**** BEGIN LOGGING AT Sun Mar 12 10:59:56 2006 Mar 12 11:16:00 great Mar 12 11:16:12 hh.org mail is broken *again* Mar 12 11:24:53 jnc: The patch is no longer needed for zd1211, so including it breaks wpa for zd1211 ;) Mar 12 11:41:40 morning Mar 12 11:42:13 hi all Mar 12 11:43:56 hey reenoo & Mardy Mar 12 11:49:01 hi reenoo, Mardy, koen & everyone else Mar 12 11:49:14 hey pH5 Mar 12 11:50:02 morning Mar 12 11:50:21 hey NAiL Mar 12 11:50:25 hey NAiL Mar 12 11:54:25 hi Mardy, koen, pH5, NAiL Mar 12 12:05:27 reenoo: could arange that the familiar people add an account for familiar-dev or familiar-orphans to OE bugzilla? Mar 12 12:06:27 anybody planning to build x today? Mar 12 12:06:29 http://www.handhelds.org/hypermail/oe/current/6970.html Mar 12 12:07:07 I was planning to build glibc today :) Mar 12 12:07:15 koen: great, go for it! Mar 12 12:07:18 * koen waits for the mail from bero Mar 12 12:11:12 pH5: haven't got around to respond to that yet. would you mind adding a .conf to prefer the old stuff? Mar 12 12:12:20 reenoo: like an example local.conf? Mar 12 12:12:49 koen: not sure. I'm not entirely convinced of the concept of adding bugzilla accounts for mailing-lists (password reminders etc.)... Mar 12 12:13:26 pH5: more like preferred-versions-x-pre-xorg-7.conf or something Mar 12 12:13:49 reenoo: ok, that's a good idea. Mar 12 12:23:57 hi ph5 Mar 12 12:23:58 hi reenoo Mar 12 12:24:03 hey pb_ Mar 12 12:24:19 hey pb_ Mar 12 12:24:21 reenoo: or some other account where people can assign bugs for familiar-stuff Mar 12 13:14:59 hey Mar 12 13:15:42 hey zecke Mar 12 13:16:19 koen: I did NOT manage to brick my NSLU2 Mar 12 13:16:31 heh, did you try to? Mar 12 13:17:27 koen: I'll talk to france, maybe we can hook procmail into the forwarding and filter out unwanted stuff like password reminders Mar 12 13:17:34 hey zecke Mar 12 13:17:54 NAiL: actually not, but now I have no reason to open the case Mar 12 13:18:23 hmm, removing R84 is a good reason? ;) Mar 12 13:18:47 NAiL: are you more in favor of unsoldering it, or 'braking' it? Mar 12 13:19:35 NAiL: why was it underclocked? Did they only pay for the lower clocked CPUs but Intel could not deliver them? Mar 12 13:19:50 Actually, I've unsoldered it on all my three slugs Mar 12 13:21:05 I don't know. Maybe they were concerned about the heat. I haven't run into any problems with heat except when I stacked two deunderclocked slugs on top of eachother Mar 12 13:25:41 03koen 07org.oe.dev * r5478d200... 10/packages/linux/ (ep93xx-kernel/defconfig ep93xx-kernel_2.6.15+2.6.16-rc5.bb): Mar 12 13:25:44 ep93xx-kernel: update .bb and defconfig to match derevo16 Mar 12 13:25:45 * MaverickCrunch is enabled now Mar 12 13:27:08 okay cya later Mar 12 13:27:11 I'm going to the gym now Mar 12 13:27:33 enjoy Mar 12 13:27:59 NAiL: do you know how I can run my SLUG in Little Endian mode? Mar 12 13:28:38 Yes Mar 12 13:28:54 I'm running LE Debian on the slug right next to me just now Mar 12 13:29:07 okay I will get back to you then Mar 12 13:29:10 cya Mar 12 13:52:02 grrrrrr why qemu don't build ;( ... Mar 12 13:52:07 | ERROR: "gcc" looks like gcc 4.x Mar 12 13:52:07 | QEMU is known to have problems when compiled with gcc 4.x Mar 12 13:52:07 | It is recommended that you use gcc 3.x to build QEMU Mar 12 13:52:07 | To use this compiler anyway, configure with --disable-gcc-check Mar 12 13:52:07 | FATAL: oe_runconf failed Mar 12 13:53:27 gremlin[it]: I would have hoped the error message was self explanatory Mar 12 13:54:20 pb_ yes true ... i mean 1) why today i need quemu, 2) why it try to build with a wrong gcc :) Mar 12 13:55:54 it's trying to build with your native gcc, which I guess is gcc 4. Mar 12 13:56:01 qemu is needed by glibc nowadays. Mar 12 13:56:16 you can avoid that by turning off binary locale generation in local.conf. Mar 12 13:57:27 I have apt-get install'ed qemu and set ASSUME_PROVIDED = "qemu-native" in local.conf. Mar 12 13:57:50 ah, good thinking Mar 12 13:58:40 pb_ i have fedora core 4 ..with gcc 4.0.2 .... Mar 12 13:59:38 mhh how turn off locale generation ? Mar 12 14:02:27 BINARY_LOCALE_ARCHS = "" or some such Mar 12 14:03:17 ok thanks pb_ Mar 12 14:03:27 ENABLE_BINARY_LOCALE_GENERATION = "0" Mar 12 14:03:50 ah, whoops Mar 12 14:03:54 thanks reenoo :) Mar 12 14:04:50 (or, BINARY_LOCALE_ARCHES = "" - note the E in ARCHES) Mar 12 14:05:35 we need both so we can default to disabling that stuff and still keeping track of what's supposed to be the content of BINARY_LOCALE_ARCHES Mar 12 14:05:50 s/keeping/keep/ Mar 12 14:25:31 re Mar 12 14:26:02 looks like I have another day with broken OE workin copy ;( Mar 12 14:26:08 monotone: misuse: 1512 missing files Mar 12 14:26:33 price for 389 fresh changesets? Mar 12 14:31:10 monotone revert --missing Mar 12 14:31:21 tried Mar 12 14:32:44 heh.. monotone: misuse: rename target 'packages/dropbear/dropbear/configure.patch' already exists Mar 12 14:32:47 etc Mar 12 14:33:27 if I remove that file and do mt update then 1512 missing again Mar 12 14:34:07 does monotone revert . work? Mar 12 14:34:16 dpes not report bugs Mar 12 14:34:34 but I have all 'missing' files on disk Mar 12 14:34:58 weird Mar 12 14:35:02 * koen looks at njs Mar 12 14:35:18 looks like today is a day when ALL scm tools are not friendly to me Mar 12 14:36:57 http://pastebin.com/597856 Mar 12 14:37:37 Can I drop the jan2006prebreakage line now? Everything back to normal? Mar 12 14:39:34 Laibsch: cant tell - my recent .dev build worked but not tested it on device Mar 12 14:41:06 OK, I'll just give it a shot. If things break; I'll restart. Robert P.J. Day mentioned about problems going from prebreakage revision to the current branch. Mar 12 14:41:58 R. Dj is good at finding problems Mar 12 14:43:02 sigh... looks like the Zaurus line is stagnating Mar 12 14:43:18 luke-jr_: why? Mar 12 14:43:22 luke-jr_: I rather won't buy zaurus Mar 12 14:43:38 hrw: what would you recommend? Mar 12 14:44:25 Laibsch: there's no real improvements since the C700 Mar 12 14:44:34 Laibsch: dont know yet - I would like to get portrait pda to use it as pda+stuff not for stuff+maybe-pda Mar 12 14:44:42 Oh, I see. you mean hardware. Mar 12 14:45:42 Laibsch: they use same screen since c700 in all clamshells - sl6000 got better one Mar 12 14:45:44 What else would he mean? don't we all use something OE-based? =p Mar 12 14:46:39 I thought your line about stagnating Z line was referring to OE. Mar 12 14:47:06 hrw: what is a "portrait pda"? Mar 12 14:47:28 guys... Mar 12 14:47:35 this still isn't #zaurus Mar 12 14:47:55 Laibsch: collie/poodle/hx4700/h2210/h3600/h3800/h3900/h5400 that kind Mar 12 14:48:38 OK, now I understood. Mar 12 14:50:45 have to go Mar 12 14:50:47 have to go Mar 12 14:52:54 reenoo: it's a developer discussion channel ;) Mar 12 14:57:32 hi Mar 12 14:57:56 hey Bernardo Mar 12 14:58:10 hey koen Mar 12 15:15:20 hi, anyone aware about what happened to bitbake that it does not accept the require keyword any more? Can't build dropbear without changing require to include Mar 12 15:38:59 re Mar 12 15:39:14 zecke: wb Mar 12 15:39:36 zecke: bitbake r394 says that oour gcc .bb are shite Mar 12 15:40:17 zecke: http://dominion.kabel.utwente.nl/koen/pda/files/bb-error.log Mar 12 15:41:06 koen: doesn't look like my ERROR output Mar 12 15:41:51 koen:copy your local.conf into my homedir on ewi Mar 12 15:42:19 koen: this smells like someone changed bitbake.conf Mar 12 15:42:34 koen: and didn't test what happens when the variable is not set Mar 12 15:43:09 zecke: local.conf is in /tmp now Mar 12 15:44:03 eFfeM: 'require' has been added to bitbake since 6 months or so. i guess you need a more recent versio Mar 12 15:44:08 hi guys Mar 12 15:44:40 eFfeM: you can even get a released version that includes require Mar 12 15:45:12 zecke, mickey|: require is broken in a bitbake revision between 380 and 394 Mar 12 15:45:17 not sure which one Mar 12 15:45:22 hi mickey| :) Mar 12 15:45:47 koen: the released version should have a working require :} Mar 12 15:48:15 * koen tries to build a softvfp image Mar 12 15:49:22 mickey|: did you win yesterday? Mar 12 15:49:47 hi, koen. what the softvfp exactly is ? Mar 12 15:50:29 leoncamel: a way to describe floatingpoint emulation on ARM Mar 12 15:50:30 leoncamel: rounding of floating point values is done by fuzzy logic -> soft :} Mar 12 15:50:36 morning Mar 12 15:50:37 mickey|: I'm using BitBake Build Tool Core version 1.3.2.1, bitbake version 1.3.2 and require used to work a week ago, Mar 12 15:50:39 hey chouimat Mar 12 15:51:00 eFfeM: does your error look similiar to koen's? Mar 12 15:51:02 koen: yeah, at first the other team missed one man and then the other 5 guys were pretty bad as well. we won 9:0 after 1h. Mar 12 15:52:32 koen, hmm, the softvfp is the way which gcc implemented ? or glibc implemented ? or NWFPE related ? Mar 12 15:52:39 gcc Mar 12 15:52:50 zecke: I don't get as much errors as koen, the only error I get is for dropbear: Mar 12 15:52:51 ERROR: /home/frans/OpenSlugHead/openslug/openembedded/packages/dropbear/dropbear_0.47.bb:1: unparsed line: 'dropbear.inc' while parsing /home/frans/OpenSlugHead/openslug/openembedded/packages/dropbear/dropbear_0.47.bb Mar 12 15:53:16 actually I get this four times for each bb in packages/dropbear once Mar 12 15:53:30 changing require in include was a suitable workaround Mar 12 15:53:43 koen, hmm, the glibc implemented called soft-float, right ? Mar 12 15:53:55 you're the 3rd or 4th person to encounter the 'require' problem this week Mar 12 15:54:10 eFfeM: ah someone forgot to push the dropbear.inc ;) Mar 12 15:54:33 leoncamel: it's a combination of gcc + glibc doing the softfloat stuff Mar 12 15:54:33 zecke: no, it's there Mar 12 15:54:35 koen, I mean -msoft-float.. is it ? Mar 12 15:54:57 eFfeM: try a monotone pull Mar 12 15:54:57 and softfloat can be softfpa or softvfp at the moment Mar 12 15:54:57 darn Mar 12 15:55:09 zecke, ah, ok, i don't have the right one so I cannot fix this Mar 12 15:55:24 hey mickeyl Mar 12 15:55:34 hi reenoo. how are things? Mar 12 15:55:42 koen, what is the difference between softvfp and softfpa ? Mar 12 15:56:32 mickeyl: can I talk you into having a look at OE bug #406 and hh.org bug #1378? Mar 12 15:56:43 mickeyl: busy as usual. you? Mar 12 15:56:55 leoncamel: ordering of the floating pointformat. But the most important thing is that softvfp will be used in the future and that softfpa is dead Mar 12 15:57:02 reenoo: yeah, it's pretty tough atm. + the next 8 weeks Mar 12 15:57:27 * reenoo nods Mar 12 15:57:45 koen: your local.conf does not set a MACHINE? Mar 12 15:57:49 406 is no problem, I'll take care about that next week Mar 12 15:57:56 zecke: correct, I have that in auto.conf Mar 12 15:58:13 koen, okey. thanks a lot .. :) Mar 12 15:58:25 koen: hehe, what else do you have in different files? Mar 12 15:58:48 zecke: on MACHINE="ep93xx" in auto.conf Mar 12 15:59:11 mickeyl: it seems I missed your efl demo at fosdem Mar 12 15:59:29 mickeyl: ok, great, that'd be one bug less to worry about. any idea about lincecache.py? the rest are missing dependencies I should be able to fix myself. Mar 12 16:00:04 koen: *nod* once I can breath a bit, I'll upload all the stuff to angstrom svn Mar 12 16:00:22 linecache.py even Mar 12 16:00:26 ~lart typos Mar 12 16:00:26 * ibot puts on a hockey mask and jumps out at typos Mar 12 16:00:33 mickeyl: angstrom? svn? Mar 12 16:01:48 koen, while cosscompiling with softvfp, what is the difference between gcc and glibc working ? I mean how dose the gcc and glibc working together ? Mar 12 16:02:08 reenoo: yes, linecache probably needs to go into python-io. i will do that as well next week Mar 12 16:02:32 mickeyl: excellent, thanks Mar 12 16:02:55 leoncamel: with softvfp gcc + glibc handle all floatingpoint exceptions instead of passing them to the kernel Mar 12 16:03:23 koen: hmm my zaurus build parsed just fine Mar 12 16:04:22 koen, ok, thanks. more clearly now .. :) Mar 12 16:04:47 * mickeyl admires the new python homepage Mar 12 16:07:56 koen, hmm. does the linux-2.6 mainline support hard-soft now ? Mar 12 16:08:07 hard-soft? Mar 12 16:08:31 koen, s/hard-soft/hard-float .. :) Mar 12 16:09:46 it has had hardfloat support since 2.6.0 :) Mar 12 16:10:07 koen, how about linux-2.4 series ? Mar 12 16:10:51 iirc all RMK patches have nwfpe Mar 12 16:11:26 koen, hmm, you mean the hardfloat supported in 2.6.0 is NWFPE ? Mar 12 16:11:45 nwfpe or fastfpe Mar 12 16:12:22 koen, I mean, dose it support a real hard-float excution instead of emulation in kernel space ? Mar 12 16:13:28 do you have an arm chip with a real FPU? Mar 12 16:13:30 koen, the NWFPE handles the unknow instruction exception and emulate in kernel space, then return to user space ? right ? Mar 12 16:13:54 that's how nwfpe works, yes Mar 12 16:14:29 koen, hmm, I don't know if there are real FPU for arm . and is there ? Mar 12 16:15:08 there are ARM chips with a FPU coprocessor, but not very much Mar 12 16:15:25 the cirrus ep93xx and omap 3430 have one Mar 12 16:15:51 koen, hmm, OK. I understand now .. :) great thanks .. so great thanks .. :) Mar 12 16:16:37 koen, I am confused for many months before.. :). Mar 12 16:17:46 koen, So. the softvfp is the best choice now. right ? Mar 12 16:18:52 yes, but softfpa isn't bad either Mar 12 16:20:05 koen, do you use IPP library from intel ? Mar 12 16:20:24 koen: no parsing errors here Mar 12 16:20:30 zecke|food: weird Mar 12 16:20:47 leoncamel: no Mar 12 16:21:29 koen, hmm, ok. Mar 12 16:22:05 koen, It seems there is some bug in iwmmxt toolchain which I build with bitbake .. Mar 12 16:23:22 leoncamel: I suspect tune-iwmmxt is missing '--with-float=softfp' in targetcc-arch Mar 12 16:23:49 testing iwmmxt is on my todolist, but it's at the bottom Mar 12 16:24:36 koen, hmm, I thinks I could test it in two days .. Mar 12 16:25:20 that would be cool Mar 12 16:25:25 koen, hmm, But I really set TARGET_FPU=soft in my machine config .. Mar 12 16:25:54 yes, iwmmxt doesn't work with hard(fpa)-float Mar 12 16:26:40 koen, hmm. could you provided the patch against it, so I could test it if I am free . Mar 12 16:28:56 yummy softvfp Mar 12 16:30:36 koen: hmm still parses here Mar 12 16:30:37 emte: i talked to the gcc devs to find out about that Mar 12 16:37:35 leoncamel: try http://dominion.kabel.utwente.nl/koen/pda/files/iwmmxt-softvfp.diff Mar 12 16:38:06 koen: hey, are you by chance building anything cross/cross-initial with gcc 4.1 yet? Mar 12 16:38:42 koen, OK. I will test it if I am free. :) Mar 12 16:38:56 jnc: I built gcc-cross-i 4.1, but glibc 2.3.5 didn't like the --disable-threads in gcc-cross-i .bb Mar 12 16:39:20 okay yeah, i think that massive system of bb includes needs some reworking Mar 12 16:39:36 there should be no exported variables, only config options (according to the gcc devs) Mar 12 16:39:54 good night. I will go to bed now .. :) Mar 12 16:40:09 'night leoncamel Mar 12 16:40:47 | ../sysdeps/unix/sysv/linux/bind.S:5: Error: symbol `__bind' is already defined Mar 12 16:40:55 and kdrive broke again :( Mar 12 16:41:17 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060312172902.log Mar 12 16:42:07 * jnc reads Mar 12 16:42:37 zecke: that's my fault Mar 12 16:42:40 * pH5 looks Mar 12 16:52:06 zecke: I don't understand where this comes from. There should be no references to _XlcNCompareISOLatin1 left in libX11.so. Could you show me the diet-x11 configure and build logs? Mar 12 16:53:09 silly me, I found it.. Mar 12 16:59:18 *sigh* Mar 12 16:59:20 You tried to change the assigned_to field from koen@handhelds.org to mickeyl@handhelds.org, but only the owner or submitter of the bug, or a sufficiently empowered user, may change that field. Mar 12 16:59:29 ~curse missing rights Mar 12 16:59:31 May you be reincarnated as a Windows XP administrator, missing rights ! Mar 12 17:01:30 pH5: http://ewi546.ewi.utwente.nl/OE/config.log Mar 12 17:01:40 pH5: tell me when you fetched it, so I can remove it Mar 12 17:01:49 pH5: ah okay Mar 12 17:01:55 koen: feel free to reassign 1378 to me Mar 12 17:02:08 koen: using the mickeyl@hh.org address please Mar 12 17:07:19 mickeyl: I assigned that bug to you. Mar 12 17:08:29 thanks Mar 12 17:09:53 | Xcomposite.c:226: error: 'XCOMPOSITE_VERSION' undeclared (first use in this function) Mar 12 17:09:56 errr? Mar 12 17:10:15 building gpe-image target, package libxcomposite Mar 12 17:13:54 i guess i should start over :( Mar 12 17:14:09 this is why i wanted an SDK seperated out Mar 12 17:14:44 jnc: sounds like an issue for ph5 Mar 12 17:15:03 jnc: try crosstools, crossdev maybe the suit you more? Mar 12 17:15:36 my trouble is that when i build something, then do monotone update, and a new version is out Mar 12 17:15:39 it conflicts Mar 12 17:15:47 so i have to rm-fr the whole oetmp thing Mar 12 17:15:50 and start over? Mar 12 17:16:16 jnc: get the bitbake manual, read on PREFERRED_VERSION... Mar 12 17:16:27 jnc: you can lock down any package to a version you want Mar 12 17:16:32 but i don't want to Mar 12 17:16:33 jnc: so the answer is, don't use the latest Mar 12 17:16:41 i want this to automagically work. Mar 12 17:16:53 using the latest code should not be so difficult Mar 12 17:17:15 my goal is to find problems, bugs, and fix them + provide for upstream Mar 12 17:17:27 so we end up using known versions of upstream packages for everything possible Mar 12 17:17:50 side effect is i get new software to play with on my Z Mar 12 17:18:22 jnc: sorry but whatever Mar 12 17:18:36 jnc: PREFERRED_VERSION is the answer... Mar 12 17:18:43 zecke: can you help me to find out which object file sneaked the _XlcNCompareISOLatin1 symbol into libX11.so? Mar 12 17:18:56 jnc: new version out, try it, try it again, update PREFERRED_VERSION Mar 12 17:19:00 pH5: sure Mar 12 17:19:03 "sorry but whatever" is not acceptable Mar 12 17:19:36 jnc: whatever, measuring your noise you are on the edge to my ignore list Mar 12 17:19:58 so ignore me, your responses to my concerns are not worth reading Mar 12 17:20:15 hehe Mar 12 17:20:29 i'm not saying this to be an attention whore Mar 12 17:20:40 i have concerns, and if i keep quiet about them then it is a loss for everyone Mar 12 17:20:57 unlikely Mar 12 17:22:42 pH5: Font.o Mar 12 17:23:46 jnc: the SDK is a worthwhile goal and we're working on that. opie is a problem though, since we have to write code for installing all the header files into the -dev pages. there's no install target yet that does that automatically :/ Mar 12 17:23:55 s/pages/packages/ Mar 12 17:23:56 hmm :/ Mar 12 17:24:17 qmake is a bit to blame here Mar 12 17:24:18 W/ Opie, i guess the Qt stuff is a huge part of the SDK eh? Mar 12 17:24:38 yeah, qt should be in there. i didn't bitbake meta-sdk for a while though Mar 12 17:24:55 it probably makes sense to divide meta-sdk into two or three parts anyway Mar 12 17:25:01 something to consider... Mar 12 17:25:12 mickeyl: I added a meta-toolchain yesterday Mar 12 17:25:55 would bitbake be capable of wrapping qmake built packages... i.e. i remember the gentoo way to do that was you'd write a class that handled the specific sort of build system, and it would build in a sandbox. if you had the appropriate class included, it would magically work 75% of the time Mar 12 17:26:43 i was thinking previously along the lines of just building packages for that target, without a (Opie/GPE/E/...) GUI Mar 12 17:26:50 the problem is not the qmake buildsystem per se, it's the .pro files for older software which don't include the necessary magic to tell qmake which files to install etc. Mar 12 17:27:08 the basic tools needed to begin cross compiling, then you can add the GUI capabilities to build on top of it maybe? Mar 12 17:27:12 oh Mar 12 17:27:16 * jnc reads Mar 12 17:27:56 ~seen [lala] Mar 12 17:27:58 [lala] was last seen on IRC in channel #oe, 5d 19h 52m 58s ago, saying: ';-)'. Mar 12 17:28:12 bbl, dinner time Mar 12 17:34:51 pH5: and yes Font.c calls _XlcNCompare Mar 12 17:35:11 pH5: in _XF86LoadQueryLocaleFont Mar 12 17:35:15 * pH5 totally confused, I can reproduce this Mar 12 17:35:43 the fact that I successfully built xserver-kdrive shows that I must have done something to that file before... Mar 12 17:35:45 pH5: i haven't gotten there yet in my build, what targets are you building? Mar 12 17:35:46 ~lart pH5 Mar 12 17:35:46 * ibot makes a balloon animal out of pH5 Mar 12 17:37:38 pH5: don't worry, we failed to test earlier Mar 12 17:37:43 what should I do about this? include the _XlcNCompareISOLatin1 function in the diet build or just use strncasecmp instead? Mar 12 17:38:13 pH5: ask pb_ I guess Mar 12 17:39:32 just use strncasecmp(), I think Mar 12 17:39:53 http://en.pastebin.ca/45376 Mar 12 17:39:59 I guess the glibc version of that function is locale aware anyway Mar 12 17:40:00 I'll try and add this patch Mar 12 17:40:47 righto Mar 12 17:41:07 drat, I have flat mouse batteries again Mar 12 17:41:21 * pb_ throws his mouse out of the window in disgust Mar 12 17:41:26 mine's been dying under heavy cpu load. way to go linux usb subsystem Mar 12 17:41:57 yeah, usb on linux is all a bit sucky. Mar 12 17:41:59 kind of a shame really Mar 12 17:42:38 pb_: it is close to monday anyway Mar 12 17:42:50 zecke: too close for comfort Mar 12 17:43:33 CTRL+ALT+F1 use screen, and create a text mode for GPE Mar 12 17:43:49 good plan Mar 12 17:43:53 * pb_ implements an aalib backend for gtk Mar 12 17:44:23 heh. there's a curses one... Mar 12 17:44:45 actually, maybe an easier plan is just to find a corded mouse Mar 12 17:51:41 pH5: should I search for other _Xlc occurences? Mar 12 17:52:13 pH5: _XlcCurrentLC is fine? Mar 12 17:52:34 zecke: _XlcCurrentLC has an empty stub Mar 12 17:53:03 okay Mar 12 17:53:34 zecke: I'm just building xserver-kdrive, let's see what happens this time... Mar 12 17:53:54 pH5: okay then I will try to motivate me to hack on FLUTE again Mar 12 17:54:02 * zecke gets some chocolate Mar 12 17:55:21 hmmm chocolate Mar 12 17:57:34 NOTE: package xserver-kdrive-0.0+cvs20060312: completed Mar 12 17:57:38 phew. Mar 12 17:59:10 thanks, zecke Mar 12 17:59:28 pH5: maybe that should go upstream? Mar 12 18:00:51 pH5: you are welcome. Mar 12 18:01:04 let us see where gpe-image/e-image breaks this time Mar 12 18:03:18 03pH5 07org.oe.dev * r27909d45... 10/packages/xproto/ (33 files): X11R7.0: X protocol headers Mar 12 18:03:23 03koen 07org.oe.dev * r7a68b2a3... 10/conf/distro/angstrom-2006.9.conf: angstrom 2006.9: correct some toolchain bits Mar 12 18:03:29 03pH5 07org.oe.dev * r402cc114... 10/packages/xlibs/ (34 files in 2 dirs): X11R7.0: X libraries Mar 12 18:03:36 03rw 07org.oe.dev * r21d01652... 10/packages/dropbear/ (4 files): dropbear: use "include", not "require". the file is there after all and there's no point in breaking backwards compatibility. Mar 12 18:03:37 jnc: not sure if they like adding #ifdefs upstream, perhaps adding another stub for _XlcNCompareISOLatin1 that just calls strncasecmp would be cleaner Mar 12 18:03:42 03pH5 07org.oe.dev * r09e1b4a1... 10/packages/ (6 files in 6 dirs): X11R7.0: X apps Mar 12 18:03:49 03pH5 07org.oe.dev * r629608ee... 10/conf/distro/preferred-x-versions-pre-X11R7.inc: add an include file with preferred versions for pre-X11R7.0 Mar 12 18:03:55 03pH5 07org.oe.dev * r53684d5e... 10/packages/ (3 files in 3 dirs): X11R7.0: xset and xrdb Mar 12 18:04:02 03pH5 07org.oe.dev * rb6bfaa32... 10/packages/xserver/ (2 files in 2 dirs): xserver-kdrive: add a (commented out) patch to make 20050207 build against X11R7.0 Mar 12 18:04:13 03pH5 07org.oe.dev * r30ed7914... 10/packages/xserver/ (xserver-kdrive_20060312.bb xserver-kdrive_cvs.bb): xserver-kdrive: make cvs build against X11R7.0 and add 20060312 snapshot Mar 12 18:04:21 03pH5 07org.oe.dev * r101a6a83... 10/packages/xcalibrate/xcalibrate_20060312.bb: xcalibrate: snapshot 20060312 to build against pre-X11R7.0 xlibs Mar 12 18:04:29 03pH5 07org.oe.dev * r344ea81d... 10/packages/xcalibrate/ (xcalibrate/xextproto.patch xcalibrate_cvs.bb): xcalibrate: patch cvs to build against X11R7.0 Mar 12 18:04:35 03jnc 07org.oe.dev * r63e85bb9... 10/packages/ (3 files in 2 dirs): More s:ftp.handhelds.org/...gpe.../...:${GPE_MIRROR}/g Mar 12 18:04:44 03pH5 07org.oe.dev * r8e33eeb1... 10/packages/xlibs/ (diet-x11/fix-disable-xlocale.diff diet-x11_X11R7.0-1.0.0.bb): diet-x11_X11R7.0-1.0.0: fix build for --disable-xlocale Mar 12 18:06:50 i committed that like a long time ago Mar 12 18:06:59 o_O Mar 12 18:07:37 03rw 07org.oe.dev * r975abe26... 10/ (11 files in 6 dirs): (log message trimmed) Mar 12 18:07:39 wpa-supplicant, gnutls, libgcrypt, various .confs: revert recent commits by jnc. Mar 12 18:07:41 - reverts revisions: Mar 12 18:07:43 53e14a99825447107bfefacf77e28b5718be54b8 Mar 12 18:07:45 b0535ae5d884f64c3f25212755c0dbbd607a4984 Mar 12 18:07:47 8aaa554502fb1d84fe29935d27089df09bfc3ea1 Mar 12 18:07:49 de47a33901a0f8666a237eaf5825ba25f364516f Mar 12 18:08:07 reenoo: hehe, yeah even discarded RP's commtents :} Mar 12 18:09:10 bleh Mar 12 18:09:15 ~lart t-offline Mar 12 18:09:15 * ibot rm -rf's t-offline Mar 12 18:09:36 reenoo: hehe, he even discarded RP's comments :} Mar 12 18:09:49 i don't know where you gather that from Mar 12 18:09:53 i didn't remove anything Mar 12 18:10:05 zecke: no, hh.org mail is being utterly broken again Mar 12 18:10:10 it's still in there, if you have a problem with it, set your PREFERRED_VERSION you are so keen about Mar 12 18:10:15 *cough* Mar 12 18:10:51 not sure why this is being reverted. it fixes the zdxxxx stuff Mar 12 18:11:04 reenoo: thanks for your review Mar 12 18:11:11 there was an accompanying commit to allow for the zdxxxx stuff, you may want to revert that too Mar 12 18:11:24 pb_: what is the status of your policy writings? Mar 12 18:11:25 also revert the.. well read the irc log Mar 12 18:11:42 zecke: pardon? Mar 12 18:12:12 pb_: nice try, You agreed to write some best practices down Mar 12 18:12:21 oh, did I? that was careless Mar 12 18:12:29 pb_: e.g. just because you can commit/push to every file, you shouldn't do it Mar 12 18:12:32 which best practises in particular? Mar 12 18:12:48 zecke: well. this isn't a permanent solution. I guess revisiting the way push access is granted would be in order. Mar 12 18:13:05 pb_: respect maintainership, do not touch 'core' files without getting acknowledged to do so Mar 12 18:13:09 zecke: as you may have gathered, I forgot about that. Mar 12 18:13:18 so, progress has been... limited. Mar 12 18:13:26 pb_: at least you remember it :) Mar 12 18:17:06 reenoo_: right, but I think we can train people to not commit cross over Mar 12 18:18:10 zecke: well. the current inflation of everyone and their dog getting push access doesn't scale well. Mar 12 18:19:27 whomever just rolled back my commits, they broke wpa-supplicant Mar 12 18:20:38 i don't know what was not clear about me working on gcc4 and wifi.... that has been the thing all along. does anyone have a problem with me working on that? Mar 12 18:21:22 reenoo_: I would love to avoid too complex processes Mar 12 18:21:40 reenoo_: if access is granted they should be told to limit their commits to specific areas Mar 12 18:21:49 reenoo_: if they fail, we can easily revoke access Mar 12 18:22:43 hmm Mar 12 18:23:00 NAiL: yes :) Mar 12 18:23:01 NAiL: you may want to roll back your commit :( it seems people are disapproving Mar 12 18:23:06 I've tried contacting maintainers several times. As of yet I have had no luck Mar 12 18:23:19 jnc: which commit? Mar 12 18:23:24 the zd thingy with wpa Mar 12 18:23:25 NAiL: that is the other case Mar 12 18:23:40 NAiL: my commits were overthrown by rw Mar 12 18:23:56 reenoo_: I'm not disagreeing with your revert, but oe@ has been broken for over a day now, so the mail could have been sent Mar 12 18:23:59 NAiL: generally you NSLU2 devles are too shy (nowadays) Mar 12 18:24:04 jnc: uh? ah. Yeah, I can take a look at that. Mar 12 18:24:13 zecke: yeah Mar 12 18:24:13 thanks, sorry about the confusion. Mar 12 18:24:21 i thought that everyone knew i was working on gcc4 / wifi Mar 12 18:24:42 koen: unlikely, he pushed immediately after RP's request Mar 12 18:24:57 reenoo_: did you even read the stuff i committed? Mar 12 18:25:06 jnc: I did Mar 12 18:25:16 zecke: btw, what about packages without any maintainer? Mar 12 18:25:46 NAiL: if it is in a image -> oe@hh.org and ask Mar 12 18:25:48 i didn't touch anything that was "core", did not remove any files, there's a lot that i was not doing wrong, and when it needed to be cleaned up then i go mail people and ask if removing stuff will break their stuff Mar 12 18:25:52 yeesh. Mar 12 18:26:07 NAiL: if not commit, and mail oe@hh.org Mar 12 18:26:10 wpa-supplicant-nossl has been broken for a lonnnng time Mar 12 18:26:33 hmm Mar 12 18:26:37 bit annoying that Mar 12 18:26:51 I'm on way too many mailinglists ;) Mar 12 18:27:15 i didn't make a push to the stable branch that broke stuff. it's a development branch. exactly what is wrong with my commit? Mar 12 18:28:06 NAiL: people - including me - push to files they should not touch :) Mar 12 18:28:28 hi does anybody have any help for a tar failure on the release branch for gettext-native-0.14.1-r3 Mar 12 18:28:30 NAiL: we have had some real issues with that, we need a way - besides reenoo - to identify it Mar 12 18:28:46 Harvy: pastebin it please Mar 12 18:28:52 jnc: I've layed out the reasons in my commit message. I'm not going to repeat myself. Mar 12 18:28:55 Harvy: I suspect your fetch was corrupt/incomplete Mar 12 18:29:33 zecke: yeah. I haven't been subscribed to oe@ earlier, so I haven't been asking there. But it's kinda cumbersome :-\ Mar 12 18:29:47 jnc: there's nothing to say against what you wanted to do, but you didn't do it right. Mar 12 18:29:48 * NAiL has marked his territorly clearly with MAINTAINER= ;) Mar 12 18:29:57 NAiL: well you might be lucky that the guy with competence sits in this channel Mar 12 18:30:08 reenoo_: i'm not a mindreader, sir Mar 12 18:30:11 often, they do. Mar 12 18:30:35 NAiL: in that case we might be able to skip the oe@hh.org roundtrip Mar 12 18:30:45 yup Mar 12 18:30:45 NAiL: but this is pb_'s task :) Mar 12 18:30:54 i was offering patches, and doing things the way i've been doing them for a long time. it was suggested i be given commit access, so i stop bothering people Mar 12 18:32:42 zecke: its at http://pastebin.com/598239 Mar 12 18:33:24 Harvy: remove /home/norm/stuff/stable/sources/gettext-0.14.1.tar.gz and /home/norm/stuff/stable/sources/gettext-0.14.1.tar.gz.md5 Mar 12 18:34:14 reenoo_: you have done yourself a diservice to NAiL and myself. Mar 12 18:34:38 i'm not dragging NAiL into this other than to say, i fixed stuff that was obviously broken Mar 12 18:34:47 no maintainer was fixing it. i did. Mar 12 18:35:44 jnc: note that I was mainly going after the collateral damage in the gnutls department Mar 12 18:35:44 and another dev saw my changes, and what i did was right, so they fixed another build to remove brokenness caused by the broken wpa stuff which is now (or was) fixed Mar 12 18:35:48 now it's broken again Mar 12 18:36:04 jnc: you dropped patches and whatnot Mar 12 18:36:14 so? i didn't remove them Mar 12 18:36:16 they're still in there Mar 12 18:36:22 if you have a problem, add them back in Mar 12 18:36:32 it's not hard Mar 12 18:36:51 nevermind Mar 12 18:36:58 bad practice to force changes without a few tests to see what else is broke .. Mar 12 18:37:08 * zecke watches reenoo's monolog Mar 12 18:37:14 i've tested the wpa stuff Mar 12 18:37:17 it works. gnutls works. Mar 12 18:37:40 one person isnt testing Mar 12 18:37:43 and sylpheed breaks Mar 12 18:37:51 okay so let's fix sylpheed Mar 12 18:38:03 * reenoo_ gives up Mar 12 18:38:08 speaking of which Mar 12 18:38:24 what, i'm the one who committed it, if you have a problem you should bring that to my attention Mar 12 18:38:32 zecke, you try my patch to see that it behaves ? Mar 12 18:38:56 http://bugs.treke.net/show_bug.cgi?id=735 Mar 12 18:39:02 pH5: I'm seing the XCOMPOSITE_BREAKAGE now Mar 12 18:39:26 pH5: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060312192107.log Mar 12 18:39:41 jnc: if you version bump a .bb it's, quite obviously actually, your duty to port patches. Mar 12 18:39:54 okay Mar 12 18:40:16 you don't care much for upstream, hmm? Mar 12 18:40:32 emte: do you really want to shell export HOST_NONSYSV? Mar 12 18:41:00 i need to do i not to set the default value? Mar 12 18:41:07 hmm Mar 12 18:41:21 emte: you can omit the 'export' part Mar 12 18:41:32 I actually upgraded wpa-supplicant to 0.4.7 in january... Mar 12 18:41:35 * NAiL coughs Mar 12 18:41:35 emte: also do you really want to touch fakeroot non native? Mar 12 18:42:05 NAiL: nice work on that btw Mar 12 18:42:08 zecke, i have to, the toolchain doesnt use -native Mar 12 18:42:28 it doesnt inherit from it Mar 12 18:42:49 NAiL: the only remaining bit was the dep on openssl, which is not legal to distribute linked in binary form ;) Mar 12 18:42:51 emte: hmm this means HOST_SYSV influences target configuration Mar 12 18:43:10 reenoo|afk: *sigh* Mar 12 18:43:12 uh Mar 12 18:43:15 NAiL: the gnutls stuff resolved it Mar 12 18:43:30 zecke, so i need to modify the tool chain to behave properly? i was trying to avoid any direct changes Mar 12 18:43:33 * NAiL didn't mean to say anything to reenoo|afk Mar 12 18:43:59 jnc: aha Mar 12 18:44:04 yeah Mar 12 18:44:07 so we can fin... well Mar 12 18:44:10 emte: we might need to add TARGET_NONSYSV (but I do not like it yet) Mar 12 18:44:11 we could until the revert Mar 12 18:44:24 finally distribute wpa-supplicant ssl capable binaries Mar 12 18:44:34 emte: also it is nice to touch documentation.conf Mar 12 18:44:43 emte: so you can document what HOST_NONSYSV does :) Mar 12 18:44:46 zecke, yeah i agree that shouldnt be added as well since that would/shoud be defined by distro\ Mar 12 18:45:02 jnc: yeah, but since breakage apparently ensued, something has to be improved first ;) Mar 12 18:45:15 i should probably be removed as a dev Mar 12 18:45:24 this is rediculous :( Mar 12 18:45:56 * NAiL has taken his fair share of bitching due to commits breaking stuff :-P Mar 12 18:46:05 jnc, how about working in a branch and then merging after issues are resolved ? Mar 12 18:46:25 emte: i didn't want to overstep myself Mar 12 18:46:29 just small commits Mar 12 18:46:57 jnc: as RP said, send a mail to oe@ first and wait a day for feedback Mar 12 18:47:53 hm Mar 12 18:47:53 so zecke i need to look at changing the toolchain to correctly call -native, remove the non-native test, remove the export , and touch documentaion.conf? Mar 12 18:48:25 emte: let me check toolchain first Mar 12 18:48:35 emte: do we talk about meta-toolchain? Mar 12 18:48:45 good, cuz my next question was where to find it :) Mar 12 18:49:00 emte: actually wait Mar 12 18:49:06 emte: what do you mean by toolchain? Mar 12 18:49:24 emte: no one should execute a TARGET binary on HOST (well except for qemu) Mar 12 18:49:57 just building upto bootstrap even, the first time fakeroot is built the non-native version is specified Mar 12 18:50:11 emte: I can't imagine that :) Mar 12 18:51:35 my issue is and has been that fakeroot binary built fails due to sysv-ipc, my flag tests and changes it to tcp-ipc so that fakeroot can run and build images Mar 12 18:51:56 emte: I understand that part Mar 12 18:52:09 no its not really runing a target binary Mar 12 18:52:09 emte: but fakeroot_PV.bb should not be changed Mar 12 18:52:21 yeah i agree Mar 12 18:52:25 cute Mar 12 18:52:36 cvs actually *alters* the files in my repo Mar 12 18:53:49 * emte passes koen some majic cvs cleaner Mar 12 18:54:33 pH5: Xcomposite should have a different include order Mar 12 18:58:22 zecke: how? src/Xcomposite.c includes xcompositeint.h which includes Xcomposite.h which defines XCOMPOSITE_VERSION. If you don't have an old Xcomposite.h in staging I don't see how this could fail? Mar 12 18:58:56 emte: revert the non native part, revert the bitbake.conf part, change documentation.conf Mar 12 18:59:17 emte: change python() to check if HOST_NONSYS is not Null Mar 12 18:59:39 pH5: well even if I have an old Xcomposite.h in staging, the one from the Xcomposite build should be used? Mar 12 19:00:55 should we be preferring the includes from staging or from the cross compile target Mar 12 19:01:51 zecke: now I slowly start to understand :) Mar 12 19:02:13 pH5: sorry, I just verified, the Xcomposite from the staging dir is used Mar 12 19:03:48 emte: one can write stuff like if bb.data.getVar(var, d, True) and bb.data.getVar(var,d,True) != '0': Mar 12 19:03:57 emte: this way we avoid defining the variable in bitbake.conf Mar 12 19:05:09 zecke: like this: http://en.pastebin.ca/45388 ? Mar 12 19:06:05 pH5: looks good Mar 12 19:08:30 pH5: will test on next compile :} Mar 12 19:08:54 zecke, ... doesnt python default non-null ? Mar 12 19:09:21 emte: bb.data.getVar returns Null if the variable does not exist Mar 12 19:09:23 perhaps i am mixingup my languages ... Mar 12 19:11:37 changing now Mar 12 19:11:50 emte: don't forget the reference documentation Mar 12 19:12:16 i wont Mar 12 19:16:00 :) Mar 12 19:19:35 03yuecelm 07org.oe.dev * rb9437836... 10/ (7 files in 3 dirs): Mar 12 19:19:35 jsut to make sure, i am removing HOST_NONSYSV from bitbake.conf altogether? Mar 12 19:19:39 apr-0.9.7,apr-util-0.9.7: replace unretrieveable 1.1.x versions Mar 12 19:19:41 - subversion needs only 0.9.7 Mar 12 19:19:43 - avoid execution of gen_uri_delims (place uri_delims.h directly) Mar 12 19:19:45 - add APACHE_MIRROR to conf/bitbake.conf Mar 12 19:19:52 03pH5 07org.oe.dev * r9d3987c5... 10/packages/xlibs/ (2 files in 2 dirs): libxcomposite-X11R7.0-0.2.2.2: build fix for people with an old Xcomposite.h in their staging directory Mar 12 19:20:37 pH5: thanks Mar 12 19:20:39 emte: right Mar 12 19:24:46 http://bugs.treke.net/show_bug.cgi?id=735 Mar 12 19:24:52 that what we are looking for? Mar 12 19:25:43 hei hei Mar 12 19:26:18 emte: better Mar 12 19:26:18 hi Mar 12 19:26:36 emte: now we could argue if we want to 'append' to EXTRA_OECONF Mar 12 19:26:37 where do the kernel defconfigs get copied to build/tmp/work/... ? Mar 12 19:26:44 emte: instead of just setting new values Mar 12 19:28:52 zecke, i was trying to do that before, it looked liek it wasnt valid from the examples i saw in python() Mar 12 19:29:01 ah okay Mar 12 19:29:13 I made my own kernel defconfig and put it into linux-openzaurus-2.6.15 but it won't get copied like the others Mar 12 19:29:18 but appending does make far more sense if it is valid Mar 12 19:30:53 hi Mar 12 19:30:56 hey hrw Mar 12 19:31:07 Delta_: you replaced which file with your config? Mar 12 19:31:33 dropbear 0.48.1 released after my complain mail Mar 12 19:31:53 03yuecelm 07org.oe.dev * r88f7b26b... 10/packages/subversion/subversion_1.2.0.bb: Mar 12 19:31:54 subversion-1.2.0: enable http support Mar 12 19:31:57 - add neon to DEPENDS Mar 12 19:31:59 Delta_: did you give it a new name that's not mentioned in linux-openzaurus_2.6.15.bb ? Mar 12 19:36:09 I just put a new one in, didn't replace any.. do I have to list it somewhere to be copied? Mar 12 19:36:27 Delta_: which machine you target? Mar 12 19:37:43 I made a new machine conf by copying the spitz conf and editing it.. Mar 12 19:38:09 Delta_: ok. but which it is? I'm curious Mar 12 19:38:20 the toradex colibri board Mar 12 19:38:30 it's an industrial pxa board.. Mar 12 19:38:53 bbiab Mar 12 19:39:57 you can see it at http://www.toradex.com Mar 12 19:39:57 Delta_: I would rather copy linux-openzaurus* as linux-colibri and cleaned Mar 12 19:41:39 ah found something.. one second.. Mar 12 19:46:37 03mickeyl 07org.oe.oz354fam083 * rf63123d8... 10/packages/python/ (3 files in 2 dirs): (log message trimmed) Mar 12 19:46:40 python 2.4.2 various packaging additions - thanks to respective bug reporters Mar 12 19:46:41 python-core: include traceback module Mar 12 19:46:46 python-threading: include dummy-thread, dummy-threading and _local-thread modules Mar 12 19:46:48 python-netclient: include LWP and Mozilla cookie jar modules Mar 12 19:46:50 python-codecs: include stringprep module Mar 12 19:46:52 python-distutils: include config subdirectory Mar 12 19:47:05 03mickeyl 07org.oe.dev * rcc7e934c... 10/packages/python/ (python-2.4.2-manifest.inc python_2.4.2.bb): Mar 12 19:47:10 python 2.4.2 various packaging additions - thanks to respective bug reporters Mar 12 19:47:16 python-core: include traceback module Mar 12 19:47:19 python-threading: include dummy-thread, dummy-threading and _local-thread modules Mar 12 19:47:23 python-netclient: include LWP and Mozilla cookie jar modules Mar 12 19:47:31 python-codecs: include stringprep module Mar 12 19:47:33 python-distutils: include config subdirectory Mar 12 19:48:40 NOTE: package dropbear-0.48.1-r0: task do_build: completed Mar 12 19:50:13 good Mar 12 19:50:22 hey mickeyl Mar 12 19:51:40 hi mickeyl Mar 12 19:52:19 hmm Mar 12 19:52:42 can someone kick handhelds.org? i can't seem to fetch http://ftp.handhelds.org/pub/projects/gpe/source/xserver-common-1.8.tar.gz Mar 12 19:53:17 mickeyl: mail has been down for the past 2 days as well Mar 12 19:53:33 ah, that's why it's so nice and calm :D Mar 12 19:53:45 indeed Mar 12 19:53:50 hehe Mar 12 19:53:57 can anyone hook me up with that file ? Mar 12 19:54:06 mickeyl: let me check Mar 12 19:54:13 ah google rulez Mar 12 19:54:14 got it Mar 12 19:54:14 mickeyl: familiar.handhelds.org somewhere? Mar 12 19:54:20 http://www.gtlib.gatech.edu/pub/handhelds.org/projects/gpe/source/?C=S;O=A Mar 12 19:54:32 mickeyl: wait Mar 12 19:54:38 mickeyl: you have it? Mar 12 19:54:46 yup, thanks Mar 12 19:55:24 .... Mar 12 19:55:28 mickeyl: I also have found a way (with a little help from Lennert) to build softvfp images with OE Mar 12 19:55:29 mtpull is failing... Mar 12 19:55:37 koen: cool! how's that? Mar 12 19:56:04 mickeyl: append ' -mfloat-abi=softfp' to targetcc_arch Mar 12 19:56:06 njs: warning: reconstructed manifest from reverse delta '10ea4805c8e0b2c399914bb1f7e4247f9bfde1cd' -> '045d0482440f032c4a04c458de27fc46d0e38596' has wrong id '8bf4ac642720a1022e09f612749548dc251ee24f' Mar 12 19:56:16 mickeyl: I'll have to wait for the toolchain to compile to test it Mar 12 19:56:25 hmmm...perhaps it went away on a second pull... Mar 12 19:56:25 koen: very good! Mar 12 19:56:32 mickeyl: how goes oz354 build? Mar 12 19:56:36 JustinP: i found monotone to be pretty good in self-recovery ;) Mar 12 19:56:40 03hrw 07org.oe.dev * ree702fdf... 10/packages/dropbear/ (dropbear_0.48.bb dropbear_0.48.1.bb dropbear.inc): Mar 12 19:56:42 dropbear: updated 0.48 to 0.48.1 and removed -DGNU_SOURCE from .inc Mar 12 19:56:44 - define was needed only for 0.48 to get scp compiled Mar 12 19:56:46 - 0.48.1 is upstream response for that workaround Mar 12 19:57:07 hrw: slower than I've wanted. i had severe problems with multimachine build. I fell back to my "standard" way for this release... i expect releases to be in place @ wednesday Mar 12 19:57:35 ok Mar 12 19:57:44 i would appreciate a last check before we announce Mar 12 19:57:52 to make sure i didn't goof up anything Mar 12 19:57:54 who knows... Mar 12 19:58:10 JustinP: err, how did you manage that? Mar 12 19:58:11 * mickeyl starting to distrust his dual xeon workstation Mar 12 19:58:14 mickeyl: rebuild and check the MD5SUM's Mar 12 19:58:23 mickeyl: I'm thinkin about updating some stuff before release in images: dropbear 0.48.1, 2.6.14-rc1 newer (mmc4 support) Mar 12 19:58:32 mickeyl: ok Mar 12 19:58:48 zecke: they are always differences due to dates and times being encoded in images. Mar 12 19:58:50 * hrw want cherrypicker in monotone.... Mar 12 19:58:51 s/they/there/ Mar 12 19:59:35 * mickeyl tries to apply his patches to lyx 1.4 Mar 12 19:59:52 i'll upgrade our lyx version next month Mar 12 20:00:02 now I will probably do "monotone rm -rf dropbear" in branch, copy .dev, comit Mar 12 20:00:08 hrw: *nod* Mar 12 20:01:07 OffTopic: ~lart pdfetex team for being ignorant about opentype Mar 12 20:01:24 pdfetex? Mar 12 20:01:33 heh.. long time since last time I used latex.. Mar 12 20:02:44 mickeyl: wednesday->sunday I have free evenings so 3.5.4 work... Mar 12 20:03:01 hrw: ah pretty cool. what do you plan `? Mar 12 20:03:26 mickeyl: test your images, do some updates? Mar 12 20:03:44 mickeyl: can you build recent linux-openzaurus-2.6.14-rc1 for 3.5.4? Mar 12 20:04:14 for c7x0 Mar 12 20:04:15 ? Mar 12 20:04:34 njs: I just pulled...got a long list of revs with unmet deps Mar 12 20:04:48 njs: last error was monotone: fatal: std::logic_error: database.cc:1146: invariant 'I(old_id == check)' violated Mar 12 20:05:01 njs: a second pull seems to be going fine, though... Mar 12 20:06:10 JustinP: the long list of revs is a little misleading... Mar 12 20:06:38 njs: just letting you know what it output. I can send you the full output if you like, but it seems to have fixed itself... Mar 12 20:06:59 we used to have bugs where we'd request stuff but never actually request other stuff that the first stuff depended on... so the stuff we did request would be queued up in memory, waiting for the stuff it needed, which never arrived, and then when we exited it would just get silently thrown away Mar 12 20:07:21 I see Mar 12 20:07:22 which was kinda bad, so code was added to warn whenever stuff got thrown away... which of course gets triggered if monotone bails out early Mar 12 20:07:26 well, I got a crash.... Mar 12 20:07:37 if we decide to crash because we're getting bad data, then we _want_ to throw that stuff out :-) Mar 12 20:07:40 this happened before the first rev was written Mar 12 20:07:48 nod Mar 12 20:07:48 yeah, I understand ^_^ Mar 12 20:07:56 yep, second pull worked Mar 12 20:08:05 I guess the server was perhaps in an indeterminate state.... Mar 12 20:08:13 which of course shouldn't happen Mar 12 20:08:31 I don't have any guess why the first pull would have had that problem, though; honestly I might be a little suspicious of the hardware involved... Mar 12 20:08:31 would you like the command output and MT/debug? Mar 12 20:09:10 hmm, sure Mar 12 20:09:13 yeah, blame the poor little hard-working celeron ;-) Mar 12 20:09:21 heh Mar 12 20:09:32 I doubt they'll have anything useful, but you never know :-) Mar 12 20:10:21 e-mail? Mar 12 20:10:27 sure, njs@pobox.com Mar 12 20:10:37 or monotone-devel@nongnu.org :-) Mar 12 20:11:05 * mickeyl compiles JUCE 1.27 Mar 12 20:11:31 njs: how far down is cherrypicking (a.k.a. a working monotone diff) on the monotone todo list? Mar 12 20:13:00 njs: ok, sent Mar 12 20:18:59 koen: hmm, beats me -- do you know anyone who wants to work on it? :-) Mar 12 20:19:35 * njs mostly has to stick to the critical path/organization type stuff, side-features like that are cool but depend on someone working on them Mar 12 20:20:10 koen: hm... but how set are you on a UI that involve a diff command and a patch command? Mar 12 20:20:43 njs: I could start modeling a UML use case? Mar 12 20:20:51 koen: because it is much simpler to add a command like 'monotone apply_changes_to_working_copy -r FROM -r TO' Mar 12 20:21:35 njs: and apply_changes_to_branch -r FROM -r TO ? Mar 12 20:21:44 e.g a mini-propagate Mar 12 20:22:20 *nod* a mini-propagate would be helpful for us Mar 12 20:29:04 * JustinP has been using svn merge lately. It was pretty useful. Mar 12 20:29:18 although using propagate and keeping the rev tree is nice too... Mar 12 20:29:29 (I was mostly using it as a propagate...) Mar 12 20:30:10 re Mar 12 20:30:29 mickeyl: yes - c7x0 got mmc4 cards support from 2.6.15 (oneliner patch) Mar 12 20:32:47 hrw: cool. we should use that as release kernel then Mar 12 20:35:30 mickeyl: yep - and it was tested by 'jan' from oesf forum Mar 12 20:39:18 * jnc waits for response to OE ML posting Mar 12 20:40:03 meanwhile, i'm stuck sitting here doing nothing Mar 12 20:40:05 jnc: FYI the mailinglist has been down this weekend Mar 12 20:40:20 koen: totally and officialy? Mar 12 20:40:40 great! so all this hubbub that reenoo makes by reverting my commmit, for naught Mar 12 20:40:47 * jnc goes into corner and cries Mar 12 20:41:10 hrw: well, hh.org isn't sending *any* mail since friday(ish) Mar 12 20:41:20 hrw: I don't know how official you want that to be Mar 12 20:41:31 more seriously though, could someone please, undo the damage reenoo just did. Mar 12 20:42:06 i would like to fix up the sylpheed bug he failed to mention to me earlier :( Mar 12 20:42:22 koen: any != any from oe@ Mar 12 20:46:15 jnc: which sylpheed version? Mar 12 20:46:45 johnX: reenoo did not say, he said something about my commit breaking sylpheed. and then backed out four of my commits. then left. Mar 12 20:46:54 i was not made aware of this prior to the reverts Mar 12 20:46:58 was your commit a new gnutls? Mar 12 20:47:02 yeah Mar 12 20:47:15 gnutls is in images == important Mar 12 20:47:18 jnc: have you tried building sylpheed? if it's broken, I imagine the problems will be fairly obvious. Mar 12 20:47:44 judging from reenoo's commit message, it sounds like your checkin reverted some gnutls patch that sylpheed depends on. Mar 12 20:47:50 he backed out everything though, without talking to me at all. i discussed the new gnutls 2 weeks ago Mar 12 20:47:57 the concensus was, go for it Mar 12 20:47:58 jnc: have you seen #725 bug Mar 12 20:48:07 ? Mar 12 20:48:24 johnX: i'll have a look Mar 12 20:48:50 hi aquadran Mar 12 20:48:54 jnc: on the mailing list? I don't remember seeing that discussion, but I might have accidentally overlooked it. Mar 12 20:49:01 johnX: that bug is not assigned to anyone Mar 12 20:49:05 koen: hmm, basically the same operation just in-memory, instead of to-working-copy? Mar 12 20:49:16 jnc: I submitted a .bb for sylpheed 2.2.2 which doesn't require that patch to gnutls Mar 12 20:49:19 koen: (sort of similar to the difference between current-merge, and planned-workspace-merge) Mar 12 20:49:23 pb_: i just recently subscribed to the mailing list. not on the mailinglist, no Mar 12 20:49:33 jnc: I didn't know who to assign it to Mar 12 20:49:35 i wasn't aware that the ML was so important until it was mentioned today Mar 12 20:49:42 * njs has to run Mar 12 20:49:45 johnX: you should not have to assign it to anyone Mar 12 20:49:46 cu njs Mar 12 20:49:57 yeah, the mailing list is the best place to discuss proposed changes Mar 12 20:50:04 johnX: it just means that something is broken in the way the OE folks deal with bugtracking Mar 12 20:50:13 i.e. people should be assigned bugs Mar 12 20:50:14 irc is useful as well, but it's rather too dependent on the right people watching at the right moment. Mar 12 20:50:19 hm. time to add something from .oz354fam083 to .dev... Mar 12 20:51:13 okay, well i was surely out of place adding a new gnutls... but if i make the commit anyways, then i would like to be notified if i broke something Mar 12 20:51:22 before all of my work is reverted suddenly Mar 12 20:51:48 that's really insulting, i even had in my commit messages a note that said i should be notified if i break something Mar 12 20:52:13 anyways, I think that patch to gnutls can be dropped if sylpheed 1.9.x is the only thing that needs it and if sylpheed 2.2.2 gets some testing Mar 12 20:52:28 johnX: what does that patch do? Mar 12 20:52:45 jnc: reenoo is responsible for >90% of the reverts in OE Mar 12 20:52:53 i like people to be there to say no Mar 12 20:53:01 I'm referring to the one that adds some openssl functions to gnutls Mar 12 20:53:13 but it means i am going to have to waste time on my weekend tracking down a bunch of crap Mar 12 20:53:34 instead of reenoo pinging me and saying, hey, you screwed up. fix it. Mar 12 20:53:42 which would save everyone a lot of time Mar 12 20:54:24 ~lart monotone for lack of cherrypicker Mar 12 20:54:25 * ibot farts in monotone's general direction for lack of cherrypicker Mar 12 20:55:03 or, hey how about this, reenoo actually fixing gnutls to work. Mar 12 20:55:06 no that would be too difficult Mar 12 20:55:54 i dropped the patch because i wasn't 100% sure it was necessary Mar 12 20:56:02 and as johnX mentions, it's been fixed upstream Mar 12 20:56:06 so it's not really necessary Mar 12 20:56:18 jnc: what does that patch do? Mar 12 20:56:54 jnc: it sounds like you're taking the reversion of those changes rather too personally. You shouldn't interpret it as some kind of personal slight, it just means that the changes in question were broken. If you fix the problems and re-submit the changes, no doubt that will be fine. Mar 12 20:57:28 i know it's not personal Mar 12 20:57:42 jnc: as reenoo said, the idea was okay, implementation not Mar 12 20:57:50 jnc: as for dropping the patch, you should only take it out if you are (close to) 100% certain that it is unnecessary. If you just aren't sure, you probably ought to ask the person who added it. Mar 12 20:57:50 it's wasted time though. i was going to sit here and take 4-5 hours today hacking on wifi/gcc4 stuff. Mar 12 20:59:18 jnc: as for the wasted time, I don't really understand why reverting the changes has caused you to waste any more time than if reenoo had just said "you screwed up, fix it". Mar 12 20:59:58 btw, there's two heads now, isn't it? Mar 12 21:00:05 and reverting the changes does at least allow everyone else to build sylpheed in the meantime, so it's probably a net saving of time overall. Mar 12 21:02:37 I don't have my devbox handy, so I can't check right now. Is there a maintainer on busybox? Mar 12 21:02:58 don't think so Mar 12 21:03:08 NAiL: there's only one head now :) Mar 12 21:03:15 koen: thx ;) Mar 12 21:03:30 I'd like to add sysctl support to busybox Mar 12 21:03:52 what's the impact of that? Mar 12 21:03:57 NAiL: Will you take bribes to also add losetup to busybox? Mar 12 21:04:19 mwester: hehe, to svn I can do that. How much? :-P Mar 12 21:05:54 NAiL: generally, though, the busybox config is a per-distro thing. You should feel free to bash the openslug one around any way you want, but I would suggest that you refrain from futzing with the familiar one (for example) unless there is compelling reason to do so. Mar 12 21:07:04 yeah, I know. I've split off the slugos-* config and added it there. I don't have a compelling reason other than to make the config the same ;) Mar 12 21:07:07 Curious, but's actually why I dropped by here today -- I was wondering *how* one changed the per-distro config for busybox (but y'all seemed to be a bit busy with other problems). Mar 12 21:07:55 I've tried the obvious things (like the defconfig file), but it doesn't seem to be used in the build - any pointers or suggestions to other places to look would be appreciated! Mar 12 21:08:20 NAiL: heh. busybox is sufficiently fragile (and sufficiently important) that making it the same is probably not a good enough reason. Mar 12 21:08:28 mwester: the slugos-specific config is in busybox/busybox-1.01/slugos. If you're compiling for slugos-*, that one is picked instead of the default one in busybox/busybox-1.01/ Mar 12 21:08:50 pb_: Fair enough. It's already split off, so it really doesn't matter that much for me. Mar 12 21:09:01 pb_: what reenoo did was wrong. Mar 12 21:09:02 Yeah, I found that one - it's the one for Unslung that seems to have been carefully concealed. Mar 12 21:09:20 him being right about my not being okay to commit, does not change that his actions are wrong Mar 12 21:09:27 mwester: try adding an unslung dir, and putting another defconfig in there. Mar 12 21:09:38 other commits depended on the stuff that i committed, and his reverts of my commits break that Mar 12 21:09:49 which i would have gladly pointed out to him Mar 12 21:09:51 NAil: could it be so simple?? :) I'll go try that! Thanks! Mar 12 21:09:53 should he have approached me Mar 12 21:10:07 mwester: I think it should be. np, and welcome (if it works) Mar 12 21:10:43 anyways... those changes are there somewhere in monotone i suppose Mar 12 21:10:49 if people want to dig them out, they're there Mar 12 21:10:59 i won't be working on wpa/gnutls meanwhile. Mar 12 21:11:17 too many people disagree with my working on it to the best of my ability Mar 12 21:11:42 I don't think anybody disagrees with you working on it to the best of your ability. I certainly haven't seen any evidence of that. Mar 12 21:11:45 uhm, I don't think anyone disagree with you working on it. As long as the commits don't break anything. Mar 12 21:11:53 People do disagree with checking in broken changes, but that's a different matter. Mar 12 21:12:13 (pb_ phrased it better) Mar 12 21:12:30 reverting the commits was a broken thing to do Mar 12 21:13:43 in your view, yes. But reenoo clearly felt that reverting them was the appropriate course of action to un-break sylpheed (and maybe other things), and I support his decision. Mar 12 21:13:49 (I haven't done a pull in a while) Out of curiosity is sylpheed 2.2.2 built by default now? or did a DEFAULT_PREFERENCE = "-1" get added to it? Mar 12 21:14:15 it causes bug #725 to be dead-end, breaks the commit NAiL made, and is detrimental to fixing sylpheed. Mar 12 21:14:26 It's unfortunate that this seems to have caused some additional breakage, but I guess the way to deal with that is to revert the dependent changesets as well. Mar 12 21:14:34 yeah Mar 12 21:14:42 i could have let him know about that stuff... Had he asked. Mar 12 21:15:11 johnX, sylpheed 2.2.2 doesnt exist Mar 12 21:15:42 It causes bug 725 to be less pressing, but scarcely "dead end". Mar 12 21:15:53 jnc: no, it doesn't break the commit I made. It just doesn't fix an unfixed issue. Mar 12 21:16:15 oh Mar 12 21:16:17 that's not so bad Mar 12 21:16:28 as for detrimental to fixing sylpheed, I'm not sure what you mean. Mar 12 21:16:29 (which I should have pushed a fix for a while ago, but have just forgotten about) Mar 12 21:16:53 hmm...I must be confused...I thought Xora added the .bb I posted for sylpheed 2.2.2 i posted to the bug tracker Mar 12 21:17:15 i don't know, i'm done with it. the mailing list is down, and thats where this should be discussed Mar 12 21:17:17 * emte finds it funny that this argument about gnutls exists when qemu was pushed and has broken most everything because it doesnt copile properly Mar 12 21:17:22 yeah Mar 12 21:17:25 compile* Mar 12 21:17:38 i really feel picked on, today. since there's so much in flux already Mar 12 21:18:13 if sylpheed doesn't work, then the answer is to fix it or use stable Mar 12 21:18:29 i'm happy to do either. more happy to fix things than now Mar 12 21:20:22 libxfont fails now Mar 12 21:20:32 ok, so if sylpheed 2.2.2 doesn't exist, then what does this mean: http://bugs.treke.net/show_bug.cgi?id=724 Mar 12 21:21:00 if sylpheed is just generally broken, sure, I agree with you. But, if sylpheed was broken by a change you made, saying "fix it or use stable" is not an acceptable response; the onus is on you to make it work, unless you can show that breaking sylpheed was necessary to fix some greater ill. Mar 12 21:21:10 emte: it has been commented out by koen Mar 12 21:21:28 zecke: I guess I am getting close to formulating the policy you requested. Mar 12 21:21:35 johnX, ah, i must not have that rev, i had to go back because of the qemu problem Mar 12 21:21:48 ah, ok Mar 12 21:22:25 pb_: could you include "Please stop programing in C, use Visual Basic instead" Mar 12 21:22:39 pb_: reenoo feels responsible enough to not notify me, and revert my changes, and ignore my offer to fix the problem without need for a revert Mar 12 21:22:48 zecke: that is at the top of my list Mar 12 21:22:55 pb_: thanks Mar 12 21:23:26 ~lart drivers without makefiles Mar 12 21:23:27 * ibot steals drivers without makefiles's mojo Mar 12 21:23:47 zecke, which has been commented out? qemu ? Mar 12 21:23:53 jnc: reenoo feels compelled to revert your changes since they are causing a problem. As I said, you shouldn't view the reversion as an attack on you personally: you should just fix the problem and then submit a corrected changeset. Mar 12 21:24:05 oh Mar 12 21:24:14 well that's the first time anyone has told me what is the proper thing to do Mar 12 21:24:16 thank you pb Mar 12 21:24:20 and we need a policy on reverts as well Mar 12 21:24:42 i think a redundant mail server might help ... Mar 12 21:24:43 doing reverts like this is evidently pissing people off Mar 12 21:25:17 but one might aruge that less developers would mean less bugs Mar 12 21:25:20 *cough* Mar 12 21:25:42 koen|tv, you taking lessons from MS now? Mar 12 21:25:44 yes. Take the linux kernel for instance ;) Mar 12 21:25:53 koen|tv: fewer machines or packages might also mean less bugs :D Mar 12 21:26:15 let remove all g* pacakges and then all o* ones ;) Mar 12 21:26:29 lol Mar 12 21:26:30 heey Mar 12 21:26:37 * NAiL votes for removing x* Mar 12 21:26:44 yeah, who uses a GUI on their zaurus anyways? Mar 12 21:26:59 I don't have a zaurus ;) Mar 12 21:27:13 I don't use a GUI on the boxes I build for Mar 12 21:27:16 my question: besides sylpheed, what else depends on those openssl functions in gnutls? Mar 12 21:27:24 s/uses a GUI on their/owns a/ Mar 12 21:27:36 gpg can, some browsers can Mar 12 21:28:05 there are quite a few aps with gnutls support Mar 12 21:28:21 GNUtls provides some things that OpenSSL does not have, visa versa; also OpenSSL is generally more secure and well tested Mar 12 21:28:28 btw, that reminds me... Would moving separate kernel modules (eg. zd1211, pwc) into a kernel-modules dir? The dirlisting in packages/ is kinda huge. Mar 12 21:28:37 * NAiL rereads his sentence Mar 12 21:28:45 btw, that reminds me... Would moving separate kernel modules (eg. zd1211, pwc) into a kernel-modules dir make sense? The dirlisting in packages/ is kinda huge. Mar 12 21:28:55 that reminds me now, why are we using sylpheed instead of sylpheed-claws anyway? Mar 12 21:29:07 johnX: hard to say for sure; you'd have to make a list of the packages that DEPEND on gnutls and then audit each one in turn. Mar 12 21:29:15 emte: probably because no one has added sylpheed claws Mar 12 21:29:45 i would have thought it was the logical choice since claws supports modules and the reg doesnt Mar 12 21:30:13 presumably regular sylpheed works well enough for the folks who are using it. Mar 12 21:30:21 feel free to add -claws, though, if you like Mar 12 21:30:21 yeah Mar 12 21:30:41 NAiL: that sounds like a good idea Mar 12 21:30:48 need to wait for jnc to re-add a working wpa_suplicant Mar 12 21:31:11 my wireless card is pretty much useless without it Mar 12 21:31:12 hey, it was working fine Mar 12 21:31:31 *cough* Mar 12 21:31:34 hehe Mar 12 21:31:56 actually Mar 12 21:32:13 does anyone know if a fiber pcmcia card exists? Mar 12 21:32:24 i could do it that way Mar 12 21:32:24 NAiL: i've really gotten into using wpa_cli client Mar 12 21:32:45 emte: SC or LC? Mar 12 21:32:54 sc Mar 12 21:33:17 I don't actually *use* wpa, I just added it on request for some nslu2-linux user Mar 12 21:33:45 (I tested it briefly though) Mar 12 21:34:40 so um, how about putting my changes back in there and set the weight to something negative, while i test this? Mar 12 21:34:46 i wonder if zecke ran away ... need to know if http://bugs.treke.net/show_bug.cgi?id=735 needs more work or not Mar 12 21:34:46 like gcc stuff does Mar 12 21:35:08 it's going to be a pain for me to diff through the branches Mar 12 21:35:38 these are changes many devs should work on Mar 12 21:35:41 not just me Mar 12 21:35:45 * NAiL has no objections to setting DEFAULT_PREFERENCE=-1 to jnc's changes (It was an upstream upgrade, right?) Mar 12 21:35:51 yeah Mar 12 21:36:39 isnt the issue that jnc modified currently used bb ? Mar 12 21:36:58 no Mar 12 21:37:36 man this is hard to follow without the mailing list Mar 12 21:37:42 would someone who has an up to date tree do me a favor and check if the default preference for sylpheed 2.2.2 is negative? Mar 12 21:38:17 johnX: it isn't Mar 12 21:38:20 johnX: nope, Mar 12 21:38:23 johnX, http://bugs.treke.net/attachment.cgi?id=565 Mar 12 21:38:57 ok, I didn't know if Xora might have set it negative before he pushed it to .dev Mar 12 21:40:09 this is the seventh time i've had to rm -fr oetmp Mar 12 21:40:14 before completing a single build Mar 12 21:40:40 okay, so can a pull an update and expect bitbake to compile an image? Mar 12 21:41:22 i mean, i was pretty close. but then reenoo did his dance, and now i'm going backwards Mar 12 21:42:05 it delays people submitting patches to sylpheed and the new gnutls stuff Mar 12 21:42:12 emte: I don't know if sylpheed has a minimum version of gnutls that it requires Mar 12 21:42:24 johnX, yeah it does Mar 12 21:42:37 or atleast -claws does Mar 12 21:42:53 I'm trying to dig through their website to find out... Mar 12 21:43:04 been quite a while since i last compiled the non-claws version Mar 12 21:43:41 johnX, why dont you just read the configure script? Mar 12 21:43:57 emte: too logical Mar 12 21:44:01 lol Mar 12 21:44:01 * johnX reads the script Mar 12 21:44:33 +PREFERRED_VERSION_wpa_supplicant = "0.4.7" Mar 12 21:44:39 why the hell did reenoo put that back in there Mar 12 21:44:42 it doesn't do anything Mar 12 21:44:55 ~lart stupid people who don't read changes Mar 12 21:44:56 * ibot nabs the moon and broadsides stupid people who don't read changes with the sea of tranquility Mar 12 21:45:20 jnc, it does on distros taht dont have a version set ... Mar 12 21:45:27 from what i understand Mar 12 21:45:52 zaurus-clamshell-2.6.conf.... it doesn't matter because wpa_supplicant would default to that anyways Mar 12 21:46:02 think bitbake world Mar 12 21:46:13 yeah it would default to that anyways Mar 12 21:46:25 i asked RP to put it in there Mar 12 21:46:38 he doesn't have time to check that, so i was hunting it down Mar 12 21:46:42 it is not necessary Mar 12 21:49:01 i'm pissed that reenoo removed new files, via the revert. i don't take that personally Mar 12 21:49:04 it concerns me deeply Mar 12 21:49:10 no, if it was missing it would default to wpa-supplicant_0.5.1.bb or wpa-supplicant_cvs.bb Mar 12 21:49:15 03tmbinc 07org.oe.dreambox * rd6e8b5ee... 10/packages/tuxbox/ (tuxbox-tuxtxt-32bpp.bb tuxbox-tuxtxt-32bpp/32bpp.diff): tuxbox-tuxtxt-32bpp: fix patch, bump PR Mar 12 21:49:24 how many other new files are reverted and fade into the monotone history? Mar 12 21:49:41 millions :) Mar 12 21:49:44 emte: 'cvs' i think is lower than a number Mar 12 21:49:55 and there is no wpa-supplicant 0.5.1 Mar 12 21:50:00 reenoo reverted it Mar 12 21:50:03 i have one ... Mar 12 21:50:05 ah Mar 12 21:50:29 as well as some other bugfixes of mine Mar 12 21:51:55 let's wait for the ml to online again and discuss it there Mar 12 21:52:07 koen|tv: okay. can do. Mar 12 21:53:06 * emte hopes the mailing list is gonna be quicker than oe is/was Mar 12 21:54:07 why is the ml offline btw? Mar 12 21:54:14 and who can fix? :) Mar 12 21:54:30 NAiL: hh.org is being messed Mar 12 21:54:57 any news? :-P Mar 12 21:55:01 * zecke pulls hairs Mar 12 21:55:17 i've been wondering, is tehre a reason CIA is only attatched to one of the monotone servers? Mar 12 21:55:34 because it wouldn't make sense to have it attached to multiple Mar 12 21:55:42 we'd get multiple copies of the same commit Mar 12 21:55:44 (AFAIK) Mar 12 21:55:48 then the CIA rss would be usefull Mar 12 21:55:55 huh? Mar 12 21:56:14 it currently only reports the commits to vanille Mar 12 21:56:15 JustinP: check against rss before sending commit info Mar 12 21:56:34 not the other 2 servers Mar 12 21:56:35 emte: eventually if you push a commit to server xyz, then server foo gets it Mar 12 21:56:49 pushing the commit to foo first, means it works a little quicker Mar 12 21:56:55 yeah, but CIArss doesnt report those Mar 12 21:57:05 it does Mar 12 21:57:07 i think it does, eventually Mar 12 21:57:15 all server are synced every 15 minutes Mar 12 21:57:16 i wonder why i miss them then ... Mar 12 21:57:21 * JustinP has no clue what you're all talking about Mar 12 21:57:39 http://cia.navi.cx/stats/project/openembedded/.rss Mar 12 21:57:41 if you connect commit e-mails and such to multiple servers we'll get multiple copies of everything, right? Mar 12 21:57:43 koen|tv: 15 minutes? Huh.. It appears nslu2-linux isn't synced that often Mar 12 21:57:46 that the url you use ? Mar 12 21:57:53 * JustinP doesn't use RSS Mar 12 21:57:56 JustinP: apples and oranges Mar 12 21:58:05 JustinP: no email involved in the current topic Mar 12 21:58:07 NAiL: the syncer has been throwing errors latetly Mar 12 21:58:12 e-mail, RSS, whatever Mar 12 21:58:17 it's all via CIA Mar 12 21:58:24 IRC too Mar 12 21:58:32 koen|tv: that explains it. I've synced manually a few times to make sure stuff doesn't diverge too much Mar 12 21:58:39 if multiple servers send commits to it we'll see th ecommits multiple times... Mar 12 21:58:59 a push from a user and a push from a server are the same thing Mar 12 21:59:00 NAiL: any reason why you stopped serving dreambox? Mar 12 21:59:03 we're all the same monotone Mar 12 21:59:27 koen|tv: huh? stopped serving dreambox? I wouldn't know anything about that. I only touch oe.dev Mar 12 21:59:33 NAiL: and nslu2 used to run a syncer themselves Mar 12 21:59:47 aha, so it's an error on our side then Mar 12 21:59:50 oh hey, the new eBay phishing emails are kind of clever Mar 12 22:00:14 jnc: oh? Mar 12 22:00:23 NAiL: monotone: warning: protocol error while processing peer monotone.nslu2-linux.org: 'received network error: access to branch 'org.openembedded.dreambox' denied by server' Mar 12 22:00:38 NAiL: yeah, it poses as an ebay mymessage, and from someone threatening to report you to safeharbor if you don't respond Mar 12 22:00:39 * emte now wonders if i am the only one that uses/watches the CIA rss Mar 12 22:00:53 the "reply message" image link is to ourlady.cn/ws/... Mar 12 22:00:58 ;) Mar 12 22:01:02 http://cia.navi.cx/stats/project/openembedded/.rss ? Mar 12 22:01:02 jnc: I've seen those for a while now Mar 12 22:01:02 emte: no, I have a few CIA rss feeds in firefox Mar 12 22:01:22 emte, koen|tv: yes, yes, RSS, but the same system sends commit e-mails and IRC messages Mar 12 22:01:33 koen|tv, yeah that is the current RSS feed i use in yarssr Mar 12 22:01:35 rss is working for me Mar 12 22:01:51 koen|tv: humm... Well, I guess that should be forwarded to those who know a bit more about that part. It's probably not intentional. Mar 12 22:01:59 I have http:///ewi546.ewi.utwente.nl/tmp/viewmtn/branch.psp?branch=org.openembedded.dev&rss=1 in my feed reader a well Mar 12 22:02:09 i'll add taht one Mar 12 22:03:25 oh cool.. that one show the merges too Mar 12 22:03:39 JustinP, it used to be with BK that the clinet had to be configured to send emails to CIA on a commit, i dont know if this has changed on monotone Mar 12 22:04:07 monotone server do that Mar 12 22:04:53 koen|tv, that server doesnt liek me Mar 12 22:05:08 oh wait Mar 12 22:05:16 you had an extra / Mar 12 22:05:17 :) Mar 12 22:08:09 night koen|sleep Mar 12 22:12:50 ok, I think I managed to severely confuse myself about the sylpheed 2.2.2 supporting gnutls...My only excuse is that it was late when I wrote that .bb Mar 12 22:13:19 cu Mar 12 22:13:39 lol Mar 12 22:13:49 that mean you have the incorrect gnutls version> Mar 12 22:13:50 sylpheed 2.2.2's configure will take --enable-gnutls and silently ignore it Mar 12 22:13:58 then use openssl Mar 12 22:14:01 lol Mar 12 22:14:24 yeah you have to explicitly disable openssl from what i remember Mar 12 22:15:09 uhm...here's the funny part...I don't know how I got the impression that mainline sylpheed supports gnutls...but I don't think it does at all Mar 12 22:16:03 i know claws does, it uses it for some of the gpg stuff and ssl mail i think Mar 12 22:16:37 i imagine you're correct...and that's also probably how I got the impression that mainline sylpheed supported it Mar 12 22:18:21 anyways, now that gnutls got reverted back to the version that has those openssl patches...it's probably best to use sylpheed 1.9.x if you don't want openssl dragged in Mar 12 22:18:35 ...and I'll look at packaging sylpheed claws Mar 12 22:21:26 let me know if you want any help Mar 12 22:22:05 i suspect you will have to do some gtk magic to fit the screens Mar 12 22:22:29 03yuecelm 07org.oe.dev * reb88f31c... 10/packages/subversion/subversion_1.2.0.bb: subversion-1.2.0: dump PR Mar 12 22:24:17 yuecelm - who is he? Mar 12 22:24:35 no idea Mar 12 22:25:00 not sure i've ever saw him in the channel Mar 12 22:25:36 dump PR....that doesn't sound good Mar 12 22:26:31 misspelling? Mar 12 22:26:55 * emte clicks rss link :) Mar 12 22:27:19 emte: did you mean gtk magic to fit 640x480? Mar 12 22:28:07 yeah it was bump Mar 12 22:28:09 http://cia.navi.cx/stats/project/openembedded/.message/ffac2 Mar 12 22:28:16 refers to an earlier commit Mar 12 22:28:24 johnX, and smaller Mar 12 22:28:38 the familiar deveices are smaller than 640x480 Mar 12 22:29:32 does the sylpheed that's already packaged work with those? Mar 12 22:30:32 i would think so Mar 12 22:30:51 should be a matter of porting a patch or two Mar 12 22:32:31 as far as I can tell there aren't any patches that have to do with window size or really anything to do with gtk at all Mar 12 22:33:17 then perhaps its never been addressed Mar 12 22:33:39 i'll look in my feeds here in a min Mar 12 22:35:40 " Mar 12 22:35:59 mischat. anyhow, "dump PR" is an interesting way of saying that you're s/r0/r1/ Mar 12 22:36:06 :/ Mar 12 22:36:13 bump PR Mar 12 22:36:16 looks like they meant ... yeah. Mar 12 22:38:37 there's no MAINTAINER field for gnutls Mar 12 22:38:48 i'm still not seeing where i went wrong Mar 12 22:38:57 * jnc larts reenoo Mar 12 22:39:30 well it did break gpe-image from building Mar 12 22:39:44 i expect that to happen, so i can get feedback from people and fix it Mar 12 22:40:27 not sit here on a sunday evening and sift through two trees, so that i can re-commit things with a miner one liner change that should have been brought to my attention Mar 12 22:40:40 it's nobody's fault Mar 12 22:40:49 it's just wasted time. i'm not happy with duplicating my efforts Mar 12 22:41:04 you cant break a whole image ... Mar 12 22:41:06 it might be nice to test if your change keeps gpe-image from building though Mar 12 22:41:18 gpe-image wasn't building anyways Mar 12 22:41:25 did anyone notice that? Mar 12 22:41:25 no Mar 12 22:41:26 has been for me Mar 12 22:41:38 building gpe-image from scratch was broken until i fixed it Mar 12 22:41:46 now it is broken again AFAIK Mar 12 22:41:48 then again my repo was a weekish-old Mar 12 22:42:04 hmmmm Mar 12 22:42:10 e-image has been building for me... Mar 12 22:42:14 * JustinP starts a gpe-image Mar 12 22:42:58 emte: wow...sylpheed (main and claws) seems to retain most of it's functionality with a window 320x320 Mar 12 22:43:38 yeah, i would think the icons might be a bit large tho ... Mar 12 22:44:07 yeah, icons get pushed off the edge of the screen and you'd need to resort to menus for some things Mar 12 22:44:47 though you get to keep "get mail, send, compose and reply" in 320x320 Mar 12 22:45:19 and the compose window is usable? Mar 12 22:45:32 that would be a no Mar 12 22:45:52 that would need some hacking Mar 12 22:46:49 was curious about that, it seems the compose window wastes a lot of space even on a reg system Mar 12 22:47:53 that's annoying...the rest of it would almost fit on a 240x320 screen Mar 12 22:48:29 especially with smaller icons Mar 12 22:48:41 just means the programmers hardset some values Mar 12 22:48:54 i would think Mar 12 22:49:34 emte: well the design of the compose window is what keeps it from being useful on a small screen Mar 12 22:49:55 there's lots of empty space... Mar 12 22:50:48 yeah between the icons and the "header/tab window Mar 12 22:52:29 who wrote the openssl patch to gnutls Mar 12 22:52:30 hmm? Mar 12 22:52:37 does anyone know? Mar 12 22:53:03 org.openembedded.dev/packages/gnutls/gnutls-1.0.25/gnutls-openssl.patch Mar 12 22:53:05 that one Mar 12 22:53:33 why would gnutls have an openssl patch? Mar 12 22:53:47 for sylpheed, apparently Mar 12 22:53:51 reenoo wanted it Mar 12 22:54:03 i'm trying to figure out what it is doing there in the first place Mar 12 22:54:14 it doesn't make sense to me Mar 12 22:54:17 * emte doesnt recall gnutls using or requireing openssl Mar 12 22:54:41 to meet sylpheed halfway in supporting gnutls I think Mar 12 22:54:45 it's probably an openssl compatibility path Mar 12 22:54:57 patch Mar 12 22:55:16 hmm Mar 12 22:55:33 maybe that was from before when tehre was a warning with gnutls Mar 12 22:57:29 my concern is that the person writing the patch was looking at openssl source Mar 12 22:57:39 which would mean that the patch is tainted Mar 12 22:57:58 not much of a concern Mar 12 22:58:08 since it cant be used anyway Mar 12 22:58:19 in a distro taht doesnt supoprt openssl Mar 12 22:58:37 emte: the code is there for you to look at Mar 12 22:58:40 if you'd like to do so Mar 12 22:58:50 now would be a good time to save yourself from embarassment Mar 12 22:58:56 ;) Mar 12 22:59:10 this is from 23 or 25 Mar 12 22:59:35 * jnc addresses emte privately Mar 12 23:11:36 great.....file-native seems to be borked now! Mar 12 23:12:29 when was the last time it worked for you? Mar 12 23:13:48 ummmm...yesterday I would assume...I built an e-image Mar 12 23:15:54 ah...then maybe it changed again. It broke for me the other day, but I haven't updated since then Mar 12 23:20:42 any objections to me maintaining gnutls? Mar 12 23:20:50 seeing as the mailing list is down Mar 12 23:21:44 since many devs seem to have dropped off to sleep, I don't know if now would be the best time to ask that question... Mar 12 23:22:37 sure Mar 12 23:22:49 i can wait until next month to do any work on OE Mar 12 23:22:55 that's okay. it can wait a month or two Mar 12 23:23:01 man, I dunno Mar 12 23:23:15 I obviously don't have any say Mar 12 23:24:04 jnc, you can take over ownership of any package not listing a maintainer i belive Mar 12 23:24:20 i would avid changing any existing packages tho ... Mar 12 23:24:28 avoid* Mar 12 23:24:34 did you contact the current maintainer? Mar 12 23:24:39 noone has ever protested me taking maintainership of unmaintaned packages Mar 12 23:24:40 there is no current maintainer Mar 12 23:24:51 yeah, I'd say if it has no MAINTAINER set go for it...and send a msg to the list to announce it Mar 12 23:25:06 i send messages, and t... Mar 12 23:25:10 well that's broken Mar 12 23:25:14 lol Mar 12 23:25:44 reenoo throws a hissy fit because his mail client doesn't work Mar 12 23:25:46 just do it and send a msg Mar 12 23:25:47 okay what was i doing again ... Mar 12 23:25:49 it will get to us Mar 12 23:25:56 * jnc shakes fist at the magic mirror Mar 12 23:26:20 jnc, just avoid changing existing, build a new version Mar 12 23:26:28 avoid changing existing? Mar 12 23:26:34 how am i supposed to work like this Mar 12 23:26:36 ah, ok...I didn't know it didn't have a maintainer... Mar 12 23:26:51 there were no maintainers, so i changed stuff Mar 12 23:26:53 now it's reverted Mar 12 23:27:02 and i am left to sort through the tons of issues? Mar 12 23:27:14 not how i wanted to spend my sunday Mar 12 23:27:45 * JustinP knows nothing about reverts Mar 12 23:27:47 oh yeah, my taxes ... i should do those today too Mar 12 23:27:54 perhaps renoo shouldn't have done that without asking Mar 12 23:28:12 JustinP, the changes apperently broke the gpe-image Mar 12 23:28:12 anyway, I suggest making sure that what renoo says is broken works ok with your fixes Mar 12 23:28:16 i think so too. it worries me that he's doing this to other people too Mar 12 23:28:23 you can always make a patch of your earlier rev and reapply it Mar 12 23:28:38 i'm not going to fight reenoo on this, he's the older more established dev Mar 12 23:28:40 let him fix it Mar 12 23:28:58 if he doesn't want me working on unmaintained stuff, then he can do it! Mar 12 23:29:02 * jnc :/ Mar 12 23:29:06 well, if it doesn't break something I imagine there won't be complaints Mar 12 23:29:25 if you alter most library stuff something will break Mar 12 23:29:31 yeah i know that Mar 12 23:29:40 i expect it to break, be notified, and go bug hunting Mar 12 23:29:52 well, just do some cursory testing Mar 12 23:29:55 i did Mar 12 23:29:58 let gpe-image build overnight Mar 12 23:30:01 push in the morning Mar 12 23:30:08 i tried building gpe-image seven times Mar 12 23:30:13 ah Mar 12 23:30:13 it failed every single time Mar 12 23:30:28 due to things completely unrelated to gnutls Mar 12 23:30:37 or libgcrypt or anything i pushed Mar 12 23:31:04 that's what i'm pissed about, is reenoo being a total asshat and reverting a change that only affects him, really Mar 12 23:31:14 instead of thinking about other devs who are working on stuff Mar 12 23:31:19 and asking them Mar 12 23:31:59 well...perhaps your other problems are 64-bit related... Mar 12 23:32:03 i dont remember explicitly but i think reenoo uses gpe in a production enviroment Mar 12 23:32:11 had reenoo rebuilt a tree from scratch, he would not have gotten that far. he would have run into the URL breakage, which i fixed; the X11 composite header mixup, which i believe is now patched, and so on Mar 12 23:32:15 its eitehr reenoo or someone else ... Mar 12 23:32:21 I would suggest just testing the package he complained about Mar 12 23:32:27 maybe it is reenoo, he should be using stable Mar 12 23:32:32 not developer branch Mar 12 23:32:42 these are issues which i will bring up on the maling list Mar 12 23:32:44 once that's fixed Mar 12 23:32:58 so make a new copy of the bb and put a DEFAULT_PREFERENCE="-1" until it's fixed... Mar 12 23:33:12 and as I said, you can always re-path by using monotone diff -r.... etc. Mar 12 23:33:17 re-patch Mar 12 23:33:36 why do i have to push this crap again though? this is a problem with the way the OE developer process works Mar 12 23:33:57 i've tried to think of every case where i did something wrong Mar 12 23:34:05 i was wondering, do we have a new page with all the valid branches listed anymore? Mar 12 23:34:26 the only one i can think of is pushing a new version bump of wpa-supplicant, since it is already maintained... but i'd brought that up many weeks ago Mar 12 23:34:33 it had an oe bugtracker ticket filed Mar 12 23:34:36 that was getting stale Mar 12 23:34:56 i mean, this is just so rediculous. and people are like "yeah i back reenoo's decision" Mar 12 23:35:07 reenoo didn't ask me about anything. Mar 12 23:35:12 no, he shouldn't have reverted a change unless it was blatantly *wrong* Mar 12 23:35:16 esp in .dev Mar 12 23:35:33 unless it sat around for a few days Mar 12 23:35:34 * NAiL agrees on that Mar 12 23:35:38 i thought very clearly about what i wanted to push. and pushed it. and it was reverted. Mar 12 23:35:56 so push it again Mar 12 23:36:10 lol Mar 12 23:36:13 i keep thinking though, i ... Mar 12 23:36:14 i mean Mar 12 23:36:18 that wouldnt be a good idea Mar 12 23:36:20 that can't be right Mar 12 23:36:40 there is something wrong with me, from the stand point of people in OE who feel it is their job to revert changes Mar 12 23:36:53 and unless i figure that out, i'm not keen on spending any more time helping the project :/ Mar 12 23:36:58 which saddens me Mar 12 23:37:07 i want to find resolution on this within the next few months Mar 12 23:37:09 1 person did something a little stupid and mean Mar 12 23:37:11 calm down please Mar 12 23:37:24 i'm pretty calm, just vocal. it's okay Mar 12 23:37:34 it should be "resolved" fairly quickly...when the ML is up Mar 12 23:37:38 jnc can you not just fork dev, get peopel to test your changes and then merge in a week or so? Mar 12 23:37:44 * jnc prays for mailinglist tiki Mar 12 23:39:44 i've no problem pulling an extra branch to compile against Mar 12 23:39:56 emte: i really wanted to get to the qemu stuff Mar 12 23:40:00 :/ Mar 12 23:40:04 sorry man Mar 12 23:40:07 not tonight Mar 12 23:40:38 * emte curses at qemu Mar 12 23:40:58 i hope it was fixed, i am about to find out Mar 12 23:41:21 zecke's comment was a bit cryptic Mar 12 23:41:51 * johnX stable sylpheed in all of it's forms Mar 12 23:42:04 s/stable/stabs Mar 12 23:42:08 lol Mar 12 23:42:15 cmon Mar 12 23:42:19 wheres the love? Mar 12 23:42:30 who actually uses this email stuff anyways? Mar 12 23:42:43 why don't we just include xchat in the default image and call it a day? Mar 12 23:42:44 specifically i use -claws Mar 12 23:43:00 doesnt xchat already live in OZ? Mar 12 23:43:09 what's the difference? Mar 12 23:43:14 what is all this -claws business? Mar 12 23:43:32 -claws is a modified sylpheed to support modules Mar 12 23:44:08 sylpheed itself is missing quite a lot of nice features Mar 12 23:44:13 basically, I'm trying to get some version of sylpheed packaged that doesn't depend on openssl or those openssl compat patches applied to gnutls Mar 12 23:45:08 johnX, use 1.3.4 it doenst use taht patch Mar 12 23:45:09 mozilla-thunderbird might be a more real target FYI, though i sort of preferred sylpheed-claws when i had a x586 Mar 12 23:45:35 thunderbird is a bit heavy for an arm chip i'd think Mar 12 23:45:52 not sure it would even run on my h3 Mar 12 23:45:56 h36* Mar 12 23:47:25 then again -claws is using 5M ram on my host ... Mar 12 23:47:51 emte: sylpheed is a nice mail client. I'm just tired of it not building/populating/staging or requiring openssl Mar 12 23:48:31 emte: actually, that's the whole issue: 1.3.4 doesn't provide the openssl compatibility that sylpheed 1.9 requires to use ssl Mar 12 23:48:57 sylpheed no longer requires openssl atall Mar 12 23:49:10 * emte should qualify that Mar 12 23:49:34 i explicitly disable openssl and enable gnutls when i build sylpheed Mar 12 23:49:43 claws? Mar 12 23:50:03 yes, the last 3/4 times i've compiled it Mar 12 23:50:16 i am probably due for another update tho Mar 12 23:51:54 1.0.5 vs 2.0 Mar 12 23:52:38 yeah ... Mar 12 23:52:45 anyway brb Mar 12 23:53:16 ok Mar 12 23:53:36 though I am just not seeing any option in the configure script to enable or disable gnutls support for claws or mainline Mar 12 23:54:14 all I see is --disable-ssl Disable SSL support using OpenSSL Mar 12 23:56:51 http://www.handhelds.org/hypermail/oe-commits/2/0243.html Mar 12 23:56:57 apparently XorA wrote the patch Mar 12 23:57:08 or he sourced it from somewhere else, it was from him either way Mar 13 00:02:22 03jnc 07org.oe.dev * r5b023a0b... 10/packages/ (gnutls/gnutls_1.3.5.bb libgcrypt/libgcrypt_1.2.2.bb): (re-)Addition of gnutls and libgcrypt packages/bumps, taking maintainership. Please report breakage to maintainer kthx. Mar 13 00:02:46 * jnc shrugs Mar 13 00:02:53 if he doesn't like it, he can be the maintainer Mar 13 00:03:03 i want to make this stuff work Mar 13 00:03:36 emte: for the qemu issues you ran into, what's your build arch.... ia32? Mar 13 00:10:31 ok, sylpheed claws is happy to build with a .bb that I made out of the standard sylpheed .bb...but I get this error message: http://pastebin.com/598825 Mar 13 00:10:49 * jnc reads Mar 13 00:11:10 that should be pretty straightforward Mar 13 00:11:16 the manual page installation is poofing out Mar 13 00:11:58 "Directory not empty" is probably a cyrptic way of saying access denied Mar 13 00:12:13 hmm Mar 13 00:13:32 permissions are fine, plenty of disk space Mar 13 00:13:55 anyways, I'm too tired to work on this right now Mar 13 00:14:13 i mean, permission to move stuff from your /usr .... into the other thing Mar 13 00:14:22 that obviously should not succeed unless you are root Mar 13 00:14:33 and if you were root, it would have just eaten your host build machine's man page dir Mar 13 00:14:41 it says ./usr/share/man/man1 Mar 13 00:14:46 * jnc blinks Mar 13 00:14:50 ieee. it does Mar 13 00:15:11 I think it's working out of /home/bitbake/tmp/work/armv5te-linux/sylpheed-claws-2.0.0-r3/image Mar 13 00:15:20 and ./usr/share/blah/blah is relative to that Mar 13 00:15:47 but it doesn't emit logs, and I'm too tired to research it right now Mar 13 00:15:55 okay Mar 13 00:15:58 good work :) Mar 13 00:16:28 however, if this gets sorted out magically it should provide a version of sylpheed that doesn't rely on openssl or gnutls openssl compat stuff Mar 13 00:16:47 but good night for now Mar 13 00:20:52 jnc, yes Mar 13 00:21:10 i've the most compatable arch known to man Mar 13 00:21:14 :) Mar 13 00:21:17 emte: if you don't mind, could you have another whack at it? i'm on amd64 and the qemu stuff is disabled by default Mar 13 00:21:40 i would need some more specifics to take a look at it and i'm kind of lacking that on my build box :/ Mar 13 00:22:34 jnc, that might be what zecke was refering to. all he said was that koen|sleep commented it out, but i am not positive if he ment qemu Mar 13 00:23:31 starting a compile now Mar 13 00:23:59 thanks Mar 13 00:40:20 03jnc 07org.oe.dev * rc04eee12... 10/packages/wpa-supplicant/ (4 files in 2 dirs): (re-)Adding updates to wpa-supplicant, myself as secondary maintainer to handle fixes on bugtracker. This is completely untested, breakage expected. Please report ASAP. Mar 13 00:47:22 so far at gettext ... Mar 13 00:51:51 same Mar 13 00:51:51 heh Mar 13 00:52:03 and my CPU is way fewer clockspeed Mar 13 00:52:25 that would be fun, see what the difference is between machines in compiling gpe-image Mar 13 00:53:41 you should finish first Mar 13 00:53:46 i am using firefox Mar 13 00:54:04 has bad effects on python Mar 13 00:54:09 firefox in the image hmm? Mar 13 00:54:13 no Mar 13 00:54:18 firefox to do my taxes Mar 13 00:54:28 :) Mar 13 00:54:31 and that you say, affects another process? Mar 13 00:54:42 it affects python Mar 13 00:54:50 sounds weird Mar 13 00:54:56 i don't have that experience here Mar 13 00:55:14 if you want to see the effect, toss your cache, start ff, and start bitbake Mar 13 00:55:24 that's what i did Mar 13 00:55:37 takes 1/3rd to 2x longer to parse Mar 13 00:55:43 ohhh that Mar 13 00:55:49 it's just proc activity Mar 13 00:56:20 no, even when the ff process is idle/zombie its the same Mar 13 00:57:01 i've ~30 tabs running in opera and it doesnt affect pythin/bitbake Mar 13 00:57:01 you might have swap issues. what does 'swapon -s' say ? Mar 13 00:57:51 Filename Type Size Used Priority Mar 13 00:57:51 /dev/hda3 partition 120476 52 -1 Mar 13 00:57:51 /dev/hda5 partition 120416 0 -2 Mar 13 00:57:51 /dev/hda6 partition 120448 0 -3 Mar 13 00:57:51 /dev/hda7 partition 120448 0 -4 Mar 13 00:57:51 /dev/hda8 partition 120448 0 -5 Mar 13 00:57:53 /dev/hda9 partition 192740 0 -6 Mar 13 00:59:43 what on hell do you have so many swap partitions for Mar 13 00:59:56 legacy habit actually Mar 13 01:00:04 i should stop doing that Mar 13 01:01:06 you're hurting the performance of that system, in the case of using a vanilla 2.6 linux kernel Mar 13 01:01:24 geez..... Mar 13 01:01:27 not really Mar 13 01:01:57 its rare that i ever use any swap Mar 13 01:02:20 i need to move my shm threshold lower tho Mar 13 01:02:46 that has actually had the biggest impact on performance Mar 13 01:03:24 i should check my logs to see if it has peeked over 200M yet Mar 13 01:03:34 peaked* Mar 13 01:03:50 okay as a point in example here, i have a pretty much untweaked system. i compiled my own kernel, sure, but who doesn't these days Mar 13 01:03:55 it works fine. Mar 13 01:04:20 your computer, has six swap partitions, a tweaked SHM thingy, who knows what else Mar 13 01:04:26 and it's not good enough? Mar 13 01:04:49 i wouldnt say tweaked Mar 13 01:05:00 and i didnt compile this kernel Mar 13 01:05:21 shm is set to default to allocating 1/2 your ram Mar 13 01:05:47 i just dont have enough posix processes to jsutify allocating it 1gb ram Mar 13 01:06:07 i have it currently set to 750Mb i think Mar 13 01:06:30 emte: you are wrong, sir Mar 13 01:06:36 am i? Mar 13 01:06:47 about which? Mar 13 01:06:58 the default is to limit itself to _maximum_ (read extreme case) of half your physical ram not inclusive of swap Mar 13 01:07:20 with a proper swap setup, that should never happen Mar 13 01:07:34 if it did, the OOM killer would deadlock Mar 13 01:08:05 it would always assign itself 1gb ram Mar 13 01:08:24 and i changed it to 256 by the looks actually Mar 13 01:08:35 * jnc sighs Mar 13 01:08:57 but i'll change it back if you think that will help Mar 13 01:09:17 i think that you are changing defaults which are defaults for a very good, and well thought out reason Mar 13 01:09:19 i have yet to ever see it over 1% Mar 13 01:09:44 and i also think you do not understand nor take the time to understand what all those numbers mean Mar 13 01:10:00 suffice it to say, the defaults are fine please leave them alone Mar 13 01:10:36 it is common for the linux 2.6 kernel series to suck up all of your ram in cache, or something that would look very odd on other operating systems Mar 13 01:10:40 it's rather normal Mar 13 01:11:10 when i did the reasearch on whay/why shm was at 1gb the kernel forum stated it was an arbitrary setting for compatability accross systems and tehre was no way to figure out what value to set it at since they couldnt possibly know the purpose of the system Mar 13 01:11:36 s/whay/what* Mar 13 01:12:34 oh btw Mar 13 01:12:40 qemu exists and failed Mar 13 01:12:44 :( Mar 13 01:12:54 what revision are you at? Mar 13 01:13:02 (monotone rev) Mar 13 01:14:47 Revision: eb88f31cce7a934eb073586d2941438672672c43 Mar 13 01:15:03 dump PR was the last commit Mar 13 01:15:30 oh Mar 13 01:15:42 try going to c04eee126492bac75d6b727b42d5dd521f149558 Mar 13 01:16:17 monotone up -r c04eee126492bac75d6b727b42d5dd521f149558 Mar 13 01:16:17 monotone: misuse: no such revision 'c04eee126492bac75d6b727b42d5dd521f149558' Mar 13 01:16:34 hmm Mar 13 01:16:38 actually that shouldn't matter Mar 13 01:17:00 ? Mar 13 01:17:50 are you sure you rm-fr'd your oetmp and sources dirs? Mar 13 01:19:28 you shouldn't ever have to remove your sources... Mar 13 01:19:57 JustinP: due to things not being tagged properly with srcdate, yeah, you do. Mar 13 01:20:15 until the policy is fixed... i don't know what else to do about it Mar 13 01:21:06 emte: make triple sure that ENABLE_BINARY_LOCALE_GENERATION is not being set Mar 13 01:21:32 or if you want to try something crazy, explicitly set it to "0" Mar 13 01:22:13 jnc: what? srcdate? Mar 13 01:22:19 mmhm yep Mar 13 01:22:26 jnc: the sources for a date will not change Mar 13 01:22:51 ? Mar 13 01:22:59 where does ENABLE_BINARY_LOCALE_GENERATION live? Mar 13 01:23:22 emte: if you don't already know, then it shouldn't be set. Mar 13 01:23:39 and how would that effect qemu detecting gcc-4.1 on my system? Mar 13 01:23:54 the srcdate uses the very beginning of the day...and it's never in the future, so the sources can't change Mar 13 01:28:03 emte: gcc-4.1 ? Mar 13 01:28:06 what? Mar 13 01:28:15 NOTE: package glibc-2.3.5+cvs20050627-r3: task do_package: started Mar 13 01:28:15 NOTE: generation of binary locales disabled. this may break i18n! Mar 13 01:28:19 that's what i get Mar 13 01:29:17 sorry gcc-4.0.1 Mar 13 01:29:40 i have both 4.0.1 and 3.3.5 Mar 13 01:30:16 emte: hey, it shouldn't matter what version of gcc you have on your host system Mar 13 01:30:28 correct it shouldnt Mar 13 01:30:30 emte: there's a cross-initial / cross compiler built Mar 13 01:30:53 but qemu donst have proper autotool scripts so it ignores the bitbake variables Mar 13 01:31:34 it should not build Mar 13 01:32:57 * emte wonders what that comment is in referance to Mar 13 01:33:55 the qemu should not build Mar 13 01:34:05 it wouldn't be a problem if qemu did not build Mar 13 01:34:34 it doesnt build for me, bit it also kills bitbake Mar 13 01:34:40 but* Mar 13 01:34:48 i cant build anything because of it Mar 13 01:35:13 okay, first can you please post your conf to pastebin Mar 13 01:35:19 i'll look that over really quick Mar 13 01:36:49 secondly can you 'mv org.openembedded.dev o.o.d~qemutrouble; monotone --db=(your oe.db file) --branch=org.openembedded.dev co' Mar 13 01:39:04 the qemu conf? local.conf? Mar 13 01:39:10 local.conf Mar 13 01:39:22 sure, nthing special there Mar 13 01:39:30 cept my sysv patch Mar 13 01:41:48 if you prefer to make the local.conf available on a webserver somewhere, that would also be fine Mar 13 01:42:22 oe.pastebin.com Mar 13 01:44:34 http://oe.pastebin.com/598941 Mar 13 01:45:03 hmm i should remove thet host_gcc_ver flag i was playing with Mar 13 01:49:32 consider my local.conf as an example of something that does work... http://oe.pastebin.com/598946 Mar 13 01:49:35 done, restarting build Mar 13 01:50:15 taht is very little different from mine Mar 13 01:50:31 cept familiar specifies gcc 3.3.5 Mar 13 01:50:34 er Mar 13 01:50:36 3.4.5 Mar 13 01:50:40 one of the two Mar 13 01:51:11 but like i said in the end they dont matter Mar 13 01:51:34 and the BBMASK Mar 13 01:52:26 lol Mar 13 01:52:26 plus the host gcc thing, which should not matter Mar 13 01:52:41 i.e. do not set it Mar 13 01:53:04 any sufficiently new version of GCC is able to compile another version of GCC Mar 13 01:54:16 Mar 11 19:42:10 emte # default parameters Mar 13 01:54:16 Mar 11 19:42:10 emte prefix="" Mar 13 01:54:16 Mar 11 19:42:10 emte interp_prefix="/usr/gnemul/qemu-%M" Mar 13 01:54:16 Mar 11 19:42:10 emte static="no" Mar 13 01:54:16 Mar 11 19:42:10 emte cross_prefix="" Mar 13 01:54:17 Mar 11 19:42:10 emte cc="gcc" Mar 13 01:54:19 Mar 11 19:42:10 emte host_cc="gcc" Mar 13 01:54:22 thats why qemu fails Mar 13 01:55:15 it will not pass unless i relink gcc to gcc-3.3 Mar 13 01:55:27 relink? Mar 13 01:55:35 on my host Mar 13 01:55:35 you should be using update-alternatives Mar 13 01:55:43 because it does not obey bitbake Mar 13 01:55:55 qemu should never be compiled at all Mar 13 01:57:05 qemu is apperently required now for glibc locales, wasnt that what kergoth said lastnight? Mar 13 01:57:14 not required Mar 13 01:57:18 well Mar 13 01:57:25 it is required if you want all those locales Mar 13 01:57:32 it is not required to build an OE target Mar 13 01:57:40 ... Mar 13 01:57:45 you'll end up with a build that has broken i18n, not a big deal really Mar 13 01:57:49 it can be fixed later Mar 13 01:57:54 it should be nonfatal Mar 13 01:57:55 how do you compile an image without glibc? Mar 13 01:58:09 what are you on about Mar 13 01:58:23 i have no clue now Mar 13 01:58:38 set the binary locale thing Mar 13 01:58:56 i said that like a half hour ago. what am i doing wrong Mar 13 01:59:12 you told me to set it explicitly to 0 somewhere Mar 13 01:59:18 no, "" Mar 13 01:59:38 i have yet to figure out where taht is yet Mar 13 01:59:44 JustinP: actually, i'm not sure setting it to nothing would be the same as unsetting it Mar 13 02:00:03 according to the line that "disables" qemu bits, it does an or-equal assignment to "0" Mar 13 02:00:27 maybe i should have said xor-equal Mar 13 02:00:37 "if not set, set it to this" i meant Mar 13 02:00:39 emte: local.conf! Mar 13 02:04:22 nice ... Mar 13 02:04:45 emte: always.... Mar 13 02:05:33 i get lectured about doing things properly so that it doesnt affect other's configurations and this qemu problem comes along and wrecks the whole system Mar 13 02:06:19 * emte wonders if doing this at the same time i am doing my taxes is/was a great idea ... Mar 13 02:07:45 hah. that sounds like a bad time, sorry to hear Mar 13 02:07:59 i hope everything works out, taxes and nosysv patches Mar 13 02:08:24 nonsysv works perfectly Mar 13 02:08:31 jsut has to be commited Mar 13 02:08:48 well a few more tests would be nice first Mar 13 02:09:04 not sure anyone else besides myself needs it yet Mar 13 02:09:04 what init are you using Mar 13 02:09:09 bsd style init? Mar 13 02:09:26 debian, so horde would be my guess Mar 13 02:09:35 i havent looked Mar 13 02:09:51 but all the sysv stuff is "compatability scripts" Mar 13 02:10:03 doesnt help with binaries using it Mar 13 02:10:04 debian uses sysvinit Mar 13 02:10:12 i just checked my system Mar 13 02:10:13 etch isnt apperently Mar 13 02:10:34 but i'll look Mar 13 02:11:00 jnc@baker:~$ dpkg-query -s sysvinit | grep Status Mar 13 02:11:00 Status: install ok installed Mar 13 02:11:12 jnc@baker:~$ cat /etc/issue Mar 13 02:11:12 Debian GNU/Linux testing/unstable \n \l Mar 13 02:13:02 i wonder where my mind goes some days Mar 13 02:13:10 yes i am using sysvinit Mar 13 02:13:24 non my system does not support sysv-ipc Mar 13 02:13:26 no* Mar 13 02:30:28 jnc, i think you addressed a legitimate concern ... i need to change the name of that flag Mar 13 02:33:53 there is really only the sysvinit, and nonsysvinit (a.k.a. bsd style init) Mar 13 02:34:06 i'm not sure what gentoo uses, but it is not sysv and is instead dependency based Mar 13 02:34:15 i think sysvinit is the most appropriate system Mar 13 02:34:30 * emte likes the debian init Mar 13 02:34:37 even if they dont use it anymore Mar 13 02:34:56 together with a helper for init scripts, like rcconf (debian), it's n00b friendly enough Mar 13 02:35:23 emte: it may not be sysvinit exactly, in debian Mar 13 02:35:27 it is sysvinit-alike Mar 13 02:35:32 i cant actually find which packaged provides sysv-ipc ... Mar 13 02:36:19 dpkg-query -S /sbin/init Mar 13 02:36:37 its simular, but it didnt have the rcX.d stuff used only init.d and a single file that specified what was run in each runlevel Mar 13 02:36:46 quite a lot cleaner system Mar 13 02:36:55 the gentoo style? Mar 13 02:37:04 oh Mar 13 02:37:36 it's up to you for writing the config file changes to a bazillion packages Mar 13 02:38:01 not to say you shouldn't, just that it's an awful lot of work when sysvinit could be improved Mar 13 02:38:38 no, you just listed the runlevels and what script to exicute in which priority Mar 13 02:38:57 basically it just cleaned up sysvinit Mar 13 02:39:08 no, not really Mar 13 02:39:41 instead of 6 directories with symlinks you had one file with the smae info Mar 13 02:39:46 same* Mar 13 02:39:57 sounds difficult to automate Mar 13 02:40:43 not really, all the scripts were the same Mar 13 02:40:48 my build is still going FYI Mar 13 02:40:51 it's on mtd now Mar 13 02:40:55 * emte checks Mar 13 02:41:04 ... Mar 13 02:41:04 | ERROR: "gcc" looks like gcc 4.x Mar 13 02:41:04 | QEMU is known to have problems when compiled with gcc 4.x Mar 13 02:41:04 | It is recommended that you use gcc 3.x to build QEMU Mar 13 02:41:04 | To use this compiler anyway, configure with --disable-gcc-check Mar 13 02:41:05 | FATAL: oe_runconf failed Mar 13 02:41:12 same same Mar 13 02:41:17 must be a familiar thing :/ Mar 13 02:41:24 i'll use yours Mar 13 02:41:43 try building with my exact configuration, see if it works. nevermind that it is not for your target machine Mar 13 02:42:06 if that works, let me know and i will try using your config on my build box Mar 13 02:46:07 changed DL_DIR, BBFILES, and removed TMPDIR from yours Mar 13 02:46:31 just so ya know Mar 13 02:47:13 okay :) Mar 13 02:47:41 makes sense since my paths arnt yours :) Mar 13 02:47:47 i'm slightly maniacal when arriving at the topic of where things are stored on my filesystem Mar 13 02:48:08 i. want. to know. :) Mar 13 02:48:13 my conf isnt that much diff in that respect Mar 13 02:48:27 especially since it doesnt live in ~ Mar 13 02:49:59 i need to figure out how to specify multiple repos under the new bitbake ... Mar 13 02:50:16 easier to diff that way Mar 13 03:00:31 jnc: I don't know about the project conventions, but one option that tools make available is to just commit things that are controversial or known to break on a branch -- that way, one doesn't have to satisfy everyone who works on mainline, until after one has had a chance to work on things a bit, get help from multiple people, get test reports, etc. Mar 13 03:01:15 njs: thanks Mar 13 03:24:19 whee Mar 13 03:24:30 now to start my studentloan app Mar 13 03:47:21 jnc, it skips qemu with your conf Mar 13 03:48:19 i wonder if its the familiar distro that is calling it Mar 13 03:49:13 changed distro, rebuilding Mar 13 04:45:18 jnc, its familiar Mar 13 04:45:26 it just failed Mar 13 04:49:38 how do you file a bug that blocks a whole distro? Mar 13 05:16:23 * france is back (gone 21:36:25) Mar 13 05:17:42 http://bugs.treke.net/show_bug.cgi?id=738 Mar 13 05:17:58 france, that may be of significance to you Mar 13 05:18:07 just noticed you came back btw Mar 13 05:26:48 * emte wonders what happens if i try to buils oz for the h36 Mar 13 05:42:16 * emte fixes chicken Mar 13 05:47:36 emte: looking..... thanks.. Mar 13 05:57:45 not exactly good news for you Mar 13 06:14:27 emte: thanks for trying the oz config Mar 13 06:14:38 if familiar has a depend on something with qemu... hmm Mar 13 06:14:46 that's definitely a bug at this present point in time Mar 13 06:15:40 jaits a problem in angstrom as well Mar 13 06:15:43 its* Mar 13 06:15:53 just failed Mar 13 06:16:35 my guess is someone fixed it for OZ but nothing else Mar 13 06:16:56 makes sense seeing as most of the devs work on OZ Mar 13 06:17:14 emte: not horrible...there are alternate solutoins (If need have too) Mar 13 06:19:34 yeah Mar 13 06:20:05 emte: there is a solutions to every issue, just some are more work than others.... Mar 13 06:20:37 true, but qemu shouldnt have been forced into the toolchain without more testing first Mar 13 06:20:49 emte: exactly.... Mar 13 06:21:13 emte: it brings up the issue of how stable OE is and an it be trusted. Mar 13 06:21:35 emte: it is goog timeing as I am looking at all the options for familiar. Mar 13 06:21:54 s/and an/and can/ Mar 13 06:23:39 emte: qemu is working well under scrathbox, been looking at build root as well. Lots of things need to be fixed and I am looking for a firm place to stand. Mar 13 06:23:55 emte: might even fix issues in OE, too early to tell. Mar 13 06:27:12 lets see if opensimpad fails ... Mar 13 06:34:03 emte: I wonder if the testsuites were even ran? Mar 13 07:01:15 dindt make it far enough in opensim for it to get to qemu, guess it doesnt build for an arm-simpad :P Mar 13 07:04:50 hey!! the mailserver is back ... and it looks like i use it way too much :P Mar 13 07:07:48 emte: wsa the mail server down? Mar 13 07:08:22 yeah, for about the last 4 days Mar 13 07:08:41 which one? Mar 13 07:08:53 hi all Mar 13 07:08:57 mailinglist Mar 13 07:09:11 hh.org ? Mar 13 07:09:15 yup Mar 13 07:09:23 nobody told me....checking.... Mar 13 07:09:24 can anyone tell me where I'm supposed to look for rootfs images from 'bitbake gpe-image' and 'bitbake opie-image'? Mar 13 07:09:41 Farcaller, deploy Mar 13 07:09:55 emte: shout at me next time! Mar 13 07:10:13 france, i thought you knew ... Mar 13 07:10:18 emte, there's only ipk dir Mar 13 07:10:31 people have been saying for the last few days it was being fixed Mar 13 07:10:42 emte: news to me...I was wondering why it was a quiet weekend... Mar 13 07:10:46 lol Mar 13 07:11:12 yeah ... on the server Mar 13 07:11:31 jnc would have prefered it to have been running this weekend Mar 13 07:11:40 emte: my last e-mail was saturday at 5pm Mar 13 07:11:51 emte: ping me next time. :-D Mar 13 07:12:25 yeah i will do Mar 13 07:12:31 thx Mar 13 07:12:46 i take it this means your not running nagios? Mar 13 07:13:41 emte: nope, someone setup big brother, many years back, but it stopped working... Mar 13 07:13:41 or an equiv monitoring tool Mar 13 07:13:46 ah Mar 13 07:13:59 emte: it is on the todo list, for after the infrastructure stuff is finished. Mar 13 07:14:15 emte: no-one noticed that the ftp server was moved this weekend. Mar 13 07:14:43 emte: failover is a good thing... Mar 13 07:14:55 i think they probably thought it was the same issue as the mailserver Mar 13 07:15:04 if anyone did notice Mar 13 07:15:19 emte: it was only down for 30 seconds during dns propigation Mar 13 07:15:24 then again, most of us have our sources cached Mar 13 07:15:51 ah, good stuff Mar 13 07:17:01 anyway nagios is a fairly nice tool Mar 13 07:17:33 used it for a while in an adaptive irc network Mar 13 07:17:35 reading now, comparing to bb Mar 13 07:19:28 heya zecke Mar 13 07:22:21 moin Mar 13 07:22:30 emte: did you push your fakeroot changes yet? Mar 13 07:23:26 no, i dont have push access Mar 13 07:23:29 morning Mar 13 07:23:47 and the morning begins with a nasty ubuntu issue Mar 13 07:24:04 root password disclosure ? ;) Mar 13 07:24:23 yup Mar 13 07:24:25 emte: ah, sorry. I will compile and push then Mar 13 07:24:31 lol Mar 13 07:24:38 zecke, no hurry Mar 13 07:24:40 nite folks.... Mar 13 07:24:43 emte: thanks for the help Mar 13 07:24:44 night Mar 13 07:24:47 np Mar 13 07:25:04 I'm glad that I do not run any public servers Mar 13 07:25:26 zecke, i should probably change the flag more accuratly to HOST_NONSYSV_IPC Mar 13 07:25:37 it could be confused with sysvinit Mar 13 07:25:37 * france is away: Away Mar 13 07:26:07 altho if i dont change it, it could double both ways ... Mar 13 07:27:45 zecke, my newest bug tho is a bit more important then my sysv patch ... http://bugs.openembedded.org/show_bug.cgi?id=738 Mar 13 07:28:20 as far as i can tell OZ is the only distro that compiles now Mar 13 07:28:46 moin moin Mar 13 07:32:29 emte: qemu? Mar 13 07:35:36 france, yeah Mar 13 07:36:29 * france is back (gone 00:10:51) Mar 13 07:36:45 emte: I have been running it for a while under scratchbox Mar 13 07:36:47 works really well Mar 13 07:36:51 even compiles qmail, out of the box. Mar 13 07:37:12 no tweaking to the files, not external arm board needed. Mar 13 07:37:15 I was surprised Mar 13 07:37:28 yeah, its current state in OE it detects gcc4.x and bitbake fails Mar 13 07:38:12 there is code to bypass it but it seems it only work on/for OZ Mar 13 07:39:05 i am not a fan of emus Mar 13 07:39:36 but the only ones i've used were hardware and software ones for microchip and altera devices Mar 13 07:39:37 emte: okay, I will leave to Uni now Mar 13 07:39:59 me neither, it is why I have an arm board, that I run ./configure on along with the testsuites when I cross build Mar 13 07:39:59 zecke, have fun in class Mar 13 07:40:02 zecke: hi Mar 13 07:42:32 anyway I have to get up in 3 hours, I better call it anight. Mar 13 07:42:35 a night Mar 13 07:42:42 * france is away: Away Mar 13 07:42:47 yeah, i've class tomorrow too Mar 13 07:42:50 night Mar 13 07:42:52 :-) Mar 13 08:01:33 :( Mar 13 08:01:46 * emte chases after the email i shoundt have jsut sent ... Mar 13 08:02:50 oh well Mar 13 08:03:19 emte: it happens Mar 13 08:03:22 lol Mar 13 08:03:28 did you read it yet? Mar 13 08:03:34 lol Mar 13 08:03:47 nope, just doing clean-up of stuff,now that I am getting e-mail again Mar 13 08:04:35 where did you send it to? Mar 13 08:04:45 oe@handhelds.org Mar 13 08:04:50 boss? girlfriend? Mar 13 08:04:54 heh, not too bad. Mar 13 08:05:25 i think it was a little bit better than my last one ... Mar 13 08:05:32 but not much ... Mar 13 08:07:54 heh, ti will get better with time Mar 13 08:08:19 yeah, i need to cripple sylpheed so i cant send email Mar 13 08:08:30 lol Mar 13 08:08:40 I still have not seen your e-mail, it can not be that bad Mar 13 08:08:44 * france is back (gone 00:26:03) Mar 13 08:09:04 i just picked it up off tyeh list so you should get it soon Mar 13 08:09:09 the* Mar 13 08:11:46 * emte is going to bed Mar 13 08:11:49 night Mar 13 08:11:52 lol Mar 13 08:12:05 that is what I have been saying, but just one more thing to be done Mar 13 08:12:18 i finished my taxes earlier Mar 13 08:12:24 very nice Mar 13 08:12:25 so only sleep now Mar 13 08:12:29 heh Mar 13 08:42:00 morning Mar 13 08:43:47 morning Mar 13 08:50:28 * france is away: Away Mar 13 08:55:52 morning Mar 13 08:56:49 hi CoreDump|home Mar 13 09:09:18 hi greentux Mar 13 09:26:22 morning all Mar 13 09:29:11 morning Mar 13 09:42:40 hmmm gnutls_1.3.5's configure fails with "libgcrypt was not found. You may want to get it from" Mar 13 09:48:51 hrw|work: ping Mar 13 09:48:54 morning folks Mar 13 09:49:44 hey mickeyl Mar 13 09:49:44 koen|sleep: ping Mar 13 09:49:55 hi XorA Mar 13 09:50:31 hail mickeyl Mar 13 09:50:35 hey pb_ Mar 13 09:54:15 good morning folks Mar 13 09:55:28 morning chris144 Mar 13 09:55:37 hi mickeyl Mar 13 09:55:48 is anyone familiar with multimachine.inc ? Mar 13 09:59:14 hrw|work: cheers for that input on offlineimap Mar 13 10:00:08 re Mar 13 10:00:16 hi zecke_ Mar 13 10:00:22 are you familiar with multimachine.inc? Mar 13 10:01:13 moi? Mar 13 10:01:15 non Mar 13 10:01:21 tant pis. Mar 13 10:01:27 i'll wait for hrw and koen then Mar 13 10:01:35 hrw|work: stop working Mar 13 10:01:39 koen|sleep: stop sleeping Mar 13 10:01:41 heh Mar 13 10:01:50 ph5 come here Mar 13 10:02:02 I'n watchting too much Chapel Show... Mar 13 10:02:11 "ph5 come here, *bitch*" Mar 13 10:09:03 * hrw|work was on briefing Mar 13 10:09:09 hi mickeyl Mar 13 10:09:22 hi hrw|work, i have a problem understanding multimachine.inc's effects Mar 13 10:09:32 can you have a look at http://pastebin.com/599423 Mar 13 10:09:43 looking.. Mar 13 10:09:44 it strikes backwards to me to see -native packages in armv5te subdirectory Mar 13 10:09:49 is that necessary? Mar 13 10:10:08 (NOTE: I'm using the latest bitbake) Mar 13 10:10:11 it happens Mar 13 10:10:21 ok, so it's nothing to worry about? Mar 13 10:10:32 yep Mar 13 10:10:45 excellent. build -> resume Mar 13 10:11:23 koen is multimachine expert Mar 13 10:14:07 zecke_: good morning Mar 13 10:16:07 good morning Mar 13 10:16:13 gnutls_1.3.5 is missinga dependency to libgcrypt. Manually building libgcrypt fixes the configure of gnutls. Can't fix ATM Mar 13 10:16:15 hi Bernardo Mar 13 10:17:57 * Bernardo wonders if neon can be made to use gnutls instead of openssl Mar 13 10:19:00 hrw|work: I build this weekend a .bb for a 802.11g usb key (for the cxx00), but won't submit it yet Mar 13 10:19:21 have to find out why it loads and seems to work, but wifi won't communicate Mar 13 10:22:13 for 2.4 or 2.6? Mar 13 10:22:19 fwiw libxcomposite-X11R7.0-0.2.2.2-r0 b0rks on compile but I believe that is known? Mar 13 10:22:30 CoreDump|home: up-to-date tree? Mar 13 10:22:43 yep, only a fe minutes old Mar 13 10:22:47 few Mar 13 10:23:15 monotone: 196 | 1156 | 0 | 0 | 0 Mar 13 10:23:48 monotone: already up to date at c04eee126492bac75d6b727b42d5dd521f149558 Mar 13 10:24:12 CoreDump|home: that should have been fixed Mar 13 10:25:17 hrw|work: 2.4 Mar 13 10:25:23 CoreDump|home: the next failure will be libxfonts Mar 13 10:25:33 with the inlaws there didn't have the time to build for 2.6 Mar 13 10:25:35 ~http://openembedded.pastebin.com/599454 Mar 13 10:25:51 jnc: ping Mar 13 10:26:15 zecke_: next fail for me is xdyinfo Mar 13 10:26:23 03freyther 07org.oe.dev * r024b0b68... 10/ (2 files in 2 dirs): (log message trimmed) Mar 13 10:26:25 conf/documentation.conf: Mar 13 10:26:27 Document the HOST_NONSYSV variable Mar 13 10:26:29 introduced by emte Mar 13 10:26:31 packages/fakeroot/fakeroot-native: Mar 13 10:26:33 There are systems without SYSV stuff. Introduce Mar 13 10:26:33 CoreDump|home: please post more Mar 13 10:26:35 a HOST_NONSYSV variable if you have such a host Mar 13 10:26:41 CoreDump|home: I'm interested in the include order Mar 13 10:26:56 zecke_: I'll upload the logs, hold on Mar 13 10:27:10 CoreDump|home: just one complete arm-linux-gcc line Mar 13 10:27:39 CoreDump|home: ph5 had a patch that looked good and pushed it to change the order of includes Mar 13 10:27:56 sec Mar 13 10:28:27 zecke_: http://openembedded.pastebin.com/599456 Mar 13 10:30:55 CoreDump|home: hmm could you try rebuilding? Mar 13 10:31:00 sure Mar 13 10:31:16 CoreDump|home: the issue is ${STAGING_INCDIR} shouldn't be that early Mar 13 10:31:32 (just the xcomposite package needs rebuilding) Mar 13 10:32:05 ah wait, this lib failed for me before. It was the reason I pulled in latest changes. I guess my $WORKDIR is still b0rked Mar 13 10:34:02 well, bbnew xlibs/libxcomposite_X11R7.0-0.2.2.2.bb didn't help. Which package exactly should I rebuild? Mar 13 10:34:26 hmm this should fix it Mar 13 10:34:56 FYI bbnew does a bitbake -c clean first Mar 13 10:36:27 http://openembedded.pastebin.com/599461 Mar 13 10:36:53 yep, the order of -I changed in the new build Mar 13 10:45:01 ~lart bbshell for depends bug Mar 13 10:45:02 * ibot shoves a crumpet down bbshell's throat, happy now?! Huh? Want some JAM with that? for depends bug Mar 13 10:59:21 * Bernardo weeps **** ENDING LOGGING AT Mon Mar 13 10:59:57 2006