'. Sep 03 11:38:36 a, no Sep 03 11:39:52 pH5: it looks like you can insert the branchname into the checkout Sep 03 11:40:00 http://www.kernel.org/pub/software/scm/git/docs/git-checkout.html Sep 03 11:40:27 gah Sep 03 11:40:33 checkout != checkout-index Sep 03 11:41:32 cu Sep 03 11:45:27 * pH5 gives up Sep 03 11:45:33 I'll just check it out manually Sep 03 11:45:51 :( Sep 03 11:46:19 I can really see how 'easy' and 'user-friendly' git is Sep 03 11:48:48 * koen tries to locate the books he'll need tomorrow Sep 03 12:36:08 anyone good at generating bitmasks? Sep 03 12:46:15 morning a;; Sep 03 12:46:28 morning all even :) Sep 03 12:47:45 hey! Sep 03 12:47:54 hrw|gone: If the .oz345x branch has a qemu-native.inc, it will probably benefit from my FILESPATH changes... Sep 03 12:48:10 koen: cairo bug noted, we should drop the patch Sep 03 12:51:26 * koen puzzled Sep 03 12:51:33 koen: yes? Sep 03 12:51:39 I think I have 1.2.4 + fixed patch on my hx Sep 03 12:51:56 koen: But it didn't work for others? Sep 03 12:55:13 does it matter if the apps were built against 1.2.2? Sep 03 12:56:57 RP: did o-hand test that patch with OABI or EABI? Sep 03 13:03:16 koen: OABI Sep 03 13:03:40 koen: It shouldn't matter what they were built against Sep 03 13:03:48 ok Sep 03 13:04:11 Xw100 + cairo 1.2.4-r1 on Ångström seems to work Sep 03 13:04:17 (on a ipaq hx4700) Sep 03 13:06:57 koen: hmm :-/ Sep 03 13:08:20 and xw100 seems to be 'faster' as xfbdev on the hx Sep 03 13:08:47 koen: Did cario seem any faster? Sep 03 13:09:30 gtk feels faster, but's that subjective Sep 03 13:09:50 and iirc XorA|gone is on oz/softfloat and had the issues as well Sep 03 13:11:26 hey mikearthur Sep 03 13:11:58 hey koen Sep 03 13:15:15 Good Morning Sep 03 13:16:21 hey B_Lizzard Sep 03 13:16:44 koen: I'm tempted just to drop the patch for now... Sep 03 13:16:46 morning Sep 03 13:17:06 RP: that's the quickest way to fix it Sep 03 13:17:14 I have a question to ask... Does the version of glibc used by OE have to be the same version as in the host computer? (Say, building glibc-2.4 needs a system with libc 2.4) Sep 03 13:17:32 Cause maybe that's where my problem lies Sep 03 13:17:45 B_Lizzard: no, my system doesn't even have a GNU libc Sep 03 13:18:10 Damn... Well, it was worth a try. Sep 03 13:18:24 pkg-config is driving me nuts Sep 03 13:18:32 B_Lizzard: feel free to pastebin your error Sep 03 13:18:35 ~pastebin Sep 03 13:18:41 somebody said pastebin was a place to paste your stuff without flooding the channel - try http://pastebin.com/ (BROKEN AND SUCKING NUTS), or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com, or http://bzflag.pastebin.ca/, or http://paste.lisp.org/ for the lisp/scheme nuts, or http://paste-it.net Sep 03 13:20:13 koen: could you do a "grep PKG_CONFIG_MIN_VERSION /usr/share/aclocal/pkg.m4" on ewi for me? Sep 03 13:20:47 koen@bitbake:~$ grep PKG_CONFIG_MIN_VERSION /usr/share/aclocal/pkg.m4 Sep 03 13:20:48 koen@bitbake:~$ Sep 03 13:20:55 thought so Sep 03 13:21:19 could you also grep pkg.m4 in staging? Sep 03 13:22:13 03rpurdie 07org.oe.dev * r1f5634cf... 10/ (3 files in 3 dirs): cairo 1.2.4: Drop broken patch pending a rethink Sep 03 13:22:26 Well.... Koen heard my problem yesterday...Basicly, I set up OE on my system yesterday, and I tried to build nano, just for fun... The toolchain is built ok, but when it reaches the point to build glibc, it says that "Preffered version 2.4 does not exist", even though it does... It then continues to build 2.3.5+cvs, which fails, probably because I use GCC4. When I do "bitbake -b ../org.openembedded.dev/packages/glibc/glibc_2.4.bb" it builds ok! Sep 03 13:23:41 I tried both bitbake branches, and the problem persists. Sep 03 13:23:48 Both 1.4 and 1.6 Sep 03 13:24:46 koen@bitbake:~$ grep PKG_CONFIG_MIN /data/build/koen/OE/build/tmp/angstrom/staging/i686-linux/share/aclocal/pkg.m4 Sep 03 13:24:46 koen@bitbake:~$ Sep 03 13:25:00 I've deleted ../build/tmp many times, updated everything... Still, nothing Sep 03 13:25:22 B_Lizzard: e.g. what machine+distribution do you use? Sep 03 13:25:38 B_Lizzard: e.g. the GNU toolchain bits need to be carefully chosen Sep 03 13:25:47 My machine is a normal P4, Distro is Ubuntu Dapper Sep 03 13:25:53 B_Lizzard: gcc, glibc and binutils need to match Sep 03 13:26:07 B_Lizzard: what machine+distribution do you want to build? Sep 03 13:26:20 B_Lizzard: in theory the host system should have no influence at all Sep 03 13:26:26 Jornada6xx (SH3) and Jlime Sep 03 13:26:51 and the DISTRO and MACHINE you have set exist? Sep 03 13:27:17 As far as I know, local.conf is just fine Sep 03 13:27:18 koen: thanks Sep 03 13:27:37 I've followed a guide by Kristoffer, if you know him Sep 03 13:28:00 B_Lizzard: The DISTRO should set the preferred versions of gcc, glibc and binutils Sep 03 13:28:01 koen: same here but orbit2-native is bitching "error: possibly undefined macro: PKG_CONFIG_MIN_VERSION Sep 03 13:28:01 " Sep 03 13:28:10 B_Lizzard: I have not met him, but he is here as well :) Sep 03 13:28:33 Yes, it does... They seem to be higher than what my host system provides Sep 03 13:28:49 B_Lizzard: that really shouldn't matter Sep 03 13:28:55 But that should not be a factor Sep 03 13:29:00 B_Lizzard: :) Sep 03 13:29:06 :) Sep 03 13:29:20 I'm stumped Sep 03 13:29:39 Why doesn't bitbake find glibc 2.4? Sep 03 13:29:43 B_Lizzard: I have little clue about what versions of gcc,glibc,binutils works best for SuperH Sep 03 13:30:03 B_Lizzard: a) a bug, and b) it needs a special glibc-initial (to support NTPL) Sep 03 13:30:26 B_Lizzard: and somehow probably only angstrom (distribution) sets that Sep 03 13:30:28 The problem is not one of compilation errors... My toolchain seems to be fine. Sep 03 13:30:40 I'll look into it... Sep 03 13:31:01 ~seen Kristoffer Sep 03 13:31:07 So, I should look into setting glibc-initial in the distro file? Sep 03 13:31:12 kristoffer was last seen on IRC in channel #oe, 5d 19h 15m 5s ago, saying: 'nevermind, big duh'. Sep 03 13:31:31 B_Lizzard: we could check your local.conf for an obvious error Sep 03 13:31:41 B_Lizzard: but then it might be best to ask Kristoffer/Ken Sep 03 13:32:16 The distro and device files should be ok... Kris has been working with them for some time. Sep 03 13:32:59 ah okay :) Sep 03 13:33:16 and you are using jlime-donkey? Sep 03 13:33:22 Yes, yes Sep 03 13:33:25 :) Sep 03 13:35:24 Should I post my local.conf in pastebin? Sep 03 13:35:36 I'm not sure if it'll help or anything... Sep 03 13:36:03 B_Lizzard: no, I assume it is correct Sep 03 13:36:11 I sadly have no clue Sep 03 13:36:20 Neither do I... Sep 03 13:36:40 I'll delete everything and restart Sep 03 13:37:05 I assume having everything in /opt/oe doesn't hurt? Sep 03 13:40:05 Ok, I'll just start again... Sep 03 13:40:15 Thanks for your interest, zecke Sep 03 13:40:26 It's very much appreciated Sep 03 13:43:13 03koen 07org.oe.dev * r3596a639... 10/ (1 packages/networkmanager/networkmanager_0.6.4.bb): networkmanager: RDEPEND on iproute2 Sep 03 13:44:51 B_Lizzard: I'm comparing the angstrom and jlime configuration ATM Sep 03 13:45:28 I looked into it yesterday... I don't remember any tragic differences. Sep 03 13:46:09 Still... Sep 03 13:46:20 http://pastebin.ca/159762 Sep 03 13:46:36 I can't find that in the jlime conf, but I have not used find yet Sep 03 13:47:23 Hmmm... Sep 03 13:47:59 by defining that your "Nothing provides..." should go away Sep 03 13:48:37 I'll try that Sep 03 13:49:07 I think that I should change "arm-linux-libc-for-gcc" to "sh3-linux-libc-for-gcc" Sep 03 13:50:25 I'm getting this bug: http://bugs.openembedded.org/show_bug.cgi?id=647. Any pointers what might cause it? Sep 03 13:50:48 B_Lizzard: the 'first' line in the pastebin should already work Sep 03 13:50:56 Yes Sep 03 13:51:09 CoreDump|home: eek, autohell :) Sep 03 13:51:17 :\ Sep 03 13:51:22 Thanks for that, zecke Sep 03 13:51:48 B_Lizzard: find out if it works :) Sep 03 13:52:07 B_Lizzard: then talk to Kristoffer to add it (if it worked) Sep 03 13:52:15 I deleted /opt/oe, so I should be back up in 20 minutes or so Sep 03 13:52:20 Will do Sep 03 13:52:42 :P Sep 03 14:05:21 hi all Sep 03 14:10:39 hey mr_nice_slacker Sep 03 14:11:29 koen: hi Sep 03 14:16:02 That "hi" was deadly Sep 03 14:17:35 B_Lizzard: hehe - could be Sep 03 14:20:35 :) Sep 03 14:43:59 RP: Im on OZ/Softfloat and cairo patch was busted for me Sep 03 14:55:19 B_Lizzard: is it working now? Sep 03 14:56:35 I'm building.... Sep 03 14:56:45 I still don't know Sep 03 14:57:10 My connection has been giving me crap Sep 03 14:57:16 But I think this is it Sep 03 14:58:51 XorA|gone: Strange as it worked on that for Jorn :-/ Sep 03 15:00:11 RP: :-( Sep 03 15:00:36 RP: so I guess its a complex interaction with various bits then Sep 03 15:00:56 I blamed it on w100 driver when I first saw it, but then I reverted to Xfbdev and it still happened Sep 03 15:01:35 koen|away: yay, embedded abiword is fixed in CVS HEAD, Ill have to see if I can pull patch in Sep 03 15:07:37 XorA: I guess so :-/ Sep 03 15:11:32 RP: or some wierd x86_64 build fault :-( Sep 03 15:13:22 RP: at least after the 25th I will have a x86 build host to play with :-) Sep 03 15:26:51 http://www.abiword.org/mailinglists/abiword-dev/2006/Sep/0014.html Sep 03 15:27:01 "One of the reasons for this change is so that I can Sep 03 15:27:01 easily build different versions of AbiWord directly from CVS using Sep 03 15:27:01 OpenEmbedded for performance testing." Sep 03 15:27:19 we should stick such quotes on the webpage somewhere Sep 03 15:28:32 koen|away: nice Sep 03 15:30:53 * XorA is just wasting time until fireworks Sep 03 15:31:20 * koen|away is waiting for the junkfood to get delivered Sep 03 15:32:05 koen|away: just built abiword embedded, so I will see the differences Sep 03 15:32:14 ah time to drive into town, bye Sep 03 15:33:01 RP: do you think you can make it to OEDEM? Sep 03 15:35:05 koen|away: psst Sep 03 15:35:20 zecke: yes? Sep 03 15:35:24 koen|away: he said yes already, do not give him the oppurtunity to reconsider it :) Sep 03 15:35:34 ah, I must have missed that Sep 03 15:36:28 ah, food Sep 03 15:36:40 * koen|away hears the doorbell Sep 03 16:26:00 koen|away: Yes, its likely I will be accompanied by tf as well Sep 03 16:27:02 ah, cool Sep 03 16:27:20 I won't bitch about abiword this time ;) Sep 03 16:28:08 koen|away: I'm sure he didn't mind and it did get a few of the issues resolved :) Sep 03 16:28:24 :) Sep 03 16:31:58 btw, Xw100 works real nice on my hx Sep 03 16:32:30 koen: cool. I'll have to give it a go on the c7x0 :) Sep 03 16:33:28 later Sep 03 17:04:39 hi all ... if I have to use tmake in a receipt to create the .pro file; in which do_??? () {} should I put it? Sep 03 17:05:08 and do I have to add something to my PATH var - as tmake can not be found? Sep 03 17:06:26 => find ./ -name tmake found it @ several places - but nowhere in staging Sep 03 17:08:31 http://bash.org/?674136 Sep 03 17:09:08 ljp: wasn't it enough to fuck up Qtopia's buildsystem? why did you fuck up QtE as well? Sep 03 17:10:06 zecke: to get more paid support Sep 03 17:11:56 koen: actually I do not think that this is the case Sep 03 17:13:06 zecke: trolltech bought novell, redhat, imendio and nokia and is subtly encouraging people to switch to gtk? Sep 03 17:14:01 koen: haeh? Sep 03 17:15:04 or they don't have an infinite number of monkeys and typewriters anymore Sep 03 17:17:10 :-D Sep 03 17:17:34 actually the latest Qt4.2 snapshot is the fastest Qt4 ever Sep 03 17:17:59 qt4 is pretty awesome Sep 03 17:18:22 it's just that the /e version seems to be made by incompetent idiots Sep 03 17:18:32 koen: i heard qt4 is pretty buggy Sep 03 17:18:43 pavlix: 'pretty' is relative Sep 03 17:18:58 pavlix: but it has a open issue tracker with patches+etc Sep 03 17:19:22 zecke: psi developers are holding back new psi release because of the bugs Sep 03 17:19:36 koen: oh well, I know some of these 'incompetent idiots' and I wouldn't label them as such ;) Sep 03 17:19:41 zecke: do they incorporate 3rd party patches? Sep 03 17:20:03 ... did the tmake questions trigger this - or is it not related? Sep 03 17:20:06 pavlix: sure, if you grant permissions Sep 03 17:20:16 pavlix: maybe :) Sep 03 17:20:24 pgfeller: which tmake question? Sep 03 17:20:42 zecke: if I have to use tmake in a receipt to create the .pro file; in which do_??? () {} should I put it? Sep 03 17:20:48 and do I have to add something to my PATH var - as tmake can not be found? Sep 03 17:20:57 => find ./ -name tmake found it @ several places - but nowhere in staging Sep 03 17:21:01 pgfeller: a) I would use qmake Sep 03 17:21:05 pgfeller: b) do_configure Sep 03 17:21:16 pgfeller: and you need to put something like tmake-native in your DEPENDS Sep 03 17:21:27 pgfeller: and write a tmake-native similiar to the qmake-native one Sep 03 17:21:35 zecke: :-) Sep 03 17:21:41 zecke: thanks Sep 03 17:21:48 pgfeller: in most cases qmake is a drop-in replacement though Sep 03 17:22:02 zecke: will try qmake Sep 03 17:22:47 koen: I wonder if releasing the greenphone means Qtopia Phone will be GPLed Sep 03 17:23:13 zecke: did the devil order skis and warm clothing? Sep 03 17:23:14 koen: otherwise a HTC + a Gtk+ license is about the same price Sep 03 17:23:26 zecke: I wouldn't hold your breath :) Sep 03 17:23:35 you have to pay for gtk nowadays? Sep 03 17:23:37 sheesh Sep 03 17:23:38 ggilbert: hey Sep 03 17:23:45 I should stop now :) Sep 03 17:23:50 koen: that is the point ;) Sep 03 17:24:15 * zecke would wish Trolltech had a clear road map Sep 03 17:24:45 a ROAD map? Sep 03 17:26:08 ;) Sep 03 17:26:17 zecke: I think I'll get a hotel near the Tierpark and kurfurstendam, since the metro stations there can get me to the U6 Sep 03 17:27:04 koen: City-Center West? Sep 03 17:27:17 that's one of the hotels on the list Sep 03 17:27:42 koen: hmm, I have no clue about good hotels Sep 03 17:28:46 * koen neither Sep 03 17:28:57 luckily I don't have high standards :) Sep 03 17:33:59 I was wondering about hotels and what the best option was Sep 03 17:34:44 a) crash at my place Sep 03 17:34:57 b) get a hotel near a metro station of line 'U6' Sep 03 17:37:53 c) get a hotel near a line that connects to U6 Sep 03 17:38:55 http://dominion.kabel.utwente.nl/koen/berlin-subway.pdf <- map Sep 03 17:39:03 It'll be a hotel given tf will be with me Sep 03 17:44:06 Tempelhof is on the U6? Sep 03 17:44:35 yes Sep 03 17:49:07 if oe_runmake can not find uic, what depends do I have to add? Sep 03 17:50:13 RP: yes, and it is 'very' near to where I life Sep 03 17:50:23 pgfeller: that depends :) Sep 03 17:50:41 RP: We will be either near 'Alt Mariendorf' Sep 03 17:50:53 zecke: :-D ... make: /stuff/tmp/staging/arm-linux/qt2/bin/uic: Command not found Sep 03 17:50:58 RP: or 'Oranienburger Tor' but we are not yet sure, as new office space is not yet rented Sep 03 17:51:20 pgfeller: take a look at qmake.bbclass and qmake-base.bbclass Sep 03 17:51:40 zecke: thanks ... 'll do Sep 03 17:52:16 zecke: ok, that's helpful to know, thanks :) Sep 03 17:53:34 I'll try and look up flights etc tomorrow... Sep 03 18:03:12 zecke: ah - I missunderstood; now I inherit qmake (removed the DEPENDS = "qmake-native") - then it compiles Sep 03 18:05:52 I almost feel guilty to ask - but what provides libstdc? Sep 03 18:09:13 pgfeller: most likely gcc(-cross) Sep 03 18:10:34 koen: thanks ... if I can not link with the following error: ... undefined reference to `std::ios_base::Init::~Init()' ... do I have a problem with the x-compiler then? Sep 03 18:10:50 keon: from log.do_compile Sep 03 18:14:36 zecke: did you see https://svn.wiisard.org/WIISARD/trunk/meta/bitbake-svn-port-swb.patch already? Sep 03 18:16:09 in case you wonder: https://svn.wiisard.org/WIISARD/trunk/meta/OpenEmbedded/packages/ Sep 03 18:23:26 pgfeller: gcc should 'provide' it Sep 03 18:24:10 zecke: so I have to add a depend, or inherit something to stage it? Sep 03 18:24:35 pgfeller: no :) Sep 03 18:24:41 pgfeller: base.bbclass should do that for you Sep 03 18:24:56 pgfeller: and that class is inherited by default Sep 03 18:25:08 zecke: do I have to inherit base - or is that done automatically? Sep 03 18:25:13 ;) Sep 03 18:25:31 my yedi force 'forsaw' this question Sep 03 18:25:34 pH5: I get less shm troubles when logged in as root it seems Sep 03 18:25:34 zecke: I see ... :-) Sep 03 18:25:53 pgfeller: BUT Sep 03 18:26:06 pgfeller: qmake.bbclass is doing a stupidity Sep 03 18:26:23 pgfeller: LIBS-=-lstdc++ LIBS+=-lsupc++ you might suffer from this one Sep 03 18:28:04 zecke: I deleted the cache ... @ the moment bitbake is processing the receipts; when done I'll pastebin the do_compile() log ... OK? Sep 03 18:28:23 pgfeller: no, why do you think cleaning the caches is needed? Sep 03 18:29:23 zecke: as you said it should be done by default I thought maybe I confused the thing somehow with my .bb (as I have to experiment a lot) Sep 03 18:30:37 zecke: http://pastebin.ca/159954 Sep 03 18:32:46 looks indeed like the libsupc/libstdc problem Sep 03 18:34:30 what can I do to overcome this? - I don't feel qualified to modify qmake.bbclass ;) Sep 03 18:35:00 CPPFLAGS += "-lstdc++" or somthing like that Sep 03 18:40:04 koen: LDFLAGS += "-lstdc++" did the trick - thanks. Sep 03 18:40:44 * koen wants a 4way MPCore dev board Sep 03 18:41:54 hmm I should consider pushing my PALMTOP_NEED_STL patch Sep 03 18:55:17 any uClibc gurus around? Sep 03 18:56:30 d'oh Sep 03 18:59:08 * pgfeller smashes his head on the table Sep 03 18:59:57 you as well? Sep 03 19:00:24 maybe I need a new hobby :-/ ...any hints what I do wrong now?: http://pastebin.ca/159986 Sep 03 19:00:32 blaster8: :-) Sep 03 19:01:18 blaster8: yep - have a hard time - unfortunately I'm not a guru, so it is unlikely I can assist you Sep 03 19:01:41 03koen 07org.oe.oz354x * r12853516... 10/ (1 packages/libsndfile/libsndfile1_1.0.11.bb): Sep 03 19:01:41 libsndfile: improve packages, from .dev Sep 03 19:01:41 applied changes from 3596a639cb106771bb7608dc97bc4ad1d5c21e7c Sep 03 19:01:41 through ae090523c163ff0d1e234b0d89d1450a2466b4b4 Sep 03 19:01:42 distinct lack of gurus, I reckon ;) Sep 03 19:01:46 03koen 07org.oe.dev * rae090523... 10/ (1 packages/libsndfile/libsndfile1_1.0.16.bb): libsndfile: slightly improve packaging Sep 03 19:02:12 blaster8: rare specimen they are Sep 03 19:02:12 * koen hugs mtn pluck Sep 03 19:02:51 * koen is waiting for a new uclibc release before trying to build it Sep 03 19:03:13 when is that going to come? Sep 03 19:03:21 it's been a long time since the last release... Sep 03 19:03:24 hello Sep 03 19:03:42 hello pavlix Sep 03 19:04:02 how do i get a compiler for mipsel/uclibc working? Sep 03 19:04:43 when bitbaking any package, it says i don't have a virtual package Sep 03 19:04:51 with lots of tinkering, usually ;) Sep 03 19:05:25 pavlix: I'm a newbe - but can't you configure this in local.conf? Sep 03 19:05:53 pavlix: its generally a good idea to put the exact error message up somewhere. no offense, but you dont know enough about oe yet to draw conclusions from those messages yourself Sep 03 19:06:22 kergoth: true, I'm just building another think, just to test... Sep 03 19:06:24 * koen stabs -dev package dependencies Sep 03 19:06:30 pavlix: are you using your own cross-compiler or trying to get OE to build one? Sep 03 19:06:34 or rather, the lack there off Sep 03 19:07:03 blaster8: i was letting oe build one, but, is it even possible? Sep 03 19:07:41 yes, but as far as I understand it, you will need a meta package describing the cross-compiling environment you want to set up Sep 03 19:08:38 blaster8: looks like you're right Sep 03 19:09:04 Alternatively, build a toolchain of your own (patched crosstool perhaps?) and follow this: http://www.openembedded.org/user-manual&dpage=ch05s05 Sep 03 19:09:52 but if you want to use long term or for a project involving more than just you, I'd suggest figuring out a meta package Sep 03 19:10:44 blaster8: it would be use by other people too Sep 03 19:10:59 time to write a meta package, then :) Sep 03 19:11:05 * koen looks at conf/machine/asuswl500g.conf Sep 03 19:11:22 blaster8: and even someone here asked me to report back Sep 03 19:11:24 http://pastebin.ca/159986 ... google thinkgs there's no main() function - but the code should be a stable one (I just would like to practice to create .bb's :-( ) Sep 03 19:11:59 conf/machine/wl500g.conf rather Sep 03 19:12:01 blaster8: you just need a appropriate machine config file Sep 03 19:12:12 got it Sep 03 19:12:21 koen: thanks Sep 03 19:12:22 blaster8: but OE can do without if you supply the appropriate params in local.conf Sep 03 19:12:36 pavlix: I pointed you to that a few days ago as well Sep 03 19:13:15 koen: ok, i might have been a bit confused then Sep 03 19:15:11 configure: creating ./config.status Sep 03 19:15:18 take take, OOo! Sep 03 19:16:30 ok, what's better for me to start with? prebuilt toolchain or oe-made toolchain? Sep 03 19:16:46 oe-made Sep 03 19:17:44 koen: ok, I'll try Sep 03 19:20:20 *lol* - it added my comment to the linker flags ... finaly it builds :-) Sep 03 19:24:41 thanks for pointers, koen :) Sep 03 19:27:25 * JustinP hacks freeze Sep 03 19:27:51 JustinP: you deleted mtnpatch too early Sep 03 19:28:01 JustinP: we need it for the svn mirror to work properly Sep 03 19:28:24 good thing we have an SCM :) Sep 03 19:29:01 koen: heh Sep 03 19:29:07 really....what is it used for? Sep 03 19:29:14 I didn't think anyone had ever used it ;-) Sep 03 19:29:23 * JustinP cheers Sep 03 19:32:46 yay! Sep 03 19:32:58 It looks like I can actually start the compile part of OOo Sep 03 19:33:02 so...what is it used for exactly? Sep 03 19:33:31 JustinP: with mtnpatch we can tie in the ciabot to update a svn mirror of the OE metadate Sep 03 19:33:54 @&R*&$*(@#$ Sep 03 19:34:04 * koen runs the THIRD configure Sep 03 19:34:31 * koen wonders who designed the OOo buildsys Sep 03 19:34:34 ah, so you use "svn" instead of "mtn"? Sep 03 19:35:06 JustinP: no, we got sick of all those whiners and/or people to dumb to read mtn docs Sep 03 19:35:47 I still don't quite understand..... Sep 03 19:36:14 how would you use something which parses an mtn patch and updates an mtn workspace to interoperate with svn... Sep 03 19:36:19 http://projects.linuxtogo.org/plugins/scmsvn/viewcvs.php/?root=openembedded needs to be keps up to date Sep 03 19:36:49 ciabot will feed a mtn diff to mtnpatch which will update the svn workspace Sep 03 19:37:03 that way commits map 1:1 Sep 03 19:37:12 wow Sep 03 19:37:13 a poor mans tailor Sep 03 19:37:19 there must have been a lot of whiners Sep 03 19:37:34 hmmmm.. Sep 03 19:37:46 well, glad it was being used ;-) Sep 03 19:37:50 did you resurrect it? Sep 03 19:38:32 it was being planned to be used Sep 03 19:38:44 mickeyl didn't implement it yet Sep 03 19:38:48 a Sep 03 19:43:02 ..... Sep 03 19:43:16 is there something which stops me from setting BBFILES in a file included in local.conf? Sep 03 19:44:17 include frozen-bbfiles.conf Sep 03 19:44:24 frozen-bbfiles.conf sets BBFILES Sep 03 19:44:28 but BBFILES isn't being set Sep 03 19:44:31 copy/paste works fine Sep 03 19:45:06 Are you setting it in local.conf with := ? Sep 03 19:45:14 no, I commented it out... Sep 03 19:45:23 (:= is what exactly again?) Sep 03 19:45:56 * JustinP looks in docs Sep 03 19:46:07 Its evaluated when parsing instead of afterwards Sep 03 19:46:56 ah....well that shouldn't matter then Sep 03 19:47:13 I've tried using := and = in frozen-bbfiles.conf Sep 03 19:47:14 no different Sep 03 19:47:26 I tried setting some other var and it's not being set Sep 03 19:48:23 bitbakes path handling confuses me. I have some scripts which work and therefore I've not had the need to learn/remember/// Sep 03 19:50:49 argh Sep 03 19:50:56 one include works, another refuses to Sep 03 19:51:42 got this more or less from an example: install -m 644 ${WORKDIR}/mileage/mileage ${D}${palmtopdir}/bin/ ... Sep 03 19:51:58 but bitbake is not happy: Sep 03 19:51:58 cannot create regular file `/stuff/tmp/work/carmilecalc-4.2.0-r0/image/opt/QtPalmtop/bin/mileage': No such file or Sep 03 19:51:59 directory Sep 03 19:52:32 ... hints to where I can find additional info on do_install? Sep 03 19:52:34 heh, bitbake is not a happy program Sep 03 19:52:58 sure it is Sep 03 19:53:06 agree :-) Sep 03 19:53:09 * T0mW continues struggling to get past dbus-glib Sep 03 19:53:11 pgfeller is trying to install a file to a directory that doesn't exist Sep 03 19:53:20 it's *install* failing, not bitbake Sep 03 19:53:31 pgfeller: install -d ${D}${palmtopdir}/bin/ Sep 03 19:53:32 pgfeller: 'install -d ${D}${palmtopdir}/bin/ ' Sep 03 19:53:38 :) Sep 03 19:53:40 * JustinP giggles Sep 03 19:53:48 or...not... Sep 03 19:53:51 pretend that never happened Sep 03 19:54:10 koen: thanks Sep 03 19:54:18 JustinP: np - thank anyhow. Sep 03 19:54:35 * koen declares the OOo buildsystem to be 'batshit insane' Sep 03 19:54:53 ERROR: Nothing provides dependency virtual/mipsel-linux-uclibc-gcc Sep 03 19:54:53 ERROR: dependency virtual/mipsel-linux-uclibc-gcc (for gcc) not satisfied Sep 03 19:54:53 NOTE: no buildable providers for gcc Sep 03 19:55:23 does it mean I have to write a recipe for mipsel-linux-uclibc-gcc or just add id somewhere as PROVIDES? Sep 03 19:57:09 'Brian's DTrace probes can observe the state of Schroedinger's cat, but he won't tell because he prefers leaving it as a teaching example.' Sep 03 19:57:26 hehe Sep 03 19:57:36 * JustinP says "fuck it" and appends to local.conf Sep 03 19:57:49 not that this code really works yet anyway Sep 03 19:58:07 PREFERRED_PROVIDER_virtual/mipsel-linux-uclibc-gcc = "gcc-cross" or something like that Sep 03 20:00:47 PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}gcc:gcc-cross" Sep 03 20:00:52 this is in the default one Sep 03 20:02:02 that sometimes doesn't work Sep 03 20:02:07 I don't know why Sep 03 20:02:12 * koen considers it a bug Sep 03 20:03:15 koen: i'm still not sure about the files in org.openembedded.dev/conf Sep 03 20:03:25 how are they loaded? Sep 03 20:03:36 * RP suspects its a bug too... Sep 03 20:03:41 pavlix, they are added to your local.conf time at runtime Sep 03 20:04:02 Kristoffer: and it depends on some of the variables, right? Sep 03 20:04:08 check the bottom part of conf/bitbake.conf Sep 03 20:04:13 depends on what distro is set too Sep 03 20:05:08 pavlix@localhost ~/oe $ bitbake gcc --debug Sep 03 20:05:08 DEBUG: CONF reading /home/pavlix/oe/org.openembedded.dev/conf/bitbake.conf Sep 03 20:05:08 DEBUG: CONF including /home/pavlix/oe/build-wl500g/conf/local.conf Sep 03 20:05:08 DEBUG: CONF including /home/pavlix/oe/org.openembedded.dev/conf/machine/wl500g.conf Sep 03 20:05:08 DEBUG: CONF including /home/pavlix/oe/org.openembedded.dev/conf/distro/generic-uclibc.conf Sep 03 20:05:08 DEBUG: CONF including /home/pavlix/oe/org.openembedded.dev/conf/distro/generic.conf Sep 03 20:05:10 DEBUG: CONF including /home/pavlix/oe/org.openembedded.dev/conf/documentation.conf Sep 03 20:05:13 DEBUG: CONF including /home/pavlix/oe/org.openembedded.dev/conf/sanity.conf Sep 03 20:05:20 this is what gets loaded Sep 03 20:05:37 Kristoffer: someone had troubles with building jlime Sep 03 20:05:53 koen, what troubles? Sep 03 20:06:05 failing to build the correct glibc Sep 03 20:06:12 ah yes, thats true Sep 03 20:06:19 koen: ok, I see what you mean Sep 03 20:06:36 koen, I had some issues with it not draging 2.4 (although it was stated).. So I added a define line to local.conf instead Sep 03 20:07:02 ah Sep 03 20:07:02 koen, It couldnt be forced to build 2.4 no matter how hard I tried Sep 03 20:07:15 Kristoffer: could you add that define to jlime-.conf? Sep 03 20:08:09 koen, yes, but the problem is that Im building for multimple platforms and that define is specific for each platform. So if i set it, I can no longer build for the other platform. But how about I put a notice inside the conf file that state due to bug you must do this.. Sep 03 20:08:54 Kristoffer: that's why overrides exist :) Sep 03 20:09:07 FOO_BLAH_sh = "something" Sep 03 20:09:20 FOO_BLAH_arm = "something else" Sep 03 20:09:24 doh, didnt even think of that.. Sep 03 20:09:29 :) Sep 03 20:09:31 I'll add it and push .. Sep 03 20:09:35 thx man Sep 03 20:10:52 * JustinP answers his question before he asks it Sep 03 20:13:40 wowo Sep 03 20:13:43 -o Sep 03 20:13:46 ok, I have, again, decided to ripout tmp dir and start over: dbus-glib problems up the whazoo. I have a working gcc-3.4.4 in the tmp dir that I would like to re-use and forego the time to rebuild it. How can I keep that gcc/binutils ? Sep 03 20:13:54 JustinP: do you know the sound of one hand clapping as well? Sep 03 20:14:21 koen, "PREFERRED_PROVIDER_sh3_virtual/sh3-linux-libc-for-gcc = "glibc-intermediate"" sound abit right? Sep 03 20:14:54 PREFERRED_PROVIDER_virtual/sh3-linux-libc-for-gcc = "glibc-intermediate" Sep 03 20:15:12 since 'sh3' is already in the virtual part Sep 03 20:15:28 thought it looked weird. thx Sep 03 20:19:50 PREFERRED_PROVIDER_virtual/sh3-linux-libc-for-gcc_sh3 should have been the correct form Sep 03 20:20:16 but as we have seen is a bit superfluous Sep 03 20:21:30 hi Sep 03 20:22:04 hey hrw|furied Sep 03 20:22:13 hi hrw|furied Sep 03 20:22:17 furied? Sep 03 20:22:31 as 'in fury/angry' Sep 03 20:22:44 my english suxx Sep 03 20:22:46 ah, OOo actually starts building something now Sep 03 20:22:52 my home machine SUXX Sep 03 20:22:58 hrw|furied: I gathered that much, I was wondeirng why? Sep 03 20:23:18 koen: it's strangely soothing Sep 03 20:23:19 I will probably stop using it until get any cash to be able to test it with other components Sep 03 20:23:48 hdd does not have badblocks on /dev/hda1 but filesystem on hda1 goes into /dev/null on each reboot... Sep 03 20:24:07 memtest86 pass ok even after 12h nonstop Sep 03 20:24:10 hmm, it fails on compiling boost somewhere due to busybox-isms Sep 03 20:24:15 http://pastebin.com/782954 Sep 03 20:24:29 I am starting from the minimum and adding needed info.... Sep 03 20:24:41 but it seems I can't Sep 03 20:24:51 ....what is d....? the docs don't say.... Sep 03 20:24:57 expand(..., d) Sep 03 20:25:13 brb (if it will get up from backup) Sep 03 20:25:18 koen, so is PREFERRED_PROVIDER_virtual/sh3-linux-libc-for-gcc = "glibc-intermediate" correct or PREFERRED_PROVIDER_virtual/sh3-linux-libc-for-gcc_sh3? Sep 03 20:25:43 both, but the first one is better Sep 03 20:25:58 ah, oki Sep 03 20:26:02 * JustinP wonders how this makes any sense Sep 03 20:27:11 at least I see that it's in python... so reading the sources won't be hard for me Sep 03 20:27:36 hrw|furied: did you boot linux 2.6.17.* on that box? Sep 03 20:27:52 If so your XFS might be f*cked Sep 03 20:29:13 is there a target that should work on all architectures and configs? a basic one to test if building works at all? Sep 03 20:29:17 ~lart stlport Sep 03 20:29:18 * ibot moos at stlport Sep 03 20:29:50 hrw|furied, it says that Zaurus is using 2.4.20 kernel. But what does openzaurus use? Sep 03 20:30:10 2.6.recent Sep 03 20:30:39 koen, does it use vanilla or vanilla + patches? Sep 03 20:30:50 the one in OE :) Sep 03 20:31:03 :D let me check Sep 03 20:32:02 * USAGE: Sep 03 20:32:03 * Source LinuxARMEnv.Set (in tcsh) or LinuxARMEnv.Set.sh (in sh) Sep 03 20:32:03 * in order to set up the build-environment variables. Sep 03 20:32:12 so OOo doesn't actually need tcsh.... Sep 03 20:33:43 pavlix: looks like your BBFILES are wrong Sep 03 20:34:03 'NOTE: Parsing finished. 0 cached, 0 parsed, 0 skipped, 0 masked. DEBUG: parsing complete' Sep 03 20:34:08 koen: yes i've forgotten Sep 03 20:34:34 koen: can I find anywhere the list of settings I *have to* put there? Sep 03 20:34:47 see local.conf.sample Sep 03 20:35:07 ibot: botsnack Sep 03 20:35:08 thanks, JustinP Sep 03 20:35:11 03kristoffer 07org.oe.dev * r43de53f0... 10/ (1 conf/distro/jlime-donkey.conf): Sep 03 20:35:11 conf/distro/jlime-donkey.conf : GTK related versions remove, bugfix Sep 03 20:35:11 * Removed all GTK, GLIB related stuff so we start bugtrack why osb-jscore Sep 03 20:35:11 doeesnt build. Sep 03 20:35:11 * Added PREFERRED_PROVIDER for glibc-intermediate so it grabs 2.4. Sep 03 20:35:12 Had to set specific defines for both archs due to some bitbake bug when Sep 03 20:35:13 using {ARCH} in define line. Sep 03 20:35:34 * JustinP mucks about in shell.py again Sep 03 20:36:14 CoreDump|home: 2.6.17.8 (which has xfs fixed) Sep 03 20:36:16 I was trying to do this all in freeze.bb but I didn't know how to access the functions I needed... Sep 03 20:36:34 hrw|furied: try running the "fixed" xfs_repair, too Sep 03 20:37:03 koen: yes, but that's full of stuff that's otherwise set by machine and distro confs, isn't it? Sep 03 20:37:04 CoreDump|home: xfs was so deeply fucked that I dropped it and now rootfs is ext3 from wednesday backup (which is shit) Sep 03 20:37:12 heh Sep 03 20:37:23 well, ext3 is slow but very stable Sep 03 20:37:33 * JustinP still uses reiserfs with no problems Sep 03 20:37:42 CoreDump|home: the problem is that this machine lost stability Sep 03 20:37:49 hmm Sep 03 20:37:59 koen, who is handling zaurus kernel commits to russell btw? Sep 03 20:38:05 during next weekend I will test with PSU from work (from working stable machine) Sep 03 20:38:10 Kristoffer: RP Sep 03 20:38:13 hmm, if I read the comments better Sep 03 20:38:16 ah goodie Sep 03 20:38:19 Kristoffer: RP is zaurus maintainer in mainline Sep 03 20:38:34 hrw|furied: could you tar.gz the working of gconf-dbus for me? Sep 03 20:38:36 hrw|furied, good to hear, regulary syncin with git then Sep 03 20:38:48 I'm not getting anywhere with my problem :( Sep 03 20:39:04 s/working/work dir Sep 03 20:39:39 CoreDump|home: workdir? I would have to rebuild it Sep 03 20:40:04 don't have one on ewi? Sep 03 20:40:25 CoreDump|home: rm_work.bbclass Sep 03 20:40:26 I clean tmp/work Sep 03 20:40:31 ah :( Sep 03 20:40:34 Kristoffer: We have got a large fraction of the Zaurus kernel code into mainline Sep 03 20:40:46 CoreDump|home: ewi has only 36G on /home Sep 03 20:40:53 ops. 60G Sep 03 20:40:57 /dev/sda2 60G 53G 6.1G 90% /home Sep 03 20:41:09 CoreDump|home: my ~/devel/ was 12G there Sep 03 20:41:11 RP, great to hear that its active development. Just took maintainership over Jornada 7xx just so I can update alot of faulty code Sep 03 20:41:21 heh Sep 03 20:41:23 RP, support for 3200 good? Sep 03 20:41:40 Kristoffer: I have a C3000 which is near enough the same hardware ;-) Sep 03 20:41:54 and the problem of problem is that I was not able to watch Aliens on tv ;( Sep 03 20:41:55 RP, haha :D always use the same devices as maintainers :) Sep 03 20:42:00 RP: your led stuff is catching on: http://ynezz.ibawizard.net/ts7250/ep93xx_leds.patch.gz Sep 03 20:42:26 could anyone else here upload the workdir of gconf-dbus somewhere? Sep 03 20:42:43 koen: I made the interface too simple ;-) Sep 03 20:42:49 koen, why is familiar still using handhelds.org kernel instead of trying to move it into vanilla? Sep 03 20:43:14 Kristoffer: I gave up on hh.org kernels a long time ago :-( Sep 03 20:43:44 RP, yeah. I basicly hate it :) trying to get all good stuff from it and make vanilla work out of the box Sep 03 20:43:54 Kristoffer: it is easier for them to maintain kernel that way Sep 03 20:44:12 Kristoffer: probably too much hacks not ready for mainline Sep 03 20:44:19 Kristoffer: Sounds like what I did with Zaurii ;-) Sep 03 20:44:20 hrw|furied, but isnt the obvious risc of it falling behind and being an unsupported branch?.. Sep 03 20:44:22 yeah most likely Sep 03 20:44:25 RP, :D Sep 03 20:44:46 Kristoffer: Although admittedly, I had to develop the 2.6 support as well :) Sep 03 20:45:25 RP, hehe. Alot more work then. The Jornada basicly had faulty support but it was more in need of a makeup Sep 03 20:46:38 koen, btw one weird idea. But how hard would it be to incorperate tgz as a package class? Sep 03 20:46:38 hh.org hasn't done badly from my efforts. LED, backlight, w100fb, rtc, sd and quite a few other things got merged with some help from me :) Sep 03 20:46:51 Kristoffer: It already is ;-) Sep 03 20:47:10 RP, I just wish they would give up and merge their stuff with vanilla and not use leftside hacks.. Sep 03 20:47:16 RP, tgz not tar.gz :D Sep 03 20:48:04 Kristoffer: The hh.org problem was that nobody was prepared to deal with making the code mainline quality Sep 03 20:48:20 Kristoffer: 'tar' already exists Sep 03 20:50:00 RP, most likely. But its a pity. Like for instance Im using alot of code from Alex Lange that is quite nice. But perfectly wasted on a side-stream branch. Sep 03 20:51:38 Kristoffer: I totally agree. If you need any advice on getting code into mainline, ask me as I have experience with most of the maintainers and a fairly good track record ;-) Sep 03 20:52:44 RP, Im trying to work my way through russells watchful eyes currently. First time I ever had to clean out whitespaces :D Im waiting for him to give me the final go ahead to submit. Where does it usually get submitted? Through russell ..? Sep 03 20:53:18 L-A-K -> patchtracker Sep 03 20:53:35 Kristoffer: It depends what you're submitting. If its arch/arm, it goes to Russell's patch tracker Sep 03 20:53:58 that &*&$%*@ patch-tracker doesn't understand gpg signing Sep 03 20:54:31 Kristoffer: As an idea of my patch turnover with the OZ project: http://www.rpsys.net/openzaurus/patches/archive/ :) Sep 03 20:54:47 RP, yeah currently everything is related to arch/arm. I've submitted patches and fixed those things he had opinions about. Once he "accepts" it though. Does he push it further or do I submit it to the patch tracker? Sep 03 20:55:10 Kristoffer: you submit it Sep 03 20:55:11 You submit to the patch tracker and he takes it from there Sep 03 20:55:36 Kristoffer: Have a look at the patch tracker as you can see the things other people have submitted Sep 03 20:55:38 RP, thats alot of patches :D how many years of work? Sep 03 20:55:41 you can scare a lot of people by showing them an linux-openzaurus.bb Sep 03 20:55:57 2.6.7 is ~2 years old, right? Sep 03 20:55:57 Kristoffer: Since June 2004 Sep 03 20:56:20 koen: Nov 2004 :) Sep 03 20:56:37 RP, nice :D Sep 03 20:56:55 * koen waits for Philippe to test the 2.6.17 netbook patches Sep 03 20:58:06 RP, currently what works and what doesnt on the Spitz? Sep 03 20:58:15 ~lart local.conf Sep 03 20:58:15 * ibot hooks into a hydrant and hoses local.conf down Sep 03 21:01:05 Kristoffer: http://wiki.openzaurus.org/KernelInfo26#SL-Cxx00 and http://rpsys.net/openzaurus/ Sep 03 21:01:23 RP: iirc overlay is the only 'missing' thing, right? Sep 03 21:01:37 koen: voltage control, cpufreq Sep 03 21:01:42 koen: that and cpu frequency scaling on pxa270 Sep 03 21:01:57 heh Sep 03 21:02:11 cpufreq is doing funky stuff for pcmcia on the hx and h2200 Sep 03 21:02:36 koen: good or bad "funky stuff"? Sep 03 21:02:40 the worst part: I wrote that pcmcia driver :) Sep 03 21:02:54 RP: resetting the card on freq changes Sep 03 21:03:03 lol Sep 03 21:03:18 koen: this would be why I haven't touched that posioned challis :) Sep 03 21:03:33 koen: Does it work on the c7x0? Sep 03 21:03:52 the ts on my h2200 is emulating a pre-mikearthur tosa-ts when doing cpufreq Sep 03 21:03:56 RP: I think it does Sep 03 21:04:08 * koen tries to locate the c700 Sep 03 21:05:15 I love the 'conservative' governor Sep 03 21:10:26 hrw|furied: could you add the update libsndfile to the feeds if you have time? Sep 03 21:10:41 how can I avoid recompiling the gcc-3.4.4 cross compiler? ASSUME_PROVIDED doesn't seem to work Sep 03 21:11:58 * CoreDump|home trashed tmp for the bazillionth time Sep 03 21:12:12 koen: is it in oz354x? Sep 03 21:12:58 I say: ASSUME_PROVIDED += "virtual/arm-linux-gcc-3.4.4", have that in my path, but bitbake still insists that it wants to build it. Sep 03 21:13:55 arm-linux-gcc-3.4.4 is in /usr/local/arm/3.4.4/bin ... Sep 03 21:14:14 Is there some other "magic" that I'm missing? Sep 03 21:19:54 did anyone mess w/ autoconf-native-2.59-r3 in .oz354x recently? Sep 03 21:20:29 ignore that =) Sep 03 21:20:37 ok, so I suppose that ASSUME_PROVIDED is broken as well Sep 03 21:21:39 zecke: you might wanna add "bzip2" to your "sanity" thingy Sep 03 21:28:16 CoreDump|home: adding bzip2-native to ASSUMED_PROVIDED Sep 03 21:30:38 damn it...why does bitbake segfault on me if I don't use the shell... Sep 03 21:31:24 segfault? Sep 03 21:33:31 zecke: yes, segfault Sep 03 21:33:40 miscompiled python? Sep 03 21:33:43 bitbake something /home/papercrane/oe/oz354.build Sep 03 21:33:46 zsh: 28512 segmentation fault bitbake something Sep 03 21:33:48 works fine with -i.... Sep 03 21:34:01 echo "build something; exit" | bitbake -i Sep 03 21:34:11 been this way for a long, long time Sep 03 21:34:18 I suppose I could try recompiling python... Sep 03 21:34:35 of course portage is working fine for me ;-) Sep 03 21:38:40 * JustinP hugs his new build-box Sep 03 21:38:57 * CoreDump|home does the same =) Sep 03 21:39:06 * kergoth too Sep 03 21:39:20 * JustinP 's build-box says thanks for the love Sep 03 21:39:41 dualcore rocks Sep 03 21:40:20 I don't have that, just an athlon64 Sep 03 21:40:28 much faster than my old athlon Sep 03 21:41:00 03freyther 07org.oe.dev * rf1283b89... 10/ (1 classes/sanity.bbclass): classes/sanity.bbclass: Check for bzip2 in the path as well Sep 03 21:41:05 CoreDump|home: done Sep 03 21:41:07 later Sep 03 21:41:13 zecke: thanks Sep 03 21:51:04 ~curse my buildbox Sep 03 21:51:06 May the fleas of a thousand camels infest your most sensitive regions, my buildbox ! Sep 03 21:52:39 AHAH!!! Sep 03 21:53:05 dies on import psyco Sep 03 21:53:40 JustinP: psyco and amd64? Sep 03 21:54:10 it was doing this before I had an amd64 Sep 03 21:54:16 and I'm not running a 64-bit OS.... Sep 03 21:54:54 whoa...psyco wasn't in my world apparently....that's probably my problem. 1.2 installed 1.5.1 available Sep 03 21:56:36 yep, that was it Sep 03 21:56:41 updated psyco and it worked Sep 03 21:56:49 strange that it worked for -i.... Sep 03 21:57:11 * JustinP runs a bitbake world on his frozen BBFILES Sep 03 21:58:51 happy you Sep 03 22:00:37 JustinP: you got freeze to work? Sep 03 22:04:02 NAiL: heh, no...I tried but it was built without the multimachine stuff in mind Sep 03 22:04:18 NAiL: I wrote my own version as a command in the shell Sep 03 22:04:28 ah, heh Sep 03 22:04:50 NAiL: I tried to do it with python in freeze.bb but I was having problems getting what I needed Sep 03 22:05:47 it's not 100%, though as I'm missing certain core native bbfiles apparently.... Sep 03 22:06:14 2/cr Sep 03 22:09:35 aha.... Sep 03 22:09:40 * JustinP wonders.... Sep 03 22:13:05 * JustinP eeks Sep 03 22:20:27 cu Sep 03 22:25:54 * JustinP thinks up a better way to do this Sep 03 22:25:58 errr....different.... Sep 03 22:26:02 going to miss -native, damn Sep 03 23:27:10 NAiL: so you were a user of freeze? Sep 03 23:27:37 openslug/debianslug/etc used to use freeze when we did a release Sep 03 23:27:50 I see Sep 03 23:28:01 I want to use it to just build what I have already built Sep 03 23:28:05 (update feed) Sep 03 23:28:09 yeah Sep 03 23:28:14 I think I have a working version done Sep 03 23:28:30 some superfluous code in it now.... Sep 03 23:28:34 but oh well Sep 03 23:28:37 just testing Sep 03 23:29:39 ~pray Sep 03 23:29:41 May the penguin be with you... Sep 03 23:31:35 pfff Sep 03 23:32:52 * JustinP tests one more time Sep 03 23:46:21 NAiL: if you're interested: http://oe.reversefold.com/bitbake-freeze.patch Sep 03 23:48:15 you got it working proper? Sep 03 23:48:53 well...it takes some manual work Sep 03 23:48:58 * JustinP explains Sep 03 23:49:10 set PKGDIR to the checkout/packages dir in your local.conf Sep 03 23:49:24 leave BBFILES including everything Sep 03 23:49:30 then: bitbake -i Sep 03 23:49:32 freezr Sep 03 23:49:34 freeze Sep 03 23:50:02 it will write out frozen-packages.conf in your local.conf (assuming your layout is the same as mine...this should also be configurable...) Sep 03 23:50:15 in the same dir as your local.conf, that is Sep 03 23:50:32 should be easy to change Sep 03 23:50:46 I had problems including the conf file so I just copied its contents into my local.conf Sep 03 23:50:52 and it's building stuff now ;-) Sep 03 23:51:49 good ;-) Sep 03 23:52:19 * NAiL can't make heads or tails of the python code anyway Sep 04 00:14:54 NAiL: well, the path thing is easy to fix if it doesn't work for you, right at the beginning of the freeze() function Sep 04 00:14:59 * JustinP leaves to pick up his brother Sep 04 01:17:23 03nail 07org.oe.dev * rfd426656... 10/ (3 files in 3 dirs): litestream: Add package. litestream is a shoutcast-compatible streamer, but tiny and a lot less dependencies Sep 04 01:17:28 03nail 07org.oe.dev * r238e9fa5... 10/ (1 packages/icecast packages/nonworking/icecast): icecast: Move to nonworking Sep 04 01:17:34 03nail 07org.oe.dev * r274d6b1d... 10/ (1 packages/meta/slugos-packages.bb): slugos-packages: Add litestream to feeds Sep 04 01:47:54 03pfalcon 07org.oe.dev * re0057fe4... 10/ (7 files in 4 dirs): Sep 04 01:47:54 rocksndiamonds: New package, a game, Boulderdash/Emerald Mine/Supaplex/Sokoban clone. Sep 04 01:47:54 * Initial draft recipe version, but already works. Sep 04 01:47:54 * Too big even for VGA devices (window size is hardcoded in game). Sep 04 01:47:54 * Data storage formats are non-optimal (pcx/uncompressed level files). Sep 04 01:47:55 * Packaging crude, need to be split up. Sep 04 01:47:57 * Nice game. **** ENDING LOGGING AT Mon Sep 04 02:59:56 2006