| 00:00.08 | pb_ | reenoo_: what does it say just before "selecting ..."? |
| 00:01.29 | reenoo_ | pb_: selected 0.99.118-r0 |
| 00:01.32 | reenoo_ | weird |
| 00:01.55 | pb_ | hm. so, it's chosen the right version, but then goes on to use the wrong file. |
| 00:03.09 | erich141 | heh I didn't work with Xlib in the last 8 years or so and I am extremely happy about this. I guess I'll better try to add rxvt than trying to remove dependecies on those functions ;) |
| 00:03.37 | pb_ | heh |
| 00:09.33 | reenoo_ | pb_: in oemake2:110 you're re-using "fn" in a for loop. is that by intention? |
| 00:11.00 | pb_ | reenoo_: oh, whoops. that would be a problem. |
| 00:11.31 | Hrw | http://www.hrw.one.pl/zaurus/oemake2 |
| 00:11.56 | reenoo_ | Hrw: make that a patch if possible |
| 00:12.04 | Hrw | ok |
| 00:12.34 | reenoo_ | pb_: currently testing the fix |
| 00:12.40 | pb_ | cool |
| 00:12.46 | Hrw | http://www.hrw.one.pl/zaurus/oemake2.diff |
| 00:12.52 | pb_ | thanks |
| 00:14.14 | reenoo_ | pb_: seems to work |
| 00:14.29 | reenoo_ | NOTE: selecting /home/oe/cvs/oe-packages/ipkg/ipkg-native_0.99.118.oe to satisfy ipkg-native |
| 00:14.37 | pb_ | reenoo_: rock! |
| 00:16.10 | Hrw | it's time for me to learn python - but qt is higher on my dev-todolist |
| 00:16.56 | reenoo_ | pb_: http://lug-bremen.de/~reno/oemake2-vernum.patch |
| 00:17.03 | pb_ | reenoo_: thanks |
| 00:19.30 | reenoo_ | hrm, ERROR: package libtool-cross-1.5-r0: task do_configure failed |
| 00:19.36 | Hrw | pb_: oemake2 is unable to build packages when starting without packagename? |
| 00:19.52 | Hrw | hrw:packages$ oemake2pb |
| 00:19.52 | Hrw | NOTE: Parsing .oe files: / |
| 00:19.52 | Hrw | nothing to build |
| 00:20.05 | Hrw | it's with 571 oe files |
| 00:20.13 | kodiak25 | <-- reading "Learning Python, 2nd Ed." atm |
| 00:20.24 | pb_ | Hrw: you need to tell it what you want to build. |
| 00:20.33 | Hrw | pb_: ok |
| 00:20.37 | reenoo_ | kodiak25: go for the python cookbook rather |
| 00:21.34 | *** part/#oe MacDigas (~dirk@ip68-100-105-119.dc.dc.cox.net) |
| 00:21.46 | Hrw | oemake, oemake2pb, oemake2hrw and oemake2 (pb->hrw + reenoo vernum patch) |
| 00:21.46 | pb_ | Hrw: eh, your percentage thing doesn't seem to work very well for me. |
| 00:21.50 | pb_ | NOTE: Parsing .oe files: / (0323/0001) [32300 %] |
| 00:22.12 | Hrw | pb_: ok |
| 00:22.26 | reenoo_ | NOTE: Parsing .oe files: \arm-linux-gcc: Permission denied |
| 00:22.28 | kodiak25 | reenoo_: I'm skimming it to get up to speed with the syntax. I have a bunch of coding exp but not in python. (at least I have a sub to Safari Online) |
| 00:22.36 | *** join/#oe MacDigas (~dirk@ip68-100-105-119.dc.dc.cox.net) |
| 00:22.36 | pb_ | reenoo_: you can ignore that. |
| 00:22.54 | Hrw | pb_: I just grabbed it from oemake - it works for me |
| 00:23.06 | Hrw | NOTE: Parsing .oe files: - (0205/0571) [35 %] |
| 00:23.25 | kodiak25 | Can I safely ignore oemake msg: package base/python-pycap-0.1.6-r0: dependency libnet-1.1.1 unsatisfied |
| 00:23.26 | reenoo_ | pb_: checking for C compiler default output file name... configure: error: C compiler cannot create executables |
| 00:23.44 | pb_ | kodiak25: depends if you want to build python-pycap or not. |
| 00:23.47 | reenoo_ | pb_: that's from libtool-cross-1.5-r0 |
| 00:23.48 | pb_ | reenoo_: what are you trying to build? |
| 00:23.50 | pb_ | ah |
| 00:24.08 | pb_ | did you build gcc-cross already? |
| 00:25.30 | Hrw | pb_: OE can create own toolchain with gcc-cross and binutils-cross? I'm good thinking? |
| 00:25.36 | pb_ | yes |
| 00:25.45 | Hrw | nice |
| 00:25.49 | kodiak25 | pb_: the dependency failure is the first message after oemake finishes parsing the .oe files. |
| 00:25.54 | *** join/#oe dc___ (~dctanner@213.208.61.2) |
| 00:26.20 | reenoo_ | pb_: hmm. no, not with the bootstrap command you gave me |
| 00:26.24 | pb_ | reenoo_: hrm. |
| 00:26.39 | pb_ | which package is wanting libtool-cross? |
| 00:27.30 | reenoo_ | pb_: ipkg-native it seems |
| 00:27.49 | pb_ | ah. that's bogus, it should be using libtool-native |
| 00:27.50 | reenoo_ | pb_: maybe I should rm -r tmp first |
| 00:28.02 | pb_ | yeah, good plan |
| 00:29.14 | erich141 | libtool.oeclass prepends libtool-cross to DEPENDS |
| 00:29.51 | erich141 | ipkg inherits libtool |
| 00:29.58 | erich141 | ipkg-native includes ipkg |
| 00:30.08 | pb_ | ah. |
| 00:30.10 | reenoo_ | fun |
| 00:30.13 | Hrw | ok - I'll go sleep. thx for help - maybe tomorrow I'll try to check why busybox fails on do_install() (it looks like ${FILESDIR} was empty |
| 00:30.14 | pb_ | didn't we fix this just the other day? |
| 00:30.26 | pb_ | I remember talking to somebody about this same issue, or at least a very similar one. |
| 00:30.56 | reenoo_ | hrm |
| 00:30.57 | erich141 | pb_: it was about changing DEPENDS=libtool to DEPENDS=libtool-native |
| 00:31.06 | pb_ | erich141: ah. |
| 00:31.23 | Hrw | bye |
| 00:31.29 | Hrw | reenoo_: 02:31 here |
| 00:31.33 | pb_ | oh well. I guess we should conditionalise the DEPENDS in libtool.oeclass to use libtool-native for natives. |
| 00:31.54 | reenoo_ | Hrw: yeah. pretty late already =) |
| 00:32.20 | pb_ | DEPENDS_prepend="${@@ if (oe.data.getVar('PACKAGES', d, 1) != ''): return ' libtool-native ' else: return ' libtool-cross ' }" would probably do it, though a bit cheesy. |
| 00:32.40 | Hrw | reenoo_: the only time for development is after 20-22:00 ;) |
| 00:33.32 | erich141 | pb_: there are some non-native oes that might not have packages defined |
| 00:33.43 | Hrw | bye |
| 00:34.32 | pb_ | yeah, true. The right answer is really to have native.oeclass set some variable that can be tested to say "is this a native package?". |
| 00:34.35 | erich141 | those providing only static libraries to be staged |
| 00:35.02 | erich141 | pb_: then we also have to always inherit native first |
| 00:35.47 | pb_ | before libtool, yes. We already have that with other classes. |
| 00:36.00 | erich141 | oh, really? I didn't know |
| 00:36.30 | pb_ | actually, no, that's not true. DEPENDS won't be expanded until parsing is complete. |
| 00:37.16 | *** join/#oe _RoLo_ (~rolo@dpc6682192001.direcpc.com) |
| 00:37.17 | pb_ | but, yeah, there are other classes that already use similar mechanisms. I think distutils and package_ipk are both in that category. |
| 00:37.47 | *** join/#oe __stigger (~jeff@dialup-206.144.220.203.acc01-mann-gos.comindico.com.au) |
| 00:38.12 | *** join/#oe MacDigas (~dirk@ip68-100-105-119.dc.dc.cox.net) |
| 00:38.35 | *** part/#oe MacDigas (~dirk@ip68-100-105-119.dc.dc.cox.net) |
| 00:39.31 | pb_ | I suppose the other option would be to match ${PN} against ".*-native", but that seems a bit gross as well. |
| 00:41.58 | reenoo_ | pb_: will building gcc-cross_3.3.3.oe serve as a workaround? or do I need to wait for the "real" fix? |
| 00:42.26 | pb_ | reenoo_: yeah, should do. |
| 00:43.06 | erich141 | pb_: we probably can require "native" to be inhertied first for native packages |
| 00:44.04 | erich141 | I wouldn't rely on the fact that DEPENDS is expanded after parsing is complete, kergoth can change it at any time ;) |
| 00:44.16 | reenoo_ | heh |
| 00:44.30 | reenoo_ | pb_: building it now |
| 00:49.31 | pb_ | erich141: would you check in that python thing that I posted to the mailing list? |
| 00:50.35 | erich141 | pb_: I'd wait till mickeyl looks at it, all python packages are still broken |
| 00:51.29 | pb_ | ok. it makes gpe-image unbuildable for me, but I guess I can wait for mickey. |
| 00:52.26 | erich141 | pb_: I modified my local copy of python.oe yesterday to get rid of this error |
| 00:52.53 | erich141 | essentially the same thing you suggested except I did it in python's configure_prepend |
| 00:53.08 | pb_ | right |
| 00:53.43 | erich141 | so gpe-image builds for me, but python-* do not (who cares about them anyway) ;) |
| 00:54.10 | pb_ | heh |
| 00:54.17 | mickeyl | *yawn* |
| 00:54.21 | erich141 | speaking about deveil |
| 00:54.24 | mickeyl | just returned from a party |
| 00:54.26 | pb_ | ah, mickey awakes |
| 00:54.29 | mickeyl | d'oh... there was hell of activity here |
| 00:54.44 | pb_ | heh |
| 00:54.54 | mickeyl | i'm going to drop off to bed soon, just wanted to assure you i have a fix for the python stuff in my brain. will do it tomorrow though ;) |
| 00:54.55 | erich141 | pb_: good news: rxvt seems to be happy with diet |
| 00:55.02 | pb_ | cool |
| 01:01.02 | reenoo_ | pb_: gcc-cross-3.3.3-r0 fails to compile :/ |
| 01:01.42 | pb_ | oh dear. |
| 01:02.24 | pb_ | did you already build gcc-initial-cross and glibc? if not, you need those first. |
| 01:05.08 | reenoo_ | heh. no. |
| 01:05.31 | pb_ | I wonder why the bootstrap process is failing so early for you. |
| 01:05.48 | pb_ | you shouldn't need ipkg-native right at the beginning. |
| 01:07.21 | pb_ | hm, does anybody here have qmail ski11z? |
| 01:07.51 | *** join/#oe _stigger (~jeff@dialup-199.145.220.203.acc01-mann-gos.comindico.com.au) |
| 01:09.12 | reenoo_ | pb_: I'm trying without the INHERIT=package_ipk now. that seemed more promising |
| 01:10.08 | pb_ | okay |
| 01:10.26 | mickeyl | g'night |
| 01:11.03 | erich141 | rxvt builds perfect, the only thing that bothers me is that familiar has rxvt with some Xft patch |
| 01:11.19 | erich141 | I wonder where did it come from |
| 01:11.51 | pb_ | it's in handhelds.org cvs |
| 01:12.18 | pb_ | in /cvs/apps/rxvt, possibly |
| 01:12.46 | erich141 | there is no patch there, only rxvt + patch applied |
| 01:13.25 | erich141 | do we just want to set SRC_URI to handhelds.org cvs? |
| 01:13.42 | pb_ | may as well |
| 01:14.12 | erich141 | heh I didn't notice: Author: pb |
| 01:14.24 | pb_ | heh |
| 01:14.41 | erich141 | I guess I know who to ask about these patches ;) |
| 01:43.14 | reenoo_ | pb_: without the ipkg stuff the bootstap process seems to work fine |
| 01:43.39 | pb_ | reenoo_: okay. sounds like a bug in package_ipk.oeclass then. |
| 01:45.40 | reenoo_ | pb_: ok. I will need some time to fully understand all this. but I'm pretty impressed so far :) |
| 01:47.49 | reenoo_ | anybody willing to share their .vimrc btw? I don't want to mess up tabs vs. spaces >:) |
| 01:56.34 | CIA-5 | 03erich 07 * r1.1045 10packages/rxvt/ (rxvt-2.7.9/include.patch rxvt_2.7.9.oe): added rxvt_2.7.9.oe (from cvs.handhelds.org based on rxvt 2.7.9 + Xft patch) |
| 01:56.59 | erich141 | hmm, how do I tell rxvt-unicode that it requires a fat X11 |
| 01:57.31 | *** join/#oe _RoLo_ (~rolo@dpc6682192001.direcpc.com) |
| 01:58.30 | pb_ | reenoo_: cool. |
| 02:00.22 | treke|home | I need a new barber |
| 02:00.52 | treke|home | the last three tiems I've tried getting my hair cut, my current one has been closed |
| 02:01.07 | reenoo_ | heh |
| 02:01.44 | erich141 | pb_: do you mind if I change meta_gpe to require rxvt instead of rxvt-unicode? |
| 02:01.52 | pb_ | sure, go ahead |
| 02:02.29 | treke|home | how about doing something like debians x-terminal-emulator? |
| 02:03.17 | erich141 | to my shame I never installed debian |
| 02:03.24 | treke|home | shame on you |
| 02:03.55 | erich141 | I tried once to install stale version in VMWare, they didn't like each other |
| 02:04.02 | pb_ | treke|home: that'd be a good idea too, but we need to select one terminal to actually ship in the image. |
| 02:05.13 | treke|home | we're planning on shipping? |
| 02:05.41 | erich141 | bk commit |
| 02:05.46 | erich141 | dang! |
| 02:06.27 | reenoo_ | hrm |
| 02:06.29 | reenoo_ | ERROR: package python-2.3.3-r0: task do_compile failed |
| 02:06.57 | CIA-5 | 03erich 07 * r1.1046 10packages/gpe/meta-gpe.oe: meta-gpe.oe: replace rxvt-unicode with rxvt since rxvt-unicode doesn't like diet-X11 |
| 02:07.02 | erich141 | reenoo_: it is expected |
| 02:07.19 | erich141 | reenoo_: mickeyl will fix tomorrow in the morning |
| 02:07.22 | reenoo_ | even with pb_'s patch? |
| 02:07.25 | reenoo_ | ah |
| 02:07.37 | erich141 | reenoo_: I didn't try pb_'s patch |
| 02:08.56 | erich141 | if you insert this line into python_2.3.3.oe in the end of "do_compile_prepend" block, it should compile: install -m 0644 pyconfig.h ${STAGING_DIR}/${BUILD_SYS}/include/python2.3/ |
| 02:09.20 | erich141 | pb_'s patch does the same if I remember right, so it should work |
| 02:15.18 | reenoo_ | trying that now |
| 02:17.04 | reenoo_ | I get a different error now but it still doesn't compile |
| 02:17.35 | erich141 | what error? |
| 02:19.08 | reenoo_ | http://lug-bremen.de/~reno/python.log |
| 02:20.45 | erich141 | heh you chose bad name for your build directory |
| 02:21.00 | reenoo_ | doh |
| 02:21.08 | erich141 | return os.path.join(inc_dir, config_h).replace( "build", "target" ) |
| 02:21.31 | erich141 | mickeyl will fix it tomorrow |
| 02:22.06 | erich141 | for now you can change python-native.oe to ignore cross-distutils.patch |
| 02:22.36 | *** join/#oe dc___ (~dctanner@213.208.61.1) |
| 02:22.39 | reenoo_ | ah |
| 02:22.41 | reenoo_ | ok |
| 02:30.06 | reenoo_ | rebuilding python-native now |
| 02:30.16 | reenoo_ | since this will take a while... |
| 02:30.27 | reenoo_ | I'm off to bed now. 'night all |
| 02:34.36 | pb_ | treke|home: for your amusement: http://handhelds.org/~tinderbox/cgi-bin/tinder.cgi?tree=oe |
| 02:38.22 | treke|home | fun |
| 02:38.41 | erich141 | looks good, except "Guilty" field is empty |
| 02:38.54 | pb_ | yeah, I don't know how to tie it into the oe source control system yet |
| 02:39.59 | erich141 | I vote for treke |
| 02:40.19 | treke|home | it would be semi accurate |
| 02:40.29 | erich141 | hehe |
| 02:40.36 | treke|home | since any build failure would be caused by one of my updated snapshots :) |
| 02:40.52 | pb_ | heh |
| 02:41.06 | *** join/#oe rolo_ (~rolo@dpc6682199104.direcpc.com) |
| 02:41.24 | pb_ | maybe I'll just have it pick one of treke, kergoth and mickeyl at random. |
| 02:41.37 | erich141 | sounds good to me |
| 02:42.31 | pb_ | treke_: home often are the snapshots updated? |
| 02:43.27 | pb_ | how often, rather |
| 02:43.55 | *** join/#oe reenoo (~r@pD95283F9.dip.t-dialin.net) |