**** BEGIN LOGGING AT Fri Jul 15 23:59:56 2005 Jul 16 00:17:35 good mornig gentlemen. Jul 16 00:17:57 my night build failed with this message : Jul 16 00:17:58 NOTE: Task failed: Fetch failed: ftp://ftp.uk.linux.org/pub/people/dwmw2/mtd/cvs/mtd-snapshot-20050615.tar.bz2 Jul 16 00:18:55 strange thing is that this file doesn't exist anyway... what should i update ? Jul 16 00:32:45 anyone living here ? Jul 16 00:56:29 mmmm... i guess it should be looking for the mtd-snapshot-20050715 instead of mtd-snapshot-200506-15... anyone knows where this is declared ? Jul 16 01:01:50 try string match seach on .bb files Jul 16 01:08:16 TES : yeah, that's the first thing i did, but doesn't help much... Jul 16 01:09:53 aaaaaaaahaaaaa ! the bb file itself was misnamed !! Jul 16 01:10:00 thanks TES ! Jul 16 01:19:04 TES: where should i report this bug ? i'm a begginer in oe ; ) Jul 16 01:20:35 sorry don't know, begginner me too :) Jul 16 01:20:52 TES: :D Jul 16 01:23:49 <[cc]smart> alan_, TES: see channel description Jul 16 01:26:41 [cc]smart: mmm... just to be sure : i should send this to bugs.openembedded.org, right ? don't know if it is really a bug, though... Jul 16 01:37:27 good morning !!! Jul 16 01:39:36 good morning gremlin[it] Jul 16 01:47:24 hi alan_ Jul 16 03:53:01 *yawn* Jul 16 03:53:42 morning Jul 16 03:54:57 tilde hail mickeyl Jul 16 03:55:08 life sucks without working xkb Jul 16 03:55:47 hail zecke Jul 16 03:55:50 hail mickeyl Jul 16 03:56:03 zecke: yeah, that would suck Jul 16 03:56:13 morning pb_ Jul 16 03:56:23 pb_: for me it is reality... Jul 16 03:56:33 zecke: ah, sucks to be you. Jul 16 03:56:42 pb_: both machines crashed hard on me (laptop and build machine) Jul 16 03:56:52 pb_: even FreeBSD got upset and chew my package database... Jul 16 03:56:58 crumbs Jul 16 03:57:01 what did you do to them? Jul 16 03:57:20 pb_: it must have been my cats Jul 16 03:57:29 ah, that old problem Jul 16 03:57:41 pb_: Debian unstable mount: can not open shared library libblkdevid.so.1 ;) Jul 16 03:57:54 yow Jul 16 03:57:55 pb_: and ubuntu has a broken xbase-clients Jul 16 03:58:39 oh dear Jul 16 03:59:10 at least I've fixed my 'locale'-issue Jul 16 03:59:23 feels like stoneage using vi on console Jul 16 03:59:23 you moved to another country, you mean? Jul 16 04:00:05 pb_: I suspect I'm in the U.K, it is dark and cold Jul 16 04:00:28 bad luck Jul 16 04:00:35 out of all the countries you could have ended up in Jul 16 04:00:49 and yes it sucks to be me: Imagine one would code with Qt and c++ Jul 16 04:00:50 eek Jul 16 04:01:11 yeah Jul 16 04:01:19 writing c++ code for qt, in vi, on the console, in the dark Jul 16 04:01:22 hardly bears thinking about Jul 16 04:02:18 and I use 'sudo chvt 1' to escape from the big fat bloated X ;) Jul 16 04:03:12 pb_: how can one run the gcc testsuites? Jul 16 04:03:17 zecke: "make check" Jul 16 04:03:35 pb_: it fails with can not execute... and I wonder if that is known Jul 16 04:03:50 for a cross compiler? Jul 16 04:04:00 pb_: yes ;) Jul 16 04:04:10 you'll need to do some magic to make it use qemu for the tests. Jul 16 04:04:20 it might suffice to just set up the binfmt_misc thing, if you have root access. Jul 16 04:04:24 ah I love magic Jul 16 04:04:37 if you lack root, you will have to patch dejagnu. I don't know how to do that, but apparently Paul Brook is the man. Jul 16 04:05:21 it would be cute to add the dejagnu stuff to oe, so that you could automatically build gcc and run the tests. Jul 16 04:05:31 pb_: right Jul 16 04:05:37 qemu-native is already packaged, I think Jul 16 04:06:36 http://www.codesourcery.com/sourcerers/paul_brook.html contains Paul's contact details. Maybe you would like to send him a note. Jul 16 04:07:29 morning bluelightning Jul 16 04:07:36 bluelightning: hey Jul 16 04:07:39 hello zecke Jul 16 04:07:41 hello mickeyl Jul 16 04:07:45 hi pb_ Jul 16 04:07:54 hi bluelightning Jul 16 04:08:25 pb_: I've to get OE to build with F BSD first : Jul 16 04:08:39 lol brackets and braces are broken too Jul 16 04:08:52 zecke: hah Jul 16 04:09:50 bbl Jul 16 04:09:52 * pb_ -> Jul 16 04:11:40 guys, i noticed an error in the name of mtd-snapshot-20050615.bb file. it should be mtd-snapshot-20050715.bb to work. I didn't report this anywhere yet cause it's not really a bitbake bug... where should i report this ? Jul 16 04:12:34 here should be ok Jul 16 04:12:46 ok. good. Jul 16 04:14:18 this bb name has to be changed, otherwise, bitbke trys to download 'ftp://ftp.uk.linux.org/pub/people/dwmw2/mtd/cvs/mtd-snapshot-20050615.tar.bz2' which does not exist... Jul 16 04:15:00 * mickeyl upgrades Jul 16 04:20:59 03mickeyl 07org.openembedded.dev * r260d3c23... 10/packages/mtd/ (4 files): monthly mtd upgrade Jul 16 04:22:50 woot monotone really works Jul 16 04:22:59 mickeyl: how much time did you waste for that? Jul 16 04:23:07 glop glop !! my first contribution to OZ/OE ! Jul 16 04:23:09 * alan proud Jul 16 04:23:43 zecke: the upgrade took me 3 hours Jul 16 04:24:02 zecke: testing internally if all the scripts still run and then changing the 4 places where changes were necessary Jul 16 04:29:49 zecke: but i'm not done yet. i want to change the notification scheme from a pull to a push model since 0.20 features hooks for when the server receives a new revision. I need to cook a lua script that triggers the ciabot_monotone.py then Jul 16 04:31:17 mickeyl: did you see my patch for the commits list? Jul 16 04:31:30 keturn: see, yes. Jul 16 04:31:32 seen Jul 16 04:32:34 keturn: will that work with 0.20 ? Jul 16 04:35:26 I just changed which things got included in the mail, I didn't change any of the monotone invocation things Jul 16 04:35:59 ah, ok Jul 16 04:42:02 öÃ? Jul 16 04:42:05 damn Jul 16 04:46:07 woot: we're in the final in the Small and Mid Size League Jul 16 04:47:51 congrats Jul 16 04:48:42 http://robocup.mi.fu-berlin.de/wm2005.html Jul 16 04:52:59 keturn: ok, i've applied the script. Will make a change now to see if it works Jul 16 04:53:15 s/script/patch/ Jul 16 04:59:34 hmm Jul 16 04:59:43 * mickeyl looks at CIA-4 Jul 16 05:00:04 ok, that didn't work Jul 16 05:03:23 Traceback (most recent call last): Jul 16 05:03:23 File "/var/www/monotone.vanille.de/scripts/ciabot_monotone.py", line 362, in ? Jul 16 05:03:23 main(sys.argv[0], [sys.argv[1]]) Jul 16 05:03:23 File "/var/www/monotone.vanille.de/scripts/ciabot_monotone.py", line 301, in main Jul 16 05:03:23 send_changes_between(lf.get_leaves(), new_leaves, m, c) Jul 16 05:03:24 File "/var/www/monotone.vanille.de/scripts/ciabot_monotone.py", line 280, in send_changes_between Jul 16 05:03:26 send_change_for(new_rev, m, c) Jul 16 05:03:28 File "/var/www/monotone.vanille.de/scripts/ciabot_monotone.py", line 248, in send_change_for Jul 16 05:03:30 for k, v in substs: Jul 16 05:03:32 ValueError: too many values to unpack Jul 16 05:06:12 03mickeyl 07org.oe.dev * rdca8fa3b... 10/packages/gphoto2/gphoto2_2.1.6.bb: gphoto: syntax fix Jul 16 05:06:38 looks like that fixed it :D Jul 16 05:07:17 the revision is wrong in the HTML link though Jul 16 05:07:23 it's the short revision instead of the long one Jul 16 05:07:51 could you fix that? Jul 16 05:08:46 the current script is @ http://monotone.vanille.de/scripts/ Jul 16 05:14:01 mickey|bbl: should be a one-liner, line 320 s/rid/full_rid/ Jul 16 05:27:44 hi Jul 16 05:30:58 hey Jul 16 05:45:53 how disable kernel image build in ditro building ? Jul 16 05:48:04 aquadran: use ASSUME_PROVIDED Jul 16 06:02:46 zecke: thanks, ASSUME_PROVIDED = "virtual/kernel" seems to work Jul 16 07:15:28 woot Jul 16 07:15:42 zecke is the ueberhacker, I patched patcher to use gpatch Jul 16 07:27:37 all hail zecke! Jul 16 07:36:59 lol --no-dereference is not known to BSD cp Jul 16 09:19:46 evening Jul 16 09:38:11 any suggestions for an ARM emulator? I want to use oe to build an image ofcourse Jul 16 10:05:04 epronk: could hack up qemu, perhaps? Jul 16 10:05:50 ugh... is there a how-to-upgrade-commands for monotone? Jul 16 10:09:20 epronk: ARMware maybe a choice Jul 16 10:10:01 epronk: http://www.csie.ntu.edu.tw/~r88052/ARMware/en/index.php Jul 16 10:11:16 thanx. I'll take a look Jul 16 10:12:40 hmmm....I'm not able to connect to monotone.vanille.de Jul 16 10:12:51 I can view it on the web but can't pull from it Jul 16 10:13:43 JustinP: Monotone 0.20? Jul 16 10:13:43 hi luke-jr_ ! Jul 16 10:13:48 alan: ... Jul 16 10:14:12 alan: Linux has been having build issues lately-- trying to get Monotone working again to try againt Jul 16 10:14:12 yep, 0.20 Jul 16 10:14:14 again* Jul 16 10:14:42 * JustinP whispers "gentoo" (compiled fine for me) Jul 16 10:15:04 it's not compiling on OE for me, though Jul 16 10:15:09 on=in Jul 16 10:15:36 ewi546.ewi.utwente.nl is working for me but not monotone.vanille.de. Perhaps it's not allowing anonymous pulls? Jul 16 10:16:05 JustinP: compiled fine for me too, but pull stopped working =p Jul 16 10:17:43 monotone.vanille.de gives me connection refused. The other one worked fine. vanille.de worked yesterday just after I upgraded Jul 16 10:17:54 i made my own build this morning ! this mean i don't need your build anymore BUT i will reaaaaally need your advice (well... latter) about compilation. Jul 16 10:18:11 alan: what are you compiling for? Jul 16 10:18:22 zaurus c7x0 Jul 16 10:18:31 alan: aww Jul 16 10:18:38 luke-jr_: ^^ Jul 16 10:18:53 alan: should I figure out how to make a Monotone branch so you can use my image definitions? Jul 16 10:19:18 alan: it worked? I tried to compile for my C3000 but the compile failed Jul 16 10:19:28 JustinP: C3000 != C7x0 =p Jul 16 10:19:31 luke-jr_: if you want to do it, then yeah, i'll sure use it !! Jul 16 10:20:07 alan: http://utopios.org/~luke-jr/programs/oz-ljr/gpl/ljr.pkgs/ has the current ones Jul 16 10:20:13 alan: I'll look into using Monotone for it Jul 16 10:20:18 JustinP: my build finished this morning by 10am (in france) Jul 16 10:20:24 luke-jr_: cool ! Jul 16 10:20:36 luke-jr_: I know...but it's not all *that* different.... Jul 16 10:20:54 alan: are you building monotone with OE? on what host system? Jul 16 10:21:06 JustinP: why would you build MT with OE? Jul 16 10:21:30 because there's a bb file...and I want it to be installable on my zaurus Jul 16 10:22:09 boost built fine Jul 16 10:22:30 luke-jr_: i'll try to contribute ! Jul 16 10:23:20 JustinP: building monotone with OE ? Jul 16 10:23:40 ho... Jul 16 10:24:02 alan: I mean bitbake. But yes, there are 2 monotone bb files in OE. One for 0.19 and 0.20. 0.20 isn't compiling for me. Jul 16 10:24:43 ho ? really ? do you want me to try ? (i have a duron 1.3ghz) Jul 16 10:25:10 alan: sure I'd love for you to try. Like I said, though, it didn't work for me. Jul 16 10:25:29 It's monotone-5 (don't know why) Jul 16 10:25:45 ok. Jul 16 10:25:52 monotone-4 is 0.19 Jul 16 10:26:08 I'm trying 0.19 now to see if perhaps it compiles for me Jul 16 10:27:15 niarf... i've forgotten once more to export BBPATH... Jul 16 10:27:38 i should write this once for all, but i don't know where... Jul 16 10:28:18 I use a shell "script" Jul 16 10:28:18 starting the process... Jul 16 10:28:25 "source setup" Jul 16 10:29:26 yeah, that is an idea... Jul 16 10:31:45 JustinP: the extension (-4, -5) is the netsync protocol number. Jul 16 10:31:49 JustinP: "ERROR: Nothing provides monotone-5" Jul 16 10:31:52 ^^' Jul 16 10:32:03 i think i have to update... Jul 16 10:33:19 hmmm Jul 16 10:33:23 what branch are you using? Jul 16 10:33:25 mmm... i have nonotone_0.19.bb and monotone_0.20.bb... Jul 16 10:33:51 org.openembedded.dev (is that what you asked for ?) Jul 16 10:33:59 Hum... did I get that wrong... Jul 16 10:34:01 strange....I have monotone-4_0.19.bb and monotont-5_0.20.bb Jul 16 10:34:16 alan: yes, I Jul 16 10:34:21 'm using the same one Jul 16 10:34:43 alan: you have something from before when I did a rename Jul 16 10:35:03 Look at the top of monotone_0.20.bb - see it's DEFAULT_PREFERENCE'ed out. Jul 16 10:35:20 ok Jul 16 10:35:43 jbowler-away: is 0.20 compiling for you? What target machine / distro? Jul 16 10:35:51 Just do a monotone sync and monotone update, it should be there (assuming JustinP is using org.openembedded.dev) Jul 16 10:36:07 JustinP: nslu2/openslug Jul 16 10:36:50 What's the failure in 0.20? Jul 16 10:38:53 yes, I'm using org.oe.dev Jul 16 10:39:04 I'm using openzaurus/spitz Jul 16 10:39:09 I'll look up the failure Jul 16 10:39:29 There are no changes after the rename, so you have the latest version if you have monotone-5_0.20.bb Jul 16 10:39:57 lots of "might be uninitialized" warnings for one... Jul 16 10:40:05 hmmm, perhaps the version of boost is too old? Jul 16 10:40:16 JustinP: i launched the compilation of monotone 0.20 for Zaurus C7x0 from my Duron 1.3ghz with mandriva LE 2005 Jul 16 10:41:11 jbowler-away: I'll upload the log so you can see the whole thing Jul 16 10:42:43 boost is 1.32.0, which works on both NSLU2 and gentoo/x86 Jul 16 10:43:08 jbowler-away: http://pear.reversefold.com/oe/log.do_compile.11905 Jul 16 10:43:47 looks like a bunch of symbol sizes changed Jul 16 10:44:07 in CryptoPP Jul 16 10:44:22 JustinP: i have to go (dam social life... :p ) sooner than scheduled... Jul 16 10:44:47 alan: fine. I'd appreciate a note about whether it worked or not whenever you get back Jul 16 10:45:00 JustinP: i'll be back in 2 hours, maybe 3... is that ok for you ? Jul 16 10:45:22 alan: I'm pretty much always connected. leave me a message with "JustinP:" at the beginning and I'll get it :-) Jul 16 10:45:58 boost-jam-native is compiling... Jul 16 10:46:07 i must go now. Jul 16 10:46:12 see you Jul 16 10:46:25 Yes, I don't get the symbol size change warnings. Jul 16 10:47:15 JustinP: boost-jam-native compiled... Jul 16 10:47:16 jbowler-away: hmmm...I'm on a fairly new compile chain (new tmp dir) and I never compiled 0.19, so i don't know what this is... I'm trying 0.19 right now Jul 16 10:47:23 just to see if it works Jul 16 10:47:53 I'm using gcc 3.4.4 too, so it isn't that. Jul 16 10:48:19 not sure what I'm using...whatever OE normally compiles for OZ Jul 16 10:48:38 The log says 3.4.4 Jul 16 10:48:43 oh ;-) Jul 16 10:49:16 The most likely cause of this is that somehow the definition of 'word' changed between two files. Jul 16 10:49:30 weird... Jul 16 10:50:24 No, just the horrible way cryptopp is configured. See work/monotone-5-0.20-r0/monotone-0.20/cryptopp/config.h Jul 16 10:50:44 I made it work for armeb, but I may have screwed it up for arm, or it may never have worked for arm. Jul 16 10:51:42 hmmmmm... Jul 16 10:52:03 so the cryptobb patch might be doing it? Jul 16 10:52:47 On armeb I debugged it by putting #warn into #if for things like WORD64_AVAILABLE, finding out which branch was taken and making sure it was correct. In this case I suspect you could do the same thing and find WORD64_AVAILABLE (or something) was defined in one file and not in another. Jul 16 10:53:32 Yes, you could try just emptying the patch file and seeing if it worked then - the whole of cryptopp defaults to little endian ;-) Jul 16 10:54:23 is the endianness different for arm? Jul 16 10:54:49 still trying 0.19 just to see if it works Jul 16 10:54:57 Well, I'm assuming OZ is using a little-endian ARM Jul 16 10:55:21 ah, it's an OS level thing? Jul 16 10:55:22 If it isn't then the monotone might build but it certainly won't work without the patch. Jul 16 10:56:10 well I'm trying 0.19 with the patch first Jul 16 10:56:11 The CPU core doesn't care about byte sex. The memory/device systems support either (if the hardware is correctly implemented). Jul 16 10:56:20 I see Jul 16 10:56:32 well I personally have no idea what endian my compiles should be Jul 16 10:57:00 The patched version gets that right - if it doesn't it will #error out. Jul 16 10:57:15 * JustinP isn't sure what you're saying.... Jul 16 10:58:18 The patch to config.h uses definitions from the linux header files. If they aren't there (because of a different OS) the config.h does #error cryptopp - unknown endianness Jul 16 10:58:31 oh I see Jul 16 10:58:53 so the patch should work then.... Jul 16 10:59:06 Yes Jul 16 10:59:15 ok Jul 16 10:59:41 if you look further in that log there are symbols in std:: with the same error Jul 16 11:00:09 and boost:: Jul 16 11:00:53 * JustinP waits further for 0.19 to compile Jul 16 11:01:27 It will probably just fail on ld as well....:-| Jul 16 11:02:10 Where are the std:: symbols? I only see CryptoPP:: changes (to the various classes). Jul 16 11:02:20 one sec Jul 16 11:02:31 in th emidst of the CryptoPP ones Jul 16 11:02:37 one near the end Jul 16 11:02:45 with a few boost ones Jul 16 11:02:54 std::vector Jul 16 11:03:04 int boost::io::detail::upper_bound_from_fstring Jul 16 11:03:06 and so on Jul 16 11:04:26 yep, 0.19 failed Jul 16 11:04:30 with the same errors Jul 16 11:05:17 Yes, ok - so that doesn't include anything from CryptoPP - so I suspect there is a toolchain problem. Jul 16 11:05:44 Maybe std::allocator changes implementation? Jul 16 11:06:22 * JustinP is not qualified to comment Jul 16 11:06:30 The monotone-commands.o is from the top level directory, the lib3rdparty is from CryptoPP Jul 16 11:07:21 The best bet is probably to compare the include paths from the relevant compiles. Jul 16 11:09:21 would this be in the log files? Jul 16 11:10:08 and compare which packages? monotone and...? Jul 16 11:11:05 glibc? Jul 16 11:11:12 boost> Jul 16 11:11:13 ? Jul 16 11:23:51 jbowler-away: if you're willing to help fix this I can give you access to whatever files you need. Jul 16 11:27:24 I can probably debug it from here, but I can't test the result (i.e. build failures are fine, but run-time is your issue...) Jul 16 11:27:33 What's the machine/distro? Jul 16 11:29:53 Hang on... ccache arm-linux-g++ -march=armv5te -mtune=xscale Jul 16 11:29:57 Is this an XScale? Jul 16 11:37:25 Anyway, the include paths look fine. The optimisations are changed from mine, but they are just the default. If you give me the machine/distro and any local.conf config necessary I can try to build it here. Jul 16 11:38:18 Well, just put up a local.conf somewhere - that defines everything. Jul 16 11:39:24 ok Jul 16 11:39:25 thanks Jul 16 11:40:10 http://pear.reversefold.com/oe/local.conf Jul 16 11:40:24 it's a spitz (C3000) and OpenZaurus. I don't know what processor it has exactly... Jul 16 11:40:46 yep PXA270 XScale Jul 16 11:43:03 well that's interesting...arm *and* xscale? Jul 16 11:43:22 This won't happen fast - I've got to build it first... Jul 16 11:43:28 I know Jul 16 11:43:37 I could make changes on my end and try again if you like Jul 16 11:43:41 I have the toolchain set up Jul 16 11:44:34 * JustinP can't wait for a good 2.6 kernel....damn suspend problems Jul 16 11:44:50 I'm off to the movies Jul 16 11:45:04 I want to factor out your system setup, also things like the parallel make. The ASSUME_PROVIDED of gcc-2.95 seems a bit scary... Jul 16 11:45:10 I'll still be signed on, just let me know if you find anything out or need me to test anything Jul 16 11:45:25 I'm going out too, just want to start the build... Jul 16 11:45:33 jbowler-away: The gcc-2.95 is needed for the 2.4 embeddix kernels that the zaurii use Jul 16 11:45:46 that Sharp created for them I should say Jul 16 11:45:55 work on a 2.6 kernel is progressing Jul 16 11:47:02 the packages should all be compiling with the newer gcc Jul 16 12:30:51 when i hit monotone pull it tries to connect to vanille.de.. i'd like to change it to utwente.nl cause its actually working here.. so where does it save the address? Jul 16 12:36:12 nevermind got it Jul 16 12:43:46 JustinP: still no error... and monotone itself is compiling... Jul 16 13:19:53 * chouimat|ibook is away: food gathering Jul 16 13:52:06 * chouimat|ibook is back. Jul 16 14:17:33 JustinP: still here ? Jul 16 14:19:40 JustinP: the build failed, and the message error is really long. Jul 16 14:19:53 JustinP: here is all i have : http://pastebin.ca/17861 Jul 16 14:20:00 time to sleep. Jul 16 14:24:40 * chouimat|ibook is away: nap Jul 16 14:33:46 alan: you are seeing the same error as JustinP Jul 16 14:34:13 jbowler-away: i don't know what his error is... Jul 16 14:35:35 Yes, but I do and yours is the same. Jul 16 14:36:01 (His log file is up there somewhere ^) Jul 16 14:36:02 ok, if you say so... ^^ Jul 16 14:36:41 The interesting question is whether I get the same error (my monotone-5 is at the do_compile step) Jul 16 14:44:17 good night all Jul 16 14:56:36 03mickeyl 07org.oe.dev * r0cbbbdf7... 10/packages/boost/ (boost-jam-native_3.1.9.bb boost-jam-native_3.1.10.bb): Jul 16 14:56:36 boost-jam-native: Jul 16 14:56:36 - upgrade to 3.1.10 Jul 16 14:56:36 - minor description reformatting Jul 16 14:56:36 - use ${STAGING_BINDIR} in do_stage() Jul 16 14:57:31 pb_: Do you've an opinion on adding the buildmachine to the OVERRIDES? Jul 16 15:00:29 hello jbowler Jul 16 15:00:34 zecke: sure, go ahead Jul 16 15:00:48 hi JustinP Jul 16 15:01:33 I got my local.conf in the wrong place - so I just built (successfully) an openzaurus DISTRO with machine nslu2 - not very useful :-( Jul 16 15:01:44 heh, oops Jul 16 15:01:59 jbowler: hey you should use .inc files Jul 16 15:02:01 btw. Jul 16 15:02:10 :-( sorry to have wasted your time Jul 16 15:02:10 In what context? Jul 16 15:02:20 JustinP: my error Jul 16 15:02:28 Anyway, it was CPU time only Jul 16 15:02:34 jbowler: i'm reordering monotone meta data a bit and found that -4 and -5 share 99% of statements. this is where we would like to use .inc files Jul 16 15:03:18 (less duplication = less metadata to maintain) Jul 16 15:03:19 mickely: that's the same as saying ${PN}_{$PV} and ${PN}_${PV+1} should use ${PN}.inc right? Jul 16 15:03:40 pb_: build_FreeBSD, build_Linux to not interfere with normal architecture overrides? Jul 16 15:03:42 jbowler: yeah, if there are enough similarities Jul 16 15:03:42 I.e. like in nslu2-kernel_foo.bb etc - they just include nslu2_kernel.inc Jul 16 15:04:12 sometimes even new revisions add to much differences, then a .inc doesn't make sense Jul 16 15:04:15 but this is a rare case Jul 16 15:04:35 Right. cp is easy but lazy Jul 16 15:04:39 *nod* Jul 16 15:04:44 i'm doing it for monotone, ok? Jul 16 15:05:02 Yes... just let me check if I have any diffs because of the quoting. Jul 16 15:05:08 *nod* Jul 16 15:05:49 No - no diffs. Humungous numbers in boost though - did you already kill those? (PACKAGES += boost-foo) Jul 16 15:06:08 not yet Jul 16 15:06:13 i have a diff for boost uncommited Jul 16 15:06:15 pushing Jul 16 15:07:12 zecke: right Jul 16 15:07:34 boost is where I ran into the nightmare quoting problems because of the need to -Dfile=\"foo.h Jul 16 15:08:40 hmm Jul 16 15:08:48 monotone: misuse: 2 missing files Jul 16 15:08:59 any idea how i can convince monotone to just grab them from the dbß Jul 16 15:09:00 ß Jul 16 15:09:04 ~lart ß Jul 16 15:09:05 * ibot changes ß's permissions to 0777 and tells the world Jul 16 15:09:12 hm, is that xfce meta image really supposed to work? or is it just meant to give an idea of what such an image should look like? Jul 16 15:09:37 Delta_: the latter one. it needs some love Jul 16 15:09:37 mickeyl: monotone revert Jul 16 15:10:04 I have the idea it doesn't really like my love.. Jul 16 15:10:14 jbowler: righto. thanks Jul 16 15:10:34 Yet another non-obvious monotone interface Jul 16 15:11:38 JustinP: I've now got TARGET_ARCH=arm, TARGET_OS=linux, MACHINE=spitz, DISTRO=openzaurus, TARGET_FPU=soft Jul 16 15:11:45 Does that look correct to you? Jul 16 15:12:02 sounds about right I think... Jul 16 15:12:22 03mickeyl 07org.oe.dev * r50712da1... 10/packages/boost/boost_1.32.0.bb: Jul 16 15:12:22 boost: Jul 16 15:12:22 - always use SOURCEFORGE_MIRROR in SRC_URI, don't use download.sf.net directly Jul 16 15:12:38 jbowler: yep, that's what it says for me Jul 16 15:14:50 argh.....configure script is failing Jul 16 15:14:56 (not monotone, emacs ;-) Jul 16 15:17:44 anyone familiar with lua ? Jul 16 15:18:29 maybe: for very basic questions Jul 16 15:18:57 i'd need someone to write a hook that calls ciabot_monotone.py, however it needs to operate a timer that fires off 1 minute after the hook. if the hook is getting called again in the meantime, the timer needs to be reset Jul 16 15:19:54 yeah, not basic enough... I know there is some job control in there but that's all. Jul 16 15:22:07 heh, ok Jul 16 15:22:24 i know how to call external programs from it, but i probably need to read and write from a file to get that timer reset thing done Jul 16 15:22:39 guess i should ask that on the monotone list Jul 16 15:22:48 ~lart monotone for not including python instead Jul 16 15:22:48 * ibot follow's monotone with a gauntlet and ... scratch ... HUMILIATION for not including python instead Jul 16 15:23:27 reading and writing a file is done by the merge hook - i.e. there are examples in there. Jul 16 15:24:09 hmm i guess i'll rather have it call an intermediate python script that does the real thing :) Jul 16 15:24:14 And, of course, there is: http://www.lua.org/manual/5.0/ Jul 16 15:27:28 lua? did I miss something? is lua a language one should know? Jul 16 15:27:43 hmm Jul 16 15:27:45 not really Jul 16 15:27:58 :) Jul 16 15:28:13 it's tiny and some guys love to embed it because of that Jul 16 15:28:19 e.g. the monotone guys Jul 16 15:28:28 * chouimat|ibook is back. Jul 16 15:29:54 ah, ok, that's nothing for me, I stick with c and obj-c than.. Jul 16 15:30:58 *nod* Jul 16 15:31:25 i master 4 languages and i think they spread enough of the spectrum. If i don't need to i won't try to learn another. I rather get more fluent in the ones i already know Jul 16 15:33:18 besides, languages are overrated... often it's the libraries that matter Jul 16 15:34:12 GNU classpath to rule them all ;) Jul 16 15:34:39 heh Jul 16 15:34:48 languages can be used to capture task specific issues and thereby reduce bugs - but that means non-general-purpose languages. Jul 16 15:35:36 *nod* the right tool for the right job. Jul 16 15:36:39 mickeyl: I propagated my quote fixes back to org.openembedded.dev, they're syncing now. Jul 16 15:36:56 jbowler: excellent Jul 16 15:37:25 btw., i'll write a mail to oe@handhelds.org to request all of you with push access to issue a command to delete the bogus zecke branch Jul 16 15:37:39 That should Holger's problems with boost and nslu2-kernel - there might be others of mine in there too. Jul 16 15:37:41 unfortunately this doesn't propagate, all people need to yank it from their db Jul 16 15:37:51 What's the magic command? Jul 16 15:38:12 s/That should/That should fix/ Jul 16 15:38:33 db kill_branch_locally Jul 16 15:38:50 hmm Jul 16 15:38:57 or --branch ? Jul 16 15:38:59 * mickeyl checks Jul 16 15:39:03 It looks like 0.20 might support an explicit exclude of a branch from the server, not sure. Jul 16 15:39:29 03rwhitby 07org.oe.nslu2-linux * rcc91df51... 10/conf/machine/nslu2.conf: Nailed CVSDATE for gnu-config for nslu2-linux Jul 16 15:39:33 03dyoung 07org.oe.nslu2-linux * rd20d02a1... 10/packages/linux/nslu2-kernel/2.6.12/nslu2-setup.c: Change it to the new serial driver model, enabling serial console Jul 16 15:40:33 yeah, monotone --db= db kill_branch_locally is right Jul 16 15:41:03 unfortunately it'll reappear if one person forgets to do it :) Jul 16 15:42:28 Yes - that's why it would be good if the server could block it. Jul 16 15:42:36 right Jul 16 15:42:45 I have two server databases and two non-server and I have to fix all of them at once. Jul 16 15:44:57 And I assume the server has to be stopped because I can hack at the db ;-) Jul 16 15:46:21 03ccsmart 07org.oe.nslu2-linux * r75cf64dd... 10/ (3 files in 2 dirs): Added new populate script for directory creation during startup. Jul 16 15:46:25 03ccsmart 07org.oe.nslu2-linux * r685decee... 10/ (5 files in 3 dirs): Moved populate-volatile to suggested place. Jul 16 15:46:28 03rwhitby 07org.oe.nslu2-linux * r64742a62... 10/packages/linux/nslu2-kernel_2.6.12.2.bb: Jul 16 15:46:29 Bumped the PR Jul 16 15:46:29 bump the pr Jul 16 15:46:32 03ccsmart 07org.oe.nslu2-linux * r42c8721f... 10/packages/ (2 files in 2 dirs): Changed to use populate-volatile.sh. Jul 16 15:46:36 03ccsmart 07org.oe.nslu2-linux * rbc6e966f... 10/packages/openslug-init/ (openslug-init-0.10/turnup openslug-init_0.10.bb): Modified to work with populate-var and allow per target (nfs,disk...) definition of ramfs objects. Jul 16 15:46:40 03ccsmart 07org.oe.nslu2-linux * rb810c896... 10/packages/openslug-init/ (openslug-init-0.10/turnup openslug-init_0.10.bb): Adding own hostname to hosts. Offered by [g2]. Jul 16 15:46:44 03mickeyl 07org.oe.dev * rbaadb7d4... 10/packages/monotone/ (monotone.inc monotone-4_0.19.bb monotone-5_0.20.bb): monotone: extract common stuff into monotone.inc Jul 16 15:46:48 03jbowler 07org.oe.nslu2-linux * r4945900d... 10/packages/ (6 files in 2 dirs): Fix the missing quotes round string values. Jul 16 15:47:11 * mickeyl removes branch from his local db's now Jul 16 15:48:32 servers restarted Jul 16 15:48:33 g'night Jul 16 15:49:34 Ok, I'm going to fix my server pump database to only pull explicit org.openembedded branches - that will stop me getting zecke back. Jul 16 16:04:21 jbowler: any progress on monotone? Jul 16 16:05:31 JustinP: it's building glibc 2.3.5 - that's progress of a sort... Jul 16 16:06:11 * france is back (gone 05:45:12) Jul 16 16:06:36 ah :-) Jul 16 16:12:39 mickey|zzZZzz: ping Jul 16 16:12:53 mickey|zzZZzz: can I create a local branch that is not synced with vanille? Jul 16 16:13:07 mickey|zzZZzz: musn't it be named org.openembedded* Jul 16 16:13:13 Hi! What do I do if I did an bitbake task-bootstrap and it did a few cvs udpates... and now the patches for those packges do not apply? Jul 16 16:13:35 martin__: update the oe database? Jul 16 16:13:46 I tried a monotone update Jul 16 16:13:46 zecke: I suggest doing something not org.openembedded Jul 16 16:13:52 zecke: I will try Jul 16 16:14:15 again Jul 16 16:14:25 martin__: did you try merging as well? If that doesn't do it try setting CVSDATE or CVSDATE_packageName in your local.conf Jul 16 16:14:41 JustinP: what is merging? Jul 16 16:15:33 zecke: I am already up to date... Jul 16 16:17:26 JustinP: do I set those to be someing a week or two ago? Jul 16 16:18:00 martin__: how are you up to date? Jul 16 16:18:14 martin__: have you made any changes locally and checked them in? Jul 16 16:18:17 I typed monotone update in my org.openemebeded.net directory Jul 16 16:18:31 did it mention anything about merging? Jul 16 16:18:53 no. it just said: $monotone update Jul 16 16:18:53 monotone: already up to date at 1e69705dffe85267b07ba8b0aba59657bf64d09c Jul 16 16:19:06 although the pull command did complain Jul 16 16:19:27 (I am still trying to get used to monotone)... monotone: warning: caught bad_decode exception decoding input from peer monotone.vanille.de: 'protocol version mismatch: wanted '4' got '5'' Jul 16 16:19:51 martin__: ..... you're using monotone 0.19, not 0.20 aren't you? Jul 16 16:20:06 you need 0.20 since last night Jul 16 16:20:10 gah Jul 16 16:20:11 ok Jul 16 16:20:12 :D Jul 16 16:20:25 will that solve the patching problem maybe? Jul 16 16:20:35 should, yes Jul 16 16:20:38 at least for gnu-config Jul 16 16:20:57 ok. that was the one with the trouble Jul 16 16:21:09 :D was this posted anywhere? Jul 16 16:26:54 JustinP: are the instructions for updating different now? Jul 16 16:30:46 hmm... it is still telling me I am already uptodate... Jul 16 16:35:47 martin__: yes Jul 16 16:36:27 zecke: which question did you answer? Jul 16 16:36:27 and you need to pull as well Jul 16 16:36:50 I got pull to work Jul 16 16:37:06 martin__: was this posted anywhere was answered Jul 16 16:37:20 monotone --db=oe.db pull monotone.vanille.de "org.openembedded.*" Jul 16 16:37:51 damn it, meld seems nice but I do most of my dev command-line only....what should I be using? Jul 16 16:38:04 zecke: just found it here http://oe.handhelds.org/cgi-bin/moin.cgi/MonotonePhraseBook Jul 16 16:39:45 JustinP: Thanks... it seems to be working Jul 16 16:48:49 any idea how long the 'updating revisions' thing will take? Jul 16 16:49:21 Between 2s and forever, depending on CPU Jul 16 16:49:42 It counts the revisions it's writing, so you can guess after the first three or four. Jul 16 16:49:44 P4-2.8 ... and it seems to be taking forever Jul 16 16:50:07 Nope, NSLU2 266MHz takes forever. P4-2.8 is quicker. Jul 16 16:50:20 oh... so if it is on 63 of 316.... Jul 16 16:50:39 Wow, that's instantaneous! Jul 16 16:50:47 and it has taken 5 mins to get here... then... Jul 16 16:51:23 just 25 minutes, you lucky person. Jul 16 16:51:37 what exactly is it doing? Jul 16 16:52:12 martin__: checking to make sure that the changes are secure Jul 16 16:52:29 Same as db check I believe. For each revision it checks that everything in that revision is there, to find out more it would be necessary to run gprof... Jul 16 16:55:18 ok. well thanks everyone for the help! Jul 16 17:30:22 someone hacked hackaday.com =-( bastards Jul 16 17:42:39 good nite Jul 16 17:42:54 jbowler: or kcachegrind + calltree Jul 16 17:42:57 jbowler: or oprofile Jul 16 17:43:10 jbowler: or grep for verifiying in the monotone sourcecode Jul 16 17:43:39 or just ^Z, gdb attach and find out where it is. Jul 16 17:44:01 or decode it's program traversal path by listening to the sounds it makes on an AM radio .... Jul 16 17:44:08 kinda easy really if you want to be a monotone developer ;-) Jul 16 17:45:31 wow my portupgrade -a is almost done as well Jul 16 17:46:40 zecke: 0.20 fixes the problem from 0.19 with branches having to all be exported, so I can safely make a test branch com.jbowler.openembedded then later propagate it into org.openembedded.nslu2-linux. Jul 16 17:47:43 do configure scripts normally get longer? This is strange...when I unpack the configure script is much shorter. After I run configure it's real long ans has syntax errors Jul 16 17:47:53 does OE or BB change config scripts? Jul 16 17:48:04 autoreconf changes configure scripts Jul 16 17:48:56 ok Jul 16 17:49:05 if I run the configure script on my host machine it works Jul 16 17:49:10 in OE it fails on bad syhntax Jul 16 17:49:13 syntax Jul 16 17:49:46 Then there is probably a bug in the configure.in or configure.am, or something they use. Jul 16 17:50:24 hmmmm Jul 16 17:51:00 does the autoreconf need to happen? Jul 16 17:53:37 No, but the configure script needs to work correctly inside the OE build environment and on all the various systems used to build OE - that's a lot more dificult than getting a configure.am working. Jul 16 17:55:12 OE has to run autoreconf in order to ensure that our .m4 file modifications make it into the generated ./configure script before it gets run. Jul 16 17:55:17 (i.e. libtool) Jul 16 17:55:35 Ah, ok - I had wondered whethter there was something like that. Jul 16 17:56:09 and oe uses the latest autotools. lots of old configure.in/configure.ac files arent compatible with the latest versions of the tools Jul 16 17:56:34 so a great deal of adding a project to oe involves just updating it to work with current autotools (which is beneficial in general, and should probably flow upstream..) Jul 16 17:56:35 heh Jul 16 17:56:38 hey kergoth Jul 16 17:56:43 ok, I generally understand that. Does anyone know why it would be breaking the configure script? I know I read somewhere that emacs' configure script isn't quite kosher so I know it's something there... Jul 16 17:56:45 erm s/deal/deal of work in/ Jul 16 17:57:00 couldnt say without _exact_ error messages Jul 16 17:57:04 otherwise we'd just be blindly guessing Jul 16 17:57:11 hey zecke Jul 16 17:57:14 kergoth: I can give you the configure log Jul 16 17:57:28 i dunno how helpful i'd be, i've been out of it for some time now Jul 16 17:57:29 would that help? it's a syntax error... Jul 16 17:57:35 feel free to throw it on a pastebin and link it here Jul 16 17:58:30 well, one is a missing file then if I "fix" that by overriding a var it's a syntax error Jul 16 17:58:41 I doubt it would help without the configure files Jul 16 17:58:42 that doesnt tell us anything Jul 16 17:59:02 I know Jul 16 17:59:06 whoever fixes it would really need to have it in front of them Jul 16 17:59:09 so they can play around Jul 16 17:59:22 configure: error: /home/papercrane/oe/build/tmp/work/emacs-21.4-r0/emacs-21.4/./getloadavg.c is missing Jul 16 17:59:28 fixing them isnt usually an exact science, since very few people really understand things like m4 quoting issues, most just play around until it works :) Jul 16 17:59:34 that's the first error I'm getting without any fooling with the bb file Jul 16 17:59:41 hehe Jul 16 17:59:48 so read the configure.in/configure.ac and see what its doing Jul 16 18:00:06 I'll try...I've never actually done anything with autotools, though Jul 16 18:00:10 i was getting perilously close to actually comprehending all that madness before i left Jul 16 18:00:17 which frightened me ;) Jul 16 18:00:54 It would be dangerous to understand why, exactly, the scripts have \\\\\\\\ in them (as opposed to 4 or 16). Jul 16 18:00:57 alternatively, check everything you've got in and let someone else poke at it Jul 16 18:02:06 I have pretty much nothing Jul 16 18:02:14 and I don't have push access Jul 16 18:02:26 I only have a minimal bb file with inherit autotools Jul 16 18:06:09 ah. Jul 16 18:06:51 yes, I'm just starting to actually get programs in oe Jul 16 18:06:58 :) Jul 16 18:11:30 well, the README for emacs says that configure.in "uses an unholy marriage of custom-baked configuration code and autoconf macros" and should work with a recent version of autoconf and m4 Jul 16 18:11:54 and when was that written? :P Jul 16 18:12:14 last week :P Jul 16 18:12:23 yeah, who knows Jul 16 18:12:55 toss autotools and use something else :P Jul 16 18:13:20 a magic wand should cure all the developers in the world Jul 16 18:13:58 "Harry Potter can do it" Jul 16 18:14:51 300000h nice memory location .... Jul 16 18:15:39 time to use ... drumm roll ... tblptr Jul 16 18:16:07 tada Jul 16 18:16:37 * emte tries to remember what i was going to configure .... Jul 16 18:17:40 Oh use micro-emacs Jul 16 18:17:44 s/Oh/Or/ Jul 16 18:18:24 The primary problem with emacs is that it expects to be able to run its own new self to build a saved image, this doesn't work cross. Jul 16 18:19:10 hmmm.... Jul 16 18:19:22 I have micro-emacs...damn thing doesn't work right Jul 16 18:19:36 locked at a low cols/rows size and Ctrl-Space doesn't set a mark Jul 16 18:20:38 great, so to get emacs working I'd have to patch it to use emacs native to make an image somehow....::sigh:: Jul 16 18:21:08 so the other option is compling on my zaurus Jul 16 18:21:42 You have to find a way to stop the build trying to make the image, then do it the first time the pseudo-lisp stuff is loaded. Jul 16 18:21:44 which I'd do...but I don't know which package includes standard headers Jul 16 18:22:15 jbowler: but there's a bunch of autoconf problems which I can't figure out... Jul 16 18:25:13 guess what version of autoconf was used to create emacs' current configure script Jul 16 18:25:16 2.13 Jul 16 18:25:26 and we're using 2.53 Jul 16 18:25:42 2.59, excuse me Jul 16 18:25:48 JustinP: was it a deliberate decision to use the csl-arm compiler for spitz/openzaurus, or did it just happen that way? Jul 16 18:26:18 jbowler: I didn't choose a compiler myself, no. Just letting OE choose it. Jul 16 18:26:33 Well, where did you get the local.conf? Jul 16 18:26:42 did I put that in there?? Jul 16 18:26:59 You put the local.conf in, somehow Jul 16 18:27:02 I don't see a csl in it Jul 16 18:27:20 it's probably from the distro or machine conf Jul 16 18:27:40 No, there is no PREFERRED_PROVIDER in the build - that's why it gets csl-arm. Jul 16 18:28:04 ummm...well *I'm* not setting it... Jul 16 18:28:05 I'm asking where you got local.conf from - it isn't in the source database. Jul 16 18:28:23 I got my example a while ago and have been changing a few things Jul 16 18:28:48 my builds seem to work fine....I installed them on the normal OZ image Jul 16 18:28:59 why, do you see something wrong in it? Jul 16 18:29:24 Lots of other distros deliberately dis-prefer it. Jul 16 18:29:35 well...::shrug:: Jul 16 18:30:54 I also don't see any files with "csl" in the conf/distro and machine directories... Jul 16 18:31:41 echo tmp/work/gcc* Jul 16 18:32:17 what I'm saying is that I see nothing about providers, preferred or otherwise, with "csl" in it Jul 16 18:32:37 kergoth: do you know anything about this csl-arm business? I'm lost Jul 16 18:33:03 It's actually PREFERRED_VERSION, sorry. Jul 16 18:33:21 well I looked for "csl" and saw nothing Jul 16 18:33:57 Yes, under some circumstances which I don't completely understand it is the default. Jul 16 18:34:17 Oh, I remember... Jul 16 18:34:26 why, what is csl? is it what's killing the monotone compile? Jul 16 18:35:03 god I sound like a clueless newbie.... Jul 16 18:36:35 I don't know yet, but I can reproduce the problem with your local.conf, using the csl-arm gcc-cross (3.4.4), but NSLU2 (which works) uses just 3.4.4 Jul 16 18:36:55 hmmm Jul 16 18:37:05 Since it's clearly a compiler/library issue, and the libraries match, I suspect the compiler. Jul 16 18:37:21 sounds likely... Jul 16 18:37:50 What I don't understand is why the build picks up that compiler in the first place. It seems to be disabled. Jul 16 18:45:05 abiword is failing with similar errors....strange... Jul 16 18:50:02 Ok, the spitz build picks up csl-arm because of this: Jul 16 18:50:05 gcc_csl-arm-20050603.bb:DEFAULT_PREFERENCE_arm = "1" Jul 16 18:50:55 So someone thinks the compiler does work on arm... Jul 16 18:51:50 JustinP: with gcc-3.4.4 monotone-5 builds: Jul 16 18:51:52 PREFERRED_VERSION_gcc-cross-initial = "3.4.4" Jul 16 18:51:52 PREFERRED_VERSION_gcc-cross = "3.4.4" Jul 16 18:52:05 (I added those two lines to my local.conf). Jul 16 18:52:32 I didn't even do a complete rebuild - just boost and monotone-5. Jul 16 18:53:51 So I think this is a problem for whoever maintains the csl-arm GCC. It looks like something changes in the included files to alter the size of those structures between the two directories (they are in header files). Jul 16 18:54:07 ok. Will it work on a build which used the csl one? Jul 16 18:55:40 I don't know. Since I didn't change the library (so far as I know) it probably will. Run the test suite (monotone-testsuite, installs into /home/monotone). Jul 16 18:57:11 It's worth checking what other spitz/openzaurus users are doing - they might have put those lines into local.conf for the whole build... Jul 16 18:58:40 any other OZ users in here? Is anyone overriding gcc not to use csl? Jul 16 19:00:06 unfortunately the C3000/1000 are very new and I don't think there are a lot of OE users with them Jul 16 19:00:20 it would be nice ot know what the OZ distributors use... Jul 16 19:16:16 03jbowler 07org.oe.nslu2-linux * r9c0cb9ee... 10/packages/install/ (files/install-sh install-native_0.1.bb): Add support for -p and -D options to install. Jul 16 19:33:59 Guess I need to propagate that one back into openembedded.dev - native-install puts 'install' in staging, that install won't bomb out with -s (unlike some build system installs). Jul 16 20:54:23 03rwhitby 07org.oe.nslu2-linux * r9750f55a... 10/packages/ (3 files in 3 dirs): Moved OpenSlug root home directory from /home/root to simply /root Jul 16 20:54:26 03rwhitby 07org.oe.nslu2-linux * r9a6d9077... 10/packages/base-files/base-files_3.0.14.bb: Fixed the OpenSlug root move Jul 16 21:34:50 03jbowler 07org.oe.nslu2-linux * rd3dcb0c3... 10/packages/meta/openslug-packages.bb: util-linux does not build on uclibc (fails in do_install step). **** ENDING LOGGING AT Sat Jul 16 23:59:56 2005