| 00:06.55 | bug1 | (i dont have a verison that works) |
| 00:11.03 | bug1 | zwelch, if you dont i would like to know |
| 00:13.21 | zwelch | hmm |
| 00:13.34 | zwelch | seems like it built them okay |
| 00:14.28 | zwelch | them == package libtool-native-1.5-r1: task do_build completed |
| 00:16.30 | zwelch | and i'd just done a fresh bk pull |
| 00:19.38 | bug1 | zwelch, damn its all me |
| 00:20.04 | zwelch | yeah, i know how that is :/ |
| 00:20.31 | zwelch | i can't build the dvi documentation, which seems to be all me too :) |
| 00:20.34 | bug1 | i guess ill try updating my oe tools, they are a week or so old |
| 00:22.34 | kergoth | bug1: that would do it. i wouldnt count on a happy build of anything libtool related with tools more than a day old. |
| 00:26.32 | ljp | i beleive you have stabbed libtool around 250 times |
| 00:27.15 | bug1 | still doesnt work :( |
| 00:29.52 | zwelch | no wonder libtool is so broken, it's been mortally wounded by developers ;) |
| 00:30.17 | kergoth | bug1: woo! the box has reappeared on the net |
| 00:30.23 | kergoth | doing a test build now |
| 00:32.00 | bug1 | kergoth: cool |
| 00:32.18 | kergoth | libtool-native built, testing a couple libtool packages to make sure its behaving and no /usr/lib lib paths leak in |
| 00:32.47 | bug1 | it should matter what autoconf/automake etc i have on my hast machine should it |
| 00:33.09 | kergoth | nope |
| 00:33.48 | bug1 | so i couldnt have messed up any files on my host |
| 00:41.17 | *** join/#oe jh_ (~user@209.137.229.132) |
| 00:46.50 | bug1 | bah, it built that time |
| 00:47.32 | bug1 | i think all i did was comment #INHERIT="package_tar package_ipkg" |
| 00:48.25 | bug1 | must have done something else |
| 01:05.54 | bug1 | and now i cant even the error |
| 01:06.03 | bug1 | and now i cant even reproduce the error |
| 01:08.41 | kergoth | hehe |
| 01:10.34 | *** join/#oe _Bala_ (~Bala@hsdbsk142-165-117-156.sasknet.sk.ca) |
| 01:30.04 | bug1 | kergoth, in tslib's zap.patch, it does |
| 01:30.20 | bug1 | + if (sqr (s->x - djt->hist [prev].x) + |
| 01:30.20 | bug1 | + sqr (s->y - djt->hist [prev].y) > djt->delta) { |
| 01:30.35 | bug1 | should it be "< djt->delta" |
| 01:30.52 | bug1 | because dejitter is trying to eliminate the small movements not the big ones |
| 01:31.23 | bug1 | it works better for me with it swapped around |
| 01:32.31 | bug1 | ./autoconfigure.sh: line 15: aclocal-1.7: command not found |
| 01:34.26 | bug1 | zwelch, did you build ipkg-native successfully in the build you just started ? |
| 01:34.42 | kergoth | sounds like its time we fix ipkg to not call autoconfigure.sh |
| 01:34.49 | kergoth | but instead use our existing autotools handling |
| 01:37.51 | bug1 | actually, false alarm, it was an old ipkg oe |
| 01:38.12 | bug1 | 118, 121 works |
| 01:55.51 | zwelch | bug1: yeah, my image built successfully... now i get to find out if it actually works |
| 02:00.07 | bug1 | zwelch, cool, i have a long way to go |
| 02:00.19 | kergoth | there, finally, libtool seems to be behaving again |
| 02:01.10 | kergoth | mithro--: verified builds from empty tmp -> successful libsdl-image stage is happy |
| 02:07.29 | kergoth | bug1: re that tslib code, yeah, it looks like that logic is backwards. variance throws out largely deviating events, not dejitter. |
| 02:36.14 | *** join/#oe jhonen[away] (~jhonen@cdm-66-76-82-127.fayt.cox-internet.com) |
| 02:36.22 | *** part/#oe jhonen[away] (~jhonen@cdm-66-76-82-127.fayt.cox-internet.com) |
| 02:41.20 | kergoth | bug1: you have cvsdate set? |
| 02:41.36 | kergoth | bug1: if so, eitehr unset it, or set CVSDATE_mtd to today. the current .oe needs new cvs, they renamed some utils |
| 02:41.41 | bug1 | CVSDATE=20040505 |
| 02:41.46 | bug1 | ok |
| 02:42.02 | kergoth | er CVSDATE_mtd-native, or whatever |
| 02:43.05 | bug1 | do i set CVSDATE_mtd-native in its .oe or in conf.local ? |
| 02:43.10 | kergoth | local.conf. |
| 02:43.14 | bug1 | cool |
| 02:50.51 | bug1 | i think i found a bug |
| 02:50.54 | bug1 | CVSDATE=20040505 |
| 02:50.55 | bug1 | CVSDATE_mtd-native=20040509 |
| 02:51.22 | bug1 | it created mtd-native-20040505-r0 |
| 02:51.38 | bug1 | should be mtd-native-20040509-r0 |
| 02:51.42 | kergoth | yep |
| 02:51.51 | kergoth | known issue. the version stays current even though it pulls old cvs |
| 02:52.00 | kergoth | not as easy thing to fix unfortunately, needs investigation |
| 02:52.16 | bug1 | fair enough |
| 02:56.18 | kergoth | bug1: do you have an account in the oe bug tracking system yet? if so, should open one for that so it doesnt get forgotten |
| 02:57.50 | bug1 | kergoth i do have an acount, have only been there a couple of times though |
| 02:59.54 | bug1 | ill try and add the bug |
| 02:59.58 | kergoth | cool |
| 03:02.36 | chouimat|Zzzzzz | night |
| 03:09.24 | bug1 | bug is added |
| 03:09.31 | kergoth | thanks |
| 03:10.09 | bug1 | np |
| 03:41.23 | mjc | greetings all |
| 03:46.17 | bug1 | woops, meant to stabs xext-cvs |
| 03:46.54 | mjc | where do I grab the more recent snapshots? |
| 03:47.19 | kergoth | of what? |
| 03:47.27 | mjc | oz |
| 03:47.52 | kergoth | there arent any. you're welcome to use oe to build a rootfs for your device and help work out the issues. |
| 03:48.08 | kergoth | 5500? talk to hrw. he's doing collie testing |
| 03:48.09 | bug1 | actually, x11-cvs stabs xext-cvs |
| 03:48.23 | mjc | and why does multisync a.)hate me and b.) refuse to connect to my z |
| 03:49.16 | bug1 | i havent tried multisync |
| 03:49.56 | bug1 | staging/arm-linux/lib/libX11.la referes to /usr/lib/libXau ! |
| 03:50.53 | mjc | think its time to give my Z a new rootfs or something |
| 03:53.18 | mjc | kergoth: poodle, I believe |
| 03:53.45 | kergoth | ah. dont know anyone testing out builds there |
| 03:53.46 | kergoth | yet |
| 03:53.48 | kergoth | bug1: directly? |
| 03:53.52 | mjc | me soon :) |
| 03:57.04 | bug1 | kergoth, i see you fixed it |
| 03:58.44 | bug1 | (my x11 oe was inheriting libtool) |
| 03:58.52 | kergoth | ah. |
| 03:59.02 | kergoth | yeah, you want the absolute latest of everything |
| 03:59.10 | kergoth | there are still remnant failures from some of our changes, working through them now |
| 03:59.15 | mjc | anyone got a copy of updater.sh? |
| 03:59.25 | kergoth | should be some floating out in openzaurus.org/official/ |
| 04:02.55 | mjc | anything new on the SD driver for 2.6? |
| 04:03.03 | mjc | or has nobody touched it yet |
| 04:03.13 | *** join/#oe T`|Hel (total@128.2.180.62) |
| 04:13.44 | treke|laptop | no one has touched it |
| 04:14.40 | mjc | sounds like it'll be fun if I evre get up to that level |
| 04:36.31 | mjc | whered oedev run off to |
| 04:37.31 | treke|laptop | oedev is obsolete |
| 04:37.50 | CIA-5 | 03kergoth 07 * r1.601 10oe/classes/base.oeclass: Correct typo in sed script in soinstall. |
| 04:37.51 | CIA-5 | 03kergoth 07 * r1.602 10oe/classes/autotools.oeclass: Exclude autopoint from autoreconf execution by default, since we don't have the gettext bits ourselves. |
| 04:39.20 | kergoth | mjc: nowadays you only need to set PATH and OEDIR. |
| 04:39.28 | kergoth | mjc: and naturally, setup local.conf. |
| 04:40.45 | kergoth | whee, found another app that uses autoconf & libtool but not automake.. have to override LIBTOOL |
| 04:45.04 | *** join/#oe T`|Heli (total@GALILEO.ACT.CS.CMU.EDU) |
| 04:48.19 | bug1 | NOTE: import patch '/home/glennm/Qonos/Software/oe/packages/libtool/libtool-1.5/prefix.patch' as 'prefix.patch' |
| 04:48.19 | bug1 | NOTE: patch with name 'prefix.patch' already applied |
| 04:49.01 | mjc | arg. and FC1 uses python2.2 |
| 04:51.01 | gb2 | kergoth: oops. i found that bug in libinstall, but i guess i ended up not fixing it in soinstall |
| 04:52.32 | kergoth | found another one |
| 04:52.41 | kergoth | it doesnt check if sonamelink == libfile |
| 04:52.47 | gb2 | oh |
| 04:52.48 | kergoth | so in the case where libblah.so.1 is the real file |
| 04:52.49 | gb2 | good point |
| 04:52.50 | kergoth | bye bye |
| 04:52.53 | gb2 | boom |
| 04:52.55 | gb2 | yeah |
| 04:53.01 | kergoth | hehe |
| 04:53.07 | gb2 | heh. interesting failure mode though :) |
| 04:53.27 | kergoth | hehe. symlink to itself blowing away the real thing |
| 04:53.28 | kergoth | fun stuff |
| 04:53.48 | gb2 | right. work tomorrow. |
| 04:54.22 | CIA-5 | 03kergoth 07 * r1.603 10oe/classes/base.oeclass: oe_libinstall bugfix: make sure the soname isnt the same as the filename of the real file. |
| 04:54.30 | CIA-5 | 03kergoth 07 * r1.1496.2.1 10packages/ (5 files in 3 dirs): Update libtool bits. libtoolize no longer requires a runtime STAGING_DIR var set. |
| 04:54.30 | CIA-5 | 03kergoth 07 * r1.1505 10packages/gnome/orbit2_2.10.2.oe: |
| 04:54.30 | CIA-5 | Merge openembedded@openembedded.bkbits.net:packages |
| 04:54.30 | CIA-5 | into linux.local:/home/kergoth/code/packages |
| 04:54.30 | CIA-5 | 03kergoth 07 * r1.600 10oe/classes/base.oeclass: Correct missing quote in oe_soinstall. |
| 04:58.29 | *** join/#oe offroadgeek (~offroadge@dsl092-250-033.sfo4.dsl.speakeasy.net) |
| 05:00.29 | *** join/#oe fpk^off (~Rel@florenz133.server4free.de) |
| 05:05.07 | kergoth | whew, there we go |
| 05:05.12 | kergoth | meta-gpe and task-bootstrap build |
| 05:05.36 | kergoth | now to verify meta-opie |
| 05:05.48 | kergoth | fix all those.. then i'll fire up a oemake -k world |
| 05:07.38 | CIA-5 | 03kergoth 07 * r1.1510 10packages/sqlite/sqlite_2.8.9.oe: Libtool usage fix for readline. |
| 05:08.21 | CIA-5 | 03kergoth 07 * r1.1511 10packages/ (12 files in 6 dirs): oe_libinstall usage fixes for packages in the meta-gpe dependency chain. |
| 05:20.23 | mjc | configure: creating ./config.status |
| 05:20.23 | mjc | /bin/sh ./i686-linux-libtool --mode=compile ccache gcc -DHAVE_CONFIG_H -I. -I/home/mjc/oe-packages-exported/tmp/base/libtool-native-1.5-r1/libtool-1.5/libltdl -I. -I/home/mjc/oe-packages-exported/tmp/staging/i686-linux/include -I/home/mjc/oe-packages-exported/tmp/staging/i686-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -c -o ltdl.lo ltdl.c |
| 05:20.27 | mjc | ./i686-linux-libtool: line 3612: syntax error near unexpected token `(' |
| 05:20.28 | mjc | that's just lovely |
| 05:20.30 | mjc | ./i686-linux-libtool: line 3612: ` $echo "*** with $libname but no candidates were found. (...for file magic test)"' |
| 05:20.43 | kergoth | ick, i havent seen that before |
| 05:27.50 | CIA-5 | 03kergoth 07 * r1.594.1.1 10oe/ (2 files in 2 dirs): Libtool handling changes. Kill libtool.oeclass, make autotools add libtool-native to deps, and hard fail on autoreconf execution failure. |
| 05:27.50 | CIA-5 | 03kergoth 07 * r1.1508 10packages/jpeg/jpeg_6b.oe: Correct the jpeg build to make use of libtool-cross, since it has no configure.in/ac. |
| 05:27.50 | CIA-5 | 03kergoth 07 * r1.1509 10packages/libtool/libtool-cross_1.5.oe: correct staging in libtool-cross |
| 05:28.38 | kergoth | heh, lagged mails |
| 05:42.57 | *** join/#oe kurre2 (~tomimo@ncircle.nullnet.fi) |
| 05:44.31 | CIA-5 | 03kergoth 07 * r1.1507 10packages/autoconf/autoconf-2.59/autoreconf-foreign.patch: Correct a typo in autoreconf. |
| 05:44.31 | CIA-5 | 03kergoth 07 * r1.1506 10packages/autoconf/ (autoconf-2.59/autoreconf-foreign.patch autoconf_2.59.oe): Apply patch to call automake --foreign by default. |
| 05:44.48 | kergoth | whee, slowly but surely the old mails are coming through |
| 05:51.35 | CIA-5 | 03kergoth 07 * r1.1512 10packages/ (7 files in 2 dirs): Kill off some references to the old nonexistant libtool oeclass. |
| 06:13.30 | CIA-5 | 03kergoth 07 * r1.1513 10packages/opie-ftplib/opie-ftplib_cvs.oe: Correct opie-ftplib usage of oe_libinstall. |
| 06:21.31 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 06:22.10 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 06:23.05 | mithro | ibot: seen kergoth |
| 06:23.08 | ibot | kergoth is currently on #uclibc #oe #elinux #handhelds.org #ol #openzaurus #tacobeam. Has said a total of 619 messages. Is idling for 7m 57s |
| 06:23.27 | kergoth | mithro: i battled libtool, and won. just the battle though, not the war |
| 06:23.34 | kergoth | whats up |
| 06:24.05 | mithro | yeah i was wondering about the libtool problem i was seeing |
| 06:24.17 | kergoth | i've done a rm -rf tmp; oemake libsdl-image; with success |
| 06:24.35 | mithro | dang it's a new day, it's going to rebuild glibc again |
| 06:25.41 | mithro | is opie cvs binary compatible with opie 1.0.3? |
| 06:26.01 | treke|ho1e | no |
| 06:26.29 | kergoth | mithro: you can set CVSDATE if you want to quit rebuilding glibc daily |
| 06:26.42 | mithro | dang |
| 06:27.18 | kergoth | CVSDATE=20040509 |
| 06:27.24 | kergoth | or CVSDATE_glibc=20040509 |
| 06:27.24 | kergoth | heh |
| 06:30.29 | treke|ho1e | wtf |
| 06:30.42 | treke|ho1e | are virus's now sending out fake virus notices now? |
| 06:32.26 | CIA-5 | 03kergoth 07 * r1.604 10oe/classes/base.oeclass: BUGFIX: initialize the value of our 'libtool' shell variable in oe_libinstall, on the off chance that the code that calls us exports it. |
| 06:33.10 | treke|ho1e | http://treke.net/~ggilbert/virus.txt |
| 06:34.51 | mithro | I got one of them as well |
| 06:35.31 | treke|ho1e | I cant figure out what the point of that message is |
| 06:36.43 | bug1 | i just sent a libtool patch to the list, prefix-patch isnt needed in libtool-cross or libtool-native as its included from regular libtool |
| 06:37.08 | mithro | brb as restarting xchat |
| 06:37.34 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 06:38.20 | mithro | kergoth: what do i have to clean to get libsdl-image to build again? |
| 06:38.37 | mithro | i'm assuming libtool and libsdl-image |
| 06:38.39 | mithro | anything else? |
| 06:38.54 | kergoth | mithro: your oe and packages trees need to be completely up to date, and i'd recommend cleaning libtool-native, jpeg, libsdl-image, and autoconf-native just to be on the safe side |
| 06:39.08 | kergoth | oemake -c clean libtool-native jpeg libsdl-image autoconf-native; oemake libsdl-image |
| 06:46.01 | bug1 | im off home |
| 06:46.29 | *** join/#oe treke|home (~ggilbert@lsanca2-ar29-4-43-192-190.lsanca2.elnk.dsl.genuity.net) |
| 06:47.46 | mithro | anyone know much about cylinder/header/sector in regard to disks? |
| 06:49.32 | *** join/#oe treke|ho1e (~ggilbert@lsanca2-ar29-4-43-192-190.lsanca2.elnk.dsl.genuity.net) |
| 07:03.56 | kergoth | hm, lrzsz busted |
| 07:09.14 | treke|ho1e | hmmm |
| 07:09.23 | treke|ho1e | I need to remember to mail this ram back next week |
| 07:12.25 | mithro | dang more likling problems |
| 07:14.11 | kergoth | mithro: ? |
| 07:14.23 | kergoth | i just tested libsdl-image a couple hours ago |
| 07:14.33 | kergoth | there are other failures floating around.. been fixing them as i find em |
| 07:16.15 | mithro | | /home/arm/openembedded/tmp.boxer/cross/bin/../lib/gcc-lib/arm-linux/3.3.3/../../../../arm-linux/bin/ld: cannot find -lSDL_ttf-2.0 |
| 07:16.31 | mithro | in python-pygame |
| 07:16.31 | kergoth | what package? |
| 07:16.36 | kergoth | k |
| 07:18.10 | kergoth | ah, yep, busted |
| 07:19.27 | CIA-5 | 03kergoth 07 * r1.1514 10packages/libsdl/libsdl-ttf_2.0.3.oe: Correct libsdl-ttf's -2.0 symlink. |
| 07:19.30 | kergoth | mithro: there, clean and rebuild libsdl-ttf |
| 07:27.18 | kergoth | treke|ho1e: you know if rt's email gateway bits are working atm? |
| 07:28.58 | treke|ho1e | hmmm. I have a funny feeling I didnt set up the aliases in postfix |
| 07:29.00 | treke|ho1e | one sec |
| 07:29.25 | kergoth | np, no rush, was just curious |
| 07:30.04 | treke|ho1e | oddly enough, I thought I got everything moved. Then when I actually try using the machine I go "damn forgot that" |
| 07:30.13 | kergoth | hehe |
| 07:30.17 | kergoth | thats always what happens |
| 07:30.33 | treke|ho1e | in my case I forgot some odd things. |
| 07:30.46 | mithro | kergoth: thanks :) |
| 07:30.55 | treke|ho1e | the machines main purpose in life is to provide me with email. I forgot to copy my maildir back over to it :p |
| 07:30.55 | kergoth | np |
| 07:31.01 | kergoth | hehe |
| 07:31.20 | treke|ho1e | luckilly it all compressed down to about 150megs so it didnt take long to upload |
| 07:32.05 | *** join/#oe schurig (~schurig@frdb-d9b963d7.pool.mediaWays.net) |
| 07:32.33 | treke|ho1e | whoohooo |
| 07:32.38 | schurig | hoho |
| 07:32.39 | treke|ho1e | cvs kmail supports threading of all folders |
| 07:32.56 | kergoth | nice |
| 07:32.58 | kergoth | hey schurig |
| 07:33.09 | mithro | woot it compiled! |
| 07:33.21 | mithro | treke|ho1e: threading of all folders? |
| 07:33.26 | treke|ho1e | mithro: yes |
| 07:33.38 | kergoth | i'd assume he means so you dont have to go into each and every folder and enable threading |
| 07:33.43 | kergoth | i hate that shit |
| 07:33.43 | treke|ho1e | yup |
| 07:34.14 | treke|ho1e | a bit of a pain when you have 450 folders total :p |
| 07:34.45 | kergoth | heh, damn. i've got .. inbox, and around 25 boxes for ml's. thats about it |
| 07:34.52 | kergoth | should probably organize someday |
| 07:35.00 | treke|ho1e | my mailing lists are organized by date range |
| 07:35.07 | kergoth | ah, cool |
| 07:35.14 | treke|ho1e | there's the current box and then the past in one month chunks |
| 07:35.21 | treke|ho1e | current is good for about 2 weeks iirc |
| 07:35.53 | kergoth | thats handy. i find i end up going oh the hell with this and doing a D .* in mutt to wipe it, then realize i removed something i wanted |
| 07:36.02 | treke|ho1e | hehe. |
| 07:36.11 | treke|ho1e | The one big mailbox thing worked ok when I was using mut |
| 07:36.21 | treke|ho1e | once I switched to imap it was just undoable |
| 07:36.24 | kergoth | limiting the view by regex is wonderful |
| 07:36.28 | kergoth | yeah, i bet |
| 07:36.47 | kergoth | you prefer kmail to evo? i havent tried kmail in ages |
| 07:36.55 | treke|ho1e | I rather like kmail |
| 07:36.57 | treke|ho1e | its improving |
| 07:37.32 | treke|ho1e | The vfolder bits are done a bit better in evolution |
| 07:37.39 | treke|ho1e | but they've existed longer |
| 07:38.29 | kergoth | i've been using imap with evo for a couple weeks at least, mostly to start using my gpg keys, being able to keep them local, you know |
| 07:39.24 | treke|ho1e | one of the guys at work seems to think the imap implementation in kmail works better than evo's. I cant tell the difference. |
| 07:39.35 | treke|ho1e | I prefer evolutions offline imap implementation over kmails |
| 07:40.43 | treke|ho1e | all in all I'd say they are about evenly matched. Both have their ups and downs |
| 07:41.06 | kergoth | cool |
| 07:41.39 | treke|ho1e | kmail is much snappier than nautilus :) |
| 07:41.52 | treke|ho1e | though that seems to be a common theme between kde and gnome apps :) |
| 07:42.15 | kergoth | i never used to like gnome apps, but now i find them more visually appealing |
| 07:42.57 | schurig | kergoth: I'm about to (finally!) start using OE as my buildroot tool :-) |
| 07:42.57 | treke|ho1e | I find them more visually appealing, but I dont really like the design philosphy that they are embracing. |
| 07:43.13 | treke|ho1e | schurig: Who'd have ever thought it would happen :) |
| 07:43.19 | kergoth | i like some of the approaches, like HAL, d-bus, etc |
| 07:43.21 | schurig | kergoth: I just re-learn all the new stuff. Congrats for oemake :-) |
| 07:43.25 | treke|ho1e | kergoth: yup. |
| 07:43.34 | treke|ho1e | kergoth: I like a lot of their technology ideas |
| 07:43.39 | kergoth | yeah |
| 07:43.51 | schurig | treke|ho1e: hey, the company I worked for is now bankrupt and I'm working in a brandly-new founded company |
| 07:43.54 | kergoth | schurig: glad to hear it. the more the merrier |
| 07:43.55 | treke|ho1e | it's mostly the ui/implementation of tem :p |
| 07:44.04 | treke|ho1e | schurig: Sorry to hear it |
| 07:44.08 | schurig | kergoth: still with TI? |
| 07:44.32 | kergoth | schurig: yeah. lost my old job in education though, that project got killed, bunch of people got let go. transferred to semiconductor now, in their OMAP linux group |
| 07:45.07 | treke|ho1e | kergoth: I'm hoping kde decides to pick up on the hal+dbus setup |
| 07:45.10 | schurig | wow, you're changing jobs faster than I change my underwear (eeeeeek!) |
| 07:45.13 | kergoth | treke|ho1e: yeah, me too. |
| 07:45.23 | kergoth | schurig: hehe. was only in that last one for a year. :\ |
| 07:45.24 | kergoth | oh well |
| 07:45.28 | kergoth | the new job seems very coo |
| 07:45.29 | kergoth | l |
| 07:45.32 | kergoth | less distro and build crap |
| 07:45.34 | kergoth | mostly kernel work |
| 07:45.59 | kergoth | schurig: how've things been going for you? hows the family? |
| 07:47.03 | schurig | kergoth: quite fine. My daughter is learning new stuff every day, never thought it would be so exciting to watch a baby develop skills and techniques |
| 07:48.01 | schurig | kergoth: what's the difference between libtool-cross and libtool-native? Except that the first one doesn't seem to work? |
| 07:48.13 | schurig | s/work/build/ |
| 07:48.27 | kergoth | they all build fine here. bug1 did report there was a multiuple patch application issue |
| 07:49.02 | kergoth | libtool-native is the one that does the real work. it has the share data, libtoolize, and our patched macros and everything |
| 07:49.09 | kergoth | libtool-cross provides one thing and one thing only. |
| 07:49.17 | kergoth | ${TARGET_PREFIX}libtool |
| 07:49.26 | kergoth | a prebuilt libtool for the target platform |
| 07:49.34 | kergoth | for packages that dont use autotools, so they dont generate their own. |
| 07:49.38 | schurig | kergoth: am am starting completely from fresh, even without a cross-compiler in my path. I did try to oemake libtool-cross, but it always tried to run arm-linux-gcc, which does not exist at this time. libtool-native did compile, thought |
| 07:49.45 | kergoth | schurig: oe builds itself a toolchain. |
| 07:49.52 | kergoth | schurig: your local.conf is incorrect. look at local.conf.sample. |
| 07:50.19 | kergoth | yes, libtool-cross needs the toolchain, because the script it generates (arm-linux-libtool) references the crosscompile tools |
| 07:50.29 | kergoth | only like 2-3 packages in our entire repo uses libtool-cross |
| 07:50.40 | kergoth | which is why its no longer in the default dependencies for libtool packages |
| 07:50.52 | kergoth | (fyi, you want your trees updated. both. today.) |
| 07:51.01 | kergoth | i just reworked this mess to suck less within the past 24 hours |
| 07:55.50 | CIA-5 | 03kergoth 07 * r1.604.1.1 10oe/classes/base.oeclass: Fix a bug in oe_libinstall as relates to the shared lib symlinking when the ELF binary has no information on the Library soname (prevent symlinking to ourselves). |
| 07:56.04 | schurig | kergoth: I did update just before I joined #oe :-) |
| 07:56.09 | kergoth | good good |
| 07:56.48 | schurig | kergoth: do you suggest gcc-3.4.0 or gcc-3.3.3 ? |
| 07:57.14 | kergoth | your call. 3.4.0 seems better, and supports optimizing for more platforms, but there -are- packages that still fail to build with it. |
| 07:57.32 | schurig | kergoth: something important? |
| 07:57.34 | kergoth | mostly c++ apps, but there are a few non-c++ ones floating around yet. we havent had time to finish fixing them all |
| 07:57.49 | kergoth | nah, most things work easily |
| 07:57.54 | schurig | oh, I use mostly c++ (Qt/Embedded, of course) |
| 07:57.56 | kergoth | and the fix for ones that are broken is simple enough |
| 07:58.08 | kergoth | mickey fixed qte, and i know the meta-opie package builds |
| 07:58.23 | kergoth | s/package/meta package/ |
| 07:58.27 | kergoth | heh |
| 07:59.12 | schurig | kergoth: I'm on Qt3, so I'll have to fix it by myself ... |
| 07:59.23 | kergoth | yep |
| 07:59.43 | kergoth | they're easy fixes really. like making sure templates use self-> rather than just directly referencing its members |
| 07:59.46 | kergoth | etc |
| 07:59.47 | kergoth | your call. |
| 07:59.48 | schurig | what does this mean: $ oemake -n gcc |
| 07:59.50 | schurig | NOTE: exceptions.NameError:name 'autotools_dep_prepend' is not defined while evaluating: |
| 07:59.50 | schurig | autotools_dep_prepend(d) |
| 08:00.18 | kergoth | sounds like your oe tree isnt up to date.. it works fine here |
| 08:00.25 | kergoth | if you look at autotools.oeclass |
| 08:00.32 | kergoth | you can see the definition of autotools_dep_prepend |
| 08:00.50 | kergoth | it uses def syntax in the .oeclass, which was relatively new |
| 08:01.14 | schurig | kergoth: I see it, it's on the top of the file |
| 08:01.35 | kergoth | schurig: our inline python code in oe metadata is *everywhere* now :) |
| 08:02.01 | schurig | so my tree IS uptodate, the error must lurk somewhere else around |
| 08:02.05 | treke|ho1e | hmmm cool |
| 08:02.13 | treke|ho1e | this switch might actually support rlogin directly |
| 08:02.35 | kergoth | schurig: def syntax adds autotools_dep_prepend in this case directly into python's namespace |
| 08:02.43 | kergoth | schurig: which is why the ${@} syntax can directly call it |
| 08:03.14 | schurig | kergoth: it's data.py which fails: |
| 08:03.15 | schurig | <PROTECTED> |
| 08:03.16 | schurig | <PROTECTED> |
| 08:03.35 | schurig | BUT i have the cache on. Maybe I should switch it off ? |
| 08:03.35 | kergoth | i cant reproduce that. |
| 08:03.40 | kergoth | i have the cache on too |
| 08:03.48 | kergoth | i just ran oemake -n gcc not 30 seconds ago |
| 08:03.59 | kergoth | you *do* have python 2.3 or higher, right? |
| 08:04.03 | kergoth | we require it. |
| 08:04.03 | schurig | with an uncommented cache it works |
| 08:04.11 | schurig | python 2.3.3 |
| 08:04.14 | kergoth | odd. |
| 08:04.36 | kergoth | the cache is supposed to get marked dirty following metadata changes even to the classes |
| 08:04.42 | kergoth | shouldnt ever have to wipe it manually or disable it |
| 08:04.46 | kergoth | must be a bug in that area |
| 08:04.52 | schurig | I think that the cache might not be up-to-date. What I currently do is to make my own oe directory, where I copy those packages from oe-packages that I like into. |
| 08:05.05 | schurig | I guess I copied new stuff which the cache didn't notice. |
| 08:05.21 | kergoth | like i said, the cache is supposed to recognize new oes and oeclasses |
| 08:05.25 | kergoth | as well as changes to any of them |
| 08:05.27 | schurig | (the reason for my copying is to have something reproducible for my customers, the BK tree changes all the time) |
| 08:05.32 | kergoth | nod |
| 08:05.49 | kergoth | when i was in education at ti i maintained an internal package metadata repos in cvs with a subset of the data, only the packages we needed |
| 08:05.54 | kergoth | then tracked bk for oe itself |
| 08:06.01 | treke|ho1e | "I guess if I run opie on openzaurus, I should replac the former gui's directory with opie's directory in the root file system. Is that right?Are there extra steps involved?" |
| 08:06.08 | kergoth | treke|ho1e: ha |
| 08:06.39 | schurig | Originally I thought we would make -r1, -r2, -r3 .. etc packages and never touch "released" packages. But maybe this was a too optimistic thinking, because in reality it would be too tedious. |
| 08:07.27 | treke|ho1e | I still think that would be a good idea. So far we really haven't "released" much of anything |
| 08:07.31 | kergoth | schurig: since right now there's no way to have "local overrides" of package metadata, and considering the # of little changes necessary, i found it less of a pain to have a seperate metadata repo myself |
| 08:07.43 | *** join/#oe imk (dj@ca-temecula-cuda2-c1b-243.snbrca.adelphia.net) |
| 08:07.43 | kergoth | well, that would be -nice- and all |
| 08:07.49 | kergoth | perhaps once we actually have maintainers |
| 08:07.53 | kergoth | and do releases. |
| 08:08.41 | schurig | treke|ho1e: we could put some keyword into *.oe files, as "RELEASED=true" and make some BK-scripts that refuse to comming anything an released stuff. That way a developer would be forced to create a new *.oe file with a higher -rX release number. |
| 08:08.45 | kergoth | schurig: fyi, current patcher on your site is broken in so many ways i hardly know where to start. |
| 08:09.11 | treke|ho1e | schurig: perhaps something like the gentoo ACCEPT_KEYWORDS? |
| 08:09.26 | kergoth | schurig: the expectation is that any change that affects the user needs to inrement revision, release or not. |
| 08:09.53 | kergoth | for a real release of a distro, we'll have meta packages depending on specific versions |
| 08:11.23 | kergoth | schurig: fyi, the current OE tools themselves are about to release 1.0. all the blocking issues have been fixed. the only holdups are a logo and documentation. |
| 08:11.35 | kergoth | which are pretty major, i think i'll send out an email asking for help with docs |
| 08:11.45 | treke|ho1e | kergoth: I agree with that part. What I'm thinking is that once a release has been done with a particular ${PV}-${PR} you never edit that particular ${PR} again. |
| 08:12.08 | treke|ho1e | whether it is a binary change or not |
| 08:12.10 | kergoth | treke|ho1e: oh, you're talking about when PR becomes part of the filename. |
| 08:12.32 | kergoth | since thats when you can create a second pr and have it live next to the previous. |
| 08:12.38 | treke|ho1e | right |
| 08:13.13 | kergoth | well, the only issue i see with that is we wont be the only ones tracking our metadata |
| 08:13.26 | kergoth | and who knows how many distros we'll be buildling out of this repo |
| 08:13.40 | kergoth | do we do it as soon as anyone we know of releases something with that particular .oe? |
| 08:14.06 | kergoth | i could even make a case for every revision resulting in a new .oe |
| 08:14.25 | treke|ho1e | That would be a pain in some ways, but I could agree with that arguement |
| 08:14.44 | kergoth | could always have wrapper scripts around such things, to make it easier on us. |
| 08:14.56 | kergoth | tools for assisting metadata maintainance |
| 08:15.15 | schurig | treke|ho1e: but that would NOT help us if someone just changes patches, so maybe it's not worth the trouble |
| 08:15.24 | schurig | kergoth: fix patches |
| 08:15.30 | kergoth | ? |
| 08:15.30 | schurig | s/fix/send/ |
| 08:15.34 | kergoth | ? |
| 08:15.38 | schurig | for broken patcher, I mean |
| 08:15.42 | kergoth | oh, yeah |
| 08:15.56 | treke|ho1e | ? |
| 08:16.05 | kergoth | patcher -r fails, patcher -b of a patch then patcher -a on the same patch, without making any cahnges, fails. |
| 08:16.25 | schurig | kergoth: that's weird, because I use patcher on my 2.4 and 2.6 kernel trees |
| 08:16.40 | kergoth | i'm unable to use it to maintain a patchset for any package in oe |
| 08:16.42 | kergoth | i've tried |
| 08:16.45 | kergoth | end up reverting to diff -urNd |
| 08:16.48 | kergoth | i dunno whats up. |
| 08:16.54 | schurig | kergoth: might be some weird flags for diff and patch that bork |
| 08:16.57 | treke|ho1e | kergoth: I have an old patcher if you want it :) |
| 08:17.09 | kergoth | treke|ho1e: i've got one laying around in my ~/bin/ on a couple boxes.. |
| 08:17.12 | kergoth | just need to pull that over here |
| 08:17.20 | treke|ho1e | the one I stole from buildroot |
| 08:17.20 | kergoth | i know that old one works happily |
| 08:17.28 | kergoth | hehe, yeah, i think thats where mine came from too |
| 08:17.51 | schurig | kergoth: ok, then send me the old one, h.schurig@mn-logistik.de |
| 08:18.10 | kergoth | treke probably has it handy, mine's on a box at work |
| 08:18.26 | kergoth | grr, and when did cyrus-sasl compile start failing |
| 08:18.38 | treke|ho1e | the thing I like about always incrementing pr is that the oe's more or less become constants. Any breakage in a "released" oe should be caused by someone else in the dependency chain |
| 08:18.41 | schurig | kergoth: my autotools_dep_prepend problem is reproducible |
| 08:18.50 | kergoth | schurig: not for me :) |
| 08:18.55 | schurig | I have an oe directory with just: ./native/patcher/patcher-native.oe |
| 08:18.55 | schurig | ./native/patcher/patcher.oe |
| 08:18.59 | schurig | ./native/autoconf/autoconf-native_2.59.oe |
| 08:18.59 | schurig | ./native/autoconf/autoconf_2.59.oe |
| 08:19.02 | schurig | ./native/automake/automake-native_1.8.2.oe |
| 08:19.02 | schurig | ./native/automake/automake_1.8.2.oe |
| 08:19.05 | schurig | ./native/glibc/glibc-initial_cvs.oe |
| 08:19.05 | schurig | ./native/glibc/glibc_cvs.oe |
| 08:19.08 | schurig | ./native/binutils/binutils-cross_2.15.90.0.3.oe |
| 08:19.08 | schurig | ./native/binutils/binutils_2.15.90.0.3.oe |
| 08:19.11 | schurig | ./native/gcc/gcc-cross_3.4.0.oe |
| 08:19.12 | schurig | ./native/gcc/gcc_3.4.0.oe |
| 08:19.14 | schurig | ./native/gcc/gcc-cross-initial_3.4.0.oe |
| 08:19.16 | schurig | ./native/libtool/libtool-native_1.5.oe |
| 08:19.16 | *** mode/#oe [+o kergoth] by ChanServ |
| 08:19.18 | schurig | ./native/libtool/libtool_1.5.oe |
| 08:19.19 | *** kick/#oe [schurig!kergoth@zelda.tuxnami.org] by kergoth (kergoth) |
| 08:19.28 | *** join/#oe schurig (~schurig@frdb-d9b963d7.pool.mediaWays.net) |
| 08:19.37 | kergoth | i get the point, less flooding please |
| 08:19.42 | *** mode/#oe [-o kergoth] by kergoth |
| 08:20.00 | schurig | I enable CACHE, remove the old cache. On the first run oe make -n gcc it works, on the second run it doesn't work. |
| 08:20.17 | kergoth | thats really strange. |
| 08:20.20 | kergoth | i always have the cache enabled |
| 08:20.23 | schurig | kergoth: maybe it's important to have the list of exact *.oe files, to reproduce this. |
| 08:20.25 | kergoth | better report that to mickey |
| 08:20.26 | kergoth | yes |
| 08:20.27 | kergoth | definately |
| 08:20.30 | kergoth | perhaps tar it up |
| 08:20.32 | kergoth | or so |
| 08:20.40 | kergoth | schurig: hop over to bugs.openembedded.org |
| 08:20.42 | treke|ho1e | schurig: patcher should be arriving in your mailbox shortly |
| 08:21.02 | kergoth | can use guest/guest for now |
| 08:21.14 | kergoth | open a report on this, assign it to mickey |
| 08:22.50 | treke|ho1e | it's either 2 or 3 am |
| 08:22.57 | treke|ho1e | perhaps a soda isnt the best choice of bevereges |
| 08:23.13 | kergoth | heh, there is that |
| 08:23.25 | kergoth | schurig: l/p: schurig/schurig on bugs.openembedded.org |
| 08:24.48 | treke|ho1e | must be morning in germany |
| 08:25.11 | kergoth | 2.6.6 released |
| 08:28.24 | schurig | treke|ho1e: it is |
| 08:28.31 | schurig | treke|ho1e: what is "snatches" ? |
| 08:28.47 | kergoth | snatch. grab. get. take. obtain. |
| 08:29.09 | treke|ho1e | snatchen |
| 08:29.23 | schurig | bug submitted |
| 08:36.05 | mickeyl | morning |
| 08:36.07 | *** join/#oe zinga (~arnaudb@186.80-203-227.nextgentel.com) |
| 08:36.42 | schurig | mickeyl: moin! |
| 08:36.54 | schurig | mickeyl: finally trying to utilize OE for my device |
| 08:37.26 | mickeyl | schurig: good plan. expecting your fixes :D |
| 08:38.11 | schurig | mickeyl: for now, you can have a bug report. I just added with regard to caching to bugs.oe.org |
| 08:38.19 | mickeyl | mmh |
| 08:38.41 | mickeyl | i read the logs from this morning |
| 08:39.03 | mickeyl | if you don't trust the cache, disable it in local.conf. |
| 08:39.14 | mickeyl | will try to reproduce that this week |
| 08:41.26 | schurig | ok, right now I'm at "oemake sed-native". Now I'm at "oemake gcc-cross". All starting with a very empty setup. Will look how far I come. |
| 08:42.08 | mickeyl | Oe has been pretty stable until a few days ago ... |
| 08:42.18 | kergoth | hehe. yeah, bad timing |
| 08:42.42 | schurig | NOTE: package glibc-initial-2.3.2-r0: task do_configure started |
| 08:42.51 | schurig | <PROTECTED> |
| 08:42.53 | schurig | :-( |
| 08:42.59 | kergoth | reworking libtool stuff, switching to oe_libinstall to fix cygwin builds, etc |
| 08:43.05 | kergoth | schurig: ? |
| 08:43.11 | kergoth | glibc ships with a configure.. hmm |
| 08:43.13 | kergoth | oh |
| 08:43.59 | schurig | kergoth: the directory /usr/src/oe-tmp/base/glibc-initial-2.3.2-r0/glibc-2.3.2 exists, but is empty |
| 08:44.05 | kergoth | err? |
| 08:44.13 | kergoth | did it not download the tarball? |
| 08:44.15 | kergoth | or what? |
| 08:44.30 | schurig | 18115189 Mar 1 2003 ../oe-downloads/glibc-2.3.2.tar.gz |
| 08:44.45 | schurig | If it hadn't, then the md5 checksum test shouldn't have worked |
| 08:45.06 | kergoth | hm? i didnt ask if the download failed |
| 08:45.07 | schurig | but yes, it DID download. I hope the 18115189 bytes are correct. |
| 08:45.09 | kergoth | i asked if it tried at all |
| 08:45.18 | kergoth | because i've never seen that. |
| 08:45.26 | schurig | kergoth: why should it try then the file is already there? |
| 08:45.55 | kergoth | schurig: it tried at some point, when you downloaded the fucking thing with oe in the first place. |
| 08:46.12 | kergoth | all i asked was if it tried to do that. i didnt ask if it tried to do that NOW, did i? |
| 08:46.26 | kergoth | heh |
| 08:46.37 | schurig | kergoth: yes, that was probably during the weekend. At friday night I just started an "oemake world" on the whole oe-packages tree, this did download some megabytes ... |
| 08:47.01 | schurig | some lines above the error I see "NOTE: Unpacking /usr/src/oe/native/glibc/glibc-initial_2.3.2.oe to /usr/src/oe-tmp/base/glibc-initial-2.3.2-r0/" |
| 08:47.13 | schurig | how to you unpack an *.OE file? |
| 08:47.15 | kergoth | unpacking a .oe? |
| 08:47.17 | kergoth | what the hell |
| 08:47.30 | kergoth | i just built that successfully about an hour ago |
| 08:47.39 | kergoth | you've got a nack for finding problems i cant reproduce :) |
| 08:48.18 | kergoth | maybe someone else *cough*mickey*cough* should try a build of glibc-initial |
| 08:48.21 | kergoth | and i'll go to bed. |
| 08:49.02 | schurig | kergoth`zzz: Ok, I migth have found the problem. I my local tree I had glibc-initial-2.3.2.oe, but no glibc-2.3.2 |
| 08:49.39 | schurig | in this case the "include" statement should have thrown an error |
| 08:49.41 | kergoth`zzz | thatd do it. talk to ya later. good luck with oe, glad to see you using it again. |
| 08:49.46 | kergoth`zzz | no, it shouldnt have. |
| 08:49.51 | kergoth`zzz | oe.conf has to be able to include files that dont exist |
| 08:50.07 | kergoth`zzz | what should be done is, we need a way to start a line with '-' like in gmake to indicate you dont care about it failing |
| 08:50.14 | kergoth`zzz | nobody's gotten around to implementing that |
| 08:50.17 | kergoth`zzz | feel free :) |
| 08:50.30 | schurig | kergoth`zzz: got night |
| 08:51.44 | schurig | hehe, that was a clear case of a copyright violation: I did not copy it right :-) |
| 08:52.04 | mickeyl | heh |
| 08:52.08 | mickeyl | g'night kergoth`work |
| 08:52.18 | *** join/#oe Hrw (~hrw@sd38.internetdsl.tpnet.pl) |
| 08:52.26 | mickeyl | morning hrw |
| 08:52.34 | Hrw | hi |
| 08:53.09 | Hrw | mickeyl: if I "install -d directory" and it is empty then it is not included in package? |
| 08:55.01 | proti | hi |
| 08:55.43 | mickeyl | Hrw: isn't it? it should be if the directory is in the FILES path |
| 08:56.32 | Hrw | mickeyl: dropbear has empty /etc/dropbear/ in do_install() but it is not included in ipk - I think that this was palandar problem |
| 09:01.16 | mickeyl | let me check |
| 09:01.36 | mickeyl | mickey@r2d2:/usr/local/projects/tmp/deploy/ipk$ ipkgfiles dropbear_0.41-r2_arm.ipk |
| 09:01.36 | mickeyl | drwxr-xr-x root/root 0 2004-05-09 21:19:24 ./ |
| 09:01.36 | mickeyl | drwxr-xr-x root/root 0 2004-05-09 21:19:23 ./etc/ |
| 09:01.36 | mickeyl | drwxr-xr-x root/root 0 2004-05-09 21:19:23 ./etc/init.d/ |
| 09:01.36 | mickeyl | -rwxr-xr-x root/root 2027 2004-05-09 21:19:23 ./etc/init.d/dropbear |
| 09:01.37 | mickeyl | drwxr-xr-x root/root 0 2004-05-09 21:19:23 ./etc/dropbear/ |
| 09:01.51 | mickeyl | etc dropbear is in the package |
| 09:02.01 | mickeyl | the real reason for pendalars problem is probably like i wrote in the email |
| 09:02.12 | mickeyl | we didn't launch reconfiguring packages since yesterday nte |
| 09:02.15 | mickeyl | nite even |
| 09:02.25 | mickeyl | hence it didn't generate the key |
| 09:02.37 | mickeyl | (reconfiguring packages launches all postinst scripts) |
| 09:04.13 | Hrw | btw - I'll build opie-image using gcc 3.3.3 to check is that give me starting opie |
| 09:04.33 | proti | hrw: I've been bitten by the update-qtfondir bug friday. |
| 09:04.56 | Hrw | proti: opie-taskbar has update-qtfontdir |
| 09:05.11 | Hrw | btw - naming and contents of opie packages are strange for me |
| 09:05.18 | proti | When was it included ? |
| 09:05.46 | Hrw | proti: or at least it should - I sent patch few days ago |
| 09:06.16 | Hrw | http://home.hrw.one.pl/oe/0506_opie-taskbar.oe |
| 09:08.44 | proti | There is nothing in the bk tree. |
| 09:09.12 | Hrw | ok - I'll commit it later - now have something to do at work |
| 09:17.37 | proti | Thanks. |
| 09:19.08 | Hrw | proti: as you probably readed I have few things to do in packages/ |
| 09:39.28 | Twiun | hi guys |
| 09:43.42 | proti | hi |
| 09:45.52 | mickeyl | hi Twiun how are things |
| 09:46.15 | *** join/#oe kestrel (~kestrel@nusnet-223-175.dynip.nus.edu.sg) |
| 09:46.31 | Twiun | mickeyl: hiya, stressful, and you? |
| 09:47.52 | Hrw | hi Twiun |
| 09:48.09 | Hrw | btw - someone compiled lrzsz today? |
| 09:49.36 | mickeyl | kergoth apparantly forgot to commit autotools.patch |
| 09:49.52 | mickeyl | which is funny because he is the one who reminds me always to use bksfiles -x |
| 09:49.59 | mickeyl | Twiun: same here, but not bad |
| 09:51.28 | CIA-5 | 03mickey 07 * r1.1516 10packages/lrzsz/lrzsz-0.12.20/autotools.patch: add a (dummy) autotools patch until kergoth provides the real one |
| 09:52.22 | mickeyl | d'oh thats not enough |
| 09:52.28 | mickeyl | there's also a m4.patch missing |
| 09:52.30 | mickeyl | mmh |
| 09:54.48 | proti | Hrw: Mine just barfed : aclocal: configure.in: 260: macro `AM_GNU_GETTEXT' not found in library |
| 09:55.01 | proti | Known bug I think. |
| 09:55.55 | Hrw | yep |
| 09:56.22 | proti | Any workaround known ? |
| 09:57.09 | Hrw | proti: yes - waitiing for kergoth fix |
| 09:57.35 | Twiun | hi Hrw, proti |
| 09:57.42 | Hrw | proti: if you want to build task-bootstrap then you can remove it from it's depends |
| 09:58.11 | mickeyl | oh funny, now libtool broke also here |
| 09:58.14 | mickeyl | grrrr |
| 09:58.19 | mickeyl | all the times the libtool drama didn't affect me |
| 09:58.28 | mickeyl | now they managed to break it also for me |
| 09:58.32 | mickeyl | good job |
| 09:58.35 | mickeyl | *sigh* |
| 09:58.35 | Hrw | ;) |
| 09:58.56 | mickeyl | guess that means no OE work for me today |
| 09:59.00 | Hrw | I keep oe-tmp-after-libtool.tar for this ocassions |
| 10:00.08 | mickeyl | heh good plan |
| 10:00.46 | Hrw | now I'm building gcc 3.3.3 toolchain |
| 10:01.09 | mickeyl | | NOTE: Executing autoreconf --install --force --exclude=autopoint -I /usr/local/projects/tmp/base/libtool-native-1.5-r1/libtool-1.5/m4 |
| 10:01.09 | mickeyl | | /usr/local/projects/tmp/staging/i686-linux/bin/autoreconf: unrecognized option `--exclude=autopoint' |
| 10:01.09 | mickeyl | | Try `/usr/local/projects/tmp/staging/i686-linux/bin/autoreconf --help' for more information. |
| 10:01.09 | mickeyl | | FATAL: autoreconf execution failed. |
| 10:01.44 | schurig | mickeyl: what's "bksfiles -x" ? |
| 10:02.11 | mickeyl | bk sfiles -x shows unadded files |
| 10:02.18 | mickeyl | bk sfiles -c shows files with uncommited diffs |
| 10:02.19 | mickeyl | etc. |
| 10:02.30 | schurig | mickeyl: ah, i mostly use "bk citool", which does more or less the same |
| 10:03.02 | schurig | mickeyl: and, how's it going with your "Dr." ? |
| 10:03.55 | Hrw | schurig: but you cannot use bk citool remotely if you don;t have X11 forwarding... |
| 10:03.55 | schurig | Hrw: that's a point |
| 10:03.55 | Hrw | schurig: I work like that |
| 10:04.58 | Hrw | schurig: I'm novice in bk usage |
| 10:15.10 | mickeyl | schurig: ph.d. is progressing... slowly though :) |
| 10:15.32 | mickeyl | schurig: has harlekin finished the app for your company? |
| 10:15.58 | schurig | mickeyl: i think. He sent me an email to get in contact with me, but I wasn't here, so I'm waiting until he comes online again :-) |
| 10:16.33 | mickeyl | mickey@r2d2:/usr/local/projects/packages$ ping sources.redhat.com |
| 10:16.33 | mickeyl | PING sources.redhat.com (12.107.209.250) 56(84) bytes of data. |
| 10:16.33 | mickeyl | --- sources.redhat.com ping statistics --- |
| 10:16.33 | mickeyl | 9 packets transmitted, 0 received, 100% packet loss, time 7998ms |
| 10:16.40 | mickeyl | anyone know a mirror for glibc cvs ? |
| 10:17.42 | schurig | mickeyl: treke? Someone in the eCOS-Mailing list also said that RedHAT CVS is down. So I guess it's actually being dealt with, the eCOS people have good contacsts to RH (they actually belonged to them a year ago) |
| 10:18.03 | schurig | is glibc-cvs so much better as glibc-2.3.2 ? |
| 10:18.27 | mickeyl | yes |
| 10:18.31 | mickeyl | don't ask me details though |
| 10:18.33 | mickeyl | pb_ knows |
| 10:20.53 | Hrw | mickeyl: it probably be useful if we can have CVSDATE support for packages - so you can stick glibc_cvs with CVSDATE=20040509 (assuming you have sources from that day) and other cvspackages with CVSDATE=NOW() |
| 10:21.26 | mickeyl | we have that |
| 10:21.42 | mickeyl | echo "CVSDATE_glibc_cvs=20040509" >> conf/local.conf |
| 10:21.57 | Hrw | o.. I have to read docs/code more |
| 10:22.05 | mickeyl | i doubt its mentioned there |
| 10:22.11 | mickeyl | noone cares to write docs :( |
| 10:23.45 | Hrw | my bk copy at work: takepatch: 674 new revisions, 0 conflicts in 339 files |
| 10:25.10 | mickeyl | quite a bit |
| 10:25.27 | Hrw | rarely used |
| 10:28.28 | mithro | anyone know why i would get a |
| 10:31.21 | mithro | hrm don't worrk about it |
| 10:32.03 | mithro | how about this error |
| 10:32.05 | mithro | QPaintDevice: Must construct a QApplication before a QPaintDevice |
| 10:33.26 | mickeyl | sounds like you're playing with SDL |
| 10:33.28 | mickeyl | right? :D |
| 10:33.38 | mithro | sorry got distracted :) |
| 10:33.53 | mithro | when doing a pygame.display.init() |
| 10:34.20 | mithro | which in theory should startup the application okay |
| 10:34.35 | mithro | pygame.init() also cause the same error |
| 10:34.47 | mithro | (which make sense as it calls display.init()) |
| 10:35.17 | mickeyl | this sounds like one of the usual SDL compile problems |
| 10:35.31 | mickeyl | where the qtopia sdl main is not chosen |
| 10:35.41 | mickeyl | this normally goes via -DSDLmain=<bla> |
| 10:35.53 | mickeyl | check the compiler command lines for that |
| 10:36.26 | mithro | well umm :) |
| 10:36.29 | mithro | <PROTECTED> |
| 10:36.34 | mithro | ? |
| 10:36.52 | mickeyl | that's pretty useless |
| 10:36.59 | mithro | btw why are the sdl stuff depending on libsdl-x11 when i selected libsdl-qpe ? |
| 10:37.02 | mickeyl | check the compiler command lines when they're in action |
| 10:37.30 | mickeyl | no idea. it doesn't here when using PREFFERED_PROVIDERS |
| 10:38.06 | mithro | hrm |
| 10:38.07 | mithro | /opt/arm/openembedded/tmp.boxer/staging/arm-linux/include/SDL/SDL_main.h:53:1: warning: "main" redefined |
| 10:38.54 | mithro | i'm also getting |
| 10:38.56 | mithro | -Dmain=DL_main |
| 10:39.06 | mithro | which kinda looks wrong? |
| 10:40.01 | mickeyl | DL_main ? |
| 10:40.04 | mickeyl | that is wrong |
| 10:40.10 | mickeyl | it should be SDL_main |
| 10:40.46 | mickeyl | might be an error in how PyGame includes sdl |
| 10:40.56 | mickeyl | there is only one right way to do it |
| 10:41.46 | mithro | so what should i be looking for? |
| 10:41.47 | schurig | mickeyl: looks like you used somehere in a Makefile $VAR instead of $(VAR) |
| 10:42.03 | schurig | mithro: oh, that line was meant for you |
| 10:42.44 | schurig | mithro: when you use $VAR in a makefile and $(V) is empty, then you get <empty>AR, that is, "AR". |
| 10:44.33 | mithro | well pygame is getting the following |
| 10:44.36 | mithro | SDL=-I/opt/arm/openembedded/tmp.boxer/staging/arm-linux/include/SDL -DQWS -D_REENTRANT -Dmain=SDL_main -lSDLmain -lSDL-1.2 -L/opt/arm/openembedded/tmp.boxer/staging/arm-linux/lib -L/opt/arm/openembedded/tmp.boxer/staging/arm-linux/lib/ -lqpe -lqte -lpthread |
| 10:44.53 | mithro | so i'm guessing pygame is overriding that |
| 10:45.02 | mickeyl | look for how SDL is included |
| 10:45.16 | mickeyl | it should look like |
| 10:45.30 | mickeyl | #ifdef QWS |
| 10:45.30 | mickeyl | extern "C" int main(int argc, char *argv[]); |
| 10:45.30 | mickeyl | #include <SDL.h> |
| 10:45.30 | mickeyl | #endif |
| 10:45.41 | mickeyl | or similar |
| 10:45.54 | mickeyl | probably not main, but something else since this a python extension |
| 10:46.08 | mithro | well pygame doesn't have any specific QWS stuff? |
| 10:46.21 | mickeyl | yeah probably without the #ifdef |
| 10:47.14 | Hrw | bon appetit mickey |
| 10:53.00 | *** join/#oe pb_ (~pb@2002:3e03:42cc:1:20b:6aff:fe19:f290) |
| 10:53.35 | Hrw | hi pb |
| 10:56.19 | pb_ | hi hrw |
| 10:58.15 | mithro | i think i've figured out where i have to do something |
| 10:58.21 | mithro | but i'm not sure how to do it |
| 10:59.51 | Hrw | pb_: mickey told that you are the guy who knows best why glibc_cvs is better then glibc_2.3.2. glibc 2.3.2 has 10_cvs.patch applied which (I think) updates it to a little more up-to-date code then plain 2.3.2. is there so many changes in cvs that glibc_cvs is much better? |
| 11:04.43 | pb_ | Hrw: yes |
| 11:05.22 | Hrw | so my next toolchain will be cvs based |
| 11:05.41 | Hrw | but 1st I want to build and test gcc 3.3.3 one |
| 11:14.25 | pb_ | morning chouimat |
| 11:14.32 | chouimat | morning |
| 11:14.40 | chouimat | hi pb_ |
| 12:12.11 | pb_ | hi mickeyl |
| 12:12.17 | mickeyl | hi pb_ |
| 12:12.31 | mickeyl | i think we should extract base-files and put them into oe like initscripts |
| 12:13.06 | mickeyl | it's only a couple of files |
| 12:13.22 | mickeyl | and we want to adapt half of them anyway |
| 12:14.46 | mithro | hey mickeyl |
| 12:14.49 | mickeyl | pb_: i changed the patch gateway user, so you may or may not experience difficulties. if you have a patch handy, please submit and let us see if that still works |
| 12:15.00 | mithro | got a sec to help me out with this pygame stuff |
| 12:15.02 | mickeyl | mithro: I'm afraid I can't be of any help since I'm not familiar with pygame |
| 12:15.12 | mickeyl | i just added it to oe and got it to compile |
| 12:15.13 | mithro | i've got the pygame side done |
| 12:15.19 | mithro | i need help with the sdl side :) |
| 12:15.30 | mithro | i can't use sdl_main |
| 12:15.39 | mithro | because of the way the module works |
| 12:15.55 | mickeyl | well the pygame folks seem to have found a way through this or else it wouldn't work for them |
| 12:16.05 | mickeyl | so you have to compile it on your x86 box and then compare |
| 12:16.09 | mithro | so i kinda need to know if there is a function like the following |
| 12:16.27 | mithro | <PROTECTED> |
| 12:16.27 | mithro | <PROTECTED> |
| 12:16.27 | mithro | <PROTECTED> |
| 12:16.27 | mithro | <PROTECTED> |
| 12:16.27 | mithro | <PROTECTED> |
| 12:16.28 | mithro | <PROTECTED> |
| 12:16.30 | mithro | <PROTECTED> |
| 12:16.39 | mithro | i think we'll need another one for QWS |
| 12:16.48 | *** join/#oe mrz80 (500@caledonia.nerdc.ufl.edu) |
| 12:16.58 | mithro | thats basicly the init side |
| 12:17.31 | mickeyl | mithro: hmm. it puzzles me a bit that we have to do additional work for it to work with qws |
| 12:17.36 | mithro | the SDL_main is basically empty on linux/msdos |
| 12:17.46 | mithro | s/msdos/freebsd |
| 12:18.15 | mithro | on windows the SDL_RegisterApp does all the hard work of SDL_main |
| 12:18.29 | mickeyl | mithro: look through the qtopia specific main in SDL. seems we just have to copy the initialization there |
| 12:18.31 | mithro | and SDL_InitQuickDraw does it on MWERKS |
| 12:18.51 | mithro | if there a way we could push that into a SDL_RegisterApp like function? |
| 12:19.05 | pb_ | mickeyl: yah, I agree with what you said about base-files |
| 12:19.17 | mickeyl | pb_: ok, good. will do that |
| 12:19.34 | pb_ | I don't have a patch on hand at the moment, but I'll try submitting something later today and let you know. |
| 12:19.38 | mickeyl | k |
| 12:20.46 | mithro | mickeyl: so would moving that stuff around be okay? |
| 12:21.09 | mickeyl | mithro: i advise against generally moving it |
| 12:21.21 | mickeyl | mithro: it would seem better to me to adapt SDL_Video_Init (or so) |
| 12:21.28 | mickeyl | check if we already have a qpeapplication object |
| 12:21.32 | mickeyl | if not, create one |
| 12:21.57 | mickeyl | because we would break all qtopia sdl apps around if we would move it |
| 12:22.07 | mithro | i was going to move it |
| 12:22.11 | mithro | then call it in sdl_main |
| 12:22.28 | mithro | so all the qtopia sdl apps don't need to change |
| 12:23.51 | mickeyl | not sure if that works |
| 12:23.58 | mickeyl | the problem is that SDLmain is a static library |
| 12:24.07 | mickeyl | well it may work |
| 12:24.25 | mithro | it seems to work for windows? |
| 12:24.38 | mickeyl | on a side note, someone should go through the agawa-piro-mickey patch, review it, clean up and submit upstream |
| 12:24.49 | mickeyl | dave hedbour seems to no longer care about qtopia/sdl |
| 12:25.35 | mithro | if i get into this sdl stuff i'll have a look |
| 12:25.53 | mithro | but i can't guarrenty anything atm |
| 12:26.18 | mickeyl | fair enough |
| 12:32.22 | mithro | dang |
| 12:32.25 | mithro | <PROTECTED> |
| 12:32.25 | mithro | <PROTECTED> |
| 12:32.25 | mithro | <PROTECTED> |
| 12:33.32 | mickeyl | not necessary |
| 12:33.54 | mickeyl | you can construct the app without |
| 12:34.02 | mithro | you can? |
| 12:34.04 | mithro | how? |
| 12:34.10 | mickeyl | um |
| 12:34.28 | mickeyl | app = new QPEApplication( 1, "executable name" ) |
| 12:34.32 | mickeyl | or so |
| 12:34.38 | mithro | ahh okay |
| 12:37.37 | mickeyl | bon appetit hrw |
| 12:39.08 | mithro | whats the 1 mean? |
| 12:40.02 | *** join/#oe schurig (~schurig@p5080A20F.dip0.t-ipconnect.de) |
| 12:43.37 | pb_ | bbl |
| 12:44.36 | mickeyl | bon appetit |
| 12:44.50 | mickeyl | mithro: argc |
| 12:44.53 | mickeyl | argument count |
| 12:51.07 | mithro | ahh okay |
| 12:58.02 | Hrw | re |
| 12:58.08 | proti | re |
| 13:00.19 | *** part/#oe Hrw (~hrw@sd38.internetdsl.tpnet.pl) |
| 13:02.12 | *** join/#oe Hrw (~hrw@sd38.internetdsl.tpnet.pl) |
| 13:04.05 | mithro | okay i think i've figured this out |
| 13:10.47 | *** join/#oe Neil (3ee93@lsanca1-ar51-4-42-017-130.lsanca1.dsl-verizon.net) |
| 13:10.58 | Hrw | hi Neil |
| 13:11.10 | Neil | hello |
| 13:12.29 | mickeyl | hi neil |
| 13:17.24 | mickeyl | ibot: seen sikko |
| 13:17.27 | ibot | sikko <~chatzilla@fia136-144-100.dsl.hccnet.nl> was last seen on IRC in channel #opie, 5d 22h 54m 18s ago, saying: 'but could someone help me with this pda, if it works, it will be sold in europe!'. |
| 13:21.37 | *** join/#oe reenoo_ (~nooph@wlan-client-073.informatik.uni-bremen.de) |
| 13:21.51 | reenoo_ | afternoon |
| 13:22.43 | mickeyl | hi reenoo_ |
| 13:22.46 | mithro | is there a way to get the output of patcher? |
| 13:22.51 | mithro | ie why a patch failed? |
| 13:23.12 | mickeyl | use patch |
| 13:23.20 | mickeyl | cat <patch>|patch -p1 --dry-run |
| 13:23.23 | mickeyl | this will explain |
| 13:23.39 | mickeyl | you then can modify the offending hunks |
| 13:23.52 | reenoo_ | think that's what patcher does first as well |
| 13:24.47 | mithro | tim@gateway:/home/arm/openembedded/tmp.boxer/base/python-pygame-1.6-r0$ cat /home/arm/openembedded/packages.boxer/python/python-pygame-1.6/qpe.patch |patch -p1 --dry-run |
| 13:24.47 | mithro | patching file pygame-1.6/src/base.c |
| 13:24.47 | mithro | tim@gateway:/home/arm/openembedded/tmp.boxer/base/python-pygame-1.6-r0$ |
| 13:24.49 | mithro | ? |
| 13:25.28 | mithro | that would point to patcher should work right? |
| 13:26.51 | mickeyl | yes |
| 13:26.59 | mickeyl | note that patcher has problems when patches are already applied |
| 13:27.08 | Neil | doesn't patcher leave reject files in the sources too? |
| 13:27.09 | mickeyl | in that case i would clean the package and rebuild |
| 13:27.12 | Neil | that'd also hint at what's wrong |
| 13:27.34 | mithro | yeah i did a clean |
| 13:27.35 | mickeyl | no. patcher won't try to apply when --dry-run doesn't work |
| 13:27.42 | mickeyl | hence it doesn't leave .rej files |
| 13:27.45 | Neil | ok |
| 13:28.02 | mithro | and patcher is still complaining |
| 13:29.02 | mithro | any ideas? |
| 13:29.10 | mickeyl | no |
| 13:29.21 | mickeyl | call patcher manually and see what it tells |
| 13:29.30 | mithro | how do you call patcher? |
| 13:29.40 | mickeyl | patcher |
| 13:29.43 | mickeyl | patcher --help |
| 13:29.47 | mickeyl | gives a lot of instructions |
| 13:30.19 | mickeyl | in any case, the father of patcher is online here... should ask him when nothing works :) |
| 13:31.24 | mithro | where does patcher install? |
| 13:31.40 | mickeyl | install ? |
| 13:31.47 | mickeyl | it's in ${STAGING_BINDIR} |
| 13:31.50 | mickeyl | like all host tools |
| 13:31.51 | mithro | ahh in staging |
| 13:32.46 | mithro | i've never had to use em directly |
| 13:33.32 | mithro | tim@gateway:/home/arm/openembedded/tmp.boxer/base/python-pygame-1.6-r0$ ../../staging/i686-linux/bin/patcher --help NOTE: Eventually I will display some usage, but for now this is missing |
| 13:33.32 | mithro | tim@gateway:/home/arm/openembedded/tmp.boxer/base/python-pygame-1.6-r0$ |
| 13:33.44 | mithro | :( |
| 13:35.42 | schurig | robilad: doing some java stuff for OpenEmbedded ? |
| 13:35.58 | robilad | schurig: I'm everywhere ;) |
| 13:36.19 | schurig | me too |
| 13:37.37 | robilad | heh:) |
| 13:38.02 | mithro | schurig: i need help with patcher :) |
| 13:38.25 | robilad | i was busy in the ;ast couple of days raising some havoc about the scsl licenses from sun, since they were spinning it as open-source, whetever. i found some nice 'sco-like' provisions in it. |
| 13:38.39 | schurig | mithro: today I heard that patcher on my web page is borked (althought it works for me). I guess the patch and diff tools behave slightly different |
| 13:38.40 | *** join/#oe jh_ (~user@209.137.229.132) |
| 13:38.54 | mithro | well it patches everything else correctly |
| 13:38.56 | robilad | looks like people are starting to check their contracts :) |
| 13:38.57 | mithro | just not this patch |
| 13:39.17 | schurig | mithro: do you want to create a patch or import a patch? |
| 13:39.35 | mithro | import a patch |
| 13:39.49 | mithro | every other patch i've created so far as worked... |
| 13:39.58 | schurig | mithro: is the patch in -p1 format? e.g. would patch --dry-run -p1 <patch work without error? |
| 13:40.25 | mithro | yeah the dryrun works okay when run manually |
| 13:40.50 | schurig | then maybe you just do a "patcher -f -i patch" and see what rejects you get :-) |
| 13:41.03 | schurig | "-f" is "May the force be with you" |
| 13:41.20 | schurig | mithro: but maybe I should rename this to "-j": "May Jesus be with you" :-) |
| 13:41.59 | mithro | why does patcher try two different dry runs? |
| 13:42.27 | schurig | when I had to use "patcher -f -i ..." I usually do "vi file.rej file", fix the problems and later do a "patcher -r" to make a cleanly applying patch |
| 13:43.06 | mithro | ahh hah! |
| 13:43.26 | mithro | i was running patch from a different location then patcher was running |
| 13:43.29 | schurig | mithro: does it? When I search for "--dry-run" in my version I only find it once |
| 13:43.59 | mithro | will g++ compile C code okay? |
| 13:44.35 | schurig | mithro: you can add "patcher -ddd -i patch" to see lots of debug info |
| 13:44.56 | mickeyl | yeah, be sure to always use patcher from the same directory (usually the topdir of the src) |
| 13:45.29 | mithro | mickeyl: you wouldn't happen to know how to get setup.in to use g++ instead of gcc? |
| 13:45.30 | Neil | g++ can't compile all C code |
| 13:45.40 | schurig | mithro: it will, but it might spit out more errors warnings. And you can use more C++ismn in your source. |
| 13:45.52 | Neil | the most obvious case being C code that has variables named with C++ keywords |
| 13:46.18 | schurig | Neil: int new = TRUE; |
| 13:46.26 | Neil | heh |
| 13:47.58 | gb2 | http://david.tribble.com/text/cdiffs.htm |
| 13:48.18 | gb2 | this is a good webpage describing incompatibilities between ISO C and ISO C++ |
| 13:53.53 | mithro | dang - some of the things this module is doing C++ doesn't like |
| 13:54.09 | mithro | | src/base.cc: In function `void PyGame_RegisterQuit(void (*)())': |
| 13:54.10 | mithro | | src/base.cc:105: error: invalid conversion from `void (*)()' to `void*' |
| 13:56.13 | gb2 | mithro: that probably requires an explicit cast |
| 13:56.13 | gb2 | hrmm. |
| 13:56.16 | gb2 | it shouldn't though |
| 13:56.29 | gb2 | ofcourse. i don't know what it's actually trying |
| 13:56.52 | pb_ | re |
| 13:56.56 | gb2 | re pb |
| 13:58.17 | pb_ | schurig: I think the way patcher tries to apply the patch with -R, and bails out if that works, is bogus. It means that it will often refuse to apply a patch that just deletes lines. |
| 13:58.55 | *** join/#oe a-atwood (~xf009976@proxyle02.ext.ti.com) |
| 13:59.57 | schurig | pb_: hmm, from my comments in patcher I see: |
| 14:00.02 | schurig | # I have two options: I just revert the patch via patch -R ... |
| 14:00.02 | schurig | <PROTECTED> |
| 14:00.02 | schurig | <PROTECTED> |
| 14:00.11 | schurig | that's the only occurrence of -R |
| 14:00.39 | schurig | pb_: so I guess that your problem has been already dealt |
| 14:01.10 | pb_ | <PROTECTED> |
| 14:01.10 | pb_ | <PROTECTED> |
| 14:01.10 | pb_ | <PROTECTED> |
| 14:01.16 | pb_ | that's the bit of code I was objecting to. |
| 14:01.21 | pb_ | maybe it's already been removed. |
| 14:01.39 | schurig | pb: and, when did you download patcher? Weeks ago? |
| 14:01.50 | pb_ | I dunno. It's whatever OE downloaded for me this morning. |
| 14:02.28 | pb_ | Since patcher has no version number that I can see, it's hard to tell what vintage it actually is. |
| 14:02.43 | schurig | pb_: oh, they are downloading it from www.holgerschurig.de.... urks. |
| 14:02.47 | gb2 | patcher could use versioning |
| 14:03.07 | Hrw | schurig: maybe you should version patcher and remove nonversioned on |
| 14:03.09 | Hrw | e |
| 14:03.55 | schurig | I put one version on my private home page then I thought that my company will get bankrupt and I'm without a job ... but it's not the most up-to-date-version there |
| 14:04.07 | mickeyl | then again, you could help kergoth finishing the rewrite in python |
| 14:04.14 | mickeyl | see openembedded.bkbits.net/patcher |
| 14:04.19 | mickeyl | :D |
| 14:04.55 | mithro | means i only have 1 socket in the wall |
| 14:06.14 | schurig | mickeyl: I actually believe that kergoth should have sent patches instead of creating a fork. But hey, it's GPL, so he's free. |
| 14:06.44 | gb2 | patches turning it from perl into python? |
| 14:07.33 | schurig | gb2: ah, now I've got it, it's a rewrite, not a fork (/me has some headache today ...) |
| 14:07.40 | mickeyl | schurig: uhm... ? |
| 14:07.42 | mickeyl | ah |
| 14:07.45 | mickeyl | hehe |
| 14:07.45 | *** join/#oe reenoo (~nooph@wlan-client-282.informatik.uni-bremen.de) |
| 14:08.19 | Hrw | hi reenoo |
| 14:09.16 | reenoo | howdy |
| 14:09.40 | schurig | pb_, mithro: I've update the patcher.oe file |
| 14:09.56 | schurig | I'll version patcher on my homepage the other day and then also version the patcher.oe file |
| 14:10.00 | mithro | that would be cool |
| 14:10.00 | mithro | :) |
| 14:10.00 | mithro | (ie though it's C api) |
| 14:10.04 | mithro | dunno why you want to use an evil language like perl, but it was still nifty ;) |
| 14:10.07 | Hrw | I got problem ;( uicmoc3-native-3.3.1-r0 do_configure fails -> The specified system/compiler is not supported: oe/gcc33/tmp/base/uicmoc3-native-3.3.1-r0/qt-x11-free-3.3.1/mkspecs/linux-oe-g++ ;( |
| 14:10.28 | schurig | mithro: just read patcher. You CAN write readable code in perl. :-) |
| 14:10.31 | Hrw | mithro: I saw once C++ program with C interpreter embedded |
| 14:10.55 | mithro | libsdl takes longer to configure on my machine then to build |
| 14:11.10 | schurig | mithro: ccache? |
| 14:11.17 | mithro | yeah :) |
| 14:11.37 | Hrw | using OE without ccache is masochizm |
| 14:11.40 | mithro | but configuring takes a very long time |
| 14:12.12 | mithro | but then buildroot didn't download a new version of glibc every day ;) |
| 14:12.21 | schurig | configure should have been written in a programming language. After all, automake (or was it autoconf) requires perl anyway. Those shell monsters for libtool and configure are sick. |
| 14:12.22 | mickeyl | unfortunately ccache is of virtually no use in OE |
| 14:12.36 | schurig | mickeyl: why? |
| 14:12.45 | mithro | i wonder what size my ccache is |
| 14:12.53 | mickeyl | schurig: since a new compiles invalidates the cache |
| 14:12.55 | schurig | mickeyl: ccache -s |
| 14:13.05 | mickeyl | schurig: and OE is creating a new toolchain |
| 14:13.08 | mickeyl | all the time |
| 14:13.23 | mithro | cache hit 20 |
| 14:13.24 | mithro | cache miss 54986 |
| 14:13.33 | mithro | my cache is to small |
| 14:13.45 | Hrw | mithro: you can enlarge it |
| 14:14.04 | schurig | mickeyl: hmm, I bk pulled oa-packages and copy those packages that I want. This way I don't suffer from "XYZ broke the *.oe file under my ass" problems. And then the toolchain won't be rebuild all the time. |
| 14:14.14 | a-atwood | in 7-10 days! Guaranteed! |
| 14:14.19 | mithro | yeah done |
| 14:14.21 | schurig | cache hit 55145 |
| 14:14.21 | schurig | cache miss 57573 |
| 14:14.24 | gb2 | cache hit 274652 |
| 14:14.24 | gb2 | cache miss 288699 |
| 14:14.37 | gb2 | cache size 3.7 Gbytes |
| 14:14.37 | gb2 | max cache size 4.0 Gbytes |
| 14:14.43 | mithro | he he gb2 does alot more builds then me :) |
| 14:14.50 | Hrw | gb2: lots |
| 14:14.50 | schurig | cache size 1.2 Gbytes |
| 14:14.50 | schurig | max cache size 5.0 Gbytes |
| 14:14.54 | gb2 | i do toolchain builds over and over and over |
| 14:14.59 | schurig | but I build ONLY those things that I actually use |
| 14:15.00 | Hrw | cache size 850.4 Mbytes |
| 14:15.01 | Hrw | max cache size 976.6 Mbytes |
| 14:15.15 | reenoo | that reminds me... |
| 14:15.16 | Hrw | had it set to 1.5GiB |
| 14:15.30 | reenoo | I need new hard disks and a new PSU :/ |
| 14:16.05 | Hrw | reenoo: I want 48h days |
| 14:16.46 | gb2 | uicmoc3-native looks wrong compared to uicmoc-native. they both look like they could be updated though |
| 14:16.49 | Hrw | hmm.. I think that I can test opie/gcc333 on Z without i18n |
| 14:17.43 | mickeyl | gb2: ? |
| 14:18.02 | gb2 | mickeyl: Hrw got an error |
| 14:18.10 | gb2 | i was glancing at them to see what might have gone wrong |
| 14:18.24 | mickeyl | gb2: oh that's an artefact of your qmake work |
| 14:19.38 | mickeyl | it worked previously using the linux-g++ |
| 14:19.38 | gb2 | does the configure script call qmake itself? |
| 14:19.38 | *** join/#oe prpplague (~billybob@12.148.134.14) |
| 14:19.38 | gb2 | neither of these inherit qmake |
| 14:19.43 | mickeyl | do_configure() { |
| 14:19.43 | mickeyl | <PROTECTED> |
| 14:19.43 | mickeyl | <PROTECTED> |
| 14:19.43 | mickeyl | } |
| 14:19.56 | mickeyl | this is opie-i18n |
| 14:20.00 | mickeyl | so this should work |
| 14:20.06 | gb2 | his error was uicmoc3-native |
| 14:20.09 | mickeyl | eeek |
| 14:20.11 | mickeyl | no |
| 14:20.15 | mickeyl | that doesn't look good |
| 14:20.23 | mickeyl | ah you just used the sed razor |
| 14:20.27 | mickeyl | this is wrong |
| 14:20.33 | gb2 | huh? |
| 14:20.38 | mickeyl | opie-i18n- doesn't want to use qws linux-oe-g++ |
| 14:20.55 | mickeyl | it wants to use the non-qws linux-g++ |
| 14:21.08 | mickeyl | because these are native utilities |
| 14:21.19 | mickeyl | and we don't have a qte/server running normally |
| 14:21.37 | mickeyl | (on the build machine) |
| 14:21.50 | gb2 | is opie-i18n a native package? |
| 14:22.09 | mickeyl | opie-lrelease-native, yes |
| 14:22.12 | mickeyl | opie-lupdate-native, yes |
| 14:22.30 | gb2 | the oe-g++ was intended to support packages that inherit native |
| 14:22.33 | mickeyl | opie-i18n uses both native tools |
| 14:22.37 | gb2 | it sets the compiler correctly |
| 14:23.52 | gb2 | OE changes CXX to be either build or target compiler |
| 14:24.01 | mithro | well python-pygame has compiled |
| 14:24.03 | gb2 | and that's what the qmake.oeclass passes in |
| 14:24.38 | mithro | i also need another sdl application to test i havn't borked it |
| 14:25.28 | mithro | any suggestions? |
| 14:26.01 | mickeyl | vectoroids |
| 14:26.15 | Hrw | billiards |
| 14:26.50 | pb_ | schurig: cool |
| 14:27.15 | gb2 | mickeyl: the only change that should go in i think is removing the -DQWS from linux-oe-qmake.conf in qmake-native |
| 14:27.15 | mithro | they confirmed to work? :) |
| 14:27.32 | gb2 | mickeyl: then rely on the opie oeclass to add it for those that need it |
| 14:27.55 | gb2 | i can go ahead and do that now if you'd like |
| 14:28.03 | mithro | my sd card is slow :( |
| 14:28.14 | mithro | whats that nosync option which speeds it up? |
| 14:28.14 | mickeyl | gb2: go ahead. don't add it in opie.oecalss though, add it to palmtop.oeclass |
| 14:28.32 | gb2 | mickeyl: i thought you said we did that already |
| 14:28.39 | mickeyl | gb2: no we didn't |
| 14:28.46 | mickeyl | gb2: i said it should be done |
| 14:28.48 | gb2 | ah |
| 14:28.57 | mithro | dang it didn't link like i expected |
| 14:28.58 | mickeyl | gb2: one problem still remaining is qt vs. qte |
| 14:29.17 | gb2 | CXXFLAGS_append = " -DQWS" ought to do the trick. |
| 14:29.49 | mickeyl | hmm |
| 14:30.07 | mickeyl | i would have used QMAKEVARS_POST += "DEFINES+=QWS" |
| 14:30.26 | mickeyl | doesn't make a difference |
| 14:30.43 | mickeyl | one open issue to handle is still qt vs. qte |
| 14:31.17 | gb2 | either probably works. i guess qmakevars may be more correct |
| 14:31.42 | mithro | okay i got a question :) |
| 14:32.17 | mickeyl | qws/bla-bla-g++ templates link to qte if CONFIG=qt |
| 14:32.21 | mickeyl | bla-bla-g++ templates link to qt if CONFIG=qt |
| 14:32.26 | mickeyl | we need to handle that |
| 14:32.32 | mithro | if i do a -lSDLmain it should link the program with any function in the SDLmain libary? |
| 14:33.16 | treke|ho1e | there is an SDLMain library? |
| 14:33.30 | mithro | well SDLMain is a lib produced by SDL |
| 14:33.35 | schurig | bye bye |
| 14:33.41 | mithro | libSDLMain.a |
| 14:34.06 | gb2 | EXTRA_QMAKEVARS_POST_append = " DEFINES+=QWS" |
| 14:34.07 | treke|ho1e | I think what you want is -Dmain=SDLMain |
| 14:34.07 | mithro | i'm assuming you pull it into your application to do the main part :) |
| 14:34.15 | gb2 | i'm ending up with that to fit in with what qmake.oeclass does |
| 14:34.42 | mithro | treke|ho1e: in the SDLMain stuff i have declared a function called "initQCop" |
| 14:34.56 | mithro | i need to call this in pygame to initalise the QPEApplication stuff |
| 14:35.19 | CIA-5 | 03gbritton 07 * r1.597.1.1 10oe/classes/palmtop.oeclass: moved -DQWS from qmake-native into palmtop.oeclass |
| 14:35.38 | treke|ho1e | pygame? Sounds like your on to a loser. |
| 14:35.57 | CIA-5 | 03gbritton 07 * r1.1503.1.1 10packages/qmake/qmake-native-3.2.3/linux-oe-qmake.conf: move setting of -DQWS from qmake-native into palmtop.oeclass |
| 14:36.10 | mithro | mickeyl: any ideas? |
| 14:36.45 | Hrw | I started patching :) 1. http://home.hrw.one.pl/oe/0510_sysvinit.diff - removing update-rc.d from it |
| 14:38.04 | gb2 | mickeyl: there's a little more left to abstract in linux-oe-g++, i think just the qt vs. qte stuff wrt. libraries |
| 14:38.14 | gb2 | mickeyl: that should be straightforward to do though |
| 14:40.51 | mithro | okay i think it's my lack of C knowledge hindering me here |
| 14:42.21 | *** join/#oe fido_hl (~hl@z2421-034-174.Raab-Heim.Uni-Linz.AC.AT) |
| 14:42.31 | Hrw | can someone build modutils-collateral? I have errors like those when MACHINE is not set (but it is set to "collie" in conf/local.conf). initscripts builds fine |
| 14:45.15 | mithro | i declear "static void initQCop(int argc, char *argc)"; in SDL_qtopia_main.cc |
| 14:45.51 | mithro | then in base.c in pygame i do a "extern void initQCop(int argc, char *argc[]);" |
| 14:46.04 | mithro | (make that first line an *argc[] as well) |
| 14:46.31 | mithro | if i specifically link against SDLmain using -lSDLmain it should join the two right? |
| 14:47.01 | treke|ho1e | mickeyl: fear. kdbase 3.3 will theoretically be able to build against qt/e |
| 14:47.10 | mithro | or am i mising something? |
| 14:55.19 | mithro | any idea anyone? |
| 14:55.26 | Neil | treke|ho1e: kdebase or kdelibs? |
| 14:56.12 | pb_ | mithro: you need to omit 'static' |
| 14:56.45 | mjc | woot. building gcc 3.4 |
| 14:56.55 | mithro | from the first one? |
| 14:57.43 | pb_ | yes |
| 14:58.12 | mithro | will the function then be able to access static varibles/functions? |
| 14:58.45 | treke|ho1e | Neil: kdebase |
| 14:59.11 | Neil | that surprises me |
| 14:59.26 | pb_ | mithro: mm? |
| 15:00.20 | mithro | well the initQCop function needs to put set the static varible app to a new QPEApplication |
| 15:00.30 | mithro | s/put// |
| 15:00.49 | pb_ | that's fine |
| 15:00.52 | mithro | okay |
| 15:00.56 | Hrw | fsck... |
| 15:02.07 | mithro | and it's right that the so won't depend on the .a as the linker will pull in the stuff at build time? |
| 15:03.27 | Hrw | ha! it's good to have old OE |
| 15:03.55 | mithro | Hrw: why? :) |
| 15:04.03 | Hrw | after pulling OEDIR I get unable to build modutils-collateral |
| 15:05.16 | Hrw | mithro: at home I have up-to-date OEDIR and PKGDIR - unable to build, at work I had old OEDIR/PKGDIR - it build, pulled PKGDIR it built, pulled OEDIR - it not built |
| 15:06.32 | mithro | dang now i get |
| 15:06.36 | mithro | | /home/arm/openembedded/tmp.boxer/cross/bin/../lib/gcc-lib/arm-linux/3.3.3/../../../../arm-linux/bin/ld: cannot find -lSDL_main |
| 15:07.40 | mithro | yet SDLmain can be found in the staging directories |
| 15:08.14 | mithro | and i'm an idiot again :) |
| 15:11.48 | mithro | then i'm expecting a segfault :) |
| 15:12.25 | mithro | pb_: that didn't seem to work, i'm still getting |
| 15:12.34 | mithro | ImportError: /mnt/card/usr/lib/python2.3/site-packages/pygame/base.so: undefined symbol: initQCop |
| 15:13.16 | fido_hl | does anyone have problems with lrzsz? i get m4.patch not found. there is no m4.patch in the lrzsz-0.12.20 dir |
| 15:13.29 | Hrw | heh.. found problem |
| 15:13.41 | Hrw | fido_hl: known problem - wait for kergoth |
| 15:13.58 | fido_hl | Hrw: ok |
| 15:19.27 | mithro | pb_: any idea? |
| 15:19.40 | pb_ | mithro: what does nm have to say? |
| 15:19.55 | mithro | about base.so or libSDLmain.a? |
| 15:20.27 | pb_ | the latter |
| 15:20.51 | *** join/#oe drw (~drw@c-67-172-218-120.client.comcast.net) |
| 15:20.57 | mithro | what am i looking for :) |
| 15:21.00 | mithro | never used nm |
| 15:21.10 | pb_ | references to the function in question |
| 15:21.33 | mithro | 00000338 T _Z8initQCopiPPc |
| 15:21.49 | pb_ | that seems fine |
| 15:22.03 | pb_ | how about in base.so? |
| 15:22.54 | mithro | <PROTECTED> |
| 15:23.03 | mithro | i'm assuming U means undefined |
| 15:23.05 | gb2 | something got C vs. C++ linkage wrong |
| 15:23.24 | mithro | could i need to wrap that in extern "C" {} or something? |
| 15:23.54 | gb2 | _Z8initQCopiPPc demangles to initQCop(int, char**) |
| 15:24.19 | mithro | but why isn't it mangled in base.so? |
| 15:24.31 | gb2 | base.so probably thinks it's a C function |
| 15:25.06 | mjc | is there a listing of what goes on the oz rootfs anywhere? |
| 15:25.21 | pb_ | mithro: is base.so written in C, or C++? |
| 15:25.40 | mithro | C |
| 15:25.45 | pb_ | ah, that would be a problem |
| 15:25.59 | gb2 | the C++ library probably wants to export a C interface |
| 15:26.04 | mithro | so i need an "extern "C" {}" around the function? |
| 15:26.08 | pb_ | you either need to call it with the mangled name directly, or define it with C linkage in your other library |
| 15:26.32 | mithro | or is there some way other way to define C linkage? |
| 15:26.52 | mithro | i definatley want to export it as C function |
| 15:28.11 | Hrw | I'll commit at home |
| 15:28.46 | mithro | pb_: ? |
| 15:29.28 | Hrw | mjc: set DISTRO=openzaurus MACHINE=collie (or which zaurus you have) and build "bootstrap-image" |
| 15:30.17 | Hrw | mjc: or if you want then I can publish one of my bootstrap-images so you can see |
| 15:30.29 | gb2 | mithro: extern "C" initQCop(int c, char **v) { ... code here ... } |
| 15:30.34 | gb2 | that should do what you want |
| 15:31.19 | gb2 | in your header, you want: extern "C" { initQCop(int, char **); } |
| 15:31.41 | mjc | Hrw: ah cool, distro and machine are set in local.conf already |
| 15:31.58 | Hrw | mjc: so "oemake bootstrap-image" should be enough |
| 15:32.25 | Hrw | mjc: or "oemake opie-image" if you want to have (not-yet?)-working opie |
| 15:33.09 | mithro | yay for more evilness |
| 15:33.11 | mithro | ImportError: /usr/lib/python2.3/site-packages/pygame/base.so: undefined symbol: SDL_main |
| 15:33.16 | Hrw | mithro: ;) |
| 15:34.10 | mithro | now the question is why is it even after that? |
| 15:34.29 | mithro | as SDL_main isn't even in any of the source code |
| 15:37.50 | mithro | atleast my ccache hits are climbing |
| 15:39.12 | Hrw | mickeyl: can openssl dependency in ntp be dropped? I think that it is not needed.. |
| 15:41.05 | mickeyl | hmm |
| 15:41.24 | treke|ho1e | ntp has ssl support |
| 15:41.34 | treke|ho1e | if you disable it you can drop the dependency. |
| 15:42.00 | mickeyl | ya |
| 15:42.10 | mickeyl | i think we don't really need ssl support in ntp |
| 15:42.56 | Hrw | ok - I drop it (on my system it wasn't even compiled with ssl) |
| 15:43.09 | treke|ho1e | make sure it isnt. |
| 15:43.12 | treke|ho1e | check the oe |
| 15:43.16 | treke|ho1e | dont just remove the dependency. |
| 15:43.50 | treke|ho1e | mickeyl: I'm not particularly convinced that getting rid of openssl dependencies is that useful. Seems there are always going to be quite a few things depending on it. |
| 15:44.10 | Neil | people are fairly likely to want https anyway |
| 15:44.23 | treke|ho1e | konq, links, libetpan |
| 15:44.33 | pb_ | ssl in general is certainly useful, but I'm not sure that openssl is the right answer. |
| 15:44.41 | pb_ | it's a big chunk of code for what it does. |
| 15:44.51 | Neil | it might not be the right answer, but it's the answer that apps authors most often choose |
| 15:44.57 | treke|ho1e | pb_: it's a big chunk of code, but it's a big chunk of code that is being used |
| 15:44.58 | mickeyl | someone should really look into if matrixssl can be used instead |
| 15:46.01 | Hrw | ok - I'll look more at ntp - 1st split it to ntp and ntpdate packages (opie-systemtime needs ntpdate only not whole ntp daemon) |
| 15:49.21 | mickeyl | yep |
| 15:50.04 | Hrw | hmm.. maybe adding "-c --clean" to oebuild will be useful ;) |
| 15:50.39 | Hrw | ok. compiled nice |
| 15:50.39 | mickeyl | don't we already have oebuild clean ? |
| 15:51.31 | Hrw | but I have to oebuild clean cos/cos.oe; oebuild cos/cos.oe |
| 15:51.45 | Hrw | oebuild -c cos/cos.oe would be more useful |
| 15:52.58 | mickeyl | well |
| 15:53.03 | Hrw | when "oemake" fails on something I prefer oebuilding failed thing (after checking why) |
| 15:53.34 | mickeyl | i don't feel good with adding convenience functions to oemake when you can use other means insteda |
| 15:53.39 | mickeyl | e.g. i have a function |
| 15:53.55 | mickeyl | oenew() { oebuild clean $1 && oebuild $1 } |
| 15:54.05 | mithro | ahh figured it out |
| 15:54.19 | Hrw | ok |
| 15:54.24 | mithro | dang |
| 15:54.26 | mithro | <PROTECTED> |
| 15:54.26 | mithro | Fatal Python error: (pygame parachute) Segmentation Fault |
| 15:54.26 | mithro | Aborted |
| 15:54.57 | mickeyl | sounds like gdb time |
| 15:55.01 | mithro | yeah |
| 15:55.11 | mithro | anyone got a prebuilt gdb package? |
| 15:55.34 | mickeyl | not here |
| 15:55.59 | mickeyl | i really would test this on x86 prior to on your Z |
| 15:55.59 | mithro | btw would parseing initQCoq(1, "test") into a initQCop(int argc, char* argv[]) cause a segfault? |
| 15:56.19 | mickeyl | could be |
| 15:56.29 | mickeyl | "test" actually is char* |
| 15:56.32 | mickeyl | not an array |
| 15:56.40 | mickeyl | so you should reserve an array there |
| 15:56.48 | *** join/#oe Disconnect (~dis@3ffe:bc0:91:0:203:6dff:fe1a:c5b6) |
| 15:56.48 | mithro | how do i do that? :) |
| 15:56.54 | mickeyl | no idea |
| 15:56.57 | mickeyl | better grab a C book |
| 15:57.37 | Neil | char *myArray[1]; myArray[0] = "test"; |
| 15:58.25 | Neil | that'll give you a constness warning though |
| 15:58.56 | Neil | const char *myArray[1]; myArray[0] = "test"; should work better |
| 16:04.25 | Hrw | btw.. mickeyl - I tested apmd more precise on Z - it needs fixing - after resume keyboard is gone (random chars generating) |
| 16:04.37 | Hrw | bye mickeyl |
| 16:06.06 | pb_ | mithro: yes, passing a char* when a char*[] is expected would certainly cause a segfault. |
| 16:06.42 | pb_ | you probably want something like: char *myargv[] = { "test", NULL }; initQCop(2, myargv); |
| 16:06.55 | pb_ | er, but with 1 rather than 2 |
| 16:08.55 | mithro | is that NULL important? |
| 16:08.58 | pb_ | no, not really |
| 16:09.12 | mithro | so the stuff Neil said would work? |
| 16:09.14 | pb_ | if you pass argc as 1, the callee should only look at the first element of the array |
| 16:09.28 | pb_ | yup, what Neil said would be fine too |
| 16:09.47 | Neil | unless the function you're passing to specifically requires a null-terminated array, the null is unnecessary and what I said should work too |
| 16:09.56 | mithro | lousy gdb taking long time to compile |
| 16:10.14 | mithro | does QPEApplication require a null-terminated array? |
| 16:10.34 | Neil | I doubt it, assuming that first argument is the count |
| 16:10.59 | pb_ | writing it as char *array[] = { ...}; is fractionally more efficient than char *array[1]; array[0] = ..., but in initialisation code I doubt you will see much of an impact |
| 16:11.00 | Neil | oh, QApp? then it's definitely not required |
| 16:11.02 | CIA-5 | 03kergoth 07 * r1.1515.1.1 10packages/lrzsz/lrzsz-0.12.20/ (autotools.patch m4.patch): Add the m4 and autotools patches for lrzsz. Build still fails with @CATALOGS@ not being sed substituted on the makefiles. |
| 16:11.03 | CIA-5 | 03kergoth 07 * r1.1518 10packages/BitKeeper/deleted/.del-autotools.patch: |
| 16:11.03 | CIA-5 | Merge openembedded@openembedded.bkbits.net:packages |
| 16:11.03 | CIA-5 | into linux.local:/home/kergoth/code/packages |
| 16:11.20 | mithro | morning kergoth |
| 16:11.48 | pb_ | hey kergoth |
| 16:11.54 | kergoth | morning, bbl, heading into the office... i need to start going to bed earlier to get back on a slightly more sane schedule |
| 16:12.09 | Hrw | hi kergoth |
| 16:12.43 | *** join/#oe treke_ (~ggilbert@ip230-225.outside.westmont.edu) |
| 16:13.59 | pb_ | kergoth: heh |
| 16:16.28 | mithro | dang gdb doesn't compile |
| 16:18.11 | Hrw | bye |
| 16:19.22 | mithro | well i got 11 minutes before i need to be in bed |
| 16:20.25 | mithro | #0 0x40265034 in strchr () from /lib/libc.so.6 |
| 16:20.25 | mithro | #1 0x402667d8 in strrchr () from /lib/libc.so.6 |
| 16:20.25 | mithro | #2 0x403a7960 in _fini () from /usr/lib/python2.3/site-packages/pygame/base.so |
| 16:20.45 | mithro | i need to compile with debuging on right? |
| 16:22.13 | mithro | which oe is going to strip right? |
| 16:22.36 | pb_ | yes. |
| 16:22.49 | mithro | how do i stop that? |
| 16:23.13 | pb_ | set INHIBIT_PACKAGE_STRIP in local.conf |
| 16:24.23 | mithro | to what? |
| 16:24.53 | pb_ | anything |
| 16:25.06 | pb_ | oh, no, actually it has to be "1" |
| 16:25.14 | pb_ | that seems a bit bogus |
| 16:25.21 | mithro | well i set it to 1 |
| 16:25.25 | mithro | so it was a luck guess |
| 16:26.30 | pb_ | heh |
| 16:26.49 | mithro | it's still saying there arn't any debuging symbols? |
| 16:27.15 | mithro | how do i find out if the binary is stripped? |
| 16:28.04 | mithro | the binary definatley isn't striped :( |
| 16:29.55 | pb_ | must have been built without -g |
| 16:30.04 | mithro | i added it in |
| 16:30.14 | mithro | is argc suppose to be 1 or 0? |
| 16:30.33 | pb_ | 1 |
| 16:31.49 | mithro | yay! |
| 16:31.52 | mithro | i figured it out :) |
| 16:31.59 | pb_ | cool |
| 16:32.07 | mithro | it helps if you use the args you are creating :) |
| 16:32.14 | pb_ | heh |
| 16:32.33 | mithro | how to figure out how to use pygame again :) |
| 16:34.33 | mithro | now i just need to find that sdl.sh script treke gave me |
| 16:38.37 | mithro | wooot! |
| 16:38.46 | mithro | pygame on zaurus! |
| 16:39.12 | gb2 | was stripping breaking it? |
| 16:40.05 | mithro | no |
| 16:40.18 | mithro | it was the fact it's 2:10am in the morning breaking it :) |
| 16:40.55 | gb2 | ah |
| 16:41.16 | *** join/#oe ShALLaX (~ufifiytt@cpc1-ware1-3-0-cust130.lutn.cable.ntl.com) |
| 16:41.35 | gb2 | the various autoperformed things in OE should probably be audited and redone in a consistant manner. |
| 16:42.00 | jh_ | mithro: nice! i was thinking about trying that over the weekend. let us know when your first game .oe is available :) |
| 16:43.13 | *** join/#oe qfh (~qfh@ppp-82-135-0-153.mnet-online.de) |
| 16:43.54 | ShALLaX | hell.. oe sure takes up a lot of room whilst building opie-image |
| 16:44.40 | treke_ | not nearly as much as when building world |
| 16:44.51 | ShALLaX | hehe |
| 16:44.52 | ShALLaX | doh |
| 16:44.53 | pb_ | ShALLaX: indeed, probably a whole $10 worth of hard disk or something. |
| 16:44.54 | ShALLaX | pcmcia-cs build tailed |
| 16:45.07 | ShALLaX | pb_: not the price i paid for these hdds |
| 16:45.07 | ShALLaX | hehe |
| 16:45.13 | pb_ | heh |
| 16:45.15 | treke_ | sounds like you overpaid :p |
| 16:45.22 | ShALLaX | nah, sata |
| 16:45.32 | gb2 | it's about 2.5-3gigs for a glibc opie-world iirc |
| 16:45.32 | mithro | okay now i need to prep a patch |
| 16:45.42 | ShALLaX | im up to 4.5 on opie-image |
| 16:45.55 | treke_ | so that would be $4.5 for my sata drive :) |
| 16:45.57 | ShALLaX | why is pcmcia-cs trying to compile against ipod-2.4.24? |
| 16:46.06 | *** join/#oe SubNuke (~subnuke@nh-londndry-c41-bg1a-32-219.lndnnh.adelphia.net) |
| 16:46.23 | mithro | okay i've done a bk add on the new files |
| 16:46.24 | treke_ | ShALLaX: it's needs to build against some kernel |
| 16:46.31 | ShALLaX | why an ipod one, though |
| 16:46.37 | treke_ | ShALLaX: If you dont tell it which one, it picks on its own |
| 16:46.37 | mithro | what do i have to do to commit them again bk ci -y"message" ? |
| 16:46.37 | ShALLaX | it fails for me :\ |
| 16:46.43 | pb_ | ShALLaX: heh, right. and a 120G SATA drive costs what, $100? |
| 16:46.57 | treke_ | lemme look |
| 16:47.02 | treke_ | I have a receipt right here :p |
| 16:47.20 | ShALLaX | pb: it was more like a 160 that costs $180 |
| 16:47.20 | ShALLaX | :\ |
| 16:47.23 | ShALLaX | and two of them at that |
| 16:47.24 | ShALLaX | raided |
| 16:47.24 | pb_ | ShALLaX: what machine are you targetting? |
| 16:47.31 | ShALLaX | pb_: collie 5500 |
| 16:47.42 | pb_ | and you set MACHINE=collie in local.conf? |
| 16:47.48 | ShALLaX | i believe so, yes |
| 16:48.02 | pb_ | odd. that should certainly cause it to prefer openzaurus-sa for the kernel. |
| 16:48.02 | ShALLaX | shallax@Silver conf $ cat local.conf | grep MACH |
| 16:48.02 | ShALLaX | MACHINE = "collie" |
| 16:48.23 | pb_ | make sure that oe/conf/collie.conf says |
| 16:48.23 | pb_ | PREFERRED_PROVIDERS_append = " virtual/kernel:openzaurus-sa" |
| 16:48.31 | gb2 | uclibc opie-image only came out as 2.2gigs |
| 16:48.41 | ShALLaX | yep |
| 16:48.48 | ShALLaX | the problem is.. i had no DISTRO= set |
| 16:48.58 | ShALLaX | i take it i should have set that to "openzaurus"? |
| 16:49.13 | pb_ | I don't think DISTRO should be all that significant here. but, yes, you should set it to openzaurus |
| 16:49.50 | ShALLaX | if i delete packages/tmp/* will that restore it to a default state? |
| 16:50.18 | *** join/#oe chouimat (~dieu@r2351064.cidc.net) |
| 16:50.59 | pb_ | yes |
| 16:51.05 | ShALLaX | oki |
| 16:51.37 | mithro | the chimp demo seems to run okay on the zaurus |
| 16:53.10 | mithro | even found works well |
| 16:53.13 | mithro | anyway i need to be in bed |
| 16:53.20 | mithro | will submit patches tommorrow |
| 16:53.32 | jh_ | some debugging of the patcher going on? libtool-native prefix.patch is applied twice |
| 16:55.17 | pb_ | I think Glenn McGrath sent a patch for that. |
| 16:55.40 | mithro | have a gnight people |
| 16:55.49 | mithro | i'll try and package up some python games ;) |
| 16:57.55 | jh_ | ah, yep, you're right. thx. |
| 16:58.33 | *** join/#oe GpF (~merdos@195.56.59.27) |
| 16:58.49 | GpF | hi! |
| 17:00.38 | *** join/#oe [2synth] (~synth@pcp858545pcs.ptchar01.fl.comcast.net) |
| 17:05.25 | *** join/#oe mjc (mjc@sartre.ispvip.biz) [NETSPLIT VICTIM] |
| 17:05.30 | *** join/#oe proti (seb@sbernard.net1.nerim.net) [NETSPLIT VICTIM] |
| 17:07.38 | treke_ | hehe the case of the lost loopback is solved :) |
| 17:15.02 | ShALLaX | heh |
| 17:15.14 | ShALLaX | now for the case of my lost sanity |
| 17:17.15 | mickeyl | heh |
| 17:17.38 | *** join/#oe treke_ (~ggilbert@ip230-225.outside.westmont.edu) |
| 17:18.05 | *** join/#oe prpplague (~billybob@12.148.134.14) |
| 17:21.46 | *** join/#oe stigger (~jeff@dip-220-235-100-96.nsw.westnet.com.au) |
| 17:23.01 | *** join/#oe Toi|work (~pleemans@oia-gw1.be.jnj.com) |
| 17:31.35 | Neil | oops |
| 17:31.48 | Neil | just forgot to redirect ipkg-make-index |
| 17:31.59 | Neil | and now I see why it's been suggested that the Sources line be omitted |
| 17:32.15 | Neil | those glibc sources add up |
| 17:32.27 | Neil | a bunch of useless file urls to patches |
| 17:39.55 | treke_ | yeah. ipkg-make-index needs to be patched to init those |
| 17:40.07 | treke_ | removing them from the ipks probably shouldnt be done :) |
| 17:40.41 | treke_ | though they'd probably still get stored once on the device, so maybe the SRC_URI isnt the best thing in the world |
| 17:41.31 | ShALLaX | http://gentoox.shallax.com/mrtg/temp.html |
| 17:41.35 | ShALLaX | its hell being in this room |
| 17:41.36 | ShALLaX | heh |
| 17:42.09 | ShALLaX | 4 xboxes + 1 pc kicking out 45C |
| 17:42.10 | ShALLaX | L\ |
| 17:42.11 | ShALLaX | :\ |
| 17:45.22 | treke_ | hehe |
| 17:45.37 | treke_ | I dont even want to know how warm the server room here is |
| 17:45.41 | treke_ | I was sweating like a pig |
| 17:51.41 | *** join/#oe jochen (~jochen@pD9E0106A.dip.t-dialin.net) |
| 18:00.00 | mickeyl | cvs experts online? can I tell cvs log to only consider events after a given date ? |
| 18:01.17 | gb2 | yes |
| 18:01.27 | mickeyl | cool. happen to know the syntax offhand? |
| 18:02.03 | mickeyl | nm, got the info page |
| 18:02.07 | gb2 | manpage suggests -ddates where dates is semicolon separated dates, or a range of d1<d2 or d2>d1 |
| 18:02.42 | mickeyl | cool, thanks |
| 18:07.31 | *** join/#oe kergoth (kergoth@zelda.tuxnami.org) |
| 18:08.48 | gb2 | afternoon kergoth |
| 18:08.56 | kergoth | hey |
| 18:10.20 | kergoth | damn monday. |
| 18:12.31 | gb2 | heh, you're the third person i've seen complain about monday within about 5 minutes |
| 18:13.47 | jochen | gb2: yes, this has been a really bad Monday... |
| 18:14.07 | prpplague | gb2: well you can make that 4 |
| 18:14.15 | gb2 | heh |
| 18:14.17 | prpplague | gb2: its a crappy monday here too |
| 18:14.30 | kergoth | eep, i have a code review in 45 minutes i forgot about |
| 18:14.31 | kergoth | hmm |
| 18:15.32 | prpplague | gb2: my cable internet was cut off this morning, apparent, they haven't been charging my credit card for the last 6 months |
| 18:16.03 | kergoth | ack |
| 18:17.43 | *** join/#oe pb_ (~pb@cpc2-cmbg4-3-0-cust80.cmbg.cable.ntl.com) |
| 18:17.58 | prpplague | kergoth: pissed my off, since they acted like it was my fault |
| 18:18.16 | prpplague | kergoth: had to get a supervisor on the phone to get it straight |
| 18:18.42 | pb_ | re |
| 18:21.20 | pb_ | mickeyl: Failure: <Fault 1: "exceptions.IOError:[Errno 13] Permission denied: '/tmp/tmpfile'"> |
| 18:22.41 | *** join/#oe Disconnect (~dis@pcp701542pcs.bowie01.md.comcast.net) |
| 18:23.22 | mickeyl | ah right |
| 18:23.23 | mickeyl | hold on |
| 18:24.17 | mickeyl | pb_: try again |
| 18:24.44 | pb_ | patch: **** Can't create file /home/pb_/tmp/ppPypd0Z : No such file or directory |
| 18:26.08 | *** join/#oe dirk_ (~chatzilla@do13.in-dsl.de) |
| 18:26.13 | CIA-5 | 03pb_ 07 * r1.1515.2.1 10packages/ (BitKeeper/etc/logging_ok glib-2.0/glib-2.0-native_2.4.1.oe): add missing binaries to staging\ |
| 18:26.13 | CIA-5 | 03pb_ 07 * r1.1519 10packages/: |
| 18:26.13 | CIA-5 | Merge bk://openembedded@openembedded.bkbits.net/packages\ |
| 18:26.13 | CIA-5 | into gandalf.tm.informatik.uni-frankfurt.de:/home/pb_/packages\ |
| 18:26.28 | pb_ | heh |
| 18:26.36 | mickeyl | i have to do some minor adjustments, then it will be going again |
| 18:26.55 | pb_ | mickeyl: cool, thanks |
| 18:27.11 | mickeyl | did some cleanup to make it use your name |
| 18:27.21 | mickeyl | so we can blame you more easily ;) |
| 18:27.27 | pb_ | hehe |
| 18:27.36 | kergoth | heh, BK_USER? |
| 18:30.54 | mickeyl | no |
| 18:30.55 | mickeyl | didn't know about that thing |
| 18:30.57 | mickeyl | just added a new user |
| 18:30.59 | mickeyl | it's better anyway since he has his own repository now :) |
| 18:31.02 | pb_ | kergoth: how did lrzsz get so broken? is that fallout from the latest round of autofoo? |
| 18:31.37 | *** join/#oe florian (~fuchs@pD9E5F6A7.dip.t-dialin.net) |
| 18:31.40 | florian | hi all |
| 18:31.44 | pb_ | hi florian |
| 18:32.03 | kergoth | pb_: yes, the autoreconf was actually failing horribly and we never knew :) |
| 18:32.11 | kergoth | it just *happened* that it built properly without regen |
| 18:32.14 | pb_ | doh |
| 18:32.31 | pb_ | now I'm getting "aclocal: configure.in: 260: macro `AM_GNU_GETTEXT' not found in library" |
| 18:32.35 | kergoth | yeah |
| 18:32.43 | florian | looks like libtool is broken again... |
| 18:32.44 | pb_ | is automake not installing its bits into aclocal or something? |
| 18:32.53 | kergoth | well, thats gettext. |
| 18:32.56 | kergoth | we dont have that |
| 18:32.59 | pb_ | ah |
| 18:33.04 | kergoth | well, we do, but in a crapload of individual m4 patches |
| 18:33.11 | pb_ | well, we do, it's in gettext/gettext-native |
| 18:33.13 | kergoth | i added an lrzsz m4 patch to add the missing macros |
| 18:33.15 | kergoth | ah |
| 18:33.32 | kergoth | does it work? :) |
| 18:33.37 | pb_ | heh. let me find out. |
| 18:33.47 | kergoth | lrzsz still fails due to @CATALOGS@ not being subst'd |
| 18:33.50 | kergoth | which is weird |
| 18:33.53 | kergoth | maybe i stole the wrong macros |
| 18:33.53 | pb_ | odd |
| 18:34.02 | kergoth | AC_PO_SUBDIRS (iirc) is supposed to do that subst |
| 18:34.05 | kergoth | or something like that |
| 18:34.09 | kergoth | probably AM, not ACX |
| 18:34.12 | pb_ | florian: what's the latest breakage? |
| 18:34.47 | pb_ | kergoth: did Michael fix up the list to accept your signed messages? |
| 18:35.02 | florian | pb_: libtool-native, configure: |
| 18:35.06 | florian | pb_: /home/fuchs/Projekte/gpe/openembedded/oetmp/staging/i686-linux/bin/autoreconf: unrecognized option `--exclude=libtoolize' |
| 18:35.14 | kergoth | florian: you need to wipe and rebuild autoconf-native |
| 18:35.17 | kergoth | assuming you're up to date |
| 18:35.47 | florian | kergoth: great - i just started this :-) |
| 18:36.06 | florian | kergoth: as up to date as cvs is... |
| 18:36.41 | *** join/#oe darkschneider (~gab@81-208-36-80.fastres.net) |
| 18:37.09 | kergoth | pb_: he will, need to let him know what the mime type in question is |
| 18:37.13 | mjc | woohoo |
| 18:37.15 | pb_ | righto |
| 18:37.47 | gb2 | does autoreconf take args in the form of --exclude=foo --exclude=bar or --exclude=foo,bar ? |
| 18:37.54 | kergoth | comma seperated |
| 18:38.04 | mjc | long time no oe... but now I'm back |
| 18:38.10 | gb2 | we're already passing in --exclude=autopoint aren't we? |
| 18:38.26 | kergoth | yes, if you override EXTRA_AUTORECONf you'd damn well better put it in yourself if you're a gettext build |
| 18:38.33 | gb2 | heh |
| 18:38.47 | pb_ | heh. I've never even heard of autopoint. |
| 18:38.49 | pb_ | what does that one do? |
| 18:38.50 | kergoth | gettext. |
| 18:38.53 | pb_ | ah |
| 18:39.09 | pb_ | oh yes. "Copies standard gettext infrastructure files into a source package." |
| 18:39.23 | kergoth | yeah, i got bit badly by it |
| 18:39.32 | pb_ | what happened? |
| 18:39.40 | kergoth | one of the packages called my desktop one, and it was a different incompatible version than the stuff used in the source tree |
| 18:39.43 | kergoth | bam. |
| 18:39.50 | pb_ | ouch |
| 18:40.06 | kergoth | gettext suffers from various version incompatibility issues, much like auto* and libtool |
| 18:40.11 | pb_ | sux0r |
| 18:40.14 | kergoth | can of worms i dont feel like opening today :) |
| 18:40.17 | kergoth | but one day.. |
| 18:40.20 | pb_ | right |
| 18:40.22 | florian | better, but now compile fails.... |
| 18:40.32 | kergoth | mjc: good stuff, that |
| 18:40.34 | pb_ | florian: oh dear. what's the new error? |
| 18:40.47 | mjc | kergoth: heck yes :) |
| 18:41.25 | kergoth | i really, really need to write some wrapper scripts around clearcase's tools |
| 18:41.27 | mjc | kergoth: I made a -Q9 ogg of their newest CD if you want it... it's about 130MB |
| 18:41.33 | florian | pb_: configure: error: source directory already configured |
| 18:41.37 | kergoth | mjc: ah, nice, yeah |
| 18:41.54 | pb_ | florian: odd |
| 18:41.54 | kergoth | florian: mithro saw that |
| 18:41.58 | florian | looks like oemake clean is not sufficient |
| 18:41.58 | kergoth | florian: then it just went away |
| 18:42.03 | kergoth | i could never reproduce it |
| 18:42.23 | pb_ | it usually happens if you have symlinks in the path to where you're building, I think |
| 18:42.41 | kergoth | ah |
| 18:42.44 | kergoth | odd |
| 18:43.51 | pb_ | configure does something bizarre like "if [ -f $src_dir/Makefile ] && [ $src_dir != `pwd` ]; then blow up; fi", which loses if `pwd` expands to something different to what it was expecting. |
| 18:44.13 | kergoth | ahh right. you'd think it could normalize those paths to make comparison slightly less bad |
| 18:44.25 | pb_ | you'd think that, but you'd apparently be mistaken |
| 18:44.29 | kergoth | naturally |
| 18:44.35 | pb_ | well, installing gettext-native has changed the error. |
| 18:44.37 | mjc | kergoth: check msg :p |
| 18:44.41 | pb_ | autoreconf: configure.in: AM_GNU_GETTEXT is used, but not AM_GNU_GETTEXT_VERSIONconfigure:312: error: do not use LIBOBJS directly, use AC_LIBOBJ (see section `AC_LIBOBJ vs LIBOBJS' |
| 18:44.47 | pb_ | seems that I still lose. |
| 18:45.02 | mjc | should just s/LIBOBJS/AC_LIBOBJS/ |
| 18:45.08 | mjc | does the trick for that I believe |
| 18:45.14 | kergoth | pb_: hm... i thought i fixed that |
| 18:45.18 | kergoth | no, its slightly more complex |
| 18:45.24 | kergoth | LIBOBJS="blah.o blah2.o" |
| 18:45.25 | kergoth | becomes |
| 18:45.29 | kergoth | AC_LIBOBJ([blah]) |
| 18:45.32 | kergoth | AC_LIBOBJ([blah2]) |
| 18:45.37 | mjc | kergoth: ah, hm |
| 18:45.37 | kergoth | but still not exactly tough |
| 18:45.51 | mjc | I was close ;) |
| 18:45.58 | pb_ | maybe my lrzsz build dir is outdated. |
| 18:46.05 | kergoth | i couldve sworn i pushed a fix for that |
| 18:46.15 | pb_ | trying again |
| 18:46.18 | florian | kergoth, pb_: configure checks if there is a file "config.status" and bails out if it exists. |
| 18:46.27 | *** join/#oe bipolar (~bflong@146.145.26.90) |
| 18:46.32 | pb_ | suck, same error |
| 18:46.47 | pb_ | florian: which package is this? |
| 18:47.02 | florian | pb_: libtool-native |
| 18:47.25 | pb_ | florian: we build that one in srcdir, so config.status is expected to exist. |
| 18:47.57 | pb_ | if test "`cd $srcdir && pwd`" != "`pwd`" |
| 18:47.59 | kergoth | pb_: hmm, the autotools patch for lrzsz fixed a libobj usage, but maybe i missed oen |
| 18:47.59 | pb_ | must be that line going wrong. |
| 18:48.36 | mjc | ) |
| 18:49.15 | florian | pb_: sure... this is after configure should have finished. why does this run again if oe tries to build the source? |
| 18:49.16 | pb_ | kergoth: ah, there seems to be a LIBOBJS lurking in acinclude.m4 |
| 18:49.38 | pb_ | florian: it should be harmless to run it again. you need to figure out why it's confused about the directories. |
| 18:50.32 | florian | pb_: yes, that's true... but... |
| 18:50.43 | pb_ | kergoth: AC_REPLACE_GNU_GETOPT would appear to be the guilty party. do you have any idea where that's from? |
| 18:51.30 | mjc | listens to Oomph - Nichts (Ist Kälter als Deine Liebe |
| 18:51.36 | mjc | blarg |
| 18:51.42 | kergoth | pb_: hmm, no idea |
| 18:51.50 | kergoth | pb_: grep in the m4.patch? :) |
| 18:51.54 | mjc | http://fiasco.node41.com/mjc/mp3/scene/Oomph_-_Wahrheit_oder_Pflicht_Limited_Edition-CD-2004-MOD/11_oomph_-_nichts_(ist_kaelter_als_deine_liebe)-mod.mp3 if you don't think the CD is worht buying :p |
| 18:52.08 | pb_ | mjc: I think we could manage with just a little less detail about what you're listening to |
| 18:52.15 | kergoth | hehe |
| 18:52.19 | mickeyl | heh |
| 18:52.20 | mjc | pb_: sorry, bad past ethe 2nd time |
| 18:52.36 | mjc | pb_: and as for the detail, well, sure : |
| 18:52.37 | kergoth | i should pick that cd up actually. and wumpscut's new one |
| 18:53.15 | mjc | kergoth: it's like $26 after shipping on cduniverse |
| 18:54.20 | pb_ | kergoth: doh, my copy of the .oe was out of date. |
| 18:54.52 | kergoth | pb_: dunno if you noticed, but we only call autoreconf once now |
| 18:54.57 | kergoth | rather than on each subdir manually |
| 18:55.01 | kergoth | should speed up some configure passes |
| 18:55.03 | pb_ | ah, cool |
| 18:55.05 | kergoth | hopefully it doesnt break anything |
| 18:55.20 | pb_ | heh, now it blows up in do_compile with the @CATALOGS@ thing. |
| 18:55.32 | pb_ | so I guess I'm at least at the same level as you :-) |
| 18:55.33 | kergoth | yeah, thats where i was. didnt have time to fix |
| 18:57.06 | pb_ | I wonder why that one var isn't getting substituted. |
| 18:57.54 | *** join/#oe gb2 (~gbritton@fog.sekrit.org) |
| 19:01.39 | mjc | kergoth: god I hate liking this CDso much :p |
| 19:03.48 | jochen | mjc: not bad :-) |
| 19:04.11 | *** join/#oe reenoo_ (~r@pD9E17A71.dip.t-dialin.net) |
| 19:04.33 | mjc | jochen: good compiling music ;) |
| 19:07.11 | mjc | ditto :p |
| 19:07.22 | reenoo_ | re |
| 19:10.26 | pb_ | hi reenoo_ |
| 19:13.15 | *** join/#oe Hrw (~szczepan@c53-11.icpnet.pl) |
| 19:13.17 | Hrw | hi |
| 19:14.15 | florian | Ok, got rid of libtool... now it is building glibc |
| 19:14.34 | Hrw | is lrzsz in compilable state? |
| 19:15.44 | Hrw | ok - I saw in log |
| 19:18.21 | treke_ | it is so monday. |
| 19:24.27 | reenoo_ | hi all |
| 19:25.17 | Hrw | hi reenoo |
| 19:26.22 | Hrw | reenoo_: can you get a look at last modutils changes which I made? |
| 19:27.58 | *** join/#oe florian_ (~fuchs@pD9E5FD32.dip.t-dialin.net) |
| 19:30.33 | reenoo_ | Hrw: where do I find your changes then? |
| 19:31.00 | Hrw | reenoo_: in bk tree |
| 19:31.22 | Hrw | ops.. I had to push them.. heh.. novice ini bk I'm |
| 19:31.47 | CIA-5 | 03szczepan 07 * r1.1520 10packages/ (7 files in 5 dirs): Pack of changes to get task-bootstrap works better. |
| 19:31.55 | Hrw | now better |
| 19:32.33 | *** join/#oe woglinde (woglinde@pD9E7DF04.dip.t-dialin.net) |
| 19:33.25 | Hrw | ;) |
| 19:33.58 | reenoo_ | Hrw: I'll consult the bk web interface then... |
| 19:34.22 | Twiun | howdy |
| 19:34.29 | Hrw | reenoo_: http://openembedded.bkbits.net:8080/packages/cset@1.1520?nav=index.html|ChangeSet@-1h |
| 19:35.01 | Hrw | now needs to look in pcmcia-cs why it needs psmisc |
| 19:35.01 | gb2 | that changeset breaks dropbear |
| 19:35.14 | gb2 | it adds references to patches |
| 19:35.14 | Hrw | gb2: I'll look |
| 19:35.17 | gb2 | but doesn't add the patches |
| 19:35.33 | Hrw | ok |
| 19:36.23 | Hrw | I'll push it in |
| 19:37.38 | CIA-5 | 03szczepan 07 * r1.1521 10packages/ (dropbear/dropbear-0.41/initd.patch ntp/ntp_4.1.2.oe): |
| 19:37.39 | CIA-5 | ntp_4.1.2.oe: |
| 19:37.39 | CIA-5 | <PROTECTED> |
| 19:37.39 | CIA-5 | initd.patch: |
| 19:37.39 | CIA-5 | <PROTECTED> |
| 19:37.58 | *** join/#oe bipolar (~bflong@146.145.26.90) |
| 19:38.05 | gb2 | there were 2 patches you added to dropbear weren't there? |
| 19:38.21 | Hrw | only I - initd.patch |
| 19:39.00 | gb2 | oh |
| 19:39.02 | gb2 | nevermind |
| 19:39.10 | gb2 | 2 +'s in the diff |
| 19:39.12 | gb2 | 1 - |
| 19:39.19 | gb2 | since you added the \ to the end of the line |
| 19:39.48 | reenoo_ | Hrw: the FILESDIR statements look OK now. why did you change SRC_URI? |
| 19:40.07 | Hrw | reenoo_: modutils-collateral? it doesn't want to build without it |
| 19:42.06 | Hrw | reenoo_: FILESDIR I get from modutils-cross.oe, /etc/modules(.conf) are installed by oe_machinstall() |
| 19:44.34 | Hrw | heh.. task-bootstrap doesn't need psmisc already. but pcmcia-cs needs it for fuser |
| 19:45.05 | gb2 | that modutils looks a little screwy |
| 19:45.08 | gb2 | hrm. |
| 19:46.23 | gb2 | you don't really need the SRC_URI you have there |
| 19:46.42 | gb2 | you don't need any really, but.. hrmmm |
| 19:47.14 | Hrw | hi Pendalar |
| 19:47.33 | gb2 | oh wow, oe_machinstall is such a hack, heh. |
| 19:48.07 | pb_ | if modutils-collateral is shipping files from FILESDIR, it should have them in SRC_URI. |
| 19:49.12 | Pendalar | mickey|dinner: I thought that the issue and issue.net should just simply say something like "Welcome to {DISTRO} {version}" blah rather than Debian ;-) |
| 19:49.23 | gb2 | pb: yeah. it's shipping files from N subdirs |
| 19:49.31 | Hrw | pb_: "file://${FILESDIR}/${MACHINE}/modules" will be ok? |
| 19:49.32 | gb2 | it has arch specific subdirs |
| 19:49.56 | pb_ | Hrw: yes, that would be ok |
| 19:50.08 | Hrw | ok |
| 19:50.27 | Hrw | dropbear by default disallow remote passwordless root login |
| 19:50.39 | pb_ | that sounds like correct behaviour |
| 19:50.51 | Pendalar | treke? |
| 19:51.03 | Hrw | yes - but we don't say it for user |
| 19:51.26 | gb2 | oe_machinstall will make it mean file://${FILESDIR}/..overrides../modules |
| 19:51.51 | gb2 | whichever the first override is, that's what gets shipped. |
| 19:52.00 | gb2 | first one that matches rather. |
| 19:52.10 | pb_ | yeah. if we ever have a DISTRO called collie, we'll be in all kinds of trouble. |
| 19:53.09 | treke_ | mach_collie distro_collie |
| 19:53.10 | treke_ | :) |
| 19:53.17 | pb_ | we possibly ought to try to clean up the namespaces a bit there. It's getting a little uncomfortable having to remember to keep DISTRO, PN and MACHINE apart. |
| 19:53.31 | treke_ | and pn could be itslef |
| 19:53.57 | gb2 | it would also suck if someone decided to ship a file called "image" or "install" and listed it in SRC_URI |
| 19:54.08 | pb_ | why would that lose? |
| 19:54.14 | gb2 | since it'd just get copied into ${WORKDIR} |
| 19:54.29 | pb_ | oh, whoops, yeah. I forgot about that. |
| 19:54.30 | gb2 | and oe creates ${WORKDIR}/image during install process |
| 19:54.39 | pb_ | we should probably put the source in ${WORKDIR}/source/... |
| 19:54.49 | gb2 | yeah. |
| 19:55.02 | gb2 | might even be worthwhile doing ${WORKDIR}/build/ |
| 19:55.02 | Hrw | <PROTECTED> |
| 19:55.04 | gb2 | too |
| 19:55.40 | pb_ | yeah, though that might get confusing. not all packages can build outside srcdir. |
| 19:56.02 | pb_ | I suppose we could make a symlink farm in those cases, but it seems too much. |
| 19:56.03 | Pendalar | treke_: not sure if it matters now but /dev and /proc were not empty |
| 19:56.19 | pb_ | Pendalar: in the rootfs, or at runtime? |
| 19:56.34 | Pendalar | hmm runtime |
| 19:56.37 | pb_ | /dev is overmounted with a ramdisk during bootup, so what you see on a running system is not what the kernel sees initially |
| 19:56.54 | Pendalar | ah true.. have to reflash then |
| 19:57.19 | Pendalar | since there's been plenty of patches I thought I'd ditch it and get a new bootstrap-image |
| 19:57.24 | treke_ | look in tmp/rootfs |
| 19:57.27 | treke_ | it should be there |
| 19:57.40 | Hrw | http://home.hrw.one.pl/oe/0510_modutils-collateral.oe will be ok? |
| 19:58.25 | gb2 | pb: perhaps just /build/ then and put everything into there |
| 19:58.38 | Hrw | Pendalar: you are trying to boot collie? |
| 19:58.46 | Pendalar | already did |
| 19:59.00 | Pendalar | used bootstrap-image a couple of days ago.. reported to the ML |
| 19:59.08 | Hrw | Pendalar: oh.. i readed |
| 20:00.14 | gb2 | pb: treat ${WORKDIR} itself as owned by OE, nothing should touch it directly. |
| 20:00.27 | gb2 | imo, ${FILESDIR} should also be that way then. |
| 20:01.17 | Hrw | 3rd reflash today |
| 20:01.43 | gb2 | pb: it might also be nice to eliminate it from the SRC_URI's as well |
| 20:02.18 | gb2 | file:// always be relative to ${FILESDIR} |
| 20:02.20 | Hrw | gb2, pb_: can you look at http://home.hrw.one.pl/oe/0510_modutils-collateral.oe diff? |
| 20:02.58 | gb2 | Hrw: that should work ok for now. |
| 20:03.15 | treke_ | any non filesdir paths are fundamentally broken |
| 20:03.17 | Hrw | so I;ll commit it |
| 20:03.33 | gb2 | treke: you were agreeing with my idea? |
| 20:03.34 | Hrw | but after 4th bootstrap boot on my Z |
| 20:03.37 | treke_ | gb2: Yes |
| 20:04.29 | Hrw | shit. dropbear keygenerating really should be rather moved to init.d from postinst ;( |
| 20:04.45 | treke_ | Hrw: why? |
| 20:05.46 | Hrw | treke_: I just booted fresh bootstrap-image and keys are not created without manually starting dropbear.postinst |
| 20:05.48 | gb2 | this is probably going to be an overhaul of do_unpack |
| 20:06.21 | treke_ | Hrw: That's a bug in postinst handling. |
| 20:06.29 | treke_ | THe solution is to properly run them |
| 20:07.25 | pb_ | gb2: yeah, agreed. |
| 20:08.09 | pb_ | gb2: what I'd actually like is for .oes never to mention FILESDIR at all. If they want to ship files, they should use the copies that do_fetch puts in ${WORKDIR}. |
| 20:08.27 | Hrw | hmm.. after adding file://${FILESDIR}/${MACHINE}/modutils likes into SRC_URI I can't build this package (using oemake or oebuild) |
| 20:08.50 | gb2 | what error? |
| 20:08.58 | gb2 | pb: yeah, agreed. |
| 20:09.05 | Hrw | NOTE: Unpacking /home/szczepan/zaurus/build/oe/packages/modutils/files/modules to /home/szczepan/zaurus/build/oe/tmp/base/modutils-collateral-1.0-r0/ |
| 20:09.09 | Hrw | cp: cannot stat `/home/szczepan/zaurus/build/oe/packages/modutils/files/modules': No such file or directory |
| 20:09.12 | Hrw | NOTE: Task failed: cp /home/szczepan/zaurus/build/oe/packages/modutils/files/modules /home/szczepan/zaurus/build/oe/tmp/base/modutils-collateral-1.0-r0/./ execution failed |
| 20:09.31 | Hrw | argh... bad patch - sorry |
| 20:09.40 | gb2 | yeah. |
| 20:09.43 | gb2 | bad patch :) |
| 20:10.40 | Hrw | now works |
| 20:12.30 | CIA-5 | 03szczepan 07 * r1.1522 10packages/modutils/modutils-collateral.oe: |
| 20:12.30 | CIA-5 | modutils-collateral.oe: |
| 20:12.30 | CIA-5 | <PROTECTED> |
| 20:15.41 | Hrw | Pendalar: I'm writing reply to your bootraport now |
| 20:15.44 | kergoth | pb_, gb2: i agree that oes shouldnt reference FILESDIR in their functions. however, i dont like the idea of making file:// relative to FILESDIR. |
| 20:15.58 | pb_ | kergoth: why's that? |
| 20:16.00 | gb2 | kergoth: where else would you get files from? |
| 20:16.05 | kergoth | pb_, gb2: its conceivable that I could have an nfs mounted SRC_URI mirror, accessed via MIRRORS. |
| 20:16.21 | pb_ | kergoth: we could allow file:/// to address an absolute local URI, as an escape hatch. |
| 20:16.22 | Pendalar | Hrw well at this point a lot of it has been fixed hopefully |
| 20:16.35 | gb2 | hrm. |
| 20:16.39 | Pendalar | according to other posts anyways |
| 20:16.52 | Hrw | Pendalar: "07 output needs formatting for the zaurus screen" - you mean that "ls -l" works like on 80 columns screen? |
| 20:16.55 | kergoth | pb_: ah, so relative file:// paths are relative to FILESDIR, but absolute is absolute, period. |
| 20:16.59 | kergoth | pb_: that i wouldnt be opposed to' |
| 20:17.04 | pb_ | right |
| 20:17.07 | gb2 | i like that |
| 20:17.18 | treke_ | that actually makes sense in some ways |
| 20:17.21 | gb2 | yeah |
| 20:17.29 | kergoth | so, first step, lets kill FILESDIR usage in favor of using the WORKDIR installed versions |
| 20:17.32 | pb_ | treke_: try not to sound so surprised |
| 20:17.36 | treke_ | gives the impression that filesdir is the cwd |
| 20:17.37 | kergoth | right? |
| 20:17.42 | Pendalar | Hrw that's what I was thinking |
| 20:17.44 | gb2 | kergoth: did you see my comments about workdir pollution? |
| 20:17.47 | Hrw | Pendalar: ok |
| 20:17.58 | treke_ | gb2: one small hitch |
| 20:18.08 | Pendalar | it'd had a ton of wasted space between fields, and doing |more would allow more than readable to pass by |
| 20:18.15 | kergoth | gb2: yeah, i agree that it should probably unpack to a specific area there |
| 20:18.16 | treke_ | gb2: we do on occaision override filesdir, probably need to make sure we can still do that |
| 20:18.17 | Pendalar | the # of lines to hold was off |
| 20:18.31 | gb2 | treke: i was thinking about that |
| 20:18.50 | kergoth | how we set FILESPATh and FILESDIR neednt change |
| 20:18.56 | gb2 | treke: we typically set it manually when we have parent and children .oe's |
| 20:18.58 | treke_ | yup |
| 20:19.02 | kergoth | just file:// behavior and where we -use- FILESDIR |
| 20:19.24 | pb_ | kergoth: agreed (about the first step you mentioned) |
| 20:19.26 | gb2 | and that overriding has never felt very clean |
| 20:19.38 | Hrw | Pendalar: ok |
| 20:19.39 | treke_ | kergoth: just want to point it out. Wouldnt want gb2 to make the change one way and forget that the value in filesdir isnt read only |
| 20:19.44 | kergoth | nod |
| 20:19.47 | gb2 | it's the sort of thing that ought to be hidden from the user (author of the .oe's) |
| 20:20.14 | gb2 | treke: i don't think any of this should change until all of it is thought out. |
| 20:20.22 | treke_ | gb2: yeah, it's always been a bit of a hack |
| 20:20.24 | kergoth | gb2: well, you could have base.oeclass use 'INHERITED_PN' to mangle it by re-which'ing FILESPATH.. |
| 20:20.36 | kergoth | e.g. INHERITED_PN=autotools in autotools-native |
| 20:20.38 | kergoth | or something |
| 20:20.55 | gb2 | hrm. |
| 20:20.56 | treke_ | gb2: the way we do it right now doesnt actually work properly iirc |
| 20:21.07 | Pendalar | Hrw: btw I get digest version and am leaving to go to work.. so don't expect a quick reply ;-) |
| 20:21.12 | gb2 | treke: we hack around it right now |
| 20:21.18 | kergoth | err actually, thatd be more annoying that it seems. we have to change PN -in- the expansion of FILESPATH |
| 20:21.24 | kergoth | once it gets epxanded, we're lost |
| 20:21.25 | Pendalar | bbl *sigh* |
| 20:21.58 | kergoth | so an anonymous function in base.oeclass that does a metadata deepcopy of FILESPATH, sets PN to an alternate value, then expands and sets |
| 20:23.17 | gb2 | kergoth: won't that endup getting set to whatever PN is at the point when base.oeclass is loaded (the top level) |
| 20:23.43 | gb2 | where what we want is the deepest .oe that's included |
| 20:23.50 | kergoth | yes, but i'm talking about setting a variable before you inherit base, to something like INHERIT_PN |
| 20:24.03 | kergoth | leaving the control in the hands of those doing the inclusion |
| 20:25.20 | gb2 | kergoth: as the original PN before any inheriting? |
| 20:25.29 | kergoth | ? |
| 20:25.29 | *** join/#oe dirk_ (~dirk@do13.in-dsl.de) |
| 20:25.33 | gb2 | i think you lost me |
| 20:25.33 | kergoth | inheriting doesnt change pn. |
| 20:25.40 | kergoth | my package name is autotools-native. |
| 20:25.50 | kergoth | i need to expand FILESPATh with a PN of autotools. |
| 20:25.57 | gb2 | yep |
| 20:26.07 | kergoth | so i set INHERITED_PN=autotools before i inherit anything |
| 20:26.13 | kergoth | and base.oeclass expands it with that |
| 20:26.54 | pb_ | couldn't we just have oe.conf define FILESPATH in terms of INHERITED_PN, and default the latter to the same as ${PN}? |
| 20:27.05 | kergoth | yeah, thatd do |
| 20:27.05 | *** join/#oe kolla (~kolla@smurf.home.uninett.no) |
| 20:27.09 | gb2 | how does this interoperate with cascaded inheritance |
| 20:27.17 | kergoth | its irrelevent. |
| 20:27.22 | pb_ | gb2: doesn't interact with it at all |
| 20:27.27 | kergoth | FILESDIR does a which on FILESPATH at *expansion time* |
| 20:27.36 | kergoth | FILESPATH expands based on PN at *expansion time* |
| 20:27.40 | kergoth | nothing is expanded during inheritence. |
| 20:27.43 | kergoth | unless they use := |
| 20:28.15 | gb2 | yeah, that part i get. |
| 20:28.47 | pb_ | which bit is it that you don't get? |
| 20:29.02 | gb2 | so, if we define FILESPATH=${INHERITED_PN}-based |
| 20:29.30 | gb2 | but if we inherit twice.. i guess we endup with the deepest |
| 20:30.08 | kergoth | hm? |
| 20:30.12 | kergoth | the .oe sets INHERITED_PN |
| 20:30.30 | gb2 | oh. |
| 20:30.33 | kergoth | not the oeclass |
| 20:30.36 | kergoth | i'm autotools-native |
| 20:30.38 | kergoth | i set it to autotools |
| 20:30.43 | kergoth | ta da, life is good |
| 20:30.51 | gb2 | i thought you were refering to something done automatically when inheriting from a parent .oe |
| 20:30.58 | gb2 | rather than explicitly |
| 20:31.15 | kergoth | 1) you dont inherit a .oe, you include it. there are differences, though minor ones. 2) no |
| 20:31.18 | kergoth | :) |
| 20:31.20 | gb2 | true. yeah. |
| 20:31.38 | gb2 | inherited is perhaps a poor name then. |
| 20:31.45 | kergoth | ah, indeed. |
| 20:32.01 | gb2 | PARENT_PN perhaps |
| 20:32.21 | kergoth | parent is confusing. |
| 20:32.27 | kergoth | i include him. why is he the parent? |
| 20:32.35 | gb2 | true. |
| 20:32.50 | kergoth | hehe. FILESPN |
| 20:33.10 | treke_ | if I inherit opie.oe, why not set the variable OPIE_FILESDIR |
| 20:33.28 | treke_ | if I want a file out of it I specify file:///${OPIE_FILES}/foo.patch |
| 20:33.30 | kergoth | because you arent setting his filesdir |
| 20:33.32 | kergoth | you're setting your own |
| 20:33.49 | kergoth | oh, i see, you want -him- to have a seperate var for his filesdir? |
| 20:33.51 | kergoth | thats bad. |
| 20:33.56 | gb2 | gcc, gcc-cross, gcc-cross-initial, all share the same filesdir |
| 20:34.56 | gb2 | FILES_PN isn't bad. |
| 20:35.36 | treke_ | I guess the problem becomes, the references in the inherited src_uri would point to the wrong place |
| 20:36.08 | gb2 | treke: that's the problem all this is attempting to solve, yeah. |
| 20:37.23 | gb2 | ofcourse. you could in gcc.oe just set FILES_PN=gcc and endup being set. |
| 20:37.46 | gb2 | rather than in gcc-cross and gcc-cross-initial |
| 20:38.48 | kergoth | course that means he cant include gcc, then reference his own FILESDIR files |
| 20:38.59 | kergoth | unless he sets FILE_PN himself for his items |
| 20:39.15 | gb2 | kergoth: but there is only one filesdir for all 3 .oe's |
| 20:39.49 | pb_ | true, but it's not inconceivable someone could make a new gcc-derived .oe and want his own FILESDIR |
| 20:39.50 | kergoth | right, its all or nothing |
| 20:40.02 | kergoth | its arguable |
| 20:40.05 | pb_ | still, the includer can always override FILES_PN after including, so it doesn't seem like a big deal |
| 20:40.14 | gb2 | pb: that's true |
| 20:40.24 | kergoth | pb_: if its file:///, that wont work |
| 20:40.31 | kergoth | because it isnt expanded at expansion time |
| 20:40.34 | kergoth | but at fetch time. |
| 20:40.45 | kergoth | err s/\/\/\//\/\// |
| 20:40.48 | kergoth | file:// |
| 20:41.00 | kergoth | no ${FILESDIR}, per hwat we were just discussing |
| 20:41.07 | kergoth | thatll make it impossible to have more than one filesdir |
| 20:41.18 | pb_ | right, yeah |
| 20:41.24 | mjc | kergoth: a lot of the opie stuff won't build atm |
| 20:41.28 | pb_ | I think that's a limitation that one would just have to accept. |
| 20:41.32 | gb2 | all or nothing sounds good. |
| 20:41.35 | mjc | dependencies |
| 20:41.46 | gb2 | you're either part of one pack of .oe's, or you're not. |
| 20:41.48 | gb2 | you shouldn't span |
| 20:42.12 | kergoth | mjc: i built meta-opie last night |
| 20:42.15 | mjc | hm |
| 20:42.38 | kergoth | trees up to date? |
| 20:42.57 | mjc | was as of this morning |
| 20:43.02 | kergoth | odd |
| 20:43.06 | mjc | I'm pulling again but there aren't any changes to opie* since |
| 20:43.12 | mjc | modutils, ntp, sysvinit... |
| 20:43.18 | kergoth | gb2: uicmoc3-native is somehow picking up linux-oe-g++ btw |
| 20:43.33 | Hrw | talking about opie... opie-irdaapplet needs change in DEPENDS/RDEPENDS... but libopenobex wasn;t building last tried so it breaks meta-opie |
| 20:43.35 | kergoth | gb2: and its using its own (qt sources) mkspecs, not ours, which doesnt have that |
| 20:43.56 | gb2 | kergoth: there's a default QMAKESPEC var |
| 20:44.15 | gb2 | that's probably why |
| 20:44.15 | kergoth | should we unset it for that package, or something? |
| 20:44.22 | gb2 | that's the quick fix |
| 20:46.40 | mjc | NOTE: no buildable providers for libopie2 |
| 20:46.40 | mjc | ERROR: dependency libopie2 (for meta-opie) not satisfied |
| 20:47.07 | gb2 | kergoth: it could probably also import our specs |
| 20:47.28 | gb2 | qte is another one that's kinda like this too. |
| 20:47.43 | gb2 | no.. importing our specs won't work unless you're using our qmake |
| 20:47.59 | gb2 | since we pass vars in specially |
| 20:49.40 | mickeyl | mjc: probably something failing earlier? read a few lines above... |
| 20:50.04 | mjc | believe that was the first error but I'll look |
| 20:50.04 | gb2 | importing our config, and using the EXTRA_OEMAKE from qmake.oeclass would work. i'm not sure what the cleanest solution to that would be. |
| 20:50.24 | mjc | ah doh, sorry.. no brain today :p |
| 20:50.42 | mjc | In file included from /home/mjc/packages/tmp/staging/arm-linux/include/opie2/oapplication.h:36, |
| 20:50.47 | mjc | <PROTECTED> |
| 20:50.48 | mjc | /home/mjc/packages/tmp/staging/arm-linux/include/qpe/qpeapplication.h: In static member function `static void QPEApplication::showWidget(QWidget*, bool)': |
| 20:52.14 | mickeyl | ... |
| 20:52.23 | mjc | yes, that's an odd one indeed |
| 20:52.29 | mickeyl | you ommited the error |
| 20:52.38 | mjc | dah, sorry |
| 20:52.49 | mjc | wasn't paying attention |
| 20:53.25 | mjc | /home/mjc/packages/tmp/staging/arm-linux/include/qpe/qpeapplication.h:218: error: `qt_maxWindowRect' undeclared (first use this function) |
| 20:53.35 | mjc | /home/mjc/packages/tmp/staging/arm-linux/include/qpe/qpeapplication.h: In static member function `static void QPEApplication::setCurrentRotation(int)': |
| 20:53.38 | mjc | /home/mjc/packages/tmp/staging/arm-linux/include/qpe/qpeapplication.h:261: error: 'class QWidget' has no member named 'qwsDisplay' |
| 20:53.46 | mickeyl | yahoo |
| 20:53.52 | gb2 | when did you pull? |
| 20:53.53 | mickeyl | artefact of recent qmake changes |
| 20:53.56 | mjc | two minutes ago |
| 20:53.56 | gb2 | yeah |
| 20:53.58 | gb2 | i fixed that |
| 20:54.06 | mjc | [mjc@dvdburner packages]$ bk pull |
| 20:54.06 | mjc | Pull bk://openembedded.bkbits.net/packages |
| 20:54.06 | mjc | <PROTECTED> |
| 20:54.06 | mjc | Nothing to pull. |
| 20:54.08 | gb2 | then mickeyl told me to do it slightly different. hrm. |
| 20:54.13 | kergoth | heh, that looks like a missing QWS, eh? |
| 20:54.26 | mickeyl | yea |
| 20:54.38 | mickeyl | let me see |
| 20:54.49 | gb2 | NOTE: patch with name 'prefix.patch' already applied |
| 20:54.54 | mickeyl | ok, no biggie |
| 20:55.07 | gb2 | mickeyl: how are you fixing it? |
| 20:55.35 | mickeyl | libopie2 inherits qmake, not palmtop ;) |
| 20:55.39 | mjc | heh. |
| 20:55.42 | mickeyl | the conclusion is up to you :) |
| 20:55.50 | gb2 | heh |
| 20:57.19 | Hrw | pb_: btw running "ipkg configure" in 1st boot - I reflashed, logged, started "ipkg configure" from shell and got nothing |
| 20:57.30 | kergoth | "got nothing"? |
| 20:57.55 | gb2 | kergoth: which libtool should have prefix.patch, the main one, or just native? |
| 20:58.01 | kergoth | all of them |
| 20:58.07 | gb2 | so just the main one |
| 20:58.10 | pb_ | Hrw: check the states in the status file in the generated rootfs |
| 20:58.15 | Hrw | pb_: ok |
| 20:58.22 | kergoth | the target one doesnt need it, technically, but, if libtool-native is built before it, aclocal grabs those .m4 macros, breaking one of its makefile rules |
| 20:58.31 | kergoth | so i apply it regardless, to ensure that at least the behavior is deterministic |
| 20:58.36 | Hrw | pb_: Status: install ok installed |
| 20:58.42 | pb_ | Hrw: that's incorrect. |
| 20:58.53 | gb2 | kergoth: including it in all of them breaks since the sub ones just append to the SRC_URI list |
| 20:58.55 | Hrw | ok |
| 20:58.58 | pb_ | if the package is not yet configured, it should say 'install ok unpacked'. |
| 20:59.06 | pb_ | it sounds like image_ipk.oeclass is malfunctioning. |
| 20:59.10 | CIA-5 | 03mickeyl 07 * r1.1523 10packages/: |
| 20:59.10 | CIA-5 | Merge bk://openembedded@openembedded.bkbits.net/packages\ |
| 20:59.10 | CIA-5 | into gandalf.tm.informatik.uni-frankfurt.de:/usr/local/projects/packages\ |
| 20:59.10 | CIA-5 | 03mickeyl 07 * r1.1518.1.1 10packages/libopie/libopie2_cvs.oe: |
| 20:59.11 | CIA-5 | define QWS since this is no longer assumed by qmake.oeclass,\ |
| 20:59.11 | CIA-5 | don't overwrite automatic detection of wireless extension version\ |
| 20:59.28 | mickeyl | mjc: bkpull and rebuild opie |
| 20:59.31 | mickeyl | s/opie/libopie2/ |
| 21:00.21 | kergoth | gb2: yep, remove it from native an cross |
| 21:00.25 | mjc | pulling. |
| 21:00.28 | gb2 | yeah, that's what i'm doing. |
| 21:00.38 | kergoth | good |
| 21:00.41 | kergoth | night Hrw |
| 21:00.41 | mjc | night Hrw |
| 21:00.46 | gb2 | i'm just being good and build-testing libtool-native before committing :) |
| 21:00.53 | kergoth | hehe |
| 21:01.00 | kergoth | mine built, oddly enough |
| 21:01.08 | kergoth | maybe my pathcer is being less bastard-like |
| 21:01.12 | kergoth | patcher |
| 21:01.19 | gb2 | kergoth: who knows.. you probably have a different version than i do |
| 21:02.42 | *** join/#oe badalex (~badalex@cpe-66-1-177-91.ut.sprintbbd.net) |
| 21:03.40 | CIA-5 | 03gbritton 07 * r1.1522.1.1 10packages/libtool/ (libtool-cross_1.5.oe libtool-native_1.5.oe): remove duplicate application of a libtool patch |
| 21:03.43 | mjc | mickeyl: no change |
| 21:04.00 | mickeyl | odd |
| 21:04.09 | mickeyl | can't test right now. i'm on different things |
| 21:04.46 | gb2 | i've started a test build |
| 21:04.56 | gb2 | it'll take 10min though to get that far |
| 21:05.02 | mickeyl | k, good |
| 21:06.32 | gb2 | mjc: did you oemake -c clean libopie2? |
| 21:06.42 | mjc | ah no good call |
| 21:06.59 | mickeyl | hey what did you think i meant with _re_built heh? |
| 21:07.05 | mjc | used to portage or rpm doing that for me |
| 21:07.06 | mickeyl | s/t/d/ |
| 21:07.50 | gb2 | might be good to throw an error if you try to re-build something without cleaning. |
| 21:08.04 | mjc | warning |
| 21:08.15 | gb2 | users ignore warnings |
| 21:08.22 | gb2 | error out. allow it with an option though |
| 21:08.29 | mjc | yeah, that would wokr |
| 21:08.30 | mjc | work |
| 21:08.33 | gb2 | still warn even with the option :) |
| 21:08.52 | mjc | mickeyl: yeah it's compiling now ;) |
| 21:09.04 | gb2 | the stamps don't entirely fit well into supporting that though. |
| 21:09.53 | ShALLaX | hmm |
| 21:09.58 | ShALLaX | after ive run oemake opie-image |
| 21:10.01 | ShALLaX | where is the image put |
| 21:10.02 | ShALLaX | ;p |
| 21:10.15 | kergoth | ls tmp/deploy/ |
| 21:10.27 | ShALLaX | top notch |
| 21:10.30 | ShALLaX | how about kernel |
| 21:10.44 | kergoth | isnt deployed outside of the kernel tarball or ipk at the moment |
| 21:10.46 | ShALLaX | or do i have to build that seperately |
| 21:10.51 | ShALLaX | ahh |
| 21:10.56 | kergoth | ls tmp/deploy/ipk/ |
| 21:11.11 | ShALLaX | so just unpack the ipk for the kernel? |
| 21:11.48 | mickeyl | might as well grab it from tmp/rootfs/boot/ |
| 21:12.01 | mjc | hey ShALLaX ;) |
| 21:12.07 | ShALLaX | heya |
| 21:12.12 | ShALLaX | do i know you? ;p |
| 21:12.19 | mjc | mjc@g.o |
| 21:12.28 | ShALLaX | ooh ;p |
| 21:12.40 | ShALLaX | i dont have a boot in the rootfs dir ;p |
| 21:12.48 | mickeyl | zaurus ? |
| 21:12.52 | ShALLaX | yep |
| 21:12.55 | mickeyl | i see |
| 21:13.02 | mickeyl | have to grab it out of the building dir then |
| 21:13.17 | ShALLaX | hehe, which is where? |
| 21:13.19 | mickeyl | $TMPDIR/base/openzaurus-bla/linux-bla/arch/arm/boot/zImage |
| 21:13.32 | mjc | opie-tabmanager died now |
| 21:13.32 | ShALLaX | 2.6.5?! |
| 21:13.42 | mickeyl | ? |
| 21:13.56 | mjc | same error as libopie2 |
| 21:14.18 | ShALLaX | 2.6.5 is ported? |
| 21:14.31 | gb2 | some of those errors will want to be fixed by inheriting opie or palmtop instead of qmake |
| 21:14.32 | mickeyl | started |
| 21:14.34 | mickeyl | not very much |
| 21:14.38 | mickeyl | feel free to help |
| 21:14.39 | ShALLaX | thats the only kernel i have in that dir.. |
| 21:14.40 | mickeyl | help is needed |
| 21:14.47 | ShALLaX | shallax@Silver openzaurus-2.6.5-jl3-r0 $ ls |
| 21:14.47 | ShALLaX | linux-2.6.5 patch-2.6.5-jl3.diff temp |
| 21:15.59 | gb2 | what MACHINE did you set in your local.conf ? |
| 21:16.09 | mjc | gb2: it's inheriting opie |
| 21:16.30 | gb2 | mjc: that's odd. |
| 21:16.36 | ShALLaX | collie |
| 21:16.37 | mjc | changed it to palmtop to see what it does |
| 21:16.49 | ShALLaX | MACHINE = "collie" |
| 21:17.01 | gb2 | opie inherits palmtop |
| 21:17.15 | gb2 | EXTRA_QMAKEVARS_POST_append = " DEFINES+=QWS" |
| 21:17.24 | gb2 | and palmtop does that |
| 21:17.33 | gb2 | mjc: when did you last pull in oe? |
| 21:17.40 | mjc | after mickey said to |
| 21:17.48 | mjc | pull packages, I pulled everything |
| 21:17.52 | mjc | damn enter key |
| 21:17.58 | gb2 | does your oe/classes/palmtop.oeclass have the line i pasted? |
| 21:18.06 | mjc | ..yeah. |
| 21:18.41 | gb2 | odd. |
| 21:18.48 | florian_ | ok, next one |
| 21:18.51 | gb2 | i'm almost done building a cross toolchain |
| 21:18.56 | gb2 | so i'll be able to test that soon |
| 21:19.13 | florian_ | modutils |
| 21:19.34 | florian_ | lex.l: In function `yylex': |
| 21:19.34 | florian_ | lex.l:429: error: `yytext_ptr' undeclared (first use in this function) |
| 21:19.40 | gb2 | eek |
| 21:19.47 | gb2 | lex stuff |
| 21:19.53 | florian_ | hehe |
| 21:19.54 | woglinde | florian moduitls I guess |
| 21:20.11 | florian_ | woglinde: yep |
| 21:20.24 | woglinde | florian I have a patch for this |
| 21:20.48 | florian_ | woglinde: is it in cvs? |
| 21:20.58 | woglinde | floarin nope |
| 21:21.06 | woglinde | floarin in an other cvs |
| 21:21.09 | woglinde | mom |
| 21:22.00 | kergoth | change flow is develop -> bk -> hh cvs |
| 21:22.02 | kergoth | florian_: debian? |
| 21:22.02 | woglinde | http://cvs.tuxbox.org/cgi-bin/viewcvs.cgi/*checkout*/tuxbox/cdk/Patches/modutils.diff?rev=HEAD&content-type=text/plain |
| 21:22.17 | woglinde | kergoth yes |
| 21:22.35 | kergoth | you can install bison-old or whatever |
| 21:22.38 | kergoth | pb would recall teh specifics |
| 21:23.10 | *** join/#oe T` (~total@RAMY8.RES.cmu.edu) |
| 21:23.36 | ShALLaX | hehe welp, my image doesnt seem to work |
| 21:25.00 | florian_ | kergoth: yes |
| 21:25.44 | kergoth | bleh, need food and caffinated beverages |
| 21:26.27 | gb2 | mjc: my build is building libqpe-opie now |
| 21:26.29 | gb2 | almost there |
| 21:26.33 | mjc | k |
| 21:26.36 | mickey|brb | got it btw. |
| 21:26.50 | mickeyl | i bet we use DEFINES= in EXTRA_QMAKEVARS_POST in tabmanager |
| 21:26.52 | mickeyl | which is wrong |
| 21:26.56 | florian_ | woglinde: ok, thanks - i'll try |
| 21:26.58 | mickeyl | should always use += and -= |
| 21:27.01 | gb2 | mickeyl: that was my first guess |
| 21:27.12 | gb2 | we don't |
| 21:28.18 | mickeyl | mjc: try making that EXTRA_QMAKEVARS_POST_append |
| 21:28.41 | gb2 | huh? |
| 21:28.46 | mickeyl | oh we don't use it at all in tabmanager |
| 21:28.47 | mjc | looks like it was mixing up old and new stuff |
| 21:28.59 | mjc | oemake -c clean of each opie app seems ot be doing the trick |
| 21:29.06 | mickeyl | ah i confused that with opie-login |
| 21:29.07 | gb2 | oh. |
| 21:29.16 | gb2 | mjc: did you get tabmanager to work? |
| 21:29.46 | mickeyl | don't expect dependencies to magically clean up staging... |
| 21:29.51 | mjc | yeah |
| 21:29.52 | gb2 | mickeyl: i think some of the apps depend on libopie2 but don't have a dependency listed. |
| 21:30.02 | mjc | oemake -c clean opie-tabmanager ; oemake opie-tabmanager |
| 21:30.16 | mickeyl | gb2: which ones? |
| 21:30.28 | gb2 | mickeyl: i don't have a list right now. i'll try to generate one though |
| 21:30.34 | mickeyl | gb2: good. not all opie-apps need libopie2 |
| 21:31.00 | gb2 | mickeyl: i noticed some opie applets tried to build during oemake -k when i had libopie2 broken |
| 21:31.32 | mickeyl | gb2: which is the expected behaviour for things not using libopie2. |
| 21:31.46 | gb2 | mickeyl: they failed because they tried to include opie2/ headers |
| 21:32.03 | mickeyl | ah headers... |
| 21:32.10 | mickeyl | that's right. i forgot about headers |
| 21:32.26 | mickeyl | we use a templates in every applet |
| 21:32.35 | mickeyl | so that of course needs libopie2 headers |
| 21:32.50 | mickeyl | might as well depend on libopie2 in opie.oeclass then |
| 21:33.09 | gb2 | it seems reasonable |
| 21:33.28 | gb2 | i can generate a list though if you'd like |
| 21:33.35 | gb2 | i figured out an easy way to generate one. |
| 21:33.40 | mickeyl | how? |
| 21:33.54 | gb2 | nuke libopie2, OEMASK it out, try to build opie-image |
| 21:33.58 | gb2 | note failures |
| 21:34.04 | mickeyl | ah heh cool idea |
| 21:34.38 | gb2 | running. |
| 21:34.43 | mickeyl | well if i add that depend to libopie2 i have to go through everything opie anyway |
| 21:34.47 | mickeyl | to remove it from their DEPENDS |
| 21:34.55 | mickeyl | s/libopie2/opie.oeclass/ |
| 21:34.58 | gb2 | that's easy though |
| 21:35.08 | mickeyl | bit of shell magic? |
| 21:35.10 | gb2 | grep |
| 21:35.11 | gb2 | :) |
| 21:35.22 | mickeyl | well... but the line has to be removed if empty |
| 21:35.43 | mickeyl | or at least should |
| 21:35.57 | gb2 | i tend to: bk edit `find magic...`; vim `find magic...` |
| 21:35.59 | gb2 | go through all teh buffers |
| 21:36.08 | gb2 | then bk sfiles -c ... |
| 21:36.14 | mickeyl | well... that's over 70 files... |
| 21:36.18 | gb2 | yeah |
| 21:36.27 | mickeyl | i'll see if i can do that with regexp |
| 21:36.31 | gb2 | heh |
| 21:39.59 | gb2 | mickeyl: i've gotten very good at repetative editing in vim: /libopie<enter> then either "dd" or "dw" depending on whether it has other deps or is its own. :w :n, repeat |
| 21:43.14 | mickeyl | heh |
| 21:43.18 | mickeyl | sounds like a task for you? |
| 21:43.20 | mickeyl | will you? |
| 21:44.07 | gb2 | heh |
| 21:44.14 | gb2 | not now |
| 21:44.34 | gb2 | i'm actually doing work work right now |
| 21:44.43 | mickeyl | sure |
| 21:46.11 | proti | mickeyl: What is the task exactly. Sounds some task for perl. |
| 21:46.36 | mickeyl | for each line in *opie*.oe |
| 21:46.41 | mickeyl | look at DEPENDS= |
| 21:46.48 | mickeyl | remove libopie* from that line |
| 21:46.52 | mickeyl | erase the line of empty |
| 21:46.56 | mickeyl | end loop |
| 21:47.20 | mickeyl | libopie* can be either libopie2 or libopiecore2, libopieui2, etc. |
| 21:47.36 | proti | all of them ? No exeption ? |
| 21:47.47 | mickeyl | yeah |
| 21:47.52 | mickeyl | all in *opie*.oe |
| 21:48.09 | mickeyl | well |
| 21:48.12 | mickeyl | rather opie-*.oe |
| 21:48.22 | mickeyl | we don't want to do that for qte-for-opie etc. :) |
| 21:48.37 | proti | Sounds doable. I try. Should be one or two liner. Let me look. |
| 21:48.48 | gb2 | perl pie.. mmmm |
| 21:48.49 | mickeyl | cool |
| 21:52.11 | jochen | less-381-r0: checking for working terminal libraries... Cannot find terminal libraries - configure failed |
| 21:52.11 | jochen | FATAL: oe_runconf failed |
| 21:58.41 | kergoth`food | hm, oemake should have an option to go back to cleaning an entire dependency chain |
| 21:58.54 | kergoth`food | for the case where you arent sure what dep is hosing a package, and want to just wipe the chain and rebuild |
| 22:01.17 | mickeyl | that'd be handy |
| 22:01.52 | gb2 | it's 58 files that need edits |
| 22:02.04 | kergoth`food | definately sounds like a job for perl |
| 22:02.17 | gb2 | could eliminate depends on virtual/libc too |
| 22:02.27 | mickeyl | yea |
| 22:02.33 | kergoth`food | and patcher-native, a few of those are still floating around, mostly in -native's |
| 22:03.42 | gb2 | patcher/patcher.oe:DEPENDS_prepend_delete = "patcher " |
| 22:03.43 | gb2 | heh |
| 22:05.05 | kergoth`food | heh, guess it was trying to prevent a recursion. thats an olddd remnant |
| 22:05.09 | gb2 | yeah |
| 22:05.17 | gb2 | back before patcher-native |
| 22:05.28 | gb2 | those recursions used to hurt a lot |
| 22:05.50 | gb2 | mickeyl: |
| 22:05.51 | gb2 | opie-clockapplet-1.1.4-cvs-20040509-r0 |
| 22:05.51 | gb2 | opie-irdaapplet-1.1.4-cvs-20040509-r0 |
| 22:05.51 | gb2 | opie-libqrsync-1.1.4-cvs-20040509-r0 |
| 22:05.51 | gb2 | opie-screenshotapplet-1.1.4-cvs-20040509-r0 |
| 22:05.51 | gb2 | opie-volumeapplet-1.1.4-cvs-20040509-r0 |
| 22:05.52 | gb2 | opie-vtapplet-1.1.4-cvs-20040509-r0 |
| 22:06.06 | gb2 | that's the list of attempted but failed builds with libopie2 masked |
| 22:06.36 | gb2 | make[2]: *** No rule to make target `@CATALOGS@', needed by `all-yes'. Stop. |
| 22:11.46 | proti | mickeyl: Ok here you go |
| 22:11.55 | proti | #!/usr/bin/perl -n -i~ |
| 22:11.57 | proti | if (/DEPENDS/) { s/(libopie.*?)([\s"'])/\2/g; } |
| 22:11.57 | proti | next if /DEPENDS\s*=\s*["']\s*['"]/oi; |
| 22:11.57 | proti | print $_; |
| 22:12.05 | proti | In a file. |
| 22:12.11 | proti | chmod a+x of file.pl |
| 22:12.21 | proti | ./file.pl opie-*/*.oe |
| 22:12.30 | proti | It makes a backup too. |
| 22:12.37 | proti | -i~ |
| 22:12.45 | mickeyl | cool. thanks. |
| 22:13.02 | kergoth`food | gb2: heh, dunno w hats up there. the PO_SUBDIRS macro is supposed to be doing that replacement |
| 22:13.34 | proti | mickeyl: Don't go if several libopie. |
| 22:13.52 | mickeyl | ah. will call it a number of times then |
| 22:14.04 | proti | rather if (/DEPENDS/) { s/(libopie.*?)//g; } |
| 22:14.20 | proti | No won't work if libopie not at the end. |
| 22:14.30 | kergoth`food | eh? that .*? will greedy grab libopie till the eol |
| 22:14.31 | kergoth`food | bad |
| 22:14.47 | proti | Yes my mistake |
| 22:14.49 | kergoth`food | use something other than ., or maatch the beginning and end of the word |
| 22:14.55 | kergoth`food | blah, cant type |
| 22:15.12 | proti | Don't know if one depends is a word or not. |
| 22:15.39 | kergoth`food | perl's definition of a word is a block of non-whitespace characters |
| 22:15.44 | kergoth`food | which applies here |
| 22:16.52 | proti | if (/DEPENDS/) { s/(libopie\w*?)//g; } |
| 22:16.52 | proti | It takes the closing ' or " too. |
| 22:17.44 | proti | mickeyl: I think the first version is ok. |
| 22:17.48 | kergoth`food | hehe |
| 22:18.27 | proti | I stay with 'if (/DEPENDS/) { s/(libopie.*?)([\s"'])/\2/g; } |
| 22:21.11 | proti | I checked : DEPENDS="a b libopieui c d libopiecore libopiecheck e" gives me DEPENDS="a b c d e" |
| 22:21.25 | pb_ | gb2: yah, that lrzsz thing is weird. |
| 22:21.45 | pb_ | gb2: the sed command in the generated configure looks right, but it doesn't seem to substitute CATALOGS for some odd reason. |
| 22:21.57 | mickeyl | kergoth`food: what was that magic alias to see all differences in modified files? |
| 22:22.01 | kergoth`food | is it actually calling that macro? |
| 22:22.07 | kergoth`food | mickeyl: there are two ways |
| 22:22.15 | kergoth`food | one is dog slow, thats bk -r diffs |
| 22:22.19 | kergoth`food | which is in the bk docs |
| 22:22.23 | kergoth`food | (*cough*idiocy*coughs*) |
| 22:22.27 | mickeyl | heh |
| 22:22.29 | kergoth`food | the better one is bk sfiles -c | bk diffs -u |
| 22:22.35 | mickeyl | ah right |
| 22:22.36 | mickeyl | thanks |
| 22:22.37 | kergoth`food | np |
| 22:22.38 | kergoth`food | er |
| 22:22.41 | kergoth`food | bk diffs -u - |
| 22:22.44 | kergoth`food | heh |
| 22:22.57 | pb_ | florian_: that lex error usually means your flex is too new. on debian, need to install flex-old. |
| 22:23.15 | woglinde | pb or my postet patch |
| 22:23.37 | florian_ | woglinde: this didn't help |
| 22:23.52 | woglinde | florian? |
| 22:24.09 | florian_ | woglinde: don't ask me why... i just tried |
| 22:24.12 | woglinde | cannt believe that |
| 22:24.49 | mickeyl | g'night |
| 22:25.08 | proti | mickeyl: did the script worked ? |
| 22:25.11 | pb_ | woglinde: if your patch fixes the bug, please mail it to the list |
| 22:25.11 | mickey|zzZZzz | yeah |
| 22:25.13 | pb_ | mickey|zzZZzz: good night |
| 22:25.37 | woglinde | florian which version says your flex? |
| 22:25.51 | proti | mickey|zzZZzz: Goodnight. |
| 22:25.52 | woglinde | pb now I am not sure it works |
| 22:26.02 | proti | Bedtime for me too. |
| 22:26.07 | gb2 | oh right.. i can't build opie-image without a MACHINE= because of that modutils change. hrm. |
| 22:26.28 | pb_ | gb2: I don't think it makes any sense to build any -image without a MACHINE= |
| 22:26.40 | pb_ | you should be able to build meta-opie, if you just want to compile the apps. |
| 22:26.44 | kergoth`food | agreed. meta packages sure, but images..? |
| 22:28.30 | gb2 | yeah |
| 22:28.50 | gb2 | that was more me mumbling to myself, not complaining. |
| 22:28.58 | kergoth`food | also, can someone explain why the gpe and opie meta packages are in their dirs, not meta? |
| 22:29.24 | kergoth`food | its convenient being able to 'ls meta/' to see what your options are |
| 22:29.35 | pb_ | just historical, I think. |
| 22:29.46 | kergoth`food | i'll move em over then, if people dont mind |
| 22:29.49 | pb_ | certainly gpe/meta-gpe.oe pre-dates the meta/ directory, and I think opie does too. |
| 22:29.52 | pb_ | sure, go for it |
| 22:29.57 | kergoth`food | k |
| 22:30.02 | pb_ | it's a bit wacky having gpe-image in meta/ and meta-gpe in gpe/, anyway |
| 22:32.33 | gb2 | meta-opie also won't do anything much for me since it still walks TOPDIR |
| 22:32.38 | gb2 | which isn't my build dir |
| 22:32.42 | pb_ | ah |
| 22:33.05 | kergoth`food | gb2: open a bug. |
| 22:33.11 | pb_ | oh, opie-image doesn't use meta-opie? that's kind of funky. |
| 22:33.38 | pb_ | I think the current meta-opie should be renamed to meta-opie-everything, and the pertinent parts of opie-image pulled out into a new meta-opie.oe |
| 22:35.31 | pb_ | It'd also be cool to fix up meta-opie{-everything} to do what we talked about before: poke around in the metadata for packages with SECTION=opie, rather than walking TOPDIR. |
| 22:37.46 | gb2 | uh |
| 22:37.49 | gb2 | hrm. |
| 22:38.03 | gb2 | lrzsz doesn't have AC_SUBST(CATALOGS) in its configure.in |
| 22:38.52 | gb2 | ah |
| 22:38.54 | pb_ | I think the gettext macros handle that. |
| 22:38.56 | gb2 | AM_GNU_GETTEXT is what it does i guess. |
| 22:39.01 | pb_ | right |
| 22:39.03 | gb2 | but. po/ isn't an automake file |
| 22:39.15 | gb2 | it's Makefile.in.in not Makefile.am |
| 22:39.43 | pb_ | yeah, there's special handling for that. |
| 22:40.08 | gb2 | but the @CATALOGS@ and friends only get expanded for automake Makefile.am's |
| 22:40.08 | pb_ | in any case, @...@ substitutions are an autoconf issue, not automake. |
| 22:40.16 | pb_ | hm. |
| 22:40.40 | gb2 | reading through the configure scripts, that's what it looks like |
| 22:40.51 | pb_ | find where it says: |
| 22:40.55 | pb_ | <PROTECTED> |
| 22:40.56 | pb_ | <PROTECTED> |
| 22:40.59 | pb_ | [...] |
| 22:41.07 | pb_ | sed -e "/^POTFILES =/r $ac_dir/POTFILES" -e "/^# Makevars/r $ac_given_srcdir/$ac_dir/Makevars" -e "s|@POFILES@|$POFILES|g" -e "s|@UPDATEPOFILES@|$UPDATEPOFILES|g" -e "s|@DUMMYPOFILES@|$DUMMYPOFILES|g" -e "s|@GMOFILES@|$GMOFILES|g" -e "s|@CATALOGS@|$CATALOGS|g" -e "s|@POMAKEFILEDEPS@|$POMAKEFILEDEPS|g" "$ac_dir/Makefile.in" > "$ac_dir/Makefile" |
| 22:41.15 | pb_ | I think that's the code that's meant to be taking care of this. |
| 22:41.18 | gb2 | oh. rhm. |
| 22:43.34 | gb2 | pb: it looks like the sed script only ran the first -e |
| 22:43.52 | *** join/#oe badalex (~badalex@cpe-66-1-177-91.ut.sprintbbd.net) |
| 22:43.56 | kergoth`food | heh, wonder why that line does just seperate teh s commands with ; in a single -e |
| 22:43.58 | kergoth`food | er doesnt |
| 22:43.59 | pb_ | gb2: that's what it looks like to me also, but I can't see why that would be the case |
| 22:44.19 | kergoth | bleh |
| 22:44.31 | pb_ | kergoth`food: yah. amusingly enough, that seems to be what older versions did (or at least, the one that ships in the lrzsz tarball) |
| 22:44.32 | florian_ | good night |
| 22:44.32 | gb2 | pb: look at configure.in |
| 22:44.37 | gb2 | pb: look at the bottom of it. |
| 22:44.46 | pb_ | florian_: night |
| 22:44.58 | pb_ | !! |
| 22:45.04 | pb_ | gb2: good grief |
| 22:45.08 | gb2 | pb: that's what threw me off |
| 22:45.13 | gb2 | thinking automake didn't do it |
| 22:45.18 | pb_ | heh |
| 22:45.23 | gb2 | we go and clobber automake's work :) |
| 22:45.25 | pb_ | well, that certainly solves the mystery |
| 22:45.50 | gb2 | pb: my guess is some debian person broke it |
| 22:45.51 | gb2 | ;) |
| 22:46.03 | pb_ | doh |
| 22:46.05 | pb_ | ~lart debian |
| 22:46.10 | gb2 | based purely on the other lines there |
| 22:46.19 | pb_ | yeah |
| 22:46.51 | pb_ | oh well. wanna make a patch to remove that bogus stuff? |
| 22:47.13 | gb2 | sure. |
| 22:48.30 | pb_ | cool |
| 22:49.11 | pb_ | heh |
| 22:50.50 | woglinde | night |
| 22:52.33 | kergoth | wtf were they thinking |
| 22:52.44 | gb2 | pb: that fixed it |
| 22:52.48 | pb_ | gb2: rock |
| 22:52.51 | kergoth | nice |
| 22:52.53 | pb_ | kergoth: the mind boggles |
| 22:52.57 | kergoth | indeed |
| 22:53.28 | pb_ | clearly their brains have been fried by years of slow serial transfers, or something |
| 22:53.36 | kergoth | hehe |
| 22:54.22 | pb_ | though I guess we ought to package kermit if we want to talk about slow transfers. |
| 22:54.57 | kergoth | actually, that wouldnt be a bad idea. use a pda to download kernels to u-boot on arm boards |
| 22:55.00 | kergoth | hehe |
| 22:55.10 | pb_ | heh, right |
| 22:55.52 | kergoth | heh, we have c-kermit in oe now |
| 22:56.15 | CIA-5 | 03gbritton 07 * r1.1525 10packages/lrzsz/ (lrzsz-0.12.20/po-fix.patch lrzsz_0.12.20.oe): add patch to lrzsz to fix debianizing breakage of po |
| 22:56.23 | pb_ | oh, heh, so we do. |
| 22:56.33 | pb_ | I wonder what's different in gkermit compared to regular ckermit |
| 22:56.46 | pb_ | ah, apt knows. |
| 22:56.49 | kergoth | ckermit's buildsystem is a big pile of crap, i remember that much |
| 22:56.56 | pb_ | "The non-free package ckermit adds connection |
| 22:56.56 | pb_ | <PROTECTED> |
| 22:57.01 | kergoth | ah. |
| 22:57.09 | kergoth | non-free, wonder what license its under |
| 22:57.44 | pb_ | good question. let's find out. |
| 22:58.02 | kergoth | and set the LICENSE field in the oe while you're at it :) |
| 22:58.11 | pb_ | of course |
| 22:59.05 | pb_ | "This is the new C-Kermit 7.0 and 8.0 license." |
| 22:59.05 | pb_ | fear |
| 23:00.10 | pb_ | ah, I guess this is why it's non-free. |
| 23:00.14 | pb_ | "The |
| 23:00.14 | pb_ | <PROTECTED> |
| 23:00.14 | pb_ | <PROTECTED> |
| 23:00.14 | pb_ | <PROTECTED> |
| 23:00.28 | kergoth | eep |
| 23:01.26 | pb_ | oh, and it also says you can only distribute it in a distro that has an OSI-approved license. I'm pretty sure that would fail the DFSG test as well. |
| 23:02.49 | pb_ | I'm not sure what to put in the LICENSE field for this. |
| 23:05.49 | kergoth | icky |
| 23:05.50 | kergoth | heh |
| 23:05.56 | kergoth | dunno |
| 23:06.13 | pb_ | I suppose we could copy COPYING.TXT to some website, then include the url |
| 23:06.43 | pb_ | Putting the whole text of the licence in LICENSE seems unreasonable. |
| 23:07.00 | mjc | qt doesn't like me |
| 23:07.18 | kergoth | nod. could just say 'c-kermit 7.0/8.0' or something, and force people who really want to know the terms of that license to dig it up themselves |
| 23:07.29 | pb_ | yah, that'd work |
| 23:07.31 | kergoth | or throw out ckermit in favor of gkermit in avoidance |
| 23:07.34 | kergoth | hehe |
| 23:07.42 | pb_ | heh, that might be even better |
| 23:07.48 | pb_ | let me see what the gkermit build system is like |
| 23:09.17 | mjc | <PROTECTED> |
| 23:09.17 | mjc | <PROTECTED> |
| 23:09.17 | mjc | <PROTECTED> |
| 23:11.38 | pb_ | ah, gkermit is rather more basic. |
| 23:12.02 | pb_ | doesn't have an interactive mode, just command-line options for file transfer. |
| 23:15.39 | kergoth | ah |
| 23:15.58 | kergoth | mjc: yep. i reported that to gb2 a bit earlier, scroll up and you can see the workaround |
| 23:21.38 | mjc | I don't see it |
| 23:22.24 | kergoth | 15:43 < gb2> kergoth: there's a default QMAKESPEC var |
| 23:22.24 | kergoth | 15:43 < gb2> that's probably why |
| 23:22.24 | kergoth | 15:43 < kergoth> should we unset it for that package, or something? |
| 23:22.24 | kergoth | 15:43 < gb2> that's the quick fix |
| 23:22.29 | kergoth | try that |
| 23:22.30 | mjc | thanks |
| 23:22.34 | kergoth | havent tested it |
| 23:22.35 | kergoth | np |
| 23:24.06 | mjc | testing now |
| 23:25.33 | pb_ | gb2: regarding that DEPENDS_prepend_delete thing you mentioned earlier, maybe we should look into trying to wean .oeclasses off of _prepend/_append. |
| 23:26.41 | kergoth | heh |
| 23:26.45 | pb_ | it's mostly an issue of ordering, but I don't know offhand how many packages would need to change. |
| 23:26.46 | kergoth | theres gonna be a lot of brokenness from that |
| 23:26.47 | kergoth | yeah |
| 23:26.50 | kergoth | i agree though |
| 23:27.23 | pb_ | kergoth: when do base.oeclass and INHERIT get read, in relation to the .oe itself? |
| 23:27.31 | kergoth | pb_: i'm thinking, do something evil like store the expanded metadata from all the existing .oe's, so we know what we -want- |
| 23:27.36 | kergoth | then make some of the class changes |
| 23:27.40 | kergoth | and compare to find brokenness |
| 23:27.44 | pb_ | yeah, good plan |
| 23:28.00 | pb_ | we could probably make some h4x0red version of mickey's parse cache code for that. |
| 23:28.00 | mjc | kergoth: in the .oe? |
| 23:28.11 | mjc | caues in my shell seems not to have done anything |
| 23:28.42 | kergoth | try.. QMAKESPEC[export] = 0 |
| 23:28.44 | kergoth | in the .oe |
| 23:28.52 | kergoth | that should prevent it from emiting ours into the shell code |
| 23:32.09 | mjc | kergoth: er, no change, and yes I cleaned :P |
| 23:36.35 | treke_ | so has it been monday for everyone else? |
| 23:36.39 | kergoth | mjc: look at the run. script by the log |
| 23:36.41 | kergoth | mjc: read it |
| 23:36.47 | kergoth | mjc: see if QMAKESPEC is in there at all |
| 23:36.59 | kergoth | treke_: oh hell yes. more so than usual even |
| 23:40.03 | mjc | yeah, it is |
| 23:40.07 | mjc | export QMAKESPEC="linux-oe-g++" |
| 23:40.22 | treke_ | kergoth: same |
| 23:40.31 | treke_ | only good news is nortel figured out the switch problem |
| 23:40.36 | treke_ | we need to upgrade the OS :) |
| 23:41.34 | kergoth | mjc: oh, QMAKESPEC[export] = |
| 23:41.37 | kergoth | mjc: try that |
| 23:41.58 | *** join/#oe Disconnect (~dis@3ffe:bc0:91:0:203:6dff:fe1a:c5b6) |
| 23:42.15 | *** join/#oe mjc (mjc@sartre.ispvip.biz) |
| 23:42.29 | mjc | yeah, it is |
| 23:42.34 | mjc | export QMAKESPEC="linux-oe-g++" |
| 23:42.45 | kergoth | odd |
| 23:42.54 | mjc | think I got it |
| 23:43.02 | mjc | I put unset QMAKESPEC in do_configure |
| 23:43.07 | mjc | kinda hackist |
| 23:43.07 | kergoth | hehe, thatll do |
| 23:43.09 | mjc | hackish |
| 23:43.12 | kergoth | nod |
| 23:43.18 | kergoth | oe needs an 'unexport' command or something |
| 23:43.20 | kergoth | :) |
| 23:43.25 | mjc | heh |
| 23:43.37 | chouimat | kergoth: unexport? |
| 23:43.49 | kergoth | chouimat: yes, prevent a variable from being emitted into the shell code |
| 23:44.10 | chouimat | kergoth: ok something like local scope |
| 23:44.23 | kergoth | yep |
| 23:44.46 | kergoth | its that way by default, but on occasion we want to undo the actions of an oeclass who already exported it |
| 23:45.05 | chouimat | kergoth: VART=blah; do something; unset VART :) |
| 23:46.37 | kergoth | heh, yeah, mjc just used unset in the shell code to work around it. ugly though. |
| 23:47.55 | *** join/#oe SubNuke (~subnuke@nh-londndry-c41-bg1a-32-219.lndnnh.adelphia.net) |