**** BEGIN LOGGING AT Tue Jul 12 23:59:57 2005 Jul 13 00:19:11 hi all Jul 13 00:20:42 <__law__> hi Jul 13 00:21:16 <__law__> i have deleted a file in my monotone repo how can i get this back from server? Jul 13 00:21:38 <__law__> i only get this error monotone: misuse: 1 missing files Jul 13 00:22:23 <__law__> if taht happend with bitbake it restored the file from the server Jul 13 00:24:29 try "monotone revert " Jul 13 00:26:16 <__law__> jbowler, that works thanks Jul 13 01:13:25 can someone add orinoco-modules-cs in BOOTSTRAP_EXTRA_RDEPENDS for poodle-2.4.conf? Jul 13 01:13:32 thanks Jul 13 01:13:58 can't commit Jul 13 01:17:03 kestrel: yep Jul 13 01:17:11 good morning, btw. Jul 13 01:17:47 mornin too Jul 13 01:22:06 hi agian Jul 13 01:23:44 morning Jul 13 01:25:50 03mickeyl 07org.openembedded.dev * r347a9adf... 10/conf/machine/ (poodle-2.4.conf poodle-2.6.conf zaurus-clamshell-2.4.conf): zaurus machine configurations: add orinoco-cs to poodle, specify 0.13e for Zaurus 2.4 kernels since later versions don't work there Jul 13 01:35:55 mickeyl: thanks for that. i was looking at 0.15 orinoco for a couple of hours now ... couldn't figure out why Jul 13 01:36:36 how's 2.6 kernel on poodle? Jul 13 01:41:58 kestrel: http://www.cs.wisc.edu/~lenz/zaurus/ Jul 13 01:42:05 looks like it boots, but lots of devices are missing Jul 13 01:42:12 drivers, even Jul 13 01:42:51 good morning pb_ Jul 13 01:42:59 hey XorA Jul 13 01:46:36 mickeyl: i think you got a typo on poodle-2.4.conf... i think it should be machine/poodle-2.4.conf:PREFERRED_VERSION_orinoco-modules-cs = "0.13e" Jul 13 01:46:48 PREFERRED_VERSION_orinoco-modules-cs = "0.13e" Jul 13 01:46:56 and not PREFERRED_VERSION_orinoco-cs = "0.13e" Jul 13 01:47:00 :) Jul 13 01:47:14 and together with PREFERRED_VERSION_orinoco-modules= "0.13e" Jul 13 01:47:24 without the "cs" Jul 13 01:51:19 argh Jul 13 01:51:20 ~lart me Jul 13 01:51:20 * ibot burns mickeyl to a crisp with a laser Jul 13 01:52:04 mickeyl: good morning Jul 13 01:52:36 actually it needs to be orinoco-modules = "0.13e", because the .bb file name matters here, not the ipkg name Jul 13 01:52:42 * mickeyl fixes Jul 13 01:52:55 right... Jul 13 02:08:19 03mickeyl 07org.openembedded.dev * r8e26f209... 10/conf/machine/ (poodle-2.4.conf zaurus-clamshell-2.4.conf): zaurus-configs: fix stupid typos introduced in last commit Jul 13 02:23:24 03mickeyl 07org.openembedded.dev * rf47b4940... 10/packages/meta/opie-collections.inc: opie-collections: repair gathering OPIE_MORE_RDEPENDS depending on ROOT_FLASH_SIZE Jul 13 02:31:12 Hi, when i boot the image. the image hangs at after starting the syslogd and i dont get the login prompt. but i do get a login prompt on the lcd Jul 13 02:31:19 i dont know what the problem is ? Jul 13 03:20:39 hey Jul 13 03:20:49 Hi Zecke Jul 13 03:20:54 can I put some rule in my configure.in to always use gcc as linker? Jul 13 03:21:04 instead of changing each Makefile manually Jul 13 03:24:46 oe should be setting LD = "gcc" anyway. Jul 13 03:24:51 is that not happening for you? Jul 13 03:28:34 pb_: hey, it uses libtool+g++ to link, I want gcc and -lsupc++ Jul 13 03:29:41 ah, I see Jul 13 03:30:06 I can't think of any easy way to overcome that. Jul 13 03:30:36 I had hoped I could set CXXLINK and CXXLD from within configure.in and it would land in each Makefile.in Jul 13 03:33:46 that might work, yeah Jul 13 03:34:06 you'd probably need to do it near the end of configure.in, so that it overrides the values that autoconf chooses for itself Jul 13 03:35:34 how would I do it? Sorry I'm still an idiot of m4 Jul 13 03:35:48 AC_DEFINE(CXXLD, $(CC)) ? Jul 13 03:35:56 anyone experienced this before on a WLAN card: hermes_read_ltv() rid does not match type Jul 13 03:36:13 i'm using orinoco0.13e Jul 13 03:37:48 zecke: configure.in is just a shell script Jul 13 03:37:58 CXXLD="${CC}" Jul 13 03:38:32 ~fishslap zecke Jul 13 03:38:33 * ibot slaps zecke up side the head with a wet fish. Jul 13 03:38:52 * zecke will rely on autotools for the NeXT big thing Jul 13 03:44:40 pb_: thanks Jul 13 03:44:59 ~lart bad coders creating and copy 'n pasting insecure code everywhere Jul 13 03:45:00 * ibot executes killall -HUP bad coders creating and copy 'n pasting insecure code everywhere Jul 13 03:50:19 ~lart mozilla for not uploading firefox source Jul 13 03:50:20 * ibot pulls out his louisville slugger and uses mozilla's head to break the homerun record for not uploading firefox source Jul 13 03:58:35 is there a good alternative to sysv init? Jul 13 04:04:09 ~lart do13 Jul 13 04:04:10 * ibot slaps do13 around with a large trout Jul 13 04:04:20 ibot, that was a weak lart Jul 13 04:18:49 dang. my devboard keep on hardlocking Jul 13 04:20:12 ~lart interrupt handler Jul 13 04:20:13 * ibot runs at interrupt handler with an origami Swiss Army knife, and inflicts a nasty paper cut Jul 13 04:20:22 hi all Jul 13 04:20:29 hey bluelightning Jul 13 04:20:33 hey zecke Jul 13 04:20:35 bluelightning: how are you? Jul 13 04:20:45 zecke: not too bad, yourself? Jul 13 04:20:54 bluelightning: not too bad Jul 13 04:23:02 koobla: http://initng.thinktux.net Jul 13 04:23:20 I'm using it on my desktop atm Jul 13 04:23:38 bluelightning: thanks Jul 13 04:23:54 will it work on arm? Jul 13 04:24:31 not sure... I don't see why it couldn't be ported across though Jul 13 04:26:10 I recall a small amount of discussion a few months ago about adapting Apple's new init system (I forget its name) for use on handhelds... I don't think it went anywhere Jul 13 04:30:11 wel well - will look into it, if time allows. Jul 13 04:42:44 trying to create some device entries ; found out that if we want to create directories into /dev. it does not work. went through the source of makedevs.c and found that the function responsibe for creating directory entries does not do anything ? Jul 13 04:44:41 eJumbo: look at your init sciprs and check what they do to /dev Jul 13 04:46:08 [13:33] hey mickeyl, I lost your phone number Jul 13 04:46:08 [13:33] could you email it to me? Jul 13 04:47:09 mithro_: and give you his cell phone number Jul 13 04:47:13 (if he does not respond) Jul 13 04:47:32 mickeyl: i just tried orinoco0.13e on poodle with my own build, but i'm getting errors: hermes_read_ltv() rid does not match type. I see from the orinoco mailinglist, you said 0.15x doesn't work on ARM, but the orinoco people says it can Jul 13 04:48:14 mickeyl: have you tried compiling with gcc2.95 ? just like what was used in openzaurus build-root Jul 13 04:49:10 i am looking at image_ipk.bbclass which invokes makedevs giving device_table-minimal.txt as param Jul 13 04:50:56 i am not using devfs. Jul 13 04:55:36 kestrel: lol, that sounds like it works from monday to friday but not wednesday between 10h and 12h Jul 13 04:56:04 zecke: where you located? Jul 13 04:57:20 mithro_: in capital city (the capital city of germany) Jul 13 05:01:39 berlin? Jul 13 05:08:18 yes Jul 13 05:08:41 zecke: I'll staying in berlin for a couple of nights in about a week or so Jul 13 05:10:31 want to catch up, I'll buy you a Jul 13 05:23:39 zecke: did it work for you? Jul 13 05:24:09 * mithro pokes mickeyl you about at all? Jul 13 05:24:37 i'm getting errors: rid 0xXXX does not match type 0xXXXX Jul 13 05:24:41 hehe Jul 13 05:40:27 Hi folks, some newbie question. I need to add a package which builds in 2 steps, first native library and tools, which are used to generate additional source and then build crooss library and eventually tools. As far as I understand this is how it's split in OE as well. What would be an easy to understand package to look for example? I tried to understand gcc, but it kind of hard. Jul 13 05:41:59 <_chronic> help for oe/am newbie? final executable linking is failing through some sort of env pollution. it's trying to link against host libs (/usr/lib) instead of those in ${STAGING_LIBDIR} but can't find where ... tried explicitly setting it in do_compile(), but no luck yet... tia Jul 13 05:43:02 lazy_marmot: I don't know an easy package but try looking at kaffe Jul 13 05:48:09 zecke: Thanks! I'll look into it. One more question, when I add some library to DEPENDS in my meta package for example expat, it build file, but when I add INSTALL_PKG - it says that such ipk doesn;t exits, which is true, it was split into 3 ipk. Where can I see how the result will be packaged ? I couldn't figure this out from bb file. Jul 13 05:49:38 lazy_marmot: PACKAGES is by default ${PN} (Package name) Jul 13 05:49:57 lazy_marmot: but a bbfile is free to do stuff like PACKAGES = "one two three" Jul 13 05:50:10 lazy_marmot: and then packages one, two and three get created Jul 13 05:52:59 zecke: this is exactly how I thoight, but I couldn't see anything like this. I double check this once more. Thanks. Jul 13 05:53:19 lazy_marmot: but there is some magic involved Jul 13 05:53:44 lazy_marmot: if you just have a libfoo.so.2.3.4 'installed' a libfoo package will be built Jul 13 05:53:57 lazy_marmot: and the default packaging is PN, PN-dev, PN-doc Jul 13 05:54:48 pb_: can I set AM_CXXFLAGS from within a configure.in? Jul 13 05:55:05 zecke: I'm not sure. You might need to check the automake manual. Jul 13 05:57:24 zecke: exactly, the magic part is not obvious :) This actually was the question why library was splited to 3. Jul 13 06:02:23 hey zecke, you could email me mickeyl 's mobile? Jul 13 06:05:49 <_chronic> re oe/am ... hmmm .... config.status is reporting libdir=/usr/lib, not ${STAGING_LIBDIR} Jul 13 06:10:33 mithro: give me your email address Jul 13 06:11:08 mithro: ah I've office and cell phone number Jul 13 06:28:39 03xora 07org.openembedded.dev * r026dfb76... 10/packages/mplayer/mplayer-atty_1.1.5.bb: mplayer-atty: re-enabled its ability to render subtitles and OSD Jul 13 06:29:52 morning Jul 13 06:37:00 bblr Jul 13 06:58:52 hello? Jul 13 06:59:04 you still around zecke? Jul 13 06:59:17 afternoon Jul 13 06:59:43 how does one dial (from germany) a germany mobile phone number in internation format? Jul 13 07:00:24 mithro: +49 is the german prefix Jul 13 07:00:29 +49169.... Jul 13 07:00:44 or 0049169 Jul 13 07:00:54 so... Jul 13 07:01:04 i'm sitting in a hotel in germany Jul 13 07:01:09 i hae to dial 0 to get out Jul 13 07:01:27 0 0049... Jul 13 07:01:41 mithro: skip the +49 part Jul 13 07:01:48 mithro: and replace it with a '0' Jul 13 07:02:09 so 0 069..... Jul 13 07:02:10 mithro: +49177 -> 0177 Jul 13 07:02:27 okay, and what is mickeyl's real name :P Jul 13 07:02:33 mithro: Michael Lauer Jul 13 07:05:48 it appears to be ringing Jul 13 07:06:16 that's a good sign.. Jul 13 07:06:36 the subcriber you have call is currently unavalible :/ Jul 13 07:06:36 mithro: I can try to confirm the number :} Jul 13 07:07:01 mithro: the cell phone number is 100% correct, I've dialed it during LinuxTag Jul 13 07:07:36 maybe he switched it off.. or he's inside of some tunnel or so.. Jul 13 07:09:45 when I put in PACKAGE_ARCH = "armv5te" Jul 13 07:09:56 into machine.conf. I get internal compiler errors :( Jul 13 07:10:12 but I guess that Jul 13 07:10:23 I keep hitting the wrong keys, sorry.. Jul 13 07:10:28 * koen hums ICE, ICE, baby Jul 13 07:10:33 PACKAGE_ARCH doesn't have any effect on your compiler. Jul 13 07:10:35 but I guess that's more something for the arm-linux group.. Jul 13 07:10:47 it has not? Jul 13 07:11:07 I took it out and now it compiles correctly.. Jul 13 07:11:14 /j gpe Jul 13 07:12:36 I thought PACKAGE_ARCH is about the compiler options used when compiling packages.. Jul 13 07:12:44 what does PACKAGE_ARCH do? Jul 13 07:13:52 It sets the architecture used for packaging. Jul 13 07:15:31 so bitbake knows what to put into a package for that specific architecture? Jul 13 07:20:41 hi Jul 13 07:20:43 mithro: hey# Jul 13 07:21:06 mickeyl: that is an option as well Jul 13 07:23:56 03tmbinc 07org.openembedded.dreambox * r7151e10f... 10/packages/netkit-base/netkit-base-0.17/opendreambox/inetd.conf: adding opendreambox inetd.conf Jul 13 07:25:44 no, bitbake just puts in whatever it built. Jul 13 07:28:24 OT: We only won with five goals ahead.. something is seriously wrong Jul 13 07:28:42 RoboCup? Jul 13 07:28:44 yes Jul 13 07:29:07 and our mid-sizers did not lose yet Jul 13 07:32:17 hi Jul 13 07:33:13 mickeyl: i'll be out the front at 7:30PM right :P Jul 13 07:33:47 mithro: excellent. Parking is hard in the Breslauer Strasse, so perhaps I can pick you up "on the fly" :) Jul 13 07:34:07 yeah that would be an okay idea Jul 13 07:35:02 good. I'll be driving a silver BMW 3 series with license number F-M-8822 Jul 13 07:35:14 it's about 4:35 now, right? Jul 13 07:35:19 right. Jul 13 07:35:38 I've never really got the hang of your crazy time system ;) Jul 13 07:36:05 whats a BMW 3 series look like? 2 door, 4 door? Jul 13 07:36:20 it looks great Jul 13 07:36:21 heh Jul 13 07:36:24 no, seriously. 2 door Jul 13 07:36:33 sport car type thing? Jul 13 07:36:37 http://vanille.de/personality/car.html Jul 13 07:38:01 okay cool Jul 13 07:38:01 wow we've Xqt running here Jul 13 07:38:11 i'll see you in about 3 hours Jul 13 07:38:18 yep, see you. Jul 13 07:38:30 zecke: xqt on what machine?= Jul 13 07:40:27 mickeyl: the ROAD device Jul 13 07:47:53 zecke are you somehow involved with the robocup? Jul 13 07:52:55 morning drw Jul 13 07:53:03 Delta_: not that much as I would like to be Jul 13 07:53:16 Delta_: I'm a member of the FU Fighters (www.fu-fighters.de) Jul 13 07:53:16 hey mickeyl, zecke, etc. Jul 13 07:53:20 ~hail drw Jul 13 07:53:24 * ibot bows down to drw and chants, "I'M NOT WORTHY!!" Jul 13 07:54:18 hi ! Jul 13 07:54:42 they run on hc12's ? Jul 13 07:55:23 my diplome thesis is about a hexapod robot I am constructing.. Jul 13 07:56:07 Delta_: aeh you#re from the FU Berlin? Jul 13 07:57:09 Delta_: yes both small size and mid size use the hc12 micro controler **** BEGIN LOGGING AT Wed Jul 13 07:58:57 2005 Jul 13 07:59:05 i know they use 2 blackfin boards for dsp, and an xscale board for comms Jul 13 07:59:31 Delta_: sadly he seems uncapable of finishing... Jul 13 07:59:32 zecke: look under publications Jul 13 07:59:55 i know them personally, can forward you the mail addresses if interested Jul 13 07:59:57 I should put on a homepage about my robot, eh :) Jul 13 08:00:38 I hope to get my robot finished till december.. Jul 13 08:01:40 I am working on the pcbs for the dsp modules now..I startet the robot with a hc11, but the dsp is more powerful.. Jul 13 08:02:47 zecke: why isn't it to be finished? no time? Jul 13 08:03:12 yay blackfin Jul 13 08:03:24 hey jamey Jul 13 08:03:44 hi zecke Jul 13 08:42:58 pb_: do you know who checked the linear-h2200 code into tslib? Jul 13 08:43:24 the auotomake check is broken but mailing mvista is like a black hole Jul 13 08:43:42 zecke: dunno. does "cvs log" not give any clues? Jul 13 08:43:47 kergoth might know more. Jul 13 08:46:24 decrement_mojo( INSTANCE_OF_REF(kergoth) ); Jul 13 08:46:32 http://cvs.arm.linux.org.uk/cgi/viewcvs.cgi/tslib/configure.ac.diff?r1=1.2&r2=1.3 Jul 13 08:46:39 AC_ARG_ENABLE(dejitter) Jul 13 08:46:52 to enable/disable linear-h2200 ;) Jul 13 08:47:52 heh Jul 13 08:47:55 ~kergoth-- Jul 13 08:47:58 hey pb_ Jul 13 08:48:06 hi mithro Jul 13 08:48:07 hi zap Jul 13 08:48:11 hey zap Jul 13 08:48:13 did you ever hear anything more from that sharp guy? Jul 13 08:48:24 which sharp guy? Jul 13 08:50:27 s/sharp/sony Jul 13 08:50:27 the guy who took us out to dinner Jul 13 08:50:37 oh, Tim Bird Jul 13 08:50:53 no, I haven't heard anything from him lately, but I haven't made any effort to contact him either. Jul 13 09:18:45 * LordVan is away: off Jul 13 09:19:53 jamey: another stupid question, if you don't mind - SD is working great, but has anybody started to reverse engineer the SDIO registers? Jul 13 09:34:43 CosmicPenguin: it's more reverse engineering SDIO cards Jul 13 09:34:51 need a logic analyzer and SD connector extender Jul 13 09:34:55 but not that I know of Jul 13 09:37:59 jamey: yeah, thats what I gathered - its no fun unless you ahve something to plug in Jul 13 09:38:18 you can buy a wifi card Jul 13 09:38:35 jamey: there are a few important registers not spelled out in the simplified spec though, and unfortunately, we're part of the SD association, so I think I'm tainted by assocation Jul 13 09:38:58 oh, which registers? Jul 13 09:39:07 I guess you can't answer that on this channel Jul 13 09:39:37 Heh - they're in the simplified public spec Jul 13 09:39:47 oh, there you go Jul 13 09:40:03 if there is something you need that is not in that spec, please let me know and I'll forward it to the author Jul 13 09:40:30 Ok Jul 13 09:44:25 there is that closed source sd wifi card driver for the zaurus Jul 13 09:44:31 if you are in a country where reverse enginnering is legal that could be an intresting disection Jul 13 09:46:18 I'm not nearly as worried about that, as I am about getting a working SDIO framework Jul 13 09:54:15 mickeyl: i'll be out the front from 7:20pm, it's now 6:52pm Jul 13 09:56:09 mickeyl: what's that about @openembedded.org btw? will those be valid mail addresses? otherwise I'll use my hh.org address. fake addresses suck. Jul 13 09:57:05 reenoo_: i think it was done for some monotone reason Jul 13 09:57:16 mithro: ok. I'm leaving shortly. anything you want to me bring with me? OZ 3.5.4-pre1 on a c7x0 ? Jul 13 09:57:52 reenoo_: it was to simplify access control on the monotone server. please generate such a key for now, we will add valid email addresses shortly Jul 13 09:57:54 to easy administartion or something Jul 13 09:59:30 I'm leaving. c'ya in a couple of minutes Jul 13 09:59:56 mickey|bbl: hrm. ok. Jul 13 10:06:45 I compiled Xfbdev and matchbox and copied all of the install dir contents into my rootfs, now when I try to run matchbox-window-manager with strace it shows me: open("/opt/oe/build/tmp/staging/arm-linux/lib/v5l/fast-mult/libXau.so.0"... why does it try to open a lib from there instead of /usr/lib ? Jul 13 10:07:34 sounds like your libc6 is defective Jul 13 10:09:34 file says: libc-2.3.3.so: ELF 32-bit LSB shared object, ARM, version 1 (ARM), stripped Jul 13 10:09:59 yah, I think the defect is a bit more subtle than that. Jul 13 10:10:09 but the libc wasn't copied by be, it was put there by bitbake bootstrap-image.. Jul 13 10:10:16 if libc6 was grossly faulty, your machine wouldn't run at all. Jul 13 10:10:37 so there are wrongs paths inside of the lib? Jul 13 10:10:42 it seems. Jul 13 10:11:21 one bad rpath, and you're done for Jul 13 10:11:33 it's odd, though, because I think all of our 2.3.3 glibcs have hwcap disabled anyway. Jul 13 10:11:49 I go and look where libc is being built.. Jul 13 10:21:46 all programs I compiled have a wrong rpath: 0x0000000f (RPATH) Library rpath: [/opt/oe/build/tmp/staging/arm-linux/lib] Jul 13 10:24:00 or hmm, not all have rpath.. Jul 13 10:43:10 Does anyone know what this patch (P07-C3000-KBDDEV_041206.patch) does for the OZ Spitz kernel? It looks like it creates a keyboard device but I can't seem to get it to work. Jul 13 10:45:35 hy all, i have this error: configdialog.cpp:230: error: 'class Opie::Core::OPcmciaSystem' has no member named 'restart' Jul 13 10:46:21 where is the restart function, when is addet to opie? Jul 13 10:48:58 i have libopie2-1.2.0+cvs-20050627 Jul 13 10:54:58 hey Jul 13 10:59:06 where does the rpath for the linker gets configured? Jul 13 11:09:37 03jbowler 07org.openembedded.nslu2-linux * rf4eaaacf... 10/packages/meta/openslug-packages.bb: Jul 13 11:09:37 Move various packages which only compile with glibc to the 'linux only' Jul 13 11:09:37 list. Jul 13 11:09:41 03jbowler 07org.openembedded.nslu2-linux * r43007d43... 10/ (8 files in 8 dirs): Jul 13 11:09:41 Add install-native package and remove the install-sh from those packages Jul 13 11:09:41 which used it to get round the install -s problem - now simply add Jul 13 11:09:41 install-native to the DEPENDS for packages which attempt to use the host Jul 13 11:09:42 install with '-s' Jul 13 11:17:34 heh Jul 13 11:17:56 monotone has been running at 100% cpu load for about an hour now Jul 13 11:18:38 Is there an Eddy Pronk around? Jul 13 11:22:15 reenoo_: ./configure --without-seti ;) Jul 13 11:22:41 i'm here Jul 13 11:25:32 zecke_: heh. well.. is this expected behaviour? Jul 13 11:26:55 reenoo_: sadly yes Jul 13 11:27:10 reenoo_: if you sync once a minute things are faster Jul 13 11:27:30 sounds broken Jul 13 11:27:32 badly Jul 13 11:28:14 hello Jul 13 11:28:25 i'm trying to get the updated collie kernel working on my Z Jul 13 11:28:43 but i'm not having much success Jul 13 11:28:51 can anybody offer me some advice? Jul 13 11:28:53 markt_: #openzaurus Jul 13 11:29:01 ok, ta :) Jul 13 11:32:26 ok. monotone isn't usable Jul 13 11:32:36 can we switch to something that works? Jul 13 11:32:56 monotone 0.20? Jul 13 11:33:08 0.19 Jul 13 11:33:25 i mean, let's give 0.20 a chance Jul 13 11:34:07 epronk: http://article.gmane.org/gmane.comp.handhelds.openembedded/4993 Jul 13 11:34:25 epronk: oh where are your patches? Jul 13 11:34:48 reenoo_: I dont see the problem. There isn't time to switch to 0.20 now Jul 13 11:34:48 reenoo_: it works - sort of Jul 13 11:34:56 there will be eventually Jul 13 11:37:06 more than an hour at 100% cpu load for an initial pull of a repo that hardly contains any revision info doesn't qualify as "it works" not even "sort of" Jul 13 11:37:28 reenoo_: turn off checking data integrity Jul 13 11:38:44 reenoo_: actually I do not know if it can get any faster :} Jul 13 11:40:10 reenoo_: my support Jul 13 11:40:22 * schurig sleeps whenever he has to use monotone Jul 13 11:41:13 OE is shooting itself into the foot by persisting on using slow base software like monotone Jul 13 11:41:24 schurig: lol Jul 13 11:41:25 indeed Jul 13 11:41:53 sorry that is just ridicoulus Jul 13 11:42:03 I stopped committing to OE because to do this, I'd have to constantly sync and this is slow Jul 13 11:42:22 schurig: stop fooling yourself, you did not push with bitkeeper more often Jul 13 11:42:33 schurig: if you need an excuse look for something else Jul 13 11:42:41 heh Jul 13 11:42:42 but yes it is fucking slow Jul 13 11:43:21 zecke_: come on, I started OE Jul 13 11:43:55 zecke_: I have to agree with schurig. I have a pile of stuff waiting to be pushed. but with monotone that doesn't seem to happen within my lifetime Jul 13 11:44:23 reenoo_: the only difference is, I believe you've a pile to sync ;) Jul 13 11:44:29 zecke_: and with the repo growing I don't expect it to get any better Jul 13 11:45:00 reenoo_: Can you setup netserve? Jul 13 11:45:07 are there any better tools? darcs or others? Jul 13 11:45:20 is it slow on the server side? Jul 13 11:46:15 reenoo_: if so, I would sync with your db and sync with vanille Jul 13 11:46:23 epronk: an initial checkout takes one hour, yesterday I got updates worth of 795 kB and this took more than half an hour. REAL time, time with (almost) 100% cpu time eaten. Jul 13 11:46:33 Gentoo uses rsync for users to keep their tree up to date and cvs to commit Jul 13 11:46:57 epronk: this happens whenever you do sync against another monotone installation. Some things, like getting the files out of your database into the filesystem are ok speedwise. Jul 13 11:47:20 git/cogito is probably worth considering Jul 13 11:47:24 are monotone developers aware of this problem? Jul 13 11:47:29 epronk: some people want to have the decentralised behavior, so they are against svn/cvs Jul 13 11:47:35 jamey: no Jul 13 11:47:40 epronk: yes, but they trade speed for security Jul 13 11:48:00 jamey: I mean it was considered, but it would require someone to sync the trees of the comitters regulary Jul 13 11:48:08 jamey: git needs someone to pull, there's no equivalent to of "push" in git Jul 13 11:48:32 I do know that git pulls rather than pushes Jul 13 11:48:40 so it would change the dynamics a bit Jul 13 11:48:52 reenoo_: schurig epronk JustinP : Alternatives to bitkeeper were evaluated one month, where were you? Jul 13 11:49:23 jamey: yeah, and so we would need some central guy who pulls & merge all the time, and if this person is ill, on holiday, etc, then the central repository would not be updated. That way I know always when Linus is on holiday. Jul 13 11:49:50 schurig: what about using the get tarball option of viewmtn? Jul 13 11:49:58 schurig: as you put the stuff in svn anyway? Jul 13 11:49:59 zecke_: I was part of the evalutation, because I addressed the speed problem of monotone in that time. mickeyl said it's fast enought, he had tested it. Jul 13 11:50:18 schurig: what about mn providing anonymous svn checkouts? Jul 13 11:50:30 zecke_: I'm afraid I'm not working on OE full time. and nobody's paying me money for it either. so, stuff like a ton of university and work deadlines matter to me for obvious reasons Jul 13 11:50:35 crap.. what am i missing here? Jul 13 11:50:36 glib-genmarshal.o(.text+0x997): In function `generate_marshal': Jul 13 11:50:36 : undefined reference to `g_fprintf' Jul 13 11:51:11 reenoo_: same applies here, so I trust other evaluating Jul 13 11:51:25 nevermind Jul 13 11:51:29 zecke_: that is something that I could do, and others as well. And it's exactly what I mean with "OE shoots into his own foot". An open-source project should not scare away a scarce resource: developers. Jul 13 11:51:31 reenoo_: but in contrast if it was a bad decision - speedwise Jul 13 11:51:46 reenoo_: we need to make it faster Jul 13 11:51:52 if it is a matter of weeks, there is not really a problem Jul 13 11:52:33 monotone 0.20 would be faster, but someone would have to change scripts, so it won't happen Jul 13 11:52:39 schurig: sorry but as this is Free Software do something productive Jul 13 11:52:44 however, it would only be marginally be faster, monotone is slow by design Jul 13 11:52:59 they might consider an alternative if you gave them one Jul 13 11:52:59 schurig: and coming here daily and saying monotone sucks is not productive in any way Jul 13 11:53:12 schurig: you waste my time to upgrade the scripts by forcing me to write this now Jul 13 11:53:21 zecke_: someone other beside me said it earlier, i'm just jumping on the bandwagon Jul 13 11:53:43 schurig: and tomorrow you will do the same? and the day after tomorrow? Jul 13 11:53:48 zecke_: I'm not forcing you into anything. Don't get so personal. And then, you can always /ignore schurig :-) Jul 13 11:53:51 schurig: spend the time to migrate the viewmtn scripts Jul 13 11:54:00 schurig: done Jul 13 11:54:12 zecke_: as long as "people force me" by saying "Oh, monotone is good enought for a big project like OE". It's not. Jul 13 11:55:42 reenoo_: if you can provide netserve, I would happy to sync your work Jul 13 11:55:53 reenoo_: until we've something better Jul 13 11:56:23 but switching away from monotone would be even more stupid, the workflow has just been established Jul 13 11:56:47 people understand how to use monotone halting OE for two weeks would hurt even more Jul 13 11:57:10 and yes it is only my opinion Jul 13 11:58:32 JustinP: did you get feedback for your bbfiles? Jul 13 12:14:27 zecke_: well. let's see how it performs after an initial pull. this does however leave me a bit worried about the situation in a few months from now. how will monotone perform when there actually is revision info? do you want to switch to something else at that point and lose revision info *again*? Jul 13 12:14:49 reenoo_: yesterday the 700kb changes did really long to sync Jul 13 12:15:00 reenoo_: but now the 100kb went through in two minutes :} Jul 13 12:15:16 reenoo_: but the offer is present Jul 13 12:16:18 zecke_: the initial pull delay, couldn't that be worked around by providing snapshots? Jul 13 12:16:40 zecke_: I'm more worried about the project as a whole than about my changes to hit the repo Jul 13 12:17:05 incinerator: I sync it could be, but you've to trust the site you're getting the snapshots from Jul 13 12:17:28 incinerator: in contrast to any other distributed system, monotone does check the integrity of the imported changes Jul 13 12:18:04 hmmm, i see. Jul 13 12:18:27 i mean, i don't really have a problem with monotone, but i have an athlon-xp3200+ with 1Gig of RAM.... Jul 13 12:18:34 I've seen incredible fasts and slow syncs, but I think security is important Jul 13 12:18:39 fast enough on that machine.... Jul 13 12:18:46 sure Jul 13 12:18:54 won't argue that.... Jul 13 12:18:55 incinerator: another thing is sqlite3 and ext3 does not play well together Jul 13 12:19:18 really? i use ext3, but i guess my big supply of ram helps me there.... Jul 13 12:19:19 we even carried a pile of disks from Uni to Home to start with Unix Jul 13 12:19:59 hmmm, it still would be interesting to see if you could save some time by providing snapshots.... Jul 13 12:21:51 the snapshots could be signed with a gpg key or whatever... Jul 13 12:22:26 well, i just did a wee experiment: copied the monotone db to another directory... Jul 13 12:22:41 schurig: http://www.vanille.de/temp/monotone.vanille.de-20070708.tar.bz2 Jul 13 12:22:43 did a pull there, was just like another syncing pull, done in a couple of seconds... Jul 13 12:22:56 <- feel free to make work with monotone 0.20 Jul 13 12:23:15 btw, monotone.vanille.de is down? Jul 13 12:24:16 hmmm, pings works, but it doesn't let me connect with monotone.... Jul 13 12:24:21 monotone: read from fd 4 (peer monotone.vanille.de) failed, disconnecting Jul 13 12:24:21 real 5m7.857s Jul 13 12:24:34 incinerator: so maybe i've shot it down? Jul 13 12:24:37 well, ewi546.ewi.utwente.nl works Jul 13 12:26:01 schurig: i guess so ;-) Jul 13 12:26:13 no, it was the evil blachner *ggg* Jul 13 12:26:35 incinerator: hehe, I'm not a fan of mt, but I'm not _that_ much against it Jul 13 12:26:58 i think the concept is cool, und die rechner werden immer schneller Jul 13 12:27:35 a couple of days ago i did a montone update Jul 13 12:27:46 montone could not automatically do the merge for a file i had modified Jul 13 12:27:58 so what did it do? called xemacs for a 3diff, COOOOL Jul 13 12:28:24 * schurig points kindly to http://www.selenic.com/mercurial/ Jul 13 12:28:47 arg, kvirc sucks Jul 13 12:29:39 incinerator: konversation is better, if you want a KDE program for it that is lighter than kopete Jul 13 12:29:43 hmmm, interesting that mercurial Jul 13 12:29:52 well, i used to use gaim for irc Jul 13 12:29:53 incinerator: and if you don't need the perl scriptability of kvirc Jul 13 12:30:03 ahm, that was ksirc Jul 13 12:30:03 but now that i'm on so many channels, gaim is not usable anymore... Jul 13 12:30:37 well, kvirc is cool in a way, but sound doesn't work and the higlight text stuff is messed up Jul 13 12:30:55 highlight disappears immediately after i release the mouse button :-( Jul 13 12:31:08 well, at least the highlighted text gets copied to clipboard Jul 13 12:31:48 I like irssi Jul 13 12:32:17 well, i guess i'll try ksirc Jul 13 12:32:17 text based, but give me the most real estate Jul 13 12:32:24 gives even Jul 13 12:32:26 hehe Jul 13 12:32:47 irssi Jul 13 12:32:49 rocks Jul 13 12:32:59 my Jul 13 12:33:02 socks Jul 13 12:33:31 it's fast, it's clean, it's predictable, and you can use all the irssi scripts Jul 13 12:33:54 i don't need scripts, just nickserv identify and that's it Jul 13 12:34:28 hmmm, doesn't emacs have an irc client? Jul 13 12:34:33 ;-) Jul 13 12:35:55 now schurig, if mercurial is so cool, why don't you guys use it instead of mt? Jul 13 12:36:25 incinerator: I don't know if it's cool, I just read about it in the second-last Linux Weekly News Jul 13 12:36:30 has that too in irssi afaik Jul 13 12:36:47 incinerator: it's fast, and it supports the "hg push" command Jul 13 12:37:05 ah ok Jul 13 12:37:30 incinerator: but this isn't a good enougth test for using it in for OE, just some of me pre-requisites of a good, distributed SCM :-) Jul 13 12:40:10 <[g2]> NAiL, around ? Jul 13 12:40:13 yes Jul 13 12:40:24 <[g2]> can you do me a favor ? Jul 13 12:40:30 Don't know yet Jul 13 12:40:38 <[g2]> it's an easy one Jul 13 12:40:46 probably ;) Jul 13 12:41:18 <[g2]> one sec Jul 13 12:41:52 $ hg push Jul 13 12:41:52 <[g2]> http://groups.yahoo.com/group/nslu2-linux/message/7416 Jul 13 12:41:53 abort: can only push to ssh:// destinations currently Jul 13 12:41:54 :-) Jul 13 12:42:06 <[g2]> DOH! wrong window Jul 13 12:43:59 [g2]: try with tremor for ogg. it doesn't use (as many) floating point operations Jul 13 12:44:26 s/ogg/ogg-vorbis/ Jul 13 12:45:37 err.. nevermind.. someone replied saying exactly that already Jul 13 12:45:56 <[g2]> reenoo_, thx Jul 13 12:46:30 <[g2]> I was actually just pinging NAiL to add it to our OpenSlug packages as it built for me Jul 13 12:47:25 <[g2]> although it is very temping to go get a usb SB Jul 13 12:47:30 <[g2]> tempting Jul 13 12:47:35 <[g2]> to plug into the nslu2 Jul 13 12:49:24 yeah, same goes for a usb LCD I guess :) Jul 13 12:49:33 speaking of which.. Jul 13 12:49:40 Rar. I made uimoc-native compile under gcc4. \o/ Jul 13 12:50:43 * reenoo_ grabs a cfontz 631 to see if it works with LCDproc now Jul 13 12:50:44 i'm having linking problems again, this time it is libsdl-mixer :-( Jul 13 12:51:06 seems ld doesn't see the libdirs in staging dir Jul 13 12:55:13 zecke_: I wasn't criticizing monotone, I was just adding to the conversation. I'm fine with my pulls taking a little while. Then again I'm not a committer, so... Jul 13 12:55:40 zecke_: no I've gotten nothing back on my bb files after I posted the "bug" yesterday. Jul 13 12:55:43 JustinP: not yet, did some one review your bbfiles yet? Jul 13 12:55:49 hehe Jul 13 12:56:23 huh? Jul 13 12:56:24 JustinP: I was just pointing out that 'monotone is slow' or 'monotone 0.20 is out use it' is not productive in anyway Jul 13 12:56:41 JustinP: not yet a committer Jul 13 12:57:16 zecke_: I, however, wasn't criticizing....I was just attempting to ask for possible alternatives. Never mind it all, I'm fine with monotone. Jul 13 12:58:10 since you all talk about monotone a lot I now try it myself :) Jul 13 12:58:17 Mercury took 19 Minutes for "modified 24076 files, added 33015 changesets and 251237 new revisions" Jul 13 12:58:42 zecke_: and again, no, no one has reviewed my bb files as yet Jul 13 12:59:03 JustinP: hehe, complaining and recognizing faults/shortcomings is good Jul 13 12:59:31 schurig: and how confident no revision was altered/faked/manipulated? Jul 13 13:00:00 schurig: I can setup a Win 95 machine to be online within 15 minutes, that is faster than a debian system Jul 13 13:00:23 add that to your "skill no one will ever needs" list *ggg* Jul 13 13:00:26 schurig: and at the Chaos Computer Congress it only needed 10 seconds to get infected by virii and worms ;) Jul 13 13:00:33 bah, my typing is rubbish today Jul 13 13:01:43 zecke_: I don't need to be confident. Projects like KDE don't need this ... after all, I'd have to check the *.bb files and patch files anyway, even if I have strong security all over the place, because someone could have hacked into your computer and modified a file which you then accidentally commited. So this hack would get a legitimate certificate ... don't trust those certifcates to much then! Jul 13 13:02:28 zecke_: he could have used a backdoor of the CCC-virii ;-) Jul 13 13:03:01 schurig: my home is encrypted and be unlocked only by my usb key, I carry at a secure place... Jul 13 13:03:50 zecke_: is is true for all other people which can push to vanille.de? Jul 13 13:04:09 zecke_: using cryptoloop? Jul 13 13:04:21 reenoo_: gosh no that is insecure Jul 13 13:04:24 reenoo_: BSD GEOM Jul 13 13:04:28 zecke_: and if it would be true, could I then trust all those people to not push evil code? So, you see, the "security by certicates" doesn't buy me much, I'd not overestimate that feature Jul 13 13:05:00 schurig: no it is twofold Jul 13 13:05:06 schurig: a) trust the committers Jul 13 13:05:19 schurig: did you know that every single commits of kde goes to addresses at TT? Jul 13 13:05:25 for review and recording? Jul 13 13:06:01 zecke_: yes, used libgeoip this week and tried it with kde.org Jul 13 13:06:14 zecke_: yeah. I was just going to indicate that there are real world attacks against it. Jul 13 13:06:24 zecke_: where's your b) ?? I saw an a) Jul 13 13:07:27 zecke_: TT hosts many kde.org websites and pays for the bandwith. Now, one could make now plot an evil scheme, like "TT is controlling KDE". But I don't like the "Bild-Zeitung" that much :-) Jul 13 13:08:42 zecke_: with an SCM like Mercury, "pushers" (committers) need an ssh account at some central server, like vanille.de is for OE. That makes sure that the connection is not tampered with. Jul 13 13:08:55 zecke_: that is, if you trust SSH enought. Jul 13 13:09:36 zecke_: monotone "just" gives the added feature that it is very hard to tamper the repository itself Jul 13 13:10:18 schurig: personally, I'd consider that more of a flaw than an advantage Jul 13 13:10:28 treke-: what? Jul 13 13:11:04 treke-: sometimes I did manual modifications to an CVS repository, but mainly because CVS was to limited. Do you mean this? Jul 13 13:11:09 schurig: no it illustrates commits get reviewd by TT employees Jul 13 13:11:16 requiring an ssh accessible account on the server. More admin overhead Jul 13 13:11:31 schurig: also kde is svn (server based) Jul 13 13:11:33 zecke_: yes, and commited get's reviewed by Microsoftians, now that they have an Open Source Lab. Jul 13 13:11:46 schurig: only trusted people can commit Jul 13 13:11:50 zecke_: you can review commites to KDE, Mozilla, whatever, without hosting the repository Jul 13 13:11:52 schurig: and each commit gets reviewd Jul 13 13:12:06 zecke_: with any protocal based on SSH only trusted people can commit, true Jul 13 13:12:23 you completely miss the point Jul 13 13:12:28 zecke, considering the volume of commits, I doubt everything is well inspected Jul 13 13:12:36 schurig: a) only selected people can commit to a trusted server Jul 13 13:12:46 schurig: b) every commit gets reviewed Jul 13 13:12:55 schurig: c) no direct manipulation on server side is possible Jul 13 13:13:18 schurig: but b) and c) is not possible for distributed scm's Jul 13 13:13:33 zecke_: a) and b) can be made with any ssh-based transport mechanism, and c) is rarely needed Jul 13 13:14:00 zecke_: and any distributed SCM for an OpenSource projects tends to have one public root, so here a) is possible Jul 13 13:14:46 actually, b) can be done with any SCM, if you have the time and dullness Jul 13 13:14:50 you could probably go as far as to say any project would have a single root Jul 13 13:17:03 zecke_: oh, by the way: in Monotone the revlog's are also hashed with SHA1 Jul 13 13:17:03 zecke: b) is possible in a distributed system if you have some sort of workflow thingy Jul 13 13:18:33 maybe it depends what "review" means. If it means "I need to look at every patch, and approve and disapprove it", (the Linus' + Adjutants way), then not every SCM can do it. When it means "Trolltech looks at every patch" or "Microsoft looks at every path", then the svn-commits-mailing list or something like this is sufficient, which is possible with every SCM Jul 13 13:18:40 I'd like to point out that on monotone's front page it says is supports "integrated code review" Jul 13 13:19:21 jamey: you need to be able and trust the reporting possibilities offered from your root tree Jul 13 13:19:49 schurig: and as a member of the CCC.de I know there is no security at all Jul 13 13:20:03 zecke: all I'm saying is that if b) is a requirement, you need to route the commits through a review process before they hit the central tree as seen by everyone else Jul 13 13:20:09 zecke_: go and tell me that you're in Mensa as well :-) Jul 13 13:20:18 schurig: but you can make things more hard and less likely to occur Jul 13 13:20:23 I'm not sure I know what c) meant Jul 13 13:20:57 jamey: ssh h1.hh.org sed -i s/*/moo /cvs/opie/* ;) Jul 13 13:21:15 jamey: in CVS, you could go to a CVSROOT and directly manipulate with vi the RCVs files there. Or delete files, or move files around. Some SCMs make this hard to impossible. Jul 13 13:21:23 jamey: I would trust the server and the committing system but it was bypassed Jul 13 13:21:32 jamey: at hh.org we luckily have tripwire installed Jul 13 13:21:49 zecke, schurig: ok, I understand what you mean by c) now Jul 13 13:22:38 jamey: and, if I got zecke right, this is an essential SCM feature for zecke_ Jul 13 13:22:54 jamey: I don't care so much, because in real life this doesn't buy so much Jul 13 13:23:40 zecke_: it's funny how you didn't care about this type of stuff with bitkeeper Jul 13 13:24:27 zecke_: I mean, yeah, we all trust larry, don't we? Jul 13 13:24:54 reenoo_: hehe, right but isn't larry trustworthy Jul 13 13:24:58 you are once again making up reasons to stay with something inherently broken Jul 13 13:25:13 this time technically, not licensing wise Jul 13 13:26:00 hehe. for the most part I do trust bitmover :) Jul 13 13:26:40 reenoo_: hehe, but trustworthy computing is an issue to be solved by Free Software as well Jul 13 13:27:04 reenoo_: my only reason to stay with monotone is that the workflow with dreambox and nslu2 really seem to work Jul 13 13:27:53 zecke_: monotone's workflow isn't so different to the one of Bitkeeper, or? Jul 13 13:28:36 sounds like you're discussion performance problems in monotone? :) Jul 13 13:28:37 zecke_: I ask this because for me mercury's workflow look similar to this as well (I am still not endorsing Mercury, because it doesn't have Branches, which you seem to use/need) Jul 13 13:28:42 schurig: not too much but different enough Jul 13 13:28:45 tmbinc: yes Jul 13 13:29:22 tmbinc: well it *was* about the supposed slowness of monotone. Now it's a holy war. Jul 13 13:29:51 tmbinc: for some people monotone is too slow (e.g. me), and they get onto the nerves of people that like monotone Jul 13 13:30:03 i know this has probably been discussed over and over (and of course denied), but can't we implement a --i-trust-my-internet-connection-and-my-local-dram which just skips those checks? Jul 13 13:30:10 zecke_: what part of the workflow is different? Jul 13 13:30:21 What is the best way to handle this: "NOTE: multiple providers are available (glibc, glibc-intermediate); NOTE: consider defining PREFERRED_PROVIDER_virtual/arm-linux-libc-for-gcc" Jul 13 13:30:41 i like monotone very much, more than cvs, svn and bitkeeper (not that they are really comparable..), so regardless of that (really annoying speed problem), i vote for keeping it Jul 13 13:31:17 tmbinc: I actually like the way of how the workflow works, but I very much dislike the speed Jul 13 13:31:57 schurig: personally I used bk lclone a lot Jul 13 13:32:00 tmbinc: but then again, I'd be happy with a central SVN server as well (pity me) Jul 13 13:32:08 schurig: now you've a db so every checkout is a lclone Jul 13 13:32:33 schurig: but if you operate in a checkout it is almost the same Jul 13 13:32:58 btw, i saw in the handhelds.org wiki that familiar won't join up anström, why not? Jul 13 13:32:59 schurig: but you can do stuff like mt commit --branch="shiny.new.branch" and then do mt sync brandnew.server Jul 13 13:33:51 reenoo_: I'm twofold on the security and your bkbits.net was a good one Jul 13 13:34:38 familiar and angstrom are two different distributions Jul 13 13:35:00 well, but doesn't familiar use oe, as well? Jul 13 13:35:21 incinerator: we tend to view oe as a tool like gcc Jul 13 13:35:30 incinerator: we do not refer to the gcc distribution Jul 13 13:35:48 incinerator: in the past familiar has used several build systems. Jul 13 13:36:05 zecke_: I'm all for security. and I'd be happy with using monotone. it's just that monotone fails to get me a repo. after around 2 hours of 100% cpu load it died a couple of minutes ago Jul 13 13:36:15 incinerator: familiar 0.8.x is built from oe, yes, but familiar will not necessarily choose same versions or same packages as angstrom Jul 13 13:36:15 incinerator: IMHO keeping a clean line between the distro and build system is a sound practice. Jul 13 13:36:25 fair enough, but does that mean you'll drop the zaurus devices then? Jul 13 13:36:35 zecke_: I can do "hg clone /home/schurig/linux-2.6.hg /tmp/linux-2.6.work", which just get's the database part of /home/schurig/linux-2.6.hg Jul 13 13:36:41 incinerator: nope, no plans to drop anything Jul 13 13:36:51 incinerator: no, we plan to keep zaurus support Jul 13 13:37:08 which distro are we talking about here? Jul 13 13:37:13 france_: Gnu Compiler Collection Distribution would sound weird ;) Jul 13 13:37:14 well, i was a bit scared because e.g. the download page for 0.8.2. doesn't show the images for the zaurus Jul 13 13:37:23 zecke: :-D Jul 13 13:37:26 france and jamey talking about familiar Jul 13 13:37:29 ok Jul 13 13:37:34 you have to type in the right url manually to get to the right dir.... Jul 13 13:37:38 reenoo_: I agree with the speed issue, it needs to be solved Jul 13 13:37:39 incinerator: that can be fixed... Jul 13 13:37:42 I've been using OZ so I'm ok, then. ;-) Jul 13 13:37:47 incinerator: that sounds like an issue with the html or php Jul 13 13:37:56 so what will be different between oz and familiar then? Jul 13 13:37:56 reenoo_: but the branching with nslu2, dreambox looks like it works out Jul 13 13:38:10 or angström and familiar Jul 13 13:38:25 for zaurus boxes i mean Jul 13 13:38:33 incinerator: I can not speak for the anstrom folks, it will be up to them Jul 13 13:39:04 k Jul 13 13:39:22 what is anstrom? Jul 13 13:39:42 typo *ggg* Jul 13 13:39:42 opps, angstrom Jul 13 13:39:43 new distro derived from OZ and familiar using OE as buildsystem Jul 13 13:39:46 ah Jul 13 13:40:04 well, it seems like a merger between oz and opensimpad now.... Jul 13 13:40:12 and opensimpad Jul 13 13:40:12 jamey: so it will be the successor of OZ? Jul 13 13:40:13 treke-: some folks want to create an new distro Jul 13 13:40:29 we need more distros ! :-) Jul 13 13:40:31 JustinP: that is what I've heard Jul 13 13:40:34 JustinP: the nice thing about open source is that the users will decide Jul 13 13:40:38 Its not a new quarter if somebody doesn't make a new handhelds distro Jul 13 13:40:39 JustinP: and the devlopers Jul 13 13:40:58 JustinP: it must be getting near time to go home the typing is going. Jul 13 13:41:03 heh Jul 13 13:41:39 france_: hehe, and now try it with ö : angström ;-) Jul 13 13:41:48 schurig: :-D Jul 13 13:43:26 CosmicPenguin: :-D Jul 13 13:43:27 yes! we definately need more distros. the community isnt nearly fractured enough yet, and there's far, far too much collaboration. Jul 13 13:44:01 reenoo_: but I do not know if the needed time is a fundamental problem or an implementation issue Jul 13 13:44:11 kergoth: hey Jul 13 13:44:19 kergoth: collaboration is sometimes dangerous (in iraq ...) Jul 13 13:44:23 ~lart kergoth for checking broken automake into tslib Jul 13 13:44:23 * ibot plops kergoth into a giant vat of herring for checking broken automake into tslib Jul 13 13:44:29 uh? Jul 13 13:44:35 tslib builds flawlessly with the latest autotools Jul 13 13:44:40 kergoth: wait a second Jul 13 13:44:41 if you're on automake 1.4, god help you Jul 13 13:44:42 kergoth: scary Jul 13 13:44:46 schurig, have you figured out how to type that strange "a" letter yet? Jul 13 13:44:54 anywhere, my wife came back, so I stop computing ... Jul 13 13:45:09 zecke_: me neither. it's just that it apparently doesn't seem to work for me at all Jul 13 13:45:17 incinerator: hmm, I have a german keyboard, so no problem with the ä. But I don't have the a with a circle on top Jul 13 13:45:25 kergoth: http://cvs.arm.linux.org.uk/cgi/viewcvs.cgi/tslib/configure.ac.diff?r1=1.2&r2=1.3 Jul 13 13:45:31 kergoth: AC_ARG_ENABLE(dejitter, Jul 13 13:45:37 same for me btw Jul 13 13:45:39 kergoth: now try --disable-linear-h2200 Jul 13 13:45:41 ;) Jul 13 13:45:44 incinerator: this letter sounds like a short "oh", AFAIK Jul 13 13:45:47 don't let my dns hostname fool you *ggg* Jul 13 13:45:55 byby Jul 13 13:45:56 * kergoth chuckles Jul 13 13:46:01 later Jul 13 13:46:05 gotta love copy & paste errors Jul 13 13:46:18 what's the default encoding in this channel anyways? Jul 13 13:46:20 utf8? Jul 13 13:46:33 kergoth: mails to mvista.com never arrive, will you fix that? Jul 13 13:46:45 incinerator: no, my setup is just broken Jul 13 13:46:46 zecke: douglas lowder isnt the tslib maintainer Jul 13 13:46:49 zecke: dont contact him anymore Jul 13 13:46:54 * koen removes the nonsense change to the hh.org wiki Jul 13 13:47:02 hehe Jul 13 13:48:02 kergoth: mails never arrived anyway Jul 13 13:48:09 at least get the facts straight when changing it Jul 13 13:48:37 hehe Jul 13 13:48:45 yeah, always had to email his personal address Jul 13 13:49:03 now i am totally confused koen Jul 13 13:49:31 btw, how come the download page for 0.8.2. doesn't feature the zaurus models? Jul 13 13:50:35 because familair is largely untested on them and we don't have install instructions Jul 13 13:50:48 * zecke_ would welcome if the Familiar Linux distribution still leverages from the OpenEmbedded package database Jul 13 13:51:14 * zecke_ would also welcome more conversative versioning of packages, and better QA ;) Jul 13 13:53:27 incinerator: we need someone to test familiar on zaurus and to help write install instructions Jul 13 13:53:41 relatively easy task Jul 13 13:53:44 hm, something completely different: is it possible to maintain the meta (or image) packages as real packages, so "ipkg upgrade" will install new packages (which were added into the meta package since then)? (RDEPENDS on that metapackage to the different packages should be enough) Jul 13 13:53:51 or is there anything not working out here? Jul 13 13:54:16 tmbinc: It is possible Jul 13 13:55:14 is there any reason why this is not done? Jul 13 13:55:45 tmbinc: don't we do it? Jul 13 13:55:58 zecke, nope Jul 13 13:56:10 if I understand you, that is our plan for next familiar release Jul 13 13:56:12 tmbinc: just never been done. No one has ever cared to do so Jul 13 13:56:33 ok i'll take a look on this then Jul 13 13:56:44 What I'm understanding is that he's asking about an ipkg package equivalent to the bootstrap image contents Jul 13 13:56:51 will ipkg upgrade remove packages which depended on a package, but don't do anymore? Jul 13 13:56:57 tmbinc: nope Jul 13 13:57:20 hm ok Jul 13 13:57:28 that would be pretty broken behavior imo Jul 13 13:57:32 treke-: that has been done ever since gpe is in OE Jul 13 13:57:57 reenoo_: we have meta-gpe, but do you actually generate an ipkg for the gpe image? Jul 13 13:58:11 PACKAGES = gpe-base-depends gpe-task-base gpe-task-settings gpe-task-pim gpe-task-apps gpe-task-games Jul 13 13:58:17 tmbinc: any hints what's missing when i get this while running 'bitbake dreambox-image' http://www.24.fi/dolphin/log.do_compile.9215 Jul 13 13:58:19 so, yes Jul 13 13:58:34 reenoo_: that doesnt sound like what he's describing Jul 13 13:58:57 dolp: are you using gcc4 as host gcc? for some reason i don't know this doesn't work for glib-native Jul 13 13:59:15 reenoo_: when I looked last, the images specified a large number of packages to be installed Jul 13 13:59:19 and then the bb installed them Jul 13 13:59:33 treke-: well, i probably only have to turn my image content into a *-task thingie then Jul 13 13:59:38 and install the -task in the image Jul 13 13:59:48 and i get my desired behaviour Jul 13 14:00:07 treke-: opie-image maybe Jul 13 14:00:30 btw why is "removing unrequired packages" broken? (unless they were selected by the user, of course. well, ipkg had to track the reason for installations, then.) Jul 13 14:00:48 s/broken/broken behaviour/ Jul 13 14:00:51 tmbinc: how do i check it? 'gcc --version' says: 3.3.5 (Debian 1:3.3.5-13) Jul 13 14:00:52 tmbinc: because ipkg doesnt know whether the packages are being used Jul 13 14:01:03 tmbinc: it just knows that nothing installed depends on it Jul 13 14:01:34 treke-: gpe-image (or meta-gpe) has always worked as tmbinc just described. Jul 13 14:02:01 tmbinc: we started coding removal of packages that were only installed because of a dependence but haven't finished debugging it Jul 13 14:02:09 tmbinc: if opie-today depends on opie-todo and I install opie-today I will get both Jul 13 14:02:13 currently, it seems to remove everything. or at least last I looked Jul 13 14:02:24 dolp: hm. dunno then, sorry. works here, and the dreambox branch doesn't do anything specific with glib-native. must be broken for non-dreambox, too, then. Jul 13 14:02:27 the idea was to track whether user requested something to be installed Jul 13 14:02:47 yeah exactly. well, this sounds great, i'll wait for that then Jul 13 14:03:04 tmbinc: I could start using opie-todo and decide I want to remove opie-today. I'd be annoyed if both went away Jul 13 14:03:06 another area ripe for new developers to contribute Jul 13 14:03:16 can be debugged on a PC in an offline root Jul 13 14:03:16 crap.. Jul 13 14:03:21 jamey: I like the idea of tracking it so you could build something like deborphan Jul 13 14:03:36 yes, could use libipkg for that Jul 13 14:03:43 but doing it as part of the course of a regular remove seems broken to me Jul 13 14:04:13 another area that needs some work is improving the performance of the ipkg metadata repos Jul 13 14:04:32 well my idea was just that distributions can be upgraded with ipkg Jul 13 14:04:33 <[cc]smart> mickey|bbl: ping Jul 13 14:04:45 including the installation of new images, removal of not anymore included packages Jul 13 14:05:29 yes, it should be possible to package the distro that way Jul 13 14:05:31 kergoth: can you spare why he is not the maintainer anymore? Jul 13 14:05:55 * treke- just doesnt think removing packages is something that should be up to the distro :) Jul 13 14:06:34 treke-: but otherwise you'll end up with a full flash, after a few updates. i agree that actively installed packages should not be removed at all. Jul 13 14:06:51 what if a distribution exchanges two conflicting packages? Jul 13 14:07:11 use conflicts for that Jul 13 14:07:17 zecke: of tslib? because he wasnt working on it anymore. so he transferred maintainership to me Jul 13 14:07:33 say that package a can't be installed while package b is installed Jul 13 14:08:14 so what should be done if a new version of the distribution (or task) wants to install a conflicting image? asking the user? (the enduser probably doesn't know anything about that package, he never selected it.) Jul 13 14:08:32 conflicting package Jul 13 14:08:32 arg Jul 13 14:08:33 zecke_: am I correct that GEOM is FreeBSD only? Jul 13 14:08:41 tmbinc: yes, unless it's just a matter of package renaming. Jul 13 14:08:42 <[cc]smart> somebody has an idea on when mickeyl will be back ? or if today at all ? Jul 13 14:08:56 * zecke_ agress with treke- Distribution should decide what gets packages and shipped Jul 13 14:09:10 reenoo_: yes, and while not having looked at the code it promises to be secure Jul 13 14:09:22 tmbinc: the package manager and distribution just don't know what is being used and how Jul 13 14:09:34 treke-: the tracking of who installed package is to make it easier for user to remove packages the user is not interested in Jul 13 14:09:45 jamey: and I like that idea Jul 13 14:09:46 why? the distribution metafile (aka tasks) specify which packages are in use Jul 13 14:09:53 unless the user actively selected new packages Jul 13 14:10:01 which must be tracked (as jamey said) Jul 13 14:10:18 we don't actually track which packages are in use on a device, though I think that would be cool Jul 13 14:10:30 especially if users willing to share it upstream (anonymously) Jul 13 14:10:34 tmbinc: no it specifies what packages were installed. A difference :) Jul 13 14:10:47 could build nifty inotify client Jul 13 14:10:52 it doesnt know if I'm actually making use of the package Jul 13 14:10:53 social package usage Jul 13 14:10:57 tmbinc: hmm.. btw can it be that i had i686 in local.conf.. i doubt i shud use that Jul 13 14:10:58 jamey: that would be really really spiffy Jul 13 14:10:59 39842 people use busybox Jul 13 14:11:26 tmbinc: tho all other stuff before that compiled just fine.. Jul 13 14:11:35 guess i'll start from scratch Jul 13 14:12:01 dolp: hm. remove that i686, where did you got it from? Jul 13 14:12:13 dolp: unless you have something special, autodetection should work ok Jul 13 14:12:25 dolp: though i don't think it's related to the missing g_fprintf Jul 13 14:12:40 yep yep.. damnit.. this is annoying Jul 13 14:12:51 tmbinc: For a library package, it may have been installed as a dependency of the release, but the user might have unpackaged applications that depend on it Jul 13 14:13:10 or scripts that depend on a binary that shipped with the image Jul 13 14:13:16 "unpackaged applications"? well, that kills it all, yes. Jul 13 14:13:21 you don't want those things disappearing out from under users Jul 13 14:13:31 he would have to install the depending library then Jul 13 14:13:47 unpackaged applications bad Jul 13 14:13:57 should make it easier for user to package applications? Jul 13 14:15:05 treke-: well, then let's have a "remove orphan packages", which must be manually executed. Jul 13 14:15:16 tmbinc: that's fine by me Jul 13 14:15:25 still i think that the "upgrade" feature is useless without automatically removing packages as well Jul 13 14:15:48 but if i have to call it "ipkg upgrade; ipkg remove-orphan" that would be fine for me as well. Jul 13 14:16:45 I wouldnt even have any complaint if you worked it into the workflow of the standard upgrade process on the graphical clients Jul 13 14:17:40 btw, speaking about ipkg: are gnu tar extensions fixed in the meantime? i have a patch again 0.99.148, but i'm not sure if it's correct nor required. Jul 13 14:18:14 tmbinc: ipkg upgrade should certainly remove Replaced packages, and I think removing orphans also makes sense, once we have a way to do so Jul 13 14:18:36 I thought the gnu tar extensions had been fixed but I have not tested it lately Jul 13 14:18:45 send patch to pigi and I Jul 13 14:19:55 ok Jul 13 14:20:34 uh, do you have email addresses handy? Jul 13 14:25:27 jamey@handhelds.org and pigi@handhelds.org should work Jul 13 14:28:26 ok Jul 13 14:32:33 welp i just finished having dinner with mickey Jul 13 14:33:23 mithro: how did he taste? Jul 13 14:33:32 * CosmicPenguin coughs Jul 13 14:33:41 I'm attempting to add another package to OE and am getting errors during the install phase. Looks its trying to install some things on my root system, not in the OE "sandbox". What's the best way to fix this? Jul 13 14:34:07 JustinP: make it honor your DESTDIR override? Jul 13 14:34:41 http://www.thekompany.com/embedded/textmaker/ Jul 13 14:34:52 zecke: with not of :P Jul 13 14:35:52 zecke_: I know that much. I suppose I'll have to go diving now.... Jul 13 14:36:00 * JustinP gets his wetsuit Jul 13 14:36:18 good luck Jul 13 14:36:27 hehe. saw that earlier Jul 13 14:37:48 heh - that is the textmaker that I thought it was Jul 13 14:37:52 http://www.softmaker.de/tmp_en.htm Jul 13 14:38:01 Thats the software we put on our PIC Jul 13 14:38:46 Heh - I even have a .bb for it.. :) Jul 13 14:39:08 For the desktop version, that is Jul 13 14:39:18 hey, that's the regular textmaker Jul 13 14:39:29 tkc is selling textmaker ... didnt know that Jul 13 14:39:48 darn thing is at the sale price too Jul 13 14:41:09 might run on OZ with the compat libs Jul 13 15:28:53 mithro: hey Jul 13 15:32:23 zecke: i assume you meant mickeyl then :P Jul 13 15:32:42 well at least I know the person I keep getting mistaken for now :P Jul 13 15:32:56 :D Jul 13 15:33:33 mickeyl: our names are a bit to close i think :P Jul 13 15:33:50 that's true. throw a coin and let's see who's going to change it ;) Jul 13 15:33:51 mithro: actually I referred to you Jul 13 15:34:07 you did Jul 13 15:34:34 zecke: mickeyl said you only live a short bit out of Frankfurt, we should have organised to meet you to :P Jul 13 15:34:35 but I'm too tired keeping my eyes open Jul 13 15:34:47 mithro: holger != holger Jul 13 15:34:56 mithro: schurig you mean, my young padawan Jul 13 15:35:20 ahh Jul 13 15:35:34 people with the same name :P Jul 13 15:35:42 where you located zecke? Jul 13 15:35:43 I'm not yet losing hairs too much Jul 13 15:35:47 Berlin Jul 13 15:36:18 i'm going to most probably be in berlin for a couple of days Jul 13 15:36:34 in about a week or two Jul 13 15:36:53 mithro: you stay in hostels? or hotels? Jul 13 15:37:04 zecke: i could drop by and buy you a drink or two Jul 13 15:37:20 zecke: i've been trying to stay in hostels but they have been booked out a bit to much Jul 13 15:37:20 what package includes, say, stdio.h Jul 13 15:37:26 * JustinP thinks he's asked this before.... Jul 13 15:37:33 Berlin look okay however Jul 13 15:37:57 JustinP: glibc? Jul 13 15:38:00 i'm currently in a hotel in Frankfurt Jul 13 15:38:16 ulibc? :P Jul 13 15:38:23 no....that's what I'm asking Jul 13 15:38:32 mithro: you might be allowed to see the holy floors of ROAD Jul 13 15:38:45 zecke: which is? Jul 13 15:38:52 * zecke_ runs Xteddy on his Cell Phone Jul 13 15:38:55 what package will install it on my target system (yes I know this is sort of a distribution question) Jul 13 15:39:11 mithro: I'm doing an intern at a company building a cell phone Jul 13 15:39:30 zecke: mickeyl mentioned something about that Jul 13 15:42:04 zecke: so I'm assuming you are talking about letting me see inside the company walls then? :P Jul 13 15:42:30 go and drag some developer discount models for us ;) Jul 13 15:42:51 zecke: if so that would be pretty cool :P Jul 13 15:43:14 but I am just as happy to buy you a drink at a cafe/bar/etc :P Jul 13 15:43:49 mickeyl: i was going to go to bed, but then i decided to book accomadation for Interlarken Jul 13 15:44:37 mithro: good decision. Better safe than sorry. I went to St.Moritz for a conference and Switzerland is nearly always overbooked. Jul 13 15:45:01 i need to wait until my caffeine level has decreased to be able to get some sleep :) Jul 13 15:45:20 zecke: give me an email address and I will send you the details of when I'll be in berlin, when they are a little more solid Jul 13 15:45:37 mithro: you got my email address ;) Jul 13 15:45:40 mickeyl: doh! i guess you can fix up some of those bugs i showed you :P Jul 13 15:45:51 zecke: okay i'll reply to the email you sent me Jul 13 15:45:53 mithro: hehe Jul 13 15:46:17 well i better go to bed Jul 13 15:46:23 sleep well Jul 13 15:46:24 have to be up for breakfast tommorrow! Jul 13 15:46:38 then another load of washing then the sites of frankfurt Jul 13 15:48:33 opps Jul 13 15:48:41 mithro: if you can not sleep yet Jul 13 15:48:53 [00:46] have to be up for breakfast tommorrow! Jul 13 15:48:53 [00:46] sleep well Jul 13 15:48:53 [00:46] then another load of washing then the sites of frankfurt Jul 13 15:48:53 [00:47] zecke: i'll most probably email you in the next couple of days Jul 13 15:48:53 [00:47] * Disconnected (Connection reset by peer). Jul 13 15:49:18 zecke: you where going to say? Jul 13 15:49:45 mithro: I wanted to spare a URL to infrared hacking in hotels Jul 13 15:50:05 infrared hacking in hotels? Jul 13 15:50:34 that sounds intresting Jul 13 15:50:40 He tried ultraviolet, but everything just glowed instead Jul 13 15:50:52 argh.....I have no idea how to get this install to use the right dir. the configure looks like it's using root-based dirs and not sandbox dirs. Makes sense as the package will be installed that way, but how does BB make the install phase install into the staging area? Jul 13 15:51:06 zecke: so whats this infrared hacking then? Jul 13 15:51:27 mithro: http://www.ccc.de/congress/2004/fahrplan/event/125.en.html Jul 13 15:51:37 JustinP: we either override DESTDIRs and stuff, or install libraries by hand (recommended, to ensure staging doesn't get polluted) Jul 13 15:51:47 mithro: http://www.ccc.de/congress/2004/fahrplan/files/327-infra-red-slides.pdf Jul 13 15:53:07 cya Jul 13 15:53:09 zecke: that looks pretty cool Jul 13 15:54:51 lol, nice Jul 13 15:56:54 reminds me of these ethernet'd hotel TVs, where they had 10.0.x.x for the TVs, and, for "security" reasons, a 192.168.x.x net for management ... layer2 switched. Jul 13 15:57:02 i stopped hacking when i was able to edit all room's wakeup times. Jul 13 15:58:36 hehe Jul 13 15:59:07 JustinP: nmpc seems to need glib-2.0. is that correct ? Jul 13 16:00:11 mickeyl: you mean ncmpc? Jul 13 16:00:12 And thats a freaky ass looking cat to boot Jul 13 16:00:18 JustinP: yes Jul 13 16:00:33 JustinP: i tried to build it with an unpopulated staging and it bailed out with Jul 13 16:00:41 mickeyl: no, it should not require glib-2.0. I didn't put it in the DEPENDS.... Jul 13 16:00:48 or should it Jul 13 16:00:50 one sec Jul 13 16:00:54 sorry...:-| Jul 13 16:00:59 heh, well what you put in depends is irrelevant, when the package needs it :D Jul 13 16:01:07 | No package 'glib-2.0' found Jul 13 16:01:07 | configure: WARNING: glib-2.4 is required for NLS support! Jul 13 16:01:07 | checking for glib-2.0 >= 2.2... Package glib-2.0 was not found in the pkg-config search path. Jul 13 16:01:07 | Perhaps you should add the directory containing `glib-2.0.pc' Jul 13 16:01:18 or we say --disable-nls or something like that? Jul 13 16:01:31 seems a rather huge dependency for just a ncurses musicpd client Jul 13 16:01:35 yeah, --disable-nls may be the way to go Jul 13 16:01:41 trying... Jul 13 16:03:53 I wonder if it wouldn't be more efficient for monotone to automatically update checkouts via a mailing list or something o.o Jul 13 16:04:05 (push instead of pull updates) Jul 13 16:04:29 JustinP: --disable-nls is documented, but doesn't work :/ Jul 13 16:04:33 :-( Jul 13 16:04:38 JustinP: i'll only commit mpc for now Jul 13 16:04:47 perhaps you can convince it Jul 13 16:05:16 I guess the question is, do you want your code base changing on you while you're trying to work? Jul 13 16:05:52 mickeyl: did you try gmpc or glurp? Those *should* work... Jul 13 16:06:02 well, with mtn you can update your copy of the repo without changing your working copy Jul 13 16:06:19 not yet, i have no x stuff built - that will take a while Jul 13 16:06:22 mickeyl: can I just move the staging dir away to try this or do I need to move other things as well? Jul 13 16:06:38 JustinP: just remove glibc from staging Jul 13 16:06:40 no Jul 13 16:06:42 glib-2 Jul 13 16:06:44 :) Jul 13 16:06:49 ah Jul 13 16:06:52 ok, I'll try that Jul 13 16:07:33 i hate that. the package says it's optional and the configure script assumes otherwise :/ Jul 13 16:12:34 hmmm, yeah....compile failures if I disable NLS Jul 13 16:13:08 does it? Jul 13 16:13:09 looks like it's required Jul 13 16:13:13 it doesn't even configure here Jul 13 16:13:23 I must not have moved away glib right ;-) Jul 13 16:13:37 how should I get rid of glib in staging? Jul 13 16:13:55 rm -f tmp/staging/arm-linux/lib/libglib* ? Jul 13 16:14:06 and perhaps also the .pc file Jul 13 16:14:11 dunno where that lives Jul 13 16:14:38 arm-linux/share/pkgconfig/glib-2.0.pc Jul 13 16:14:45 righto Jul 13 16:15:55 I also removed staging/arm-linux/include/glib* Jul 13 16:16:45 yeah, looking at the code it looks like it uses GList and such, which I assume is from glib Jul 13 16:16:47 perhaps you can disable the feature by patching the configure.ac Jul 13 16:16:52 ah Jul 13 16:16:56 in that case that won't work Jul 13 16:17:21 but it would be nice to send them a note regarding that misdocumentation Jul 13 16:17:29 i'll add the glib dependency then Jul 13 16:17:34 so I suppose it needs glib-2.0 in the depends Jul 13 16:17:55 thanks for checking it for me Jul 13 16:18:54 mickeyl: fwiw, glib isn't x related at all Jul 13 16:19:18 reenoo_: I'm aware of that. gmpc and glurb need gtk+ though Jul 13 16:22:44 mickeyl: let me know when you'd committed it to the main repository. I haven't had to merge anything yet and would like to see how it works. :-) Jul 13 16:22:58 JustinP: you'll see in a couple of seconds... Jul 13 16:22:59 you've Jul 13 16:23:27 mickeyl: right, and in other news... the gmpc .bb lacks a dependency on gnome-vfs Jul 13 16:23:38 uh oh Jul 13 16:23:42 i'll add that right away Jul 13 16:23:48 03mickeyl 07org.openembedded.dev * r38c219d6... 10/packages/musicpd/ (glurp_0.11.3.bb gmpc_0.11.2.bb mpc_0.11.2.bb ncmpc_0.11.1.bb): Add console, ncurses, gtk+, and glurp musicpd players courtesy Justin Patrin. Jul 13 16:23:48 any more or is that it? Jul 13 16:24:27 mickeyl: dunno. haven't started from a clean tree Jul 13 16:24:33 k Jul 13 16:25:08 hmmm...it would be nice if thse dependencies were documented clearly... Jul 13 16:25:19 I mean on the packages's site Jul 13 16:25:28 or, say, in the INSTALL or README files (I checked) Jul 13 16:25:29 *nod* Jul 13 16:26:18 JustinP: http://qballcow.nl/index.php?s=13&css=0 Jul 13 16:26:24 mpc depends on: gtk+-2.4, gnome-vfs, libglade Jul 13 16:26:29 gmpc even Jul 13 16:26:58 well.... Jul 13 16:27:16 if in doubt you can always read the configure.ac Jul 13 16:27:45 (for pkgconfig enabled stuff anyway) Jul 13 16:27:49 thanks for checking for me. I'll be more careful in the future Jul 13 16:28:17 heh. no problem. Jul 13 16:28:24 is there any way to easily clear out my compiled packages without losing the cross-compiling environment and such? Jul 13 16:28:51 copy TMPDIR at some point of time Jul 13 16:28:54 or, rather, move away for testing Jul 13 16:28:55 then copy it back Jul 13 16:28:57 03mickeyl 07org.openembedded.dev * re7c80380... 10/packages/musicpd/gmpc_0.11.2.bb: gmpc: add dependency to gnome-vfs. thanks reenoo Jul 13 16:29:21 yeah, but onw' thtat remove the cross-compiler, binutils, etc? Jul 13 16:29:21 there's no way for a package to revert do_stage() yet. Jul 13 16:29:49 copy it right after you compiled the toolchain Jul 13 16:29:56 then you have a blueprint Jul 13 16:30:11 good idea....I suppose I'll have to do that Jul 13 16:30:18 JustinP: I was actually looking into removing gnome deps (gconf, gnome-vfs) from gmpc but a rewrite seems easier Jul 13 16:30:53 is there a target for just the cross-compile environment? Jul 13 16:30:58 JustinP: the author was kind enough to wrap the network communication stuff in a nice little lib... Jul 13 16:31:08 JustinP: bitbake glibc or so Jul 13 16:31:22 renoo_: you can always start with mpc Jul 13 16:36:58 JustinP: nah, that library should be just fine Jul 13 16:49:24 if a "bitbake" goes right through configure / compile / etc but does nothing, what is usually the cause? With one package I'm trying I renamed the bb file and this happened. Jul 13 16:51:29 JustinP: sounds like you didnt change the S vairalbe to point at the unpacked sources in workdir. Jul 13 16:55:56 kergoth: ok, that makes some sense. Can you give me an example of a bb file which does this? Jul 13 16:56:19 .. like 90% of them? Jul 13 16:56:27 grep is your freind Jul 13 16:56:37 grep "S = " */*.bb Jul 13 16:56:39 heh Jul 13 16:59:15 well, yeah, I like grep...I figured it had ${S} and I'm not sure how to grep for that Jul 13 17:00:40 argh... Jul 13 17:02:42 that doesn't seem to be helping int his instance Jul 13 17:06:13 nope, still doing nothign even when I explicitly put the directory in there it still does nothing Jul 13 17:09:24 'night all Jul 13 17:15:16 kergoth: another hint? Jul 13 18:46:34 * france is back (gone 34:26:18) Jul 13 22:17:38 * france is away: Away Jul 13 23:15:34 o.o Jul 13 23:24:26 03noodles 07org.openembedded.dev * r96aa4ca2... 10/packages/uicmoc/ (uicmoc-native-2.3.10/gcc4.patch uicmoc-native_2.3.10.bb): Fix compilation of uicmoc-native under gcc4 **** ENDING LOGGING AT Wed Jul 13 23:59:56 2005