**** BEGIN LOGGING AT Fri Aug 27 02:59:57 2010 Aug 27 03:16:42 JDuke128: depends on driver support Aug 27 04:46:11 khem , what u mean ? Aug 27 04:48:02 howdy Aug 27 05:36:50 so is this a good place to bring up something missing from "packages_timezones_files_zoneinfo"? Aug 27 06:05:28 DreamFailure: better write an email to the mailing list Aug 27 07:05:18 03Martin Jansa  07org.openembedded.dev * r7ef38c1996 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Aug 27 07:05:18 linux-openmoko-2.6.34: upgrade to 2.6.34.6 and add patch for WSOD with 242 timming Aug 27 07:05:18 Signed-off-by: Martin Jansa Aug 27 07:39:12 if I am to have my own busybox defconfig file,can I just put it in myoverlay/files/busybox/defconfig ? Aug 27 07:39:28 recipe itself is not in the overlay Aug 27 07:56:18 nevermind, got it, thought I could put the defconfig in files/busybox/busybox-1.15.3 but apparently that does not work and it needs to be one dir up Aug 27 08:07:01 good morning Aug 27 08:18:14 eFfeM_work: it will only work with bblayers and bbappend Aug 27 08:18:22 eFfeM_work: read my blog Aug 27 08:18:32 http://sakrah.homelinux.org/node/2 Aug 27 08:18:43 on how to use bblayers and bbappend Aug 27 08:19:52 sent out the big patch for bitbake world. plz review it and let me know if its ok or needs improvements. Aug 27 08:19:58 * khem time to sleep Aug 27 08:20:58 khem, thanks actually I had a problem in FILESPATHBASE Aug 27 08:21:05 bit will read your blog entry Aug 27 08:21:27 and will look at the bitbake world issue (probably this evening my time) Aug 27 08:22:01 the other issue I had was with fstab, but guess I resolved that one too Aug 27 08:24:33 khem, btw nice blog entry, we should probably put this stuff somewere in the wiki or the manual (and maybe also tell about amend.inc) Aug 27 09:40:20 khem: thx for http://sakrah.homelinux.org/node/1 Aug 27 10:22:14 03Martin Jansa  07org.openembedded.dev * rd171f9869d 10openembedded.git/recipes/linux/linux-openmoko-2.6.32_git.bb: Aug 27 10:22:14 linux-openmoko-2.6.32: upgrade to 2.6.32.21 Aug 27 10:22:14 Signed-off-by: Martin Jansa Aug 27 11:30:43 hi eFfeM Aug 27 11:32:52 i was wondering that if i should make two seperate patches for getting rid of native recipes and converting to new style staging or make just one patch, as i did with makedevs Aug 27 11:34:11 fahad, you can do that in one (typically the do_stage is in the native recipe and when combining them it will probably disappear Aug 27 11:34:29 eFfeM: right Aug 27 11:34:47 eFfeM: so, did you get a chance to take a look at the patch i sent yesterday? Aug 27 11:34:55 but please if you do an oe-stylize make that a separate patch as it is different functionality Aug 27 11:35:04 haven't tested yesterdays patch will do so later today Aug 27 11:35:28 okay, i just did the same with lemon_3.5.4 Aug 27 11:36:01 eFfeM: i'll send that as well, and maybe one or two more Aug 27 11:36:23 ok, thanks Aug 27 11:37:22 in case of lemon, the do_install in native and non native are quite similar, but the non-native one also installs the *man* folder Aug 27 11:37:37 is there a way i can merge them both? Aug 27 11:38:44 right now i have done it like this Aug 27 11:38:47 do_install() { install -d ${D}${bindir} install -m 0755 lemon ${D}${bindir} install -m 0644 lempar.c ${D}${bindir} install -d ${D}${mandir}/man1 install -m 0644 ${WORKDIR}/lemon.1 ${D}${mandir}/man1/ } do_install_virtclass-native() { install -d ${D}${bindir} install -m 0755 lemon ${D}${bindir} install -m 0644 lempar.c ${D}${bindir} } Aug 27 11:39:56 eFfeM: http://pastebin.com/tP7iFcXj Aug 27 11:40:58 just remove the do_install_virtclass-native() . it does not matter that the man pages are also installed for native Aug 27 11:41:32 right Aug 27 11:41:42 don't know what lemon is but I am somewhat upset by Aug 27 11:41:45 install -m 0644 lempar.c ${D}${bindir} Aug 27 11:41:58 it is harmless for native, but installing a .c file in a bin dir ???? Aug 27 11:43:30 ya, dosen't make any sense Aug 27 11:45:56 for now, i am leaving it as it is, but i think i should add a comment when i send the patch, so that someone who knows about it may take a look at it Aug 27 11:46:01 've sent a mail to the list Aug 27 11:46:13 oh, okay Aug 27 12:05:02 hi Aug 27 12:05:47 I'm trying to fix ncurses_5.7 for sdk build : do you think removing sdk from BBCLASSEXTEND will create some trouble (this is the only way I found to get it compiling) Aug 27 12:07:11 it seems that others have the same problem (onb tinderbox) Aug 27 12:09:21 ericben: you can enable widec only if ${BPN} == ${PN} Aug 27 12:09:50 as suggested in "[oe] how to add file to SRC_URI for target only" Aug 27 12:10:26 Ja Aug 27 12:11:15 but fixing widec build for native/sdk/nativesdk would be probably better Aug 27 12:11:27 JaMa|Gone: thanks I hadn't yet read this thread, lot of backlog to catchup ;-) Aug 27 12:11:45 it in case someone is building some app using widec against sdk libs.. Aug 27 12:11:47 yes fixing it would be great but it seems the problem is linked to ubuntu (from the comment) Aug 27 12:12:21 fails the same on gentoo and debian-5 Aug 27 12:12:32 OK so the problem is not linked to ubuntu ;è) Aug 27 12:12:41 I'm going to investigate now thanks Aug 27 12:12:52 yw Aug 27 12:21:37 ericben , think this was also discussed on the mailing list yesterday or so Aug 27 12:21:52 hi mickey| Aug 27 12:23:24 good morning pb_ Aug 27 12:23:27 mickey| hi Aug 27 12:23:45 mickey| care to comment on the minimal.conf patch I submitted yesterday or so to the ML ? Aug 27 12:25:43 eFfeM: done, thanks for reminding Aug 27 12:28:35 np Aug 27 12:30:38 03Fahad Usman  07org.openembedded.dev * rc7fe35ece6 10openembedded.git/recipes/makedevs/ (makedevs-native_1.0.0.bb makedevs_1.0.0.bb): (log message trimmed) Aug 27 12:30:38 makedevs_1.0.0: got rid of separate native recipe and removed legacy staging Aug 27 12:30:38 * converted do_stage to do_install Aug 27 12:30:38 * replaced ${STAGING_BINDIR} with ${D}${bindir} Aug 27 12:30:38 * moved the do_install to non-native recipe Aug 27 12:30:39 * added BBCLASSEXTEND = "native" Aug 27 12:30:40 * added NATIVE_INSTALL_WORKS = "1" Aug 27 12:30:49 03Frans Meulenbroeks  07org.openembedded.dev * r59bed0ad2a 10openembedded.git/recipes/freenx/nxservice_3.4.0-2.bb: Aug 27 12:30:49 nxservice: new recipe Aug 27 12:30:49 Signed-off-by: Frans Meulenbroeks Aug 27 12:30:49 03Frans Meulenbroeks  07org.openembedded.dev * r9a501f160b 10openembedded.git/recipes/kakasi/ (kakasi-native_2.3.4.bb kakasi.inc kakasi_2.3.4.bb): Aug 27 12:30:49 kakasi: merged native and target recipe, removed legacy staging Aug 27 12:30:49 not really a beaty; but we have no simple way to specify a Aug 27 12:30:50 target only patch. This seems the cleanest way to do it Aug 27 12:30:51 Signed-off-by: Frans Meulenbroeks Aug 27 12:30:51 03Frans Meulenbroeks  07org.openembedded.dev * r08f3e210b0 10openembedded.git/recipes/cdrtools/ (cdrtools-native_2.01.01_alpha77.bb cdrtools-native_3.00.bb): cdrtools-native: moved to 3.00 Aug 27 12:55:15 JaMa|GoNe, some problems could arrise: Aug 27 12:55:17 first: Aug 27 12:55:21 bitbake udev-image Aug 27 12:55:26 under SHR distro Aug 27 12:55:36 wtf is udev-image? Aug 27 12:55:44 it doesn't exist Aug 27 12:55:53 but it's for explanation only Aug 27 12:56:01 an image with udev Aug 27 12:56:13 hmm ok Aug 27 12:56:32 so bitbake udev-image should give udev Aug 27 12:56:52 but then gpsd-udev won't be shipped Aug 27 12:56:53 or: Aug 27 12:57:01 it will be Aug 27 12:57:03 bitbake gpsd Aug 27 12:57:06 will that work? Aug 27 12:57:10 without bitbaking an image Aug 27 12:57:18 because your setting is in the image Aug 27 12:57:26 and IMAGE_DEV_MANAGER is evaluated when? Aug 27 12:57:30 maybe at image creation Aug 27 12:57:35 not in gpsd.inc Aug 27 12:57:46 yes gpsd.inc doesn't care about IMAGE_DEV_MANAGER Aug 27 12:57:56 indeed Aug 27 12:58:03 image.bbclass does Aug 27 12:58:08 indeed Aug 27 12:58:17 so what will occur ? Aug 27 12:58:31 or in other words Aug 27 12:58:34 how could it work? Aug 27 12:58:46 you install the image Aug 27 12:58:51 and then you have BAD_RECOMMANDATION Aug 27 12:58:56 which remove the recomandation? Aug 27 12:59:01 which was previously set? Aug 27 12:59:22 so it would require a reinstall? Aug 27 12:59:25 it won't _install_ BAD_RECOMMANDATIONS to image Aug 27 12:59:32 ah ok Aug 27 12:59:36 but what if you do that: Aug 27 12:59:40 opkg install gpsd Aug 27 13:00:04 or better Aug 27 13:00:04 opkg upgrade Aug 27 13:00:15 with that hold in status fill it should IMHO still ignore gpsd-udev Aug 27 13:00:21 s/fill/file/ Aug 27 13:00:50 so... Aug 27 13:00:58 bitbake shr-image #build-host Aug 27 13:01:04 opkg update #device Aug 27 13:01:21 that will convey bad recomandation from the build host to the device? Aug 27 13:01:31 or does it require reinstall? Aug 27 13:01:56 so...under SHR it's fine Aug 27 13:02:09 because we only build shr-image Aug 27 13:02:12 but imagine that: Aug 27 13:02:23 bitbake devtmpfs-image Aug 27 13:02:26 bitbake udev-image Aug 27 13:02:32 bitbake devtmpfs-image Aug 27 13:02:33 etc... Aug 27 13:02:38 under something like angstrom Aug 27 13:02:43 that would be an issue Aug 27 13:02:48 no Aug 27 13:02:55 because at every bitbake the bad recommandation changes? Aug 27 13:02:58 I don't see any issue in that Aug 27 13:03:15 ah ok I think I get it Aug 27 13:03:19 every bitbake *-image creates right image wrt BAD_RECOMMANDATIONS Aug 27 13:03:21 it's set in image Aug 27 13:03:35 so it needs reinstallation to change the settings Aug 27 13:03:39 ok thanks a lot!!!!!!! Aug 27 13:04:06 it's a nice feature, at the cost of reinstallation Aug 27 13:04:07 yes it won't remove already installed package from old image (during opkg update && opkg upgrade) Aug 27 13:04:27 great for angstrom which try to support every machine,image etc... Aug 27 13:13:43 hi eFfeM: should i push the patch for lemon? Aug 27 13:14:20 i've read the mail, but couldn't conclude anything Aug 27 13:20:01 if it builds mail it, i'll test and push, we can improve the recipe if a conclusion is reached Aug 27 13:20:10 btw pushed the other stuff as you probably have seen Aug 27 13:20:19 still have cacao or so on my todo list Aug 27 13:21:56 03Frans Meulenbroeks  07org.openembedded.dev * r1a213c6d7d 10openembedded.git/conf/distro/minimal.conf: Aug 27 13:21:56 minimal.conf: added ivp4 and ipv6 as distro features Aug 27 13:21:56 minimal.conf has nfs as distro feature but not ip4 or ipv6. Aug 27 13:21:56 This makes that busybox does not generate networking applets Aug 27 13:21:56 Signed-off-by: Frans Meulenbroeks Aug 27 13:21:57 Acked-by: Michael 'Mickey' Lauer Aug 27 13:28:41 eFfeM: mailed the patch Aug 27 13:28:55 thanks for updating the recipes for me Aug 27 13:30:15 i must leave now, have a nice weekend Aug 27 13:30:32 take care Aug 27 13:34:30 ok have a nice weekend too Aug 27 13:34:39 btw saw the patch on the list but not nin patchwork :-( Aug 27 13:34:56 can't apply it easily from the list, hope it'll show up Aug 27 13:39:29 it arrived Aug 27 13:39:32 * eFfeM was too fast Aug 27 14:13:18 morning Aug 27 14:14:06 g'day kergoth Aug 27 14:14:39 03Fahad Usman  07org.openembedded.dev * r43543be831 10openembedded.git/recipes/lemon/ (lemon-native_3.5.4.bb lemon_3.5.4.bb): (log message trimmed) Aug 27 14:14:39 lemon_3.5.4: merged native and non-native recipes Aug 27 14:14:39 * removed the lemon-native_3.5.4.bb file Aug 27 14:14:39 * added NATIVE_INSTALL_WORKS = "1" Aug 27 14:14:39 * added BBCLASSEXTEND = "native" Aug 27 14:14:40 * bumped PR Aug 27 14:14:40 Signed-off-by: Fahad Usman Aug 27 14:14:41 hi Aug 27 15:23:02 03Roman I Khimov  07org.openembedded.dev * ra62f59feec 10openembedded.git/MAINTAINERS: Aug 27 15:23:02 MAINTAINERS: add an entry Aug 27 15:23:02 Signed-off-by: Roman I Khimov Aug 27 15:23:02 Acked-by: Frans Meulenbroeks Aug 27 16:08:24 hi, I wonder if I should add myself as a maintainer for the following things: Aug 27 16:08:30 wesnoth Aug 27 16:08:34 maybe some machines Aug 27 16:08:47 like htcdream (if multiples maintainers are possible) Aug 27 16:08:55 or bug 1.x Aug 27 16:09:01 (because I'm the only one who touches it) Aug 27 16:10:55 GNUtoo|laptop, i just touched one this morning. Aug 27 16:11:02 (jk) Aug 27 16:11:03 ok Aug 27 16:11:22 kgilmer, basically it boots to console Aug 27 16:11:28 so usbnet is possible Aug 27 16:11:31 X requires 2 things Aug 27 16:11:41 *calibration support Aug 27 16:12:04 *a special session.d script for handling how wm handle 2 screens Aug 27 16:12:11 because there are differences between wms Aug 27 16:12:22 like illume working fine with the default setting Aug 27 16:12:34 matchbox1 working only one one screen Aug 27 16:12:41 and matchbox 2 needing special care Aug 27 16:12:53 the issue is that I can't test Aug 27 16:13:07 I've touch issues on my 2 screens Aug 27 16:13:29 I've tried with a mouse pointer and it's horible Aug 27 16:13:34 it's ultra-not-precise Aug 27 16:13:40 I bet it's a hardware issue Aug 27 16:14:29 (firmware issue is fixed btw) Aug 27 16:14:49 (the thing where I spent nights and days git-bisecting) Aug 27 16:15:27 ok GNUtoo|laptop Aug 27 16:15:34 so you have bad lcd touchscreens? Aug 27 16:15:38 yes Aug 27 16:15:45 before they worked ok Aug 27 16:15:52 I talked to woglinde Aug 27 16:15:58 and his seem to behave ok Aug 27 16:16:11 maybe transport issues Aug 27 16:16:44 I first noticied when I wanted to integrate navit in 1.4.3 Aug 27 16:16:58 the town/address was unusable Aug 27 16:17:23 basically I wanted to do that: Aug 27 16:17:38 java->NMEA_file->navit Aug 27 16:18:22 then because of screen issue I did some other things that didn't require the screen like for instance the sensor webpage app Aug 27 16:18:56 by worked ok I remember that when calibrated with care,matchbox keyboard was usable Aug 27 16:19:43 yes with a stylus or something Aug 27 16:19:48 still takes some concentration Aug 27 16:20:03 yes but now it's unusable Aug 27 16:20:10 I always use a stylus Aug 27 16:20:38 the offset is very high Aug 27 16:20:59 and you only get some clicks when draging Aug 27 16:21:07 like less than five Aug 27 16:21:11 instead of a continuous line Aug 27 16:21:17 or somewhat continuous Aug 27 16:23:49 GNUtoo|laptop: i think you can add yourself for the things you regularly work on Aug 27 16:24:01 ok Aug 27 17:27:39 gm Aug 27 17:28:29 eFfeM: do you have a calamari board Aug 27 17:30:24 Hi kergoth Aug 27 17:30:34 hey Aug 27 17:30:43 how does bb do with python 3.0 Aug 27 17:30:52 have you tried it in past Aug 27 17:30:55 khem, yes, at work Aug 27 17:31:06 eFfeM: uboot need love Aug 27 17:31:16 eFfeM: IIRC you touched uboot for calamari last Aug 27 17:31:31 ah ok, I know I have some patches at work that I did not push yet Aug 27 17:31:33 eFfeM: the git branch doesnt seem to be there Aug 27 17:31:39 prolly it got merged into master ? Aug 27 17:32:02 yeah, at work I moved to master last week, but didn't get to fully cleaning it up and pushing it Aug 27 17:32:06 will put it on the list Aug 27 17:32:10 I was trying to reproduce your issue but am stuck here Aug 27 17:32:25 ah ok, will take actlion later tonight Aug 27 17:32:25 if you have half baked patch thats ok Aug 27 17:32:30 send it my way Aug 27 17:32:39 atleast I can compile uboot locally Aug 27 17:32:49 ok, give me 15 mins, I'll grab what I have and email it Aug 27 17:32:59 and reach to the point where I can reproduce the gcc issue Aug 27 17:33:04 thx Aug 27 17:37:40 great Aug 27 17:38:00 btw koen seemed not too happy with your patches Aug 27 17:38:26 yeah may be he misunderstood the changes Aug 27 17:39:00 dunno Aug 27 17:39:07 I have explained it and if I dont here back I will assume it yes Aug 27 17:39:42 yeah, I have a similar issue, that's why I rfc-ed the review rules Aug 27 17:40:09 it reminds of of a person( I dont want to name) who was solely responsible for gcc fork Aug 27 17:40:25 because he would reject every patch and project came to a halt Aug 27 17:40:35 and folks said alright lets changes the waters Aug 27 17:40:50 and the real gcc project got sidelined Aug 27 17:41:24 then Richard Stallman intervened and abandoned the original gcc project and requested egcs to called gcc which is what we all use today :) Aug 27 17:42:03 eglibc is somewhat similar story Aug 27 17:42:33 but thats the fun of opensource you got to have a thick skin :) and enjoy the ride Aug 27 17:43:09 * eFfeM blesses sshfs Aug 27 17:43:49 khem, it definitely helps Aug 27 17:44:02 learned a lot from it, both technical and social Aug 27 17:44:26 yes, I have worked enough on OSS that I can preach now :) Aug 27 17:45:12 the thing is we all are humans and no one is perfect so we have to always try to get the best out of other person in some way Aug 27 17:47:53 yup Aug 27 17:49:03 btw wrt calamari and u-boot: i've rebuild recently and the branch is still there, but I can't find the commit any more seems my working copy still use the same git version which ofc it fetches from my downloads Aug 27 17:49:18 can you try with SRCREV 1f932d68bf8c3bb5fec34dc5c5d654fc8614dd31 Aug 27 17:49:26 that is the current head of that branch Aug 27 17:49:37 note to self: fix u-boot Aug 27 17:49:45 hi, concerning ncurses_5.7 build failure for sdk, simply replacing sdk by nativesdk seems to fix the problem and in that case, we get a widechar ncurses in the sdk. Do you think this is an acceptable fix ? Aug 27 17:50:00 actually I'm ironing up all kind of odds and ends for my projrect so I am bumping into all kind of issues Aug 27 17:53:14 khem not sure if you saw it, can you try that SRCREV ? Aug 27 17:57:30 ericben, what uses nativesdk tho? Aug 27 17:57:47 -sdk stuff has real users currently and nativesdk doesn't yet, afaics Aug 27 17:57:59 Or are you working on moving meta-toolchain over? Aug 27 18:00:21 Tartarus: with sdk in BBCLASSEXTEND, bitbake ncurses-sdk fails : http://tinderbox.openembedded.net/public/logs/task/7450858.txt Aug 27 18:00:58 as soon as I switch to nativesdk it builds fine, but I may have totally misunderstood what is nativesdk vs sdk ! Aug 27 18:05:13 so Aug 27 18:06:38 Ah, der Aug 27 18:06:51 I'm reading a thread dated from march and titled "BBCLASSEXTEND sdk vs. nativesdk", this should help me to understand ! Aug 27 18:07:07 Here's the problem Aug 27 18:07:10 There's no virtclass-sdk Aug 27 18:07:15 in OVERRIDES Aug 27 18:07:19 Tartarus : yes Aug 27 18:07:24 SO lets fix that, brb :) Aug 27 18:07:50 what is brb ? Aug 27 18:07:53 be right back Aug 27 18:07:58 Going to take a few min to test Aug 27 18:07:59 ok :) Aug 27 18:09:09 what is strange is that using nativesdk, ncurses-sdk builds fine with widechar support, but using sdk this doesn't work Aug 27 18:09:37 So, if you drop sdk out Aug 27 18:09:42 Then it's 5.4 for ncurses-sdk Aug 27 18:10:17 which explains why that works ! stupid me :-( Aug 27 18:11:12 back in a few tenths of minute, diner time here ;-) Aug 27 18:17:14 03Lukas D. Gorris  07org.openembedded.dev * r8e472b779e 10openembedded.git/recipes/linux/ (linux-qsd8/htcleo/defconfig linux-qsd8_git.bb): linux-qsd8: bumb srcrev, update htcleo defconfig Aug 27 18:23:45 eFfeM: it was using 1f932d68bf8c3bb5fec34dc5c5d654fc8614dd31 and thats whats failing to fetch Aug 27 18:24:17 ericben: yes using nativesdk is right gix Aug 27 18:24:18 fix Aug 27 18:24:26 *thick fingers* Aug 27 18:24:56 khem: No, gdb-cross-sdk needs ncurses-sdk Aug 27 18:25:28 Tartarus: whats the difference Aug 27 18:25:40 Did gdb get transitioned? Aug 27 18:26:22 not yet but shouldnt sdk be really nativesdk Aug 27 18:26:40 No Aug 27 18:26:46 sdk is for meta-toolchain currently Aug 27 18:26:53 nativesdk isn't used really just yet Aug 27 18:27:07 poky has a better example of what's supposed to be going on Aug 27 18:27:08 Hi mickeyl Aug 27 18:27:09 but not quite done Aug 27 18:27:23 Tartarus: eventually it should all be nativesdk thats what I mean Aug 27 18:27:25 good evening khem Aug 27 18:27:46 khem: It's not quite as simple as that, but yes Aug 27 18:28:04 I know :) however in general understanding Aug 27 18:28:23 Tartarus: native should not be used at all in future I think Aug 27 18:28:45 mickeyl: who is heading meego do you know once you told the name of that guy Aug 27 18:29:34 khem: Not sure if -native fully dies or if it's just cross/sdk that become crosssdk/nativesdk Aug 27 18:29:47 And we get to share the same exact compiler/etc inside TMPDIR and for meta-toolchain Aug 27 18:29:51 khem: I'm not sure these days. ex-o-hand has some say and Marcel Holtmann wrt. Intel. Wrt. Nokia it's probably still Quim Gil Aug 27 18:29:59 khem, aarg, i have you teh wrong srcrev, that is the one from the recipe git head for that branch is b80d30546e88c70985094d81297d449b2bc59033 Aug 27 18:30:06 http://git.denx.de/?p=u-boot/u-boot-mpc85xx.git;a=commit;h=b80d30546e88c70985094d81297d449b2bc59033 Aug 27 18:30:10 Holtmann right Aug 27 18:30:14 sorry for the mistake Aug 27 18:30:20 beware though Aug 27 18:30:23 he's not a team player :( Aug 27 18:30:32 hmm Aug 27 18:30:59 but he does good code Aug 27 18:31:06 and knows what he's doing... Aug 27 18:31:14 mickeyl: cool Aug 27 18:31:17 good side Aug 27 18:32:36 mickeyl: whats your take on meego it is gonna go anywhere Aug 27 18:32:59 I wish nokia carried on with mamona Aug 27 18:33:20 eFfeM: I am trying the new rev Aug 27 18:33:27 cool Aug 27 18:33:37 will be back in 15 mins or so Aug 27 18:34:01 eFfeM: OK if it builds I will let you know here Aug 27 18:34:23 khem: it's surely going somewhere given that Intel is pushing like mad. I'm not sure wrt. community acceptance and -- more important actually -- consumer acceptance. It has to fight against Android more than anything else, and this is going to be a tough one. I'd say it depends large on how sexy the hardware is that comes with it Aug 27 18:35:06 I think it's bad that the meego interface relies on open gl Aug 27 18:35:12 yeah I thought exactly same android vs meego Aug 27 18:35:18 and arm vs intel Aug 27 18:35:38 arm is getting into atom market with cortex-a9 Aug 27 18:35:47 and intel wants to get into arm's market Aug 27 18:36:09 AMD is not the enemy no 1 anymore its ARM for intel Aug 27 18:36:13 yep Aug 27 18:36:45 but ARM is too widespread and their licencing mechanism is really nice Aug 27 18:36:53 I dont think Intel will stand that Aug 27 18:37:05 you cant put same atom chip into all devices Aug 27 18:37:34 infact intel should manufacture ARM chips that wud be awesome Aug 27 18:37:47 best chip and bes manufecturing Aug 27 18:38:11 fabrication units of intel are second to none Aug 27 18:41:31 Nokia is seriously lagging in smartphone biz Aug 27 18:41:46 and if meego fails they will loose their edge Aug 27 18:41:49 yep Aug 27 18:41:57 they know how to build decent hardware Aug 27 18:42:01 but the software has been lagging for long now Aug 27 18:42:06 it takes ages to build a great platform Aug 27 18:42:08 right Aug 27 18:42:21 I wonder why they went with meego crap Aug 27 18:42:26 and you need to continue your path, not change all the way Aug 27 18:42:27 and not use something thats out there Aug 27 18:42:38 thats right Aug 27 18:42:40 well said Aug 27 18:42:47 like android did Aug 27 18:43:05 yep. can say lots of bad things about android, but at least they're following a goal Aug 27 18:43:05 they did what they intended and they kept the best foot down Aug 27 18:43:15 sure Aug 27 18:43:19 nothing is perfect Aug 27 18:43:32 that's what Apple did as well. we can learn many lessons from OS X Aug 27 18:43:40 I have Mac fanboys all around me Aug 27 18:44:06 yes Aug 27 18:44:19 whatever you do perseverence is the key Aug 27 18:44:29 and ofcourse it has to be reasonable Aug 27 18:44:39 lunch time Aug 27 18:44:45 bbl Aug 27 18:44:58 ~bon appetit Aug 27 18:44:59 bon appetit is, like, smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito. Buen apetito Bom Apetite. buen apetito Smaklig måltid!. Hyvää ruokahalua. Bo Proveito Aug 27 18:46:08 heh you should add Mahl zeit to it as well :) Aug 27 18:46:15 I lived in the south when I was there Aug 27 18:46:44 hehe Aug 27 18:46:55 ibot: bon appetit is also Mahlzeit! Aug 27 18:46:56 mickeyl: okay Aug 27 18:46:59 ~bon appetit Aug 27 18:47:00 bon appetit is, like, smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito. Buen apetito Bom Apetite. buen apetito Smaklig måltid!. Hyvää ruokahalua. Bo Proveito Mahlzeit! Aug 27 18:47:50 enjoy your meail Aug 27 18:48:28 * dcordes likes 'Smaklig mÃ¥ltid!' best Aug 27 18:49:56 morning Aug 27 18:50:06 khem you still about? Aug 27 19:03:26 re Aug 27 19:23:00 hmmm Aug 27 19:23:03 hi Aug 27 19:23:25 I bet that syntax is wrong: Aug 27 19:24:23 BAD_RECOMMENDATIONS_htcdream += "gpsd-udev" Aug 27 19:24:34 it will += to BAD_RECOMMENDATIONS_htcdream Aug 27 19:24:48 and then BAD_RECOMMENDATIONS_htcdream will overrite BAD_RECOMMENDATIONS Aug 27 19:35:19 hmmm Aug 27 19:35:22 maybe that: Aug 27 19:35:36 BAD_RECOMMENDATIONS_htcdream += "gpsd-udev ${BAD_RECOMMENDATIONS}" Aug 27 19:39:40 khem ping Aug 27 19:45:46 anyone? Aug 27 19:46:02 is that ok: Aug 27 19:46:04 BAD_RECOMMENDATIONS_htcdream += "gpsd-udev ${BAD_RECOMMENDATIONS}" Aug 27 20:14:08 ka6sox-work: hi Aug 27 20:15:08 BAD_RECOMMENDATIONS_append_htcdream = " gpsd-udev" Aug 27 20:15:13 is what will do it Aug 27 20:15:37 ka6sox-work: do you want to reinstall the box Aug 27 20:16:30 khem, hi Aug 27 20:16:37 is that correct: Aug 27 20:16:38 khem, well...I understand that you do... Aug 27 20:16:45 BAD_RECOMMENDATIONS_htcdream += "${BAD_RECOMMENDATIONS} gpsd-udev" Aug 27 20:27:48 03Chase Maupin  07org.openembedded.dev * r1c5eb00adc 10openembedded.git/recipes/ti/matrix-gui-e_svn.bb: Aug 27 20:27:48 matrix-gui-e: change initscript start level Aug 27 20:27:48 * Set the init script to be added to list of run level Aug 27 20:27:48 5 scripts. Aug 27 20:27:48 Signed-off-by: Chase Maupin Aug 27 20:27:48 Signed-off-by: Denys Dmytriyenko Aug 27 20:27:58 03Chase Maupin  07org.openembedded.dev * r138d0aa685 10openembedded.git/recipes/ti/matrix-gui-e_svn.bb: Aug 27 20:27:58 matrix-gui-e: add dependency packages Aug 27 20:27:58 * Added dependency packages such as qt4-embedded-fonts required Aug 27 20:27:58 for the matrix-gui application to run. Aug 27 20:27:58 * Added dependency on matrix-gui-common. Aug 27 20:27:58 Signed-off-by: Chase Maupin Aug 27 20:27:59 Signed-off-by: Denys Dmytriyenko Aug 27 20:32:21 okay khem, I'll refactor garnet to Ubuntu 10.04 Lucid 64-Bit Aug 27 20:36:43 khem, did it build ? Aug 27 20:37:28 eFfeM: yes Aug 27 20:37:37 you get the problem ? Aug 27 20:37:46 eFfeM: and I am able to reproduce the issue Aug 27 20:37:51 that you saw Aug 27 20:37:52 good Aug 27 20:38:10 I will try to look into it but no promises Aug 27 20:38:20 have two birthdays to attend this weekend Aug 27 20:38:27 np, just wanted to report is, as it stands 4.3.3 works for me Aug 27 20:38:41 khem same here, one sat one sun Aug 27 20:39:11 :) Aug 27 20:47:12 eFfeM: it seems there is no effect of the patch I mentioned Aug 27 20:47:16 something is wrong Aug 27 20:49:59 kergoth: we should prune .svn dirs when generating tarballs from svn trees Aug 27 20:50:15 prolly svn export can help Aug 27 20:50:56 khem, do you know anything about gcc-4.5 being slower for c++? Aug 27 20:51:40 Crofton|work: good question, I dont know anything or seen any benchmarks which are C++ and slow Aug 27 20:51:50 Crofton|work: are you experiencing slowness Aug 27 20:52:08 I think so Aug 27 20:52:26 hmmm Aug 27 20:52:40 can you muster some numbers Aug 27 20:52:47 I'm going to run oprofile Aug 27 20:52:51 I''l try Aug 27 20:53:29 it is more I can send x samples/sec with 4.3 and y with 4.5 Aug 27 20:53:45 also, how good is the eglibc math library? Aug 27 20:53:50 hmm Aug 27 20:53:53 as in sin/cos calculation Aug 27 20:54:13 same as glibc Aug 27 20:54:14 hi eFfeM Aug 27 20:54:26 ok Aug 27 20:54:28 does somebody knows if that is correct: Aug 27 20:54:29 BAD_RECOMMENDATIONS_htcdream += "${BAD_RECOMMENDATIONS} gpsd-udev" Aug 27 20:54:38 Crofton|work: make sure to use O2 Aug 27 20:54:42 if you need perf Aug 27 20:54:50 eglibc can be compiled using -Os Aug 27 20:54:51 what about O3? Aug 27 20:54:56 which may not be fastest Aug 27 20:55:05 O3 is another beast Aug 27 20:55:13 sometimes it slows stuff down Aug 27 20:55:16 I'm using angstrom2010 Aug 27 20:55:18 with too much inlining Aug 27 20:55:21 k Aug 27 20:55:26 and cache thrashing Aug 27 20:56:18 JaMa|GoNe, hi Aug 27 20:58:42 um, is the git server up? Aug 27 21:15:12 Crofton|work: hmmm one thing Aug 27 21:15:19 yes? Aug 27 21:15:36 Crofton|work: excess precision is now supported as per standards Aug 27 21:15:48 is your code floating point stuff Aug 27 21:16:48 yeah Aug 27 21:16:50 some is Aug 27 21:17:01 03Frans Meulenbroeks  07org.openembedded.dev * r38b96d1e42 10openembedded.git/recipes/mythtv/mythplugins/ (4 files): Aug 27 21:17:01 mythplugins: mythplugins dir renamed to mythplugins-0.22 Aug 27 21:17:01 0.23 will get its own dir Aug 27 21:17:01 Signed-off-by: Frans Meulenbroeks Aug 27 21:17:02 GNUtoo|laptop: hi (was afk) Aug 27 21:17:20 do you have a link to that Aug 27 21:17:24 khem, if the patch does not wor k it explains things Aug 27 21:17:30 I don't need it to be real accurate :) Aug 27 21:17:43 hi Aug 27 21:18:11 I would suggest to use -fexcess-precision=fast to get the same precision as previous compilers by default gcc 4.5 uses -precision=standard which is slow but accurate Aug 27 21:18:14 03Frans Meulenbroeks  07org.openembedded.dev * rc3e1a6c297 10openembedded.git/recipes/mythtv/ (5 files in 2 dirs): Aug 27 21:18:15 mythplugins-023: fixed it Aug 27 21:18:15 some patches did not apply -> repaired Aug 27 21:18:15 mythflix is not supported any more -> removed Aug 27 21:18:15 mythnetvisions -> added Aug 27 21:18:15 Signed-off-by: Frans Meulenbroeks Aug 27 21:18:19 -feces? Aug 27 21:19:18 eFfeM: I see one problem with the patch I fixed it Aug 27 21:19:20 lets see Aug 27 21:19:35 gcc takes forever to build Aug 27 21:19:38 heh Aug 27 21:19:50 there was some discussion on arm-gnu a while back Aug 27 21:20:01 god gift me a photonic computer Aug 27 21:20:13 I should try --fast-math, which is supposed to tell gcc to favour speed over accuracy Aug 27 21:20:29 khem, i'm calling it a day for now, will check the commit log tomorrow Aug 27 21:20:32 hi, just to be sure: Aug 27 21:20:42 eFfeM: ok gn Aug 27 21:20:55 Jama told me that BAD_RECOMMENDATIONS_htcdream += "${BAD_RECOMMENDATIONS} gpsd-udev" could be replaced by: Aug 27 21:21:06 BAD_RECOMMENDATIONS_append_htcdream Aug 27 21:21:14 is it ok Aug 27 21:21:21 because he seem to hesitate a bit Aug 27 21:21:24 I wrote to you a while ago Aug 27 21:21:30 ah sorry Aug 27 21:21:32 I'll look Aug 27 21:21:45 I wasn't highlighted so I didn't see Aug 27 21:21:46 sorry Aug 27 21:21:54 13:12 < khem> BAD_RECOMMENDATIONS_append_htcdream = " gpsd-udev" Aug 27 21:21:54 13:13 < khem> is what will do it Aug 27 21:22:05 ok Aug 27 21:22:07 thanks a lot Aug 27 21:22:10 np Aug 27 21:22:12 ok Aug 27 21:22:46 I think you were not in room Aug 27 21:23:14 ah ok that's possible because I'm not 24/24 online and am disconnected often(3g) Aug 27 21:23:20 Crofton|work: yeah --fast-math is ok too Aug 27 21:23:52 Crofton|work: also use -fexcess-precision=standard too Aug 27 21:23:59 I need to read some more Aug 27 21:24:15 yes Aug 27 21:24:27 JaMa|GoNe, I'm stupid Aug 27 21:24:37 do you want BAD_RECOMMANDATION for om-gta02 too Aug 27 21:24:44 and om-gta01 ? Aug 27 21:25:08 Crofton|work: gcc 4.5 should generate better code for C++ actually Aug 27 21:25:31 I suspect the fp stuff is hurting me Aug 27 21:25:36 it honors restrict keywork and uses it in real sense for data flow analysis Aug 27 21:25:42 could be Aug 27 21:25:57 we know gcc generates crap for NEON Aug 27 21:26:05 hm Aug 27 21:26:09 GNUtoo|laptop: you can, but it shouldn't be needed as gpsd is usually replaced with fso-gpsd Aug 27 21:26:18 JaMa|GoNe, ah ok Aug 27 21:26:21 NEON isn't IEEEE compliant, Aug 27 21:26:25 which causes issues Aug 27 21:26:28 so I won't do it, only for dream Aug 27 21:26:31 could somebody with channel logs paste me my conversation from around 24hrs ago? Aug 27 21:26:37 even with the right options, the code is still crap :) Aug 27 21:26:45 ~logs Aug 27 21:26:46 All conversations are logged to http://ibot.rikers.org/channel, where "channel" is replaced by the URL-encoded channel name, such as %23freenode for #freenode. Lines starting with spaces are not logged. logs are updated daily Aug 27 21:26:54 these from ibot.rikers.org seem incomplete Aug 27 21:27:34 JaMa|GoNe, I'll also do that: Aug 27 21:27:38 ${@base_conditional('MACHINE', 'htcdream', 'gpsd', 'fso-gpsd', d)} \ Aug 27 21:28:07 GNUtoo|laptop: where? Aug 27 21:28:31 recipes/images/shr-image.inc Aug 27 21:28:40 I replace fso-gpsd by that Aug 27 21:28:56 in RDEPENDS_${PN}-fso = "\ Aug 27 21:29:50 GNUtoo|laptop: be aware that PREFERRED_PROVIDER_gpsd = "fso-gpsd" Aug 27 21:30:10 GNUtoo|laptop: in shr.conf so maybe fso-gpsd will be pulled somewhere anyways Aug 27 21:30:12 hmmm Aug 27 21:30:22 but probably not to RDEPENDS Aug 27 21:30:24 so can I do that: Aug 27 21:30:31 PREFERRED_PROVIDER_gpsd_htcdream = "gpsd" Aug 27 21:30:38 y Aug 27 21:30:45 ok I'll add that too Aug 27 21:31:22 GNUtoo|laptop: but then you have to check every recipe using it to be machine arch Aug 27 21:31:46 GNUtoo|laptop: which is pain and PREFERRED_PROVIDER_gpsd is maybe just wrong Aug 27 21:31:53 Crofton|work: turn off neon and use vfp unit Aug 27 21:32:04 hmmm Aug 27 21:32:05 vfp sucks Aug 27 21:32:13 I don't need IEEE :) Aug 27 21:32:18 heh Aug 27 21:32:18 I need simd Aug 27 21:32:27 ok then neon is for you Aug 27 21:32:31 right now, I am trying to get back to 4.3 Aug 27 21:32:42 then I need to look at NEON intrinsics Aug 27 21:32:52 GNUtoo|laptop: yes it's wrong fso-gps only RPROVIDES gpsd.. Aug 27 21:32:53 JaMa|GoNe, there is no virtual provider? Aug 27 21:33:09 JaMa|GoNe, ? Aug 27 21:33:15 fso-gps only RPROVIDES gpsd.. Aug 27 21:33:18 what do you mean? Aug 27 21:33:20 I'll look Aug 27 21:33:25 ah ok Aug 27 21:33:59 if there is DEPENDS = "gpsd" somewhere fso-gpsd wont be considered as provider Aug 27 21:34:15 ouch Aug 27 21:34:17 I'll look Aug 27 21:34:33 but that's probably right.. Aug 27 21:35:02 foxtrotgps/foxtrotgps.inc:RRECOMMENDS_${PN} = "gpsd" Aug 27 21:35:03 foxtrotgps/foxtrotgps.inc:RRECOMMENDS_${PN}_shr = "fso-gpsd" Aug 27 21:35:04 ouch Aug 27 21:35:05 just someone adding this PREFERRED_PROVIDER_gpsd didn't know that it won't help with RDEPENDS/image content Aug 27 21:36:20 gpsdrive/gpsdrive.inc:RDEPENDS_${PN} = "gdk-pixbuf-loader-gif gpsd" Aug 27 21:36:21 hhmmmm Aug 27 21:36:21 what is the right thing to do? Aug 27 21:36:21 fix every package? Aug 27 21:36:27 * GNUtoo|laptop looks at the manual with providers etc... Aug 27 21:37:00 GNUtoo|laptop: foxtrotpgs is right other way would be to introduce DISTRO_GPSD_MANAGER and use it instead of gpsd/fso-gpsd for all RRECOMMENDS, RDEPENDS, image content Aug 27 21:37:19 hmmm Aug 27 21:37:38 but it still has the problem with distro+machine override yes Aug 27 21:37:44 indeed Aug 27 21:37:56 mickey|zzZZzz, hi Aug 27 21:38:20 fso-gpsd depends on ogpsd Aug 27 21:38:30 and no one knows ogpsd anymore Aug 27 21:38:47 I tried adding htcdream support for it but failed Aug 27 21:39:12 the tricky issue is that the htcdream needs a resource activator that registers to fso Aug 27 21:39:20 and ogpsd already registers the GPS Aug 27 21:39:23 I tried overriding Aug 27 21:39:42 I tried reverting Aug 27 21:39:49 nothing worked Aug 27 21:39:54 and .... Aug 27 21:40:04 that work would be useless in one month Aug 27 21:40:13 because ogpsd is deprecated Aug 27 21:40:29 the gypsy api will go away Aug 27 21:54:20 khem, the excess-precision stuff appears to be x86 only Aug 27 22:11:20 Crofton|work: not really its valid with chips which have FPUs Aug 27 22:12:03 they should fix up the release notes :) Aug 27 22:12:42 oh i take it back Aug 27 22:13:06 could be thats its only effecting x86 Aug 27 22:13:13 as neon is not really an fpu Aug 27 22:13:24 which is doing inexact computations Aug 27 22:18:44 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r2db052490b 10openembedded.git/conf/distro/include/fso-autorev.inc: (log message trimmed) Aug 27 22:18:45 fso-autorev.inc: added android-rpc Aug 27 22:18:45 fso-autorev.inc is usually included in local.conf, Aug 27 22:18:45 by people or builders of the SHR distribution. Aug 27 22:18:45 android-rpc is needed by htcdream for GPS to work. Aug 27 22:18:46 Making it autorev in fso-autorev.inc is a good idea: Aug 27 22:18:46 *it make it stay in sync with other parts of fso. Aug 27 22:18:48 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r12ac96f047 10openembedded.git/recipes/gpsd/ (gpsd-2.95/libtool.patch gpsd_2.95.bb): (log message trimmed) Aug 27 22:18:48 gpsd: add 2.95 version which fixes python binding issues Aug 27 22:18:48 gpsd 2.94 is known to have python binding issues: Aug 27 22:18:48 *From #gpsd in Freenode: Aug 27 22:18:49 Aug 26 14:48:15 GNUtoo|laptop: hmm, 2.94 sounds suspicious, i think the python bindings were bit buggy at that time Aug 27 22:18:49 *And I tried a python test program available here: Aug 27 22:18:50 http://lindi.iki.fi/lindi/gpsd/gpsd-get-position3 Aug 27 22:18:50 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * rd14d8ca807 10openembedded.git/recipes/gpsd/ (files/htcdream/gpsd-default gpsd.inc): Aug 27 22:18:51 gpsd: add htcdream config and bump PR of gpsd.inc Aug 27 22:19:07 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r510367b6ca 10openembedded.git/recipes/linux/ (linux-leviathan/defconfig linux-leviathan_git.bb): (log message trimmed) Aug 27 22:19:07 linux-leviathan: changed ext2 support from m to y and removed extended attributes in ext2/3 Aug 27 22:19:07 Without that we can't mount ext2 filesystem at boot Aug 27 22:19:07 Thanks Lukas D. Gorris (aka dcordes) for pointing it out in irc(#openmoko-cdevel), Aug 27 22:19:07 While trying to help me use his initrd that is capable of booting from an ext2 Aug 27 22:19:07 loop filesystem on a vfat partition Aug 27 22:19:08 (so a Microsoft Windows user could boot our work with fastboot, for the htcdream) : Aug 27 22:38:36 meh Aug 27 22:39:11 i'm so glad this week is over, just one shitty day after another Aug 27 22:41:44 heh Aug 27 22:42:36 At least it's rained up there, yes? Aug 27 22:43:09 a bit. at least it isn't flooding again Aug 27 22:44:17 03Willer Moreira  07org.openembedded.dev * r8d6832e3fb 10openembedded.git/recipes/dspgw-utils/ (4 files in 2 dirs): Aug 27 22:44:17 dspgw-utils: Adding DSP Gateway Utilities package Aug 27 22:44:17 Signed-off-by: Rodrigo Vivi Aug 27 22:44:17 Signed-off-by: Francisco Alecrim Aug 27 22:44:17 Signed-off-by: Willer Moreira Aug 27 22:44:27 03Khem Raj  07org.openembedded.dev * r6cd42a4739 10openembedded.git/recipes/u-boot/u-boot_git.bb: Aug 27 22:44:27 u-boot_git.bb: Update SRCREV for calamari to something fetchable. Aug 27 22:44:27 Signed-off-by: Khem Raj Aug 27 22:44:29 03Khem Raj  07org.openembedded.dev * rcae75453c1 10openembedded.git/recipes/gcc/ (gcc-4.4.4.inc gcc-4.4.4/gcc-add-t-slibgcc-libgcc.patch): Aug 27 22:44:29 gcc-4.4.4: on PPC/SPE when compiling with -Os gcc needs parts of libgcc.a make that happen Aug 27 22:44:29 * The existing patch did redefine the SHLIB_MAKE_SOLINK and Aug 27 22:44:29 SHLIB_INSTALL_SOLINK but it was ineffective because the original Aug 27 22:44:29 definitions appeared after it. We needed to rearrange the order Aug 27 22:44:30 * t-slibgcc-libgcc is already there so no need to patch it. Aug 27 22:44:30 Signed-off-by: Khem Raj Aug 27 23:21:35 03Khem Raj  07org.openembedded.dev * rc8f9779101 10openembedded.git/recipes/openssh/ (6 files in 4 dirs): Aug 27 23:21:35 openssh_5.6p1: Add recipe. Aug 27 23:21:35 * Remove unused patches from files. Aug 27 23:21:35 Signed-off-by: Khem Raj **** ENDING LOGGING AT Sat Aug 28 02:59:57 2010