**** BEGIN LOGGING AT Mon Jul 21 02:59:56 2008 Jul 21 06:06:07 * * OE Bug 4447 has been created by thebohemian(AT)gmx.net Jul 21 06:06:09 * * libusb1 0.9.0: contains flags unsuitable for gcc 3.4 Jul 21 06:06:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4447 Jul 21 06:31:01 morning Jul 21 07:27:09 morning Jul 21 07:39:45 good morning Jul 21 07:53:56 hi Jul 21 07:58:16 yop florian Jul 21 07:58:25 hi all Jul 21 08:01:22 bjr Genesis, moin florian Jul 21 08:04:19 oh, ltg is unresponsive again...let's hope that's because mickeyl is moving the tree to git! Jul 21 08:06:18 I wondered what it happened Jul 21 08:31:19 RP: moin Jul 21 08:49:07 hi zecke Jul 21 08:49:11 hi florian Jul 21 08:49:42 pb__: hey. I have an X11 question :) Jul 21 08:50:06 pb__: "Host Portable Character Encoding" any idea what this is nowadays? Jul 21 09:01:17 zecke: ASCII, I guess Jul 21 09:02:24 the Host Portable Character Encoding can be any encoding in which the basic "portable character set" is invariant across all locales that xlib supports. Jul 21 09:03:54 pb__: hmm. have you ever played with the XTest extension? specially sending fake keys? Jul 21 09:04:13 pb__: I wonder how to go from a string in utf8 to the separate keysym's Jul 21 09:04:13 zecke: yah, most of the gpe input methods use that Jul 21 09:04:21 matchbox-keyboard, xstroke,e tc Jul 21 09:05:19 zecke: right, I think this is basically an insoluble problem with the current state of the art. Jul 21 09:05:33 this was exactly the issue that derailed my old "miniclipboard" project. Jul 21 09:06:16 well, not actually insoluble, just quite hard to solve Jul 21 09:07:11 I guess you could solve it for european languages easily enough, but things like chinese and japanese would be very hard. Jul 21 09:08:21 pb__: okay, so basicly I have to use my own mapping... XStringToKeySym is only working for latin1 :( Jul 21 09:08:44 pb__: it might be more easy to make people stop using chinese and japanese and settle with latin1... Jul 21 09:08:50 :-} Jul 21 09:08:59 all hail imperalism :) Jul 21 09:09:13 that said, there is no real xlib-level support for chinese and japanese input anyway. Jul 21 09:09:32 if you are in an openmoko or gpe type world then I would be tempted to ignore xlib and just make a gtk-based mechanism for poking input into applications. Jul 21 09:09:49 03john_lee 07org.oe.dev * r21fd4b09... 10/ (1 contrib/python/generate-manifest-2.5.py): Jul 21 09:09:49 generate-manifest-2.5.py: not in sync with manifest.inc Jul 21 09:09:49 * python-compression with tarfile is not in generate-manifest Jul 21 09:09:54 03john_lee 07org.oe.dev * r0957a92c... 10/ (1 contrib/python/generate-manifest-2.5.py): generate-manifest-2.5.py: pdb requires pprint Jul 21 09:09:58 03john_lee 07org.oe.dev * r37064012... 10/ (1 contrib/python/generate-manifest-2.5.py): generate-manifest-2.5.py: add python-difflib Jul 21 09:10:04 03john_lee 07org.oe.dev * r40452ea1... 10/ (1 contrib/python/generate-manifest-2.5.py): generate-manifest-2.5.py: add python-doctest Jul 21 09:10:06 johncylee: cool Jul 21 09:10:09 03john_lee 07org.oe.dev * r1111600f... 10/ (3 files in 2 dirs): Jul 21 09:10:09 python-2.5-manifest.inc: update by generate-manifest-2.5.py Jul 21 09:10:09 * bump python_2.5.2.bb PR Jul 21 09:10:48 actually, I guess programs like canna must have a way of injecting cjk characters into xlib apps. I've never looked at how that mechanism works but maybe you could reuse the same thing. Jul 21 09:11:16 re Jul 21 09:11:39 pb__: openmoko dropped gtk some time ago... Jul 21 09:11:40 pb__: well, you should be able to send the KeySym, there is kanji, hebrew, arabic... stuff defined... I have not the slightest idea how to map that though Jul 21 09:11:56 hrw: drop? we still distribute it. Jul 21 09:12:13 zecke: with e17/qtopia based image? Jul 21 09:13:06 zecke: ah, right. Jul 21 09:13:07 hrw: the factory image is still Gtk+ based and I have not heard that this is going to change Jul 21 09:13:42 hrw: we suck at handling OM2007.2 bugs and getting patches in, that is true... and we have little resources inside the company to do so... Jul 21 09:14:12 hrw: and even the ASU image contains libgtk, gpe-scap and runs the matchbox settings-daemon... no idea how you say "dropped" Gtk+ Jul 21 09:14:13 zecke: from outside it looks like om2007.2 got dropped Jul 21 09:14:44 zecke: but good to know how it looks Jul 21 09:15:23 zecke: what's the actual input problem that you are trying to solve? Jul 21 09:16:10 hrw: the truth is. most of OM2007.2 was done by o-hand.com and we have no one with us that could maintain the apps... this doesn't mean we dropped Gtk+ though. In fact the SDK will allow you to write Gtk+ or EFL apps :) Jul 21 09:16:50 pb__: for some weird reason it is 2008 and I still do Qtopia stuff. I have ported a input method to X11 and wonder how this would ever support non latin1 input :) Jul 21 09:19:24 * ant|work makes people aware of the ML broken links http://wiki.openembedded.net/index.php/Mailing_Lists (some ltg, some gmane) Jul 21 09:21:21 openembedded-users has wrong url Jul 21 09:23:38 ant|work: It's a wiki Jul 21 09:24:19 hard to say which one looks uglier... Jul 21 09:24:35 I perhaps prefer ltg because there are the month archives Jul 21 09:41:25 morning all Jul 21 09:41:57 RP: mornin' Jul 21 09:43:58 Laibsch: there is a different chronology between gitweb and openembedded-commits (gmane is not ordered ?) Jul 21 09:44:48 ant|work: hello :) need some advice Jul 21 09:45:10 Jay7: ok, but move on #angstrom Jul 21 09:56:41 zecke: heh, right, I see Jul 21 09:57:11 zecke: I guess the h4x0rs over in #freedesktop might have some better advice than me Jul 21 10:24:43 btw, is pidgin 2.4.3 buildable now in .dev? Jul 21 10:25:20 can anybody check it? Jul 21 10:50:19 hrw, you have a minute? Jul 21 10:54:50 hrw, your name is in the MAINTANERS file for Linux, I'm making a minor change to linux_2.6.25, there should be no risk but I know we are suppose to notify any MAINTAINERS before we change a recipie so I wanted to notify you. Jul 21 10:55:05 hrw, I'm just adding the following: http://www.pastebin.ca/index.php and changing the rev to r3. Jul 21 10:56:57 svolpe: show pastebin proper link Jul 21 10:57:12 at91-l9260 is your machine? Jul 21 10:58:36 hrw, opps sorry, about the link. yes, I'm adding the at91-l9260 machine file, so I will also add DEFAULT_PREFERENCE_at91-l9260 = "1" to the linux_2.6.25 Jul 21 10:58:46 http://www.pastebin.ca/1078278 Jul 21 10:59:28 svolpe: ok fot it Jul 21 10:59:36 hrw, thank you. Jul 21 11:06:54 03svolpe 07org.oe.dev * r1a162293... 10/ (3 files in 3 dirs): Jul 21 11:06:54 Added machine file for Olimex AT91SAM development board Jul 21 11:06:54 linux_2.6.25, r3 Added support for Olimex AT91SAM development board Jul 21 11:06:58 03svolpe 07org.oe.dev * r5f9ae2c9... 10/ (1 packages/linux/linux_2.6.25.bb): Added at91-l9260 to supported machines for the linux-2.6.25 kernel Jul 21 11:13:16 svolpe: read commit policy Jul 21 11:13:32 hrw, ok. Jul 21 11:16:16 ah.. pr bump was not needed Jul 21 11:17:56 hrw, I should have made that clearer in my earlier post to you that I was reving it (PR bump) as that was one of the areas I was not sure about, my fault. Jul 21 11:18:48 hrw, what criteria do you normally use for determining if PR needs bumped? Jul 21 11:19:13 svolpe: content of the package changed... Jul 21 11:19:37 zecke, the package and not the recipie. Jul 21 11:19:57 svolpe: right. the resulting packages changed (either more or less, content or a file) Jul 21 11:20:06 s/recipie/recipe Jul 21 11:20:08 svolpe: so someone doing opkg upgrade gets new version Jul 21 11:21:01 zecke, I guess this one was grey then as I added a patch but I was really adding a new machine + patch so no one would need to upgrade as only *new users* using the new machine would see the changes Jul 21 11:21:51 zecke, normally then if you would add a patch I'm assuming you would rev. it since your changing the package. Jul 21 11:22:00 svolpe: new machine added to kernel will not require PR bump - my fault too Jul 21 11:24:57 I just want to make sure I did not mess something up, the mtn notice of my push that shows on the OE mail list shows a bunch of changes I did not make, I'm assuming the extra patch data is related to the merge? Jul 21 11:25:18 the notice I'm concerned about is: merge of '1111600f6f0e98778deec94f838de80fd1db927b' Jul 21 11:26:22 the only two files I touched were conf/machine/at91-l9260.conf and package/linux/linux_2.6.25 Jul 21 11:31:20 i'm introduicing NONFREE_MIRRORS on my distrib but i wonder if we can add nonfree recipe on OE Jul 21 11:31:58 as gentoo port , we could tell user to add themself the files needed by the commercial software Jul 21 11:32:32 but a problem is that this kind of contribution could not be tested , so i suppose it should be hard to include them in OE Jul 21 11:33:48 in an other side , i think it's very boring to package commercial software , so to save time and brain of our FLOSS developer , we should perharps find a solution to share recipes Jul 21 11:33:57 Genesis: Well, we had this with nslu2 before (requiring stuff for networking) Jul 21 11:34:09 what do you think about that ? Jul 21 11:34:10 Genesis: users had to download the source themselves and place them in the DL_DIR Jul 21 11:34:29 oki and you commit recipe on the tree ? Jul 21 11:34:44 my NONFREE_MIRRORS is like DL_DIR ... Jul 21 11:34:49 Genesis: nslu2 folks did... Jul 21 11:35:02 lucky people :) Jul 21 11:35:58 i've only ~6 packages non free Jul 21 11:36:16 most for recognisation Jul 21 11:36:40 recognition Jul 21 11:37:09 audio & video recognition , audio synthesis ... Jul 21 11:37:43 try to convict my team to let them and share the software ( they'll be free =) Jul 21 11:44:28 svolpe: the merge notfication also includes the files that were merged in, not only yours Jul 21 11:45:11 Laibsch, ok, that is what I thought, I have been inactive for a while and I'm just starting to work on OE, I would hate to brake something right out of the gate :-) Jul 21 11:45:36 Laibsch, I ment to say "just starting to work on OE again" Jul 21 11:46:17 Laibsch, thank you for checking it out for me. Jul 21 11:55:31 03thebohemian 07org.oe.dev * ra153d288... 10/ (7 files in 4 dirs): Jul 21 11:55:31 libusb1 0.9.0: Added patch to allow compilation with gcc 3.4 Jul 21 11:55:31 libusb-compat 0.0.9+0.1.0-beta1: Added patch to allow compilation with gcc 3.4 Jul 21 12:55:14 mickeyl, your u-boot-openmoko commit broke u-boot-mkimage-openmoko-native_oe.bb, here is a patch to fix it: http://www.pastebin.ca/1078358 Jul 21 13:03:05 * * OE Bug 4447 has been RESOLVED (FIXED) by thebohemian(AT)gmx.net Jul 21 13:03:07 * * libusb1 0.9.0: contains flags unsuitable for gcc 3.4 Jul 21 13:03:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4447 Jul 21 14:11:29 03mickeyl 07org.oe.dev * r1f87c0f3... 10/ (1 packages/u-boot/u-boot-mkimage-openmoko-native_oe.bb): u-boot-mkimage-openmoko-native oe unbreak, thanks svolpe. Jul 21 14:18:24 mickeyl: do you know anything about s3c24xx mmc support? Jul 21 14:19:30 pb__: hi. not really (still mainly being a userland guy), other than that the we have a bunch of patches against 2.6.x sitting in our tree Jul 21 14:20:08 pb__: we have our a6k wifi module tied to this on gta02 Jul 21 14:20:18 pb__: we use mmc/sd storage using the glamo soc Jul 21 14:20:30 the last set of patches was @ http://svn.openmoko.org/trunk/src/target/kernel/patches/, nowadays we are using a complete git tree which is at git.openmoko.org Jul 21 14:30:25 hrw: ping Jul 21 14:36:43 ant|work: pong Jul 21 14:36:59 hi Jul 21 14:37:04 it's about DEPLOY_DIR_PSTAGE Jul 21 14:37:26 yes? Jul 21 14:37:26 after the recent multimachine fixes, what's your suggestion? Jul 21 14:37:39 out of /tmp ? Jul 21 14:37:53 depends how you use p-s Jul 21 14:37:58 I prefer outside of tmpdir Jul 21 14:38:14 is it multimachine-ready? Jul 21 14:38:40 yes Jul 21 14:38:44 * chouimat|work still wonder if he orders a gta02 Jul 21 14:39:14 DEPLOY_DIR_PSTAGE = "${OEROOT}/build/pstage/" I don't see mention of glibc/uclibc machine... Jul 21 14:39:36 but I never changed angstrom's default Jul 21 14:41:38 hrw: last quesion (for today ;-) Jul 21 14:41:53 do you happen to have a emergency / bootmenu-image? Jul 21 14:42:07 for? Jul 21 14:42:23 em, you know, Zaurus mtd1 Jul 21 14:42:37 ant|work: grab sharp backup from trisoft.de Jul 21 14:42:50 or conics or other place like that Jul 21 14:43:03 I mean custom ,2.6.2x (now we can compile nandlogical Jul 21 14:43:52 you know that old idea of only one kernel in mtd, with booth menu and emergency console... Jul 21 14:44:12 its sick idea a bit Jul 21 14:44:39 ;-} Jul 21 14:45:25 I'd start with bootmenu-image with uclibc Jul 21 14:45:38 integrating mtd-tools and nandlogical Jul 21 14:46:03 hopefully all fits in the 7mb partition Jul 21 14:46:57 hrw: unfortunately the initramfs-bootmenu for zaurus is 'stalled' Jul 21 14:46:58 * Jay7 is back again Jul 21 14:47:13 hrw: thx anyway Jul 21 14:54:31 ant|work: anything for zaurus is stalled Jul 21 14:54:39 less and less people use them... Jul 21 14:55:30 hrw: I'm now spend some time on zaurus Jul 21 14:55:56 but in context of having non-standart partitioning Jul 21 15:00:58 mornin from oscon Jul 21 15:01:48 mickeyl: ah right, cool. I'll take a look at your tree. Jul 21 15:01:54 I am already using your cpufreq patches :-) Jul 21 15:02:36 heh, cool Jul 21 15:02:41 we are not even using them yet Jul 21 15:02:44 heh Jul 21 15:02:50 are they working four you? Jul 21 15:02:59 our kernel experts find them too intrusive Jul 21 15:03:13 yeah, pretty well. I had to hack them a bit to add support for s3c2412. Jul 21 15:03:21 hmm 2412? Jul 21 15:03:23 what's that Jul 21 15:03:31 mickeyl: reciva stuff :) Jul 21 15:03:42 also, on s3c2410 your code seems to want to run the memory at the lowest available bus speed (i.e. 1:4:4 divisor) which is too slow for us. I had to fiddle it to use 1:2:4. Jul 21 15:04:25 pb__: it can't be slow enough... I think due the glamo we have to run it at 100mhz... but I never looked into this Jul 21 15:06:13 mickeyl: s3c2412 is basically an updated s3c2410. arm926 cpu, smaller process, a handful of new features. Jul 21 15:06:49 ... and, annoyingly, a variety of random changes in register addresses compared to the older part. for example, all the EXTINTn registers have been shifted upwards by 0x10 for no apparent reason Jul 21 15:07:22 there's also an s3c2413, which is s3c2412 plus a camera interface (probably the same block as on s3c2440, though I don't really know much about the latter chip) Jul 21 15:07:26 pb__: yes, but apparently Samsung learned their lesson Jul 21 15:07:43 pb__: which bootloader do you use? Jul 21 15:07:55 zecke: our own proprietary one :-} Jul 21 15:08:08 pb__: ah. Jul 21 15:08:42 pb__: do you use nand? do you use software ecc or the hardware one? Jul 21 15:08:52 zecke: yah, with hardware ecc Jul 21 15:09:27 pb__: next thing. how do you flash the firmware? is your bootloader doing that? Jul 21 15:09:28 mickeyl: the main advantage for us of s3c2412 compared to the '2410 is that the '2412 is about $0.50 cheaper Jul 21 15:10:09 zecke: the boards come pre-programmed from the factory, and the firmware reflashes itself (via http download) for upgrades.] Jul 21 15:10:26 the bootloader itself doesn't have any flashing capability Jul 21 15:11:18 pb__: 2412 is arm926? not arm920? Jul 21 15:11:25 pb__: thanks. I currently fix some stuff in our xserver but want to look into hardware ecc... we flash from uboot and were not able to make the software ecc code agree with the hardware... (AFAIK) :) Jul 21 15:12:12 hrw: indeed Jul 21 15:12:51 zecke: ah, that's odd. I don't remember having had any real problems when we switched from software to hardware ecc, back in the day. Jul 21 15:13:40 pb__: do you have a custom init? e.g. starting the app directly from the kernel? if not we have removed a ldconfig that makes us more happy (faster boot) Jul 21 15:13:58 zecke: funnily enough I am working on faster boot-up this week Jul 21 15:14:08 zecke: openmoko boot process is fun Jul 21 15:14:23 zecke: it currently takes us a lot longer to boot on 2.6 compared to 2.4, which is a bit sad. Jul 21 15:14:27 2:53 on normal boot is awesome Jul 21 15:14:33 hrw: not more funny than the stock OE one :) Jul 21 15:14:57 I do not remember when last time I booted OE image... Jul 21 15:15:14 zecke: right now we are using standard sysvinit but with customised scripts. I do plan to ditch sysvinit altogether though since it doesn't really buy us much. Jul 21 15:15:47 pb__: hehe, good thinking. Yeah we are in the same position (would start sysvinit for legacy reasons... e.g. user installed apps) Jul 21 15:15:56 the only benefit it brings us is being able to "ipkg install dropbear" and have it automatically run /etc/init.d/dropbear.sh, and I think there are probably other ways of achieving that effect that will avoid a 2-second startup penalty Jul 21 15:16:24 aside from sysvinit, I think our main offenders right now are mdev and insmod. Jul 21 15:16:50 pb__: oh even mdev is slow for you? we are going to move from udev to mdev to save 10 or 20 seonds... Jul 21 15:17:23 zecke: yah, we currently run "mdev -s" twice, and I think it takes about two seconds each time Jul 21 15:17:43 I hope I can get rid of the first one of those, though last time I tried that I ended up making a brick. :-} Jul 21 15:18:20 pb__: if you see business oppurtunity in selling bricks you are done :) Jul 21 15:18:44 heh Jul 21 15:19:03 right, yeah, I just did "time mdev -s". Jul 21 15:19:08 real 0m 1.18s Jul 21 15:19:08 user 0m 0.11s Jul 21 15:19:08 sys 0m 0.69s Jul 21 15:19:24 not quite as bad as I thought, but still noticeably slow Jul 21 15:19:42 anyway... on to the fun with kdrive... bbl :) Jul 21 15:19:45 enjoy Jul 21 15:19:50 has anyone looked at upstart vs sysvinit for embedded systems? Jul 21 15:20:50 cbrake: not me. I think a custom linuxrc is the "best" way. and then providing something for legacy apps Jul 21 15:21:04 right Jul 21 15:26:32 svolpe: thanks Jul 21 15:45:14 mickeyl, np, thank you for pushing it. Jul 21 15:57:07 morning Jul 21 15:57:16 g'day kergoth`work Jul 21 16:02:44 03mickeyl 07org.oe.dev * r88061bae... 10/ (1 packages/u-boot/u-boot-openmoko_git.bb): u-boot-openmoko git add missing file Jul 21 16:11:13 hmm.. I see that uboot-mkimage-native 1.3.2 from Poky needs merging Jul 21 16:17:00 and this will end life of openmoko-mkimage-native Jul 21 16:21:37 hi kergoth`work Jul 21 16:21:41 mickeyl: alive? Jul 21 16:24:52 kergoth`work: how are you? how is work? (dangerous to ask this on a monday) Jul 21 16:26:34 the job is still quite boring, but i'm turning things around. all last week i was motivated and productive, and this is two mondays in a row that i've been in a pretty good mood Jul 21 16:26:38 scary :) Jul 21 16:27:36 hrw: kind of Jul 21 16:27:50 kergoth`work: heya. glad to hear :) Jul 21 16:28:31 kergoth`work: cool Jul 21 16:28:33 all i needed was goals. it's been 5 years since i really had a direction, career wise. job hopping with no direction makes it hard to stay motivated :) Jul 21 16:28:52 bye people Jul 21 16:29:20 kergoth`work: wasn't taking over the world good enough? Jul 21 16:29:33 mickeyl: read mail on -private and respond to it Jul 21 16:29:52 *nod* Jul 21 16:30:18 mickeyl: then only xora would need to reposnd Jul 21 16:41:02 have a nice evening all Jul 21 16:53:56 bbl Jul 21 17:12:44 do we have iotop in OE? Jul 21 17:20:10 I just did (to OM stuff though...) Jul 21 17:27:35 mickeyl: ping, need some python help Jul 21 17:51:12 Great. Just updated OE, and now I get "ERROR: Unable to open conf/bitbake.conf" when it worked just fine before..... Jul 21 17:57:50 piroko: you don't export BBPATH... Jul 21 17:58:19 zecke: I am though Jul 21 17:58:34 `env` gives BBPATH=/home/jeremy/stuff/local:/home/jeremy/stuff/build/:/home/jeremy/stuff/org.openembedded.dev Jul 21 18:00:48 /bin/sh and then env :) Jul 21 18:01:06 piroko: could be wrong BBPATH or parse error on config files... (bitbake reports that badly) Jul 21 18:02:12 zecke: /bin/sh has BBPATH set too Jul 21 18:02:50 piroko: bitbake -DDD to see what it is loading, this makes config/parse error more likely Jul 21 18:03:39 zecke: Tried that too. It looks like it's trieing to load classes/multimachine.bbclass. But that file doesn't exist... Jul 21 18:03:45 *trying Jul 21 18:04:37 piroko: so it finds the OE conf/bitbake.conf but fails on that? making BBPATH wrong or missing likely again :) Jul 21 18:31:37 zecke: Apparently it's just from this update Jul 21 18:33:35 Lol. I did a `touch class/multimachine.bbclass` and it works now :P Jul 21 18:37:38 morning all Jul 21 18:46:32 what do I need to install to mount a nfs directroy? I feel like I am missing a helper proggram Jul 21 18:49:18 Crofton: mount should be enough Jul 21 18:49:25 hmm Jul 21 18:49:55 Crofton|work: but you need rpc enabled in uclibc and nfsmount enabled with busybox (if you use those) Jul 21 18:50:07 perharps you'll need nfs-utils Jul 21 18:50:09 glibc Jul 21 18:50:26 http://rafb.net/p/zv7Wub34.html Jul 21 18:50:36 showmount -e 192.168.1.11 works Jul 21 18:51:18 kernel has nfs support? Jul 21 18:51:48 (just to make sure) Jul 21 18:52:02 that is what I am wondering Jul 21 18:52:12 what do I need for nfs client support Jul 21 18:53:23 just nfs support from fs->network filesystems (maybe you need to enable sunrpc to be able to select it) Jul 21 18:53:49 you can check /proc/filesystems Jul 21 18:54:05 yep nfs and nfs4 are there Jul 21 18:55:17 showmount sees the directory, so lots of stuff is working Jul 21 18:55:49 are you using real mount or busybox-mount? Jul 21 18:57:10 looks like real mount Jul 21 18:58:48 hm. I usually mount nfs with "-o nolock" but it shouldn't matter. Jul 21 19:13:21 Crofton: did you try mount -t nfs ... ? Jul 21 19:14:28 and check "dmesg | tail" output Jul 21 19:15:40 You need portmap if you don't use "-o nolock" Jul 21 19:25:38 Hi, I know this is not the best place to ask but I have a problem with Xfbdev, I can't find any help with elsewhere. in the tslib testing apps, my touchscreen works perfectly, but in Xfbdev, I get errors like "Invalid update: 0 0 320 240" "Invalid update 0 156 320 84" "Invalid update: 0 15 320 90" "Invalid update 0 144 320 96", plenty in a sequence, when I touch the screen in Xfbdev. Anybody seen something like that? Jul 21 19:49:44 Jay7, yes to all, no real cues Jul 21 19:51:46 Jay7, it has to be something really dumb Jul 21 19:52:19 PEBKAC Jul 21 19:52:21 :D Jul 21 19:52:33 he piroko Jul 21 19:53:02 Man, diet mountain dew tastes worse than imitation mountain dew :( Jul 21 19:55:56 it takes a -lot- of getting used to. i got used to it once Jul 21 19:56:09 cant ever have regular dew though, or you wont be used to it anymore :P Jul 21 19:56:56 Lol Jul 21 19:59:17 neither will your kidney Jul 21 19:59:31 O_o Jul 21 19:59:38 That bad huh? Jul 21 20:03:45 woglinde, your alsa-oss patch breaks compilation Jul 21 20:03:59 shoragan where and why? Jul 21 20:04:54 packages/alsa/files/libio.patch Jul 21 20:04:55 +#ifdef HAVE_FILENO Jul 21 20:04:55 result->_fileno = fdc->fd; /* ugly patchy slimy kludgy hack */ Jul 21 20:04:55 +else Jul 21 20:04:55 + result->_filedes = fdc->fd; Jul 21 20:04:55 +#endif Jul 21 20:04:59 a missing # Jul 21 20:05:18 zrgs Jul 21 20:05:21 sorry Jul 21 20:05:40 will fix it asap Jul 21 20:05:44 good spot thanks Jul 21 20:05:49 thx Jul 21 20:06:11 all non uclibc will probably fail because of that :p Jul 21 20:06:17 nop Jul 21 20:06:18 e Jul 21 20:06:22 it compiled Jul 21 20:06:31 I would notice it Jul 21 20:11:26 piroko: need to consult dewknight http://www.dewknight.com/images/dew2full.jpg Jul 21 20:27:51 ~lart mtn server Jul 21 20:27:51 * ibot slaps mtn server upside and over the head with one freakishly huge killer whale named hugh Jul 21 20:28:51 * mwester wonders why the name of the killer whale is pertinent. Jul 21 20:29:01 03koen 07org.oe.dev * r145d707e... 10/ (1 packages/ffmpeg/omapfbplay/fbplay-static.diff): omapfbplay: check in updated patch Jul 21 20:37:18 hi ant Jul 21 20:37:24 hello Jul 21 20:37:51 woglinde: thx for fixing speech-dispatcher and pulseaudio Jul 21 20:38:24 ant no prob Jul 21 20:38:29 these are dependencies of navit, a nice gps proggy Jul 21 20:38:58 (Berlin is very detailed) Jul 21 20:39:23 shoragan pushed can you quikc test it on glibc? Jul 21 20:39:29 * ant__ can even see the Biergartens Jul 21 20:39:35 hehe Jul 21 20:39:51 03woglinde2 07org.oe.dev * r49118a4c... 10/ (3 files in 3 dirs): Jul 21 20:39:51 alsa-oss: fix libio.patch Jul 21 20:39:51 *stupid else vs. #else from some how dont break uclibc build Jul 21 20:39:51 so I catched not _fildes vs __filedes, fixed now to, thanks Jul 21 20:39:51 to shoragan to spot this Jul 21 20:39:52 * add LEAD_SONAME as suggested by bitbake Jul 21 20:39:54 * bump PR this time the rightway Jul 21 20:41:24 apr give me some headaches Jul 21 20:41:54 woglinde, building now Jul 21 20:56:06 shoragan seems to work right? Jul 21 20:56:32 it compiles at least ;) Jul 21 20:56:40 than it works Jul 21 20:56:49 because is the same code than before Jul 21 21:05:18 why does Fedora nfs-utils install mount.nfs, but not the one from OE? Jul 21 21:05:37 I can't find anysign that it should bui;d even Jul 21 21:21:02 * ant__ inadvertitely updated to rev 1077, crossing the fingers... Jul 21 21:23:54 * ant__ has a new issue: Error running GNU diffutils 3-way difference tool 'diff3' Jul 21 21:24:54 mtn: help required for 3-way merge Jul 21 21:24:58 ?? Jul 21 21:25:28 mtn: we'll help you, disconnecting the rebreather... Jul 21 21:26:23 woglinde: it is the effect of your patch overwriting my hack Jul 21 21:26:33 ant you change locally a file which changed upstreamed too Jul 21 21:26:41 ant wich package? Jul 21 21:26:52 speech-dispatcher Jul 21 21:26:57 ah yes Jul 21 21:27:01 maybee Jul 21 21:27:26 how would git do in this cases? Jul 21 21:27:26 try mtn revert file_in_question Jul 21 21:27:37 and then mtn up Jul 21 21:27:49 ant did you change more than this? Jul 21 21:28:15 yes, couple of lines,long tiime ago Jul 21 21:28:52 woglinde: I'm posting a pastebin Jul 21 21:29:48 then save your file before Jul 21 21:29:53 do the revert Jul 21 21:29:56 http://rafb.net/p/5fz7ik24.html Jul 21 21:29:58 then the up Jul 21 21:30:07 now I'll follow your sugg's Jul 21 21:30:12 put the lines you need back Jul 21 21:31:24 basically mtn got confused Jul 21 21:31:29 no Jul 21 21:31:36 patch Jul 21 21:31:46 ant can you please rework your patch Jul 21 21:31:52 or the patch Jul 21 21:32:03 and put it in the bugtracker Jul 21 21:32:10 then I will apply it Jul 21 21:32:19 AM_CXX flags seems a good catch Jul 21 21:32:20 woglinde: I understand, but the question is about the behaviour Jul 21 21:32:23 expected? Jul 21 21:32:27 yes Jul 21 21:32:39 git would have do the same Jul 21 21:34:37 removed my hack, rerunning script Jul 21 21:34:55 ach, stalls on mtn: connecting to monotone.openembedded.org Jul 21 21:35:03 retry Jul 21 21:35:22 hm??? Jul 21 21:35:25 mtn up Jul 21 21:35:32 should be enough Jul 21 21:36:20 my script does mtn pull and mtn update Jul 21 21:36:40 but now cannot connect... Jul 21 21:37:00 (connected 5 mins ago ) Jul 21 21:38:38 finally connected... Jul 21 21:39:08 ant ah hm Jul 21 21:39:29 you dont need to put in the bugtracker Jul 21 21:39:45 same error, let me post mtn diff Jul 21 21:39:50 ??? Jul 21 21:40:14 did you mtn revert on packages/speech-dispatcher/files/configure_fix.patch ? Jul 21 21:40:19 hi khem Jul 21 21:40:39 hurray apr compiles again Jul 21 21:40:48 woglinde: hello Jul 21 21:47:55 !curse mtn Jul 21 21:48:34 RP: hello Jul 21 21:49:51 woglinde: brute force works (removed and mtn revert --missing ) Jul 21 21:51:45 but pulling is soo slow this evening... Jul 21 21:52:34 let me check Jul 21 21:52:48 pls Jul 21 21:53:04 it may stop all ongoing pulls Jul 21 21:53:19 ant__: I am seeing slow pull here too Jul 21 21:53:47 mickeyl: the hamsters are hungry! Jul 21 21:53:48 * woglinde saw this too Jul 21 21:54:08 hmm, can't see any problems really Jul 21 21:54:12 * khem thinks mickeyl is going to turn on the git knob :) Jul 21 21:54:20 except www-data calling mtn a lot Jul 21 21:54:26 some people are browsing the repository Jul 21 21:54:30 viewmtn is so slow Jul 21 21:54:35 now flushed Jul 21 21:54:36 this'll improve with gitweb Jul 21 21:54:49 khem: very soon now Jul 21 21:54:59 great Jul 21 21:55:06 Has anyone mounted a nfs directory onto a device booting Angstrom/other OE built distr? Jul 21 21:55:08 mickeyl: thats a good news Jul 21 21:55:14 besides using nfsroot .... Jul 21 21:55:22 Crofton: I have Jul 21 21:55:27 grrrr Jul 21 21:55:31 i'm developing with nfs all the time Jul 21 21:55:40 why doesn't it work for me then Jul 21 21:55:44 slugos works great, just did it yesterday Jul 21 21:55:45 mickeyl, not nfs root Jul 21 21:55:50 no, not nfs root Jul 21 21:55:51 Crofton: I have booted angstrom with nfsroot and then nfs mounted some other directories Jul 21 21:55:56 (and not as nfs root) Jul 21 21:56:05 root@om-gta01:/local/pkg/fso/framework/framework# mount|grep nfs Jul 21 21:56:05 192.168.0.200:/local/pkg on /local/pkg type nfs (rw,vers=3,rsize=32768,wsize=32768,soft,nointr,nolock,proto=udp,timeo=7,retrans=3,sec=sys,addr=192.168.0.200) Jul 21 21:56:08 inside emebdded device mount -t nfs 192.168.1.1:/foo bar Jul 21 21:56:11 mtn: error: I/O failure while talking to peer monotone.openembedded.org, disconnecting Jul 21 21:56:29 ant__: try pulling from monotone.openembedded.net Jul 21 21:56:30 do you have mount.nfs and mount.nfs4 installed? Jul 21 21:56:52 mickeyl: its faster now thx Jul 21 21:57:19 Crofton: no, i just have task-base-nfs Jul 21 21:57:26 and whatever that pulls Jul 21 21:57:30 hmmm Jul 21 21:57:57 Nope, no special mount utils on SlugOS either Jul 21 21:58:40 03nail 07org.oe.dev * r88a5ee5d... 10/ (8 files in 4 dirs): iscsi-target: Add package iscsi-target Jul 21 21:58:46 weird Jul 21 21:59:11 task-base-nfs is portmap and the nfs module Jul 21 21:59:15 but I have that builtin Jul 21 21:59:26 and nfs appears in the /proc/filesystems Jul 21 22:00:15 what about lockd and sunrpc modules? Jul 21 22:00:48 hmm Jul 21 22:00:55 let me check log on other end Jul 21 22:01:30 I think you can get away without lockd if you use "-o nolock" or similar when mounting... but not sure about sunrpc. Jul 21 22:02:16 showmount -e host works Jul 21 22:02:55 http://www.pastebin.ca/1078852 is hte error Jul 21 22:03:13 hmm missing codepage .... Jul 21 22:04:42 hmm Jul 21 22:04:43 check dmesg Jul 21 22:04:51 usually it shows that it can't load nls-foo Jul 21 22:04:55 nothing in dmesg Jul 21 22:05:02 oh, strange Jul 21 22:05:35 hmm wrong superblock Jul 21 22:05:38 that sounds really strange Jul 21 22:05:47 I think it is a generic message Jul 21 22:06:20 strace anything revealing? Jul 21 22:06:32 * NAiL re-learns monotone the hard way Jul 21 22:06:49 question: where is defined DEPLOY_DIR_PSTAGE ? In angstrom_c7x0 I have /oe/build/tmp/angstrom/deploy/glibc/pstage Jul 21 22:07:10 but cannot find it by grepping oe tree Jul 21 22:07:13 http://www.pastebin.ca/1078855 Jul 21 22:07:31 looking now http://www.pastebin.ca/1078855 Jul 21 22:08:06 ah, dumb me...packaged-staging.bbclass Jul 21 22:08:06 basically, it looks for mount.nfs and mount.nfs4 Jul 21 22:08:21 then passes the 192 string to the kernel mount call Jul 21 22:08:24 which fails Jul 21 22:08:32 hmm, is this nfs you exported v4-only? Jul 21 22:08:38 probably Jul 21 22:08:52 can you change that? Jul 21 22:09:02 proc filesystems says I support nfs4 though Jul 21 22:09:21 I wonder if there is a kconfig option ..... Jul 21 22:10:16 Crofton|work: Does your kernel support NFS Jul 21 22:10:33 /proc/filesystems says so Jul 21 22:10:37 Crofton|work: message looks like your kernel does not support it Jul 21 22:11:20 CONFIG_NFS_FS=y Jul 21 22:11:48 ah, it is :DEPLOY_DIR_PSTAGE = "${DEPLOY_DIR}/pstage" Jul 21 22:11:53 and not += Jul 21 22:12:08 thus cannot be overwritten... Jul 21 22:13:03 or better ?= Jul 21 22:14:21 khem, set Jul 21 22:15:51 Crofton|work: hmm so kernel should be fine Jul 21 22:16:02 yeah Jul 21 22:16:08 this is driving me wacky Jul 21 22:17:24 Crofton|work: and also CONFIG_NFS_V4=y ? Jul 21 22:17:34 yes ... Jul 21 22:18:00 I have never used nfsv4 Jul 21 22:18:14 but mount -t nfs ... worked for me on angstrom Jul 21 22:18:55 I just can't see the problem Jul 21 22:19:03 hmm Jul 21 22:19:14 let me try another share on an older box Jul 21 22:19:27 Crofton|work: you can see 192.168.1.11:/home/balister/oe is nfs4 mountable on some other computer Jul 21 22:19:41 so you can be sure the nfs server side is ok Jul 21 22:20:04 done :) Jul 21 22:20:58 I'm going to nmap the servers .... Jul 21 22:21:38 nfs and rpcbind are open **** BEGIN LOGGING AT Mon Jul 21 22:24:27 2008 Jul 21 22:35:16 yay for NAiL being back :) Jul 21 22:35:56 expect a bunch of new platform support from team foonas soon ;) Jul 21 22:37:20 timtimred: like which one? Jul 21 22:37:55 we have ... a bunch of orion based devices... thecus n1200 support... and a load of others Jul 21 22:38:13 timtimred: lacie edmini home edition? Jul 21 22:38:41 hmm, we dont have that one... most of the orion NAS are pretty samey though Jul 21 22:39:20 timtimred: I know ... where I work we have a few but still stuck on the 2.6.12.6-arm1 kernel :( Jul 21 22:39:45 send us one and we can take a look... :) Jul 21 22:40:14 timtimred: no I will rather do it my self ... and get paid for doing it ;) Jul 21 22:40:48 i can put you in touch with people who can help and get it upstream for you possibly... Jul 21 22:41:43 timtimred: might be interesting but I will have to talk to my bosses ... I work for axentra the maker of the software on the edmini home edition .. we simply used the kernel that lacie sent us Jul 21 22:42:50 nice. are you running with xfs? if so, run the test suite and get scared. Jul 21 22:43:10 :) Jul 21 22:49:39 chouimat looks like it works-ish already. Jul 21 22:49:49 http://forums.nas-central.org/viewtopic.php?f=146&t=76 Jul 21 22:50:55 downside to vanilla support is the XOR and CESA are not supported. and there are various other nice to have optimisations missing. Jul 21 22:51:23 XOR support is coming, isn't it? Jul 21 22:51:31 Can't remember what CESA was tho Jul 21 22:51:39 yes lennert reckons stuff exists for it but isnt ready Jul 21 22:52:15 its the hardware crypto stuff. Jul 21 22:52:56 03mickeyl 07org.oe.dev * r655fe7c1... 10/ (3 files in 2 dirs): illume svn increase packaging granularity Jul 21 22:53:01 03mickeyl 07org.oe.dev * r25dc6a18... 10/ (4 files in 3 dirs): remove tangogps, it now has its own directory Jul 21 22:53:06 03mickeyl 07org.oe.dev * rb7c57ac5... 10/ (1 packages/freesmartphone/frameworkd_git.bb): frameworkd git increase package granularity Jul 21 22:58:58 timtimred: I had some fun with xfs a few month ago and some ugly patches for our kernel Jul 21 22:59:25 What's a reasonable delay between pulls with mtn? Jul 21 22:59:44 Pulling once a day takes incredibly long time Jul 21 23:00:05 * * OE Bug 3678 has been RESOLVED (FIXED) by bluelightning(AT)bluelightning.org Jul 21 23:00:07 * * Opie: cursor key rotation incorrect on some devices Jul 21 23:00:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3678 Jul 21 23:01:37 it recommends in a certain optimisation doc in the BSP to use it - it scared me too much though :) Jul 21 23:03:22 that and *lots* of reports of fs corruption :) Jul 21 23:04:56 ....not good, when there are alternatives. Jul 21 23:06:02 right sleep, night Jul 22 02:58:06 * * OE Bug 4448 has been created by lukas.gorris(AT)gmx.de Jul 22 02:58:08 * * htckaiser touchscreen not working in Xfbdev (x11-image) Jul 22 02:58:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4448 **** ENDING LOGGING AT Tue Jul 22 02:59:57 2008