**** BEGIN LOGGING AT Sat Mar 31 02:59:56 2007 Mar 31 04:20:22 can anyone point me to a reciepe that uses svn? Mar 31 04:56:22 03rwhitby 07org.oe.dev * r1d9f268b... 10/ (3 files in 3 dirs): slugos-init: Add support for handing boot control over to altboot. Mar 31 07:29:10 any of the developers awake? Mar 31 08:06:25 morning all! Mar 31 08:55:24 03rwhitby 07org.oe.dev * rebdffc89... 10/ (3 files in 3 dirs): slugos-init: Remove special altboot support - turnup flash is all that is required. Mar 31 09:10:22 <`CHR1S> hi does anyone know of any documentation about using oe w/ cygwin Mar 31 09:31:57 03rwhitby 07org.oe.dev * r0d8cf148... 10/ (3 files in 2 dirs): altboot: Allow for multiple machines using the same source config directories. Mar 31 09:32:22 `CHR1S: Don't. Use a virtual machine instead. Mar 31 09:32:34 colinux or vmware. Mar 31 09:33:18 yeah, with vmware player and vmware server both available for free, using cygwin for OE development is just making things hard for yourself. Mar 31 09:35:59 03koen 07org.oe.dev * r1d13f53b... 10/ (4 files in 3 dirs): openmoko-messages: add patch from Marcin to make it build again (OM bug #333) Mar 31 09:36:42 <`CHR1S> ok ill give that a try thanks Mar 31 09:38:33 np Mar 31 09:39:43 67108864 100% 13.18MB/s 0:00:04 (xfer#1, to-check=195/198) Mar 31 09:39:50 * koen loves rsync compression Mar 31 09:39:59 :) Mar 31 09:40:04 (that was enschede -> berlin) Mar 31 10:05:58 mrdata: ping Mar 31 10:11:12 hi all Mar 31 10:11:41 gumo mr_nice Mar 31 10:11:49 x29a: hi Mar 31 10:13:38 Lorn Potter in the house? I wonder if his recent activity on the ML means better support for opie? Mar 31 10:16:06 x29a: pong Mar 31 10:16:14 mr_nice: hi Mar 31 10:16:16 mrdata: hi Mar 31 10:16:26 good morning mrdata, just read&answered your email, ill try that image right now Mar 31 10:16:50 x29a: fine and i hope it will work Mar 31 10:16:53 mrdata: do you think its necessary to erase the flash prior reflash? Mar 31 10:17:05 x29a: so I will wait for that Mar 31 10:17:38 mr_nice: yes please, getting the bootloader to work would be a much sweeter solution than flashing again for two days Mar 31 10:18:14 x29a: jepp Mar 31 10:18:27 x29a: yes, because the scanning from the ohh-bootloader for jffs2 partition could end in a endless-loop Mar 31 10:18:43 mrdata: with 0x0? Mar 31 10:19:37 x29a: i not realy shure, but i remeber it would 0xff insted Mar 31 10:20:24 x29a: padding for images is also done with 0xff Mar 31 10:20:37 x29a: for jffs2 images Mar 31 10:20:51 ok, u think first 128blocks is enough? Mar 31 10:21:49 x29a: how big is one block? i have no quick answer yet Mar 31 10:21:56 x29a: the jtag software has an erase function build in which is fast so you can erase the hole flash Mar 31 10:22:33 mr_nice: yeah, i was thinking of eraseflash 128 0xFF then Mar 31 10:22:49 mrdata: i will check Mar 31 10:23:01 x29a: good luck Mar 31 10:23:48 x29a: yes, good luck Mar 31 10:37:02 mrdata: how does the bootloader work? how many 0xFF will it search for a jffs2 image? Mar 31 10:40:11 well im off for lunch anyway now so ill start a 'eraseflash 0xFF 512' which will erase 512Blocks with 0xFF adresses, hope thatll be enough Mar 31 11:05:03 x29a: i don't know, but it scan not for 0xff, the marking for a jffs2 partition begin with 0x85 0x19 Mar 31 11:11:04 x29a: then the booloader do a jffs2 structur check, this check failed and could failed, if flash is not empty Mar 31 11:16:11 x29a: 0x85 0x19 is in image, but in flash the first block should start with 0x19 0x85 0xe0 0x01, the second with 0x19 0x85 0xe0 0x01 Mar 31 11:16:25 x29a: 0x85 0x19 is in image, but in flash the first block should start with 0x19 0x85 0xe0 0x01, the second with 0x19 0x85 0xe0 0x02 Mar 31 11:23:29 x29a: the 0x1985 is called magic bitmask for jffs2 Mar 31 11:23:34 x29a: I think the sane choice is to erase the hole first chip. the ereaseflash function is really fast! after that i would use the hh.org bootloader eflash (http://www.handhelds.org/handhelds-faq/bootldr.html) function to get in a defined state! I have read about many problem where the siemens 2.5.3 erase flash function was not used before flashing the hh.org loader and linux to it Mar 31 11:24:38 mr_nice: but first he need a usable bootloader ;-) Mar 31 11:24:49 mrdata: yes I know Mar 31 11:24:55 mr_nice: while flashig Mar 31 11:25:08 mrdata: so erasing the hole first chip is a good idea I think Mar 31 11:25:26 mr_nice: while flashing, the used partition is already ereased before Mar 31 11:26:05 mr_nice: right, the whole flash is the right thing Mar 31 11:28:20 mr_nice: and flash erase with jtag can take a long time ;-) Mar 31 11:29:01 mrdata: only if you are flashing an image to it. else it is fast Mar 31 11:29:40 mrdata: I think it is using some buildin functions or so. Mar 31 11:30:46 mr_nice: you could be right, the flash has block erase functions, if she where used, it should be fast Mar 31 11:31:42 mr_nice: for me tasks with jtag, i must rebuild a image with jtag, so that bootloader will not be looped Mar 31 11:32:17 mr_nice: and this has take a long time Mar 31 11:36:42 http://rafb.net/p/XtCJvh55.html I wonder what goes wrong here. What do I ned to build suspc? Mar 31 11:37:16 what do I need to build for suspc? Mar 31 11:44:39 03pH5 07org.oe.dev * r7db23f86... 10/ (1 packages/libsexy/libsexy_0.1.11.bb): libsexy: add bugfix release 0.1.11 Mar 31 12:11:20 Laibsch can you pastbin your bb ? Mar 31 12:44:36 mrdata, mr_nice: ok, so i erased 512 blocks with the jtag-eraseblock-function (wrote 0xFF to them) then flashed mrdata 256Kbyte image at address 0 (flashmem 0 ohh_bootldr), reset device, configured minicom to 115k 8N1 and all i get is numbers (adresses?) running down and then ENDM
STKP
C19F3FF0 Mar 31 12:44:59 one block seems to be 13100 bytes aroundish Mar 31 12:45:51 131072K (128KiB) Mar 31 12:46:36 so there is only 128 blocks (i dont know what i broke by erasing block >127) Mar 31 12:47:21 for the first 16777216Byte of flash (first bank) Mar 31 12:47:35 im pretty sure that jtag cant access the second rom (if there was any) Mar 31 12:51:46 entering C19F3FF0 on google brings up that the bootloader wont startup my flash Mar 31 12:52:25 x29a: could you post the whole startup at http://pastebin.ca please Mar 31 12:52:56 x29a: so it could be compared agains mine Mar 31 12:53:17 hmm Mar 31 12:53:30 i must find out how to tell minicom where to log first Mar 31 12:53:46 or i could pipe cat /dev/ttyS0 > log.txt Mar 31 12:53:59 ill try, gotta do some dishes first tho, will you be around for a while? Mar 31 12:55:13 x29a: make minicom in console over x-server Mar 31 12:55:33 x29a: than you could make paste and copy Mar 31 12:57:09 x29a: or use [ctrl+a] [a] and use [l] to log in to a file Mar 31 12:57:18 x29a: or use [ctrl+a] [z] and use [l] to log in to a file Mar 31 13:01:38 great, now i dont get any output anymore Mar 31 13:02:10 i just read via readmem(jtag) the 262144 byte from flash, and there was no difference, so writing/reading/storing seems to work on the flash Mar 31 13:02:48 mrdata: do i need any of the cables attached to the device for hh.org bootldr? ts, sound, what they all are? Mar 31 13:03:05 i think, flash/seriall/jtag will all work without those, right? Mar 31 13:04:27 x29a: yes, onyl power supply is helpfull, for jtag you don't need serial cable Mar 31 13:05:23 x29a: but after successfully jtaged the booloader, the serial cable is neccessary, to see whats going on Mar 31 13:05:37 zecke: hi Mar 31 13:05:43 yeah, i got the result i wrote above, no i dont get any output on serial Mar 31 13:05:47 ill reset all devices Mar 31 13:07:38 x29a: befor come ENDM STKP C19F3FF9 must be other stuff, this could be helpful Mar 31 13:09:34 x29a: your simpad should be right assambled together, the whole hardware, only the case is unimportant Mar 31 13:12:39 before ENDM STKP C19F3FF9 is only a lot of hexnumbers Mar 31 13:12:55 ill try to get the log working, but right now i dont get any serial output anymore Mar 31 13:13:07 and yeah, ill attach all the cables for the next try Mar 31 13:15:01 x29a: and control your flex cable connections of the boards for correct matching Mar 31 13:15:08 mrdata: what does the connector do, which is above 18 and 8 on http://opensimpad.org/index.php/Image:SIMpad_small.jpg Mar 31 13:19:54 x29a: going lunch now, in 15 min i will be back, between 18 and 8 is nothing i can see, 18 must be loud-speacker Mar 31 13:20:41 x29a: and 8 is uart3 for dect-module or other things like bluetooth and mmc/sd stuff ;-) Mar 31 13:24:56 x29a: above 18 and 8 is the connector of the buttons and leds. Mar 31 13:27:20 x29a: which is connected to cs3 latch and sa1100 gpio0 (only the power button) Mar 31 13:51:20 ~lart sound-juicer for being broken on feisty... Mar 31 13:51:20 * ibot hereby declares sound-juicer a troll for being broken on feisty... Mar 31 13:52:18 x29a: lunch over Mar 31 13:52:39 mrdata: i just returned, ill try now again with my simpad Mar 31 13:53:16 x29a: okay, i waiting for progress Mar 31 14:00:22 morning all Mar 31 14:01:12 03pH5 07org.oe.dev * r90c92673... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: bump hx4700 and magician kernel version to 2.6.20-hh5 Mar 31 14:01:22 good morning RP Mar 31 14:02:23 mrdata: all i now get on serial is Y Mar 31 14:02:25 pardon Mar 31 14:02:27 'y' Mar 31 14:02:36 when i reset and power up again, just another 'y' Mar 31 14:04:16 x29a: than the serial interface is not setup proberly, the initial setup is not completed Mar 31 14:04:25 hey RP Mar 31 14:04:56 hi Mar 31 14:05:40 x29a: what have you changed between your lot of hexnumber outputs and now Mar 31 14:06:04 i read 262144 bytes via jtag to verify consistency Mar 31 14:06:23 now i reflashed yout bootldr again, to set the flash into same state as before Mar 31 14:06:25 x29a: the hexnumbers has tell you, that initial setup was okay Mar 31 14:06:41 oh, but why did it not go any further? Mar 31 14:06:42 x29a: this is a good idea Mar 31 14:07:14 a good question? Mar 31 14:08:09 x29a: after the C19F3FF0 the next step is looking for the right flash Mar 31 14:08:43 and that did obviously not happen Mar 31 14:09:33 sodo you think it might be the ram? Mar 31 14:09:37 x29a: yes, the bootloader do probe, witch flash is there, witch geometry, size, organisation, how many banks Mar 31 14:10:00 u mean flash (where bootloader sits) or ram? Mar 31 14:10:25 x29a: no, the flash must be found from bootloader Mar 31 14:11:46 x29a: first flash with bootloader, copy to ram, start from there, search for components on board Mar 31 14:12:45 x29a: the bootloader have not know, witch hardware is present, so he must try and test Mar 31 14:13:36 x29a: the ooh-bootloader knows also other machines Mar 31 14:14:07 x29a: but must find the right one, witch is present Mar 31 14:17:06 mrdata: http://rafb.net/p/OHZWtD31.html Mar 31 14:19:45 how to procede? Mar 31 14:22:55 pH5: Have you committed the epoch changes the OE metadata needed or not yet? Mar 31 14:23:59 RP: they are committed (both ipk and deb packages include epoch information now). Mar 31 14:24:39 pH5: ok. I'll look at adding epoch support to bitbake 1.8.x when I get a moment Mar 31 14:25:08 RP: great. it seems to work just fine. Mar 31 14:25:27 pH5: Great. It should clean up the X11RX files quite nicely :) Mar 31 14:26:35 * hrw built openmoko-image with x11r7.2 and it works Mar 31 14:27:54 x29a: on top of your list i have by me four exta lines: U3 @00000348 F04000000 and one empty Mar 31 14:28:06 oh, right Mar 31 14:28:08 is OE going to require bitbake 1.8.1 at that point? Mar 31 14:28:18 mrdata: i forgot to copy those as well Mar 31 14:28:28 mrdata: U3 Mar 31 14:28:28 @00000348 Mar 31 14:28:28 F04000000 Mar 31 14:28:33 x29a: ah, okay Mar 31 14:28:44 and now? Mar 31 14:28:51 im totally clueless Mar 31 14:29:16 x29a: no, no, we have a simpad with serial messages to us Mar 31 14:29:29 koen: We'll have to consider requireing 1.8.2 (next stable release) although we have the omoko guys to consider Mar 31 14:29:29 i think thats good ;) Mar 31 14:29:55 mrdata: wonna switch to query/different channel for further discussion? Mar 31 14:30:04 RP: the OM people are using a fixed rev anyway Mar 31 14:30:34 koen: and they plan to update OE during next month Mar 31 14:30:49 hrw: so they can update bitbake at the same time Mar 31 14:31:54 probably Mar 31 14:31:56 x29a: so witch channel would you prefer Mar 31 14:32:24 someone use gtk-webcore 20060212 or everybody use 20050430? Mar 31 14:32:33 x29a: I think it would be nice to have the conversation logged. It could help others. Mar 31 14:32:35 mrdata: i dont care, if you want to stay here im fine (for the bot and logs) but not that we interfere with hrw and koen Mar 31 14:33:03 mr_nice: yeah, that was my thought as well, we could switch over to #handhelds.org or #familiar since theres not much goin on Mar 31 14:33:27 x29a: or create #simpad channel? Mar 31 14:33:31 koen: Lets wait and see for now - there isn't an immediate problem. There are some other task dependency changes I think we'll have to make to OE.dev which will need 1.8,2 though... Mar 31 14:33:38 hrw: can we request a bot? Mar 31 14:33:48 x29a: no idea but probably Mar 31 14:33:56 The infamous quilt-native vs. do_patch problem... Mar 31 14:34:07 #simpad actually existed Mar 31 14:34:25 hrw: people using gcc4 need 20060212 Mar 31 14:34:42 x29a: that is okay, i will not disturb the work there Mar 31 14:34:44 hrw: is OZ going to have a new release or not? Mar 31 14:34:46 koen: but does it change functionality? Mar 31 14:34:57 hrw: I need to know that for the angstrom release planning Mar 31 14:35:12 koen: good question. Mar 31 14:35:22 koen: I want to give 3.5.5 for collie/2.4 atleast Mar 31 14:35:35 koen: I do not like idea of angstrom-2.4-legacy Mar 31 14:35:36 not using angstrom-legacy? Mar 31 14:36:07 koen: do you want to support people which will have legacy machines and will try to install stuff from !legacy feeds? Mar 31 14:36:24 koen: and users want OPIE images Mar 31 14:36:25 personally I don't want to support anything Mar 31 14:36:51 hrw: Users want Opie II images :) Mar 31 14:37:10 zecke: do not make me laugh ok? Mar 31 14:37:28 hrw: laughing is healthy Mar 31 14:37:32 zecke: but thx for remind. I have to resign from that project Mar 31 14:38:37 resigned Mar 31 14:39:00 koen: so who will support angstrom? users? Mar 31 14:39:02 ~users Mar 31 14:39:03 so this is what we get for sharing our unpaid volunteer work with you? complaints and accuses? well done, this clearly supports our motivation to continue working on open source projects. Mar 31 14:40:06 hrw: ideally, yes Mar 31 14:44:41 s/accuses/accusations/ Mar 31 14:46:18 ERROR: Exception:exceptions.IOError Message:[Errno 2] No such file or directory: '/home/oe/autobuild/trees/incremental/oe/packages/uboot/files/./uboot-20070311-tools_makefile_ln_sf.patch' Mar 31 14:46:29 hmm where to place the try/catch to solve this... Mar 31 14:48:53 zecke: mtn pull Mar 31 14:48:58 or svn update even Mar 31 14:49:27 hrw: yes, this was just fixed. Still BitBake did not except that -> broken Autobuild Mar 31 14:50:26 pH5: NOTE: preferred version 7.0.10 of xproto-native not available (for item xproto-native) Mar 31 14:51:45 pH5: xdpyinfo from 7.2 is lacking too Mar 31 14:53:34 * hrw go out - cu Mar 31 14:53:39 hrw: thanks, fixing now Mar 31 14:54:20 many protos are 7.0/7.1 Mar 31 14:54:30 hrw: many protos were not updated Mar 31 14:54:46 ok Mar 31 14:55:13 no plans to copy them as epoch ones? Mar 31 14:56:01 hrw: I didn't want to introduce unneeded changes. Mar 31 14:56:03 Do you think we should do that for cosmetic reasons? Mar 31 14:56:39 rather not Mar 31 14:57:10 03pH5 07org.oe.dev * rc292afe1... 10/ (1 packages/xorg-proto/xproto-native_7.0.10.bb): xproto-native: add 7.0.10 (X11R7.2) Mar 31 14:58:23 oops, actually xdpyinfo 1.0.2 was post-X11R7.2 Mar 31 14:58:34 cu Mar 31 14:58:57 03pH5 07org.oe.dev * r2b9fbd2b... 10/ (1 packages/xorg-app/xdpyinfo_1.0.2.bb): xdpyinfo: add 1.0.2 (X11R7.2) Mar 31 15:09:36 hey mickey|brunswick Mar 31 15:09:45 mickey|brunswick: say hi to the others for me :) Mar 31 15:15:52 Hello Mar 31 15:18:38 hi Mar 31 15:37:09 morning Mar 31 15:59:25 03freyther 07org.oe.dev * ra1ada670... 10/ (1 classes/patch.bbclass): Mar 31 15:59:25 classes/patch.bbclass: Make sure to raise func_failed on any exception from within Import Mar 31 15:59:25 non existing patches raised a IOError by the md5sum method which was not Mar 31 15:59:25 catched at all and lead bitbake to exit due an unhandled exception. This Mar 31 15:59:26 is bad for all autobuilders. Mar 31 16:14:51 zecke: Does do_package_qa need pax-utils-native in insane.bbclass ? Mar 31 16:15:29 ah, its a PACKAGEFUNCS :/ Mar 31 16:15:32 ignore me... Mar 31 16:16:15 /ignore RP Mar 31 16:17:58 RP: BTW: did you take a look at the memory leak bug report? Mar 31 16:18:32 RP: (I just remembered it) do you know if we generally leak or if we hit a recursion that allocates more memory before the stack the full? Mar 31 16:18:56 zecke: I glanced at it, I don't think its a general problem Mar 31 16:19:24 maybe I don't understand but I think there is something wrong with this picture: db->dbus->"network protocol converter"->dbus->application (language independence) Mar 31 16:19:51 hi zecke RP Mar 31 16:19:57 hi chouimat Mar 31 16:20:11 hi chouimat Mar 31 16:55:23 03justinp 07org.oe.dev * ra664968f... 10/ (1 packages/efl/edje-native_0.5.0.037.bb): edje-native: make sure to use the right edje-config Mar 31 17:07:59 koen: ping Mar 31 17:08:14 koen: do you remember the mtdblock/rootfs thingy? Mar 31 17:10:07 mickey|brunswick: yes Mar 31 17:11:32 mickey|brunswick: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-February/001490.html ? Mar 31 17:11:51 thank you Mar 31 17:11:53 do you have any idea for Mar 31 17:12:00 !ombug 253 Mar 31 17:12:01 * * Bug 253, Status: NEW, Created: Unknown Mar 31 17:12:01 ? Mar 31 17:12:01 * * oe(AT)hentges.net: Typo in /etc/init.d/checkroot.sh prohibits correct processing /etc/fstab Mar 31 17:12:02 * * http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=253 Mar 31 17:13:48 I'd replace $(....) with `...` Mar 31 17:15:03 mickey|brunswick: hmm Mar 31 17:15:14 mickey|brunswick: I suspect the grep returns nothing Mar 31 17:15:17 adding " around the $(...)" helps also, however i dunno whether this is correct Mar 31 17:15:34 i wonder why this hasn't been noticed yet Mar 31 17:15:35 mickey|brunswick: x$(..) = "xrw" ? Mar 31 17:15:39 it's the same in OE Mar 31 17:15:53 mickey|brunswick: I noticed it, but I though it was because I am using bash instead of busybox Mar 31 17:16:50 x$ = xrw fixes it also Mar 31 17:17:01 however I'm still not sure whether this is functioning correctly Mar 31 17:17:05 this script is insane Mar 31 17:17:09 mickey|brunswick: http://www.monkey.org/openbsd/archive/misc/0407/msg00784.html Mar 31 17:17:49 koen: exactly Mar 31 17:18:07 not double-quoting variables in sh asks for trouble Mar 31 17:18:16 I always assume that 'old' stuff is correct in OE Mar 31 17:18:21 heh Mar 31 17:18:22 don't Mar 31 17:18:33 but lately it turns out people papered over bugs Mar 31 17:18:43 e.g. not packaging libgcc for gcc non-cross Mar 31 17:18:56 so the preferred way to fix this is? Mar 31 17:19:04 x or add " around the $() Mar 31 17:19:30 double quotes + x, commit to OE Mar 31 17:19:41 mickey|brunswick: I'd use double-quotes but that's just me Mar 31 17:20:04 koen: the x is kinda overkill when you already use double-quotes =) Mar 31 17:20:12 hmm, /me tries Mar 31 17:20:18 CoreDump|afk: it is Mar 31 17:20:32 but it marks it as 'can be empty' a bit more Mar 31 17:25:35 mickey|brunswick: HXD8 is the hardware refresh for p2? Mar 31 17:28:16 no Mar 31 17:35:10 mickey|brunswick: any hint on what it is? Mar 31 17:35:25 not a phone Mar 31 17:35:53 hmm Mar 31 17:35:55 ok, double quoting fixes this bug however, /var still doesn't get mounted Mar 31 17:35:56 now it fails with Mar 31 17:36:02 "mount: /proc busy" Mar 31 17:36:06 the uboot patches say it has gsm and gps :) Mar 31 17:36:54 koen: patches never lie... Mar 31 17:44:05 hmm Mar 31 17:44:22 can anyone confirm this to be a problem on another OE supported machine as well? Mar 31 17:44:34 mickey|brunswick: I already did that :) Mar 31 17:44:54 mickey|brunswick: that's why I said you should commit the fix to OE :) Mar 31 17:46:48 well Mar 31 17:46:55 it doesn't fix the situation here Mar 31 17:47:05 can you confirm the "proc: busy" as well=? Mar 31 17:47:10 [after the fix] Mar 31 17:48:00 let me build a new kernel for the hx first Mar 31 17:48:17 oki Mar 31 17:51:54 * koen waits for the efika build to finish Mar 31 17:55:39 can we find a log of checkroot.sh? Mar 31 17:56:11 bkbits.net Mar 31 17:56:52 eeks Mar 31 17:56:55 ancient Mar 31 17:56:57 mtn annotate + mtn log ? Mar 31 17:57:01 but might be helpful Mar 31 17:58:02 * CoreDump|afk remembers messing with it in the distant past Mar 31 17:58:13 mickey|brunswick: http://ewi546.ewi.utwente.nl/tmp/checkroot-annotated.txt Mar 31 17:58:37 zecke: and no, I didn't write OE in one day Mar 31 17:58:43 hehe Mar 31 17:58:46 =) Mar 31 17:58:53 mickey|brunswick: http://www.openembedded.org/bonsai/query/file/checkroot.sh/?items=100 Mar 31 18:02:52 hmm Mar 31 18:02:55 can i see the diffs? Mar 31 18:04:28 * mickey|brunswick tries one of the other checkroot.sh versions Mar 31 18:04:37 mickey|brunswick: not (yet), only using viewmtn Mar 31 18:13:37 even using another checkroot.sh doesn't make Mar 31 18:13:47 mount: mounting proc on /proc: device or resource busy Mar 31 18:13:52 go away :/ Mar 31 18:14:24 mickey|brunswick: is proc already mounted? Mar 31 18:16:38 mount sys and proc mount in checkroot should check if sys and proc are already mounted IMO Mar 31 18:17:05 I think we can all agree on that Mar 31 18:18:44 the main problem is that mount doesn't continue Mar 31 18:18:50 it encounters the error w/ proc Mar 31 18:18:57 then doesn't mount the others Mar 31 18:19:01 i need to issue mount -a Mar 31 18:19:03 to get /var mounted Mar 31 18:19:51 right Mar 31 18:20:06 so my question Mar 31 18:20:09 did mount change in behaviour Mar 31 18:20:15 mickey|brunswick: what will happen in 2008? will FIC rename their svn tree? Mar 31 18:20:17 or how did we introduce this bug Mar 31 18:20:32 koen: no idea, i don't think that far :D Mar 31 18:20:56 * mickey|brunswick suspects udev Mar 31 18:20:59 mickey|brunswick : busybox mount is broken Mar 31 18:21:11 since when? Mar 31 18:21:15 that's why angstrom uses proper mount Mar 31 18:21:16 mickey|brunswick: I'm not seeing a mount -a in checkroot.sh... Mar 31 18:21:28 i really can't remember this to being a problem the last couple of years Mar 31 18:21:37 CoreDump|afk: was that there once? Mar 31 18:21:52 actually busybox mount supports "mount -o move" _WAY_ better than the real mount Mar 31 18:21:52 * RP wonders why it isn't seen in poky... Mar 31 18:22:01 * mwester notes that busybox has become somewhat less stable recently... Mar 31 18:22:15 mickey|brunswick: well I dunno, but why should the busy proc mount affect other scripts then? Mar 31 18:22:24 zecke: Does the parser support the idea of do_package[depends] += "file-native" ? Mar 31 18:23:35 koen: so using the full mount cures it? Mar 31 18:24:11 if that were the case, we would see the same bug on OZ Mar 31 18:24:25 RP: it could :) Mar 31 18:24:35 anyone running OZ from .dev here who could check? Mar 31 18:26:14 zecke: Would it be simple to do? I have use case :) Mar 31 18:26:31 mickey|brunswick: /proc is indeed already mounted after the "busy" message so that shouldn't affect other scripts Mar 31 18:26:53 ok Mar 31 18:26:57 so the bug is somewhere else Mar 31 18:27:05 mickey|brunswick: proper gives less problems in other areas Mar 31 18:27:11 old parser == linux2.4 but I'm a whore :) Mar 31 18:27:14 * CoreDump|afk reflashes his neo....placing a "read junk" in checkroot.sh wasn't the ideal thing to do an a device w/o enter.... Mar 31 18:28:20 RP: but it should just work Mar 31 18:28:30 RP: it is an append to a flag... Mar 31 18:28:44 RP: and what is the usecase? Mar 31 18:30:04 zecke: I want to add a "depends" flag to tasks which takes the form do_patch[depends] = "quilt-native:do_populate_staging". In the package classes, we need to append to that for say insane'.bbclass's pax-utils-native dependencies Mar 31 18:30:27 zecke: and I don't think it just works :-( Mar 31 18:31:40 hmm Mar 31 18:31:50 wth does mount -at nonfs,nonfoo,nonbar do? Mar 31 18:32:14 mount everything _except_ nfs foo bar filesystems Mar 31 18:32:24 (e.g. networking isn't up yet) Mar 31 18:32:26 zecke: you mean you apply this? http://clipgallery.blogspot.com/2007/01/truth-about-working-in-it-industry.html Mar 31 18:32:40 ah Mar 31 18:32:42 this command does not work Mar 31 18:32:47 it just doesn't mount /var Mar 31 18:33:05 chouimat: I know, but I think an average prostitute earns more :) Mar 31 18:33:25 zecke: a lot more ... Mar 31 18:33:27 ok, i got it Mar 31 18:33:34 mount -at whatever doesn't work in busybox Mar 31 18:33:48 that should be fixed =) Mar 31 18:33:54 so now to found out why did this work in earlier versions Mar 31 18:34:00 or didn't it? Mar 31 18:34:12 chouimat: and I don't agree with 10 Mar 31 18:34:16 * CoreDump|afk dunnos Mar 31 18:34:29 RP: well, can I test this with BitBake1.8? Mar 31 18:35:16 mickey|brunswick: busybox mount got rewritten multiple times Mar 31 18:35:46 * zecke emulates rms and is happy that the fuse repsonsible for the light is broken... (for several weeks) Mar 31 18:36:31 zecke: go work for xandros ... after 3 months you will agree with this one :) Mar 31 18:37:56 grrrr, this efika build is taking ages Mar 31 18:38:04 zecke: Yes, place this in package.bbclass:do_package[depends] = "file-native:do_populate_staging" Mar 31 18:38:04 do_package[depends] += "fakeroot-native:do_populate_staging" Mar 31 18:38:37 zecke: and then add a print statement to add_task in build.py which prints the depends varflag Mar 31 18:39:00 zecke: The result is a puzzling: base_do_package Mar 31 18:39:00 fakeroot-native:do_populate_staging Mar 31 18:40:19 * RP suspects feeder in ConfHandler.py doing a getVar instead of a getVarFlag Mar 31 18:40:49 !oebug 1529 Mar 31 18:40:50 * * Bug 1529, Status: NEW, Created: 2006-10-25 10:28 Mar 31 18:40:51 * * christopher.lang(AT)chl-informatik.de: mountall.sh does not mount tmpfs filesystems Mar 31 18:40:52 * * http://bugs.openembedded.org/show_bug.cgi?id=1529 Mar 31 18:42:23 mickey|brunswick : not sure that it ever worked correctly Mar 31 18:42:40 RP: give me another half an hour Mar 31 18:42:55 mickey|brunswick : i think there was a bug on this Mar 31 18:51:33 ok, thanks guys Mar 31 18:51:36 at least i found it Mar 31 18:55:01 zecke: http://www.rpsys.net/openzaurus/temp/bb-flag-appends.patch might even be a fix. See if you think its safe to commit ;-) Mar 31 18:55:09 * RP -> back later Mar 31 18:57:12 * chouimat is listening to Fear Bill Gates by Illdisposed on Kokaiinum [Amarok] Mar 31 19:01:17 RP: is safe :) Mar 31 19:04:58 night everyone Mar 31 19:06:25 koen|away: ping? Mar 31 19:07:12 03freyther 07org.oe.dev * ra73ca61d... 10/ (5 files in 3 dirs): Mar 31 19:07:12 qa/checksum: Add python code to verify a sha256,md5sum Mar 31 19:07:12 Store md5 and sha256 sums in a ini file and verify Mar 31 19:07:12 it with a local file. This code has been tested (manually) Mar 31 19:07:12 on OSX and should be integrated into either OE or BitBake Mar 31 19:12:07 03pH5 07org.oe.dev * rcd09ce62... 10/ (1 packages/linux/linux-handhelds-2.6/hx4700/defconfig): Mar 31 19:12:07 linux-handhelds-2.6: hx4700 defconfig update Mar 31 19:12:07 - compile in common battery drivers instead of hx4700_power Mar 31 19:12:07 - use common touchscreen drivers instead of hx4700_ts Mar 31 20:20:32 koen|away we need to make BitBake run on LLVM Mar 31 20:33:37 03freyther 07org.oe.dev * r9998a3bc... 10/ (1 conf/checksums.ini classes/base.bbclass): Mar 31 20:33:37 classes/base.bbclass: Add checksum.py code to verify md5/sha256 sums on do_fetch Mar 31 20:33:37 This code uses the checksum.py from contrib/qa/checksum to verify Mar 31 20:33:37 md5sum and sha256 on do_fetch task. This code would even allow to checksum Mar 31 20:33:38 certain local patches. Mar 31 20:33:42 03freyther 07org.oe.dev * rf8de7adf... 10/ (1 classes/base.bbclass conf/checksums.ini): Mar 31 20:33:44 classes/base.bbclass: Make the checksum code work and test it with file-native Mar 31 20:33:46 conf/checksums.ini: Provide an example on how this can be used Mar 31 20:39:57 hrw|gone: nunome still fails for various settings (always started with clean tmp). I am going nuts. Why won't this compile for me? Mar 31 20:39:57 hrw|gone: http://rafb.net/p/ejhFTX26.html and http://rafb.net/p/ym59Gm74.html. Mar 31 20:42:32 I will test for angstrom-2007.1 without oabi. That seems to be the only difference. Did you follow http://www.linuxtogo.org/gowiki/OpieWithAngstrom ? Mar 31 20:44:54 hi all Mar 31 20:45:12 hi florian Mar 31 21:21:57 Good night Mar 31 21:32:55 anyone have a packages repository in front of them? Mar 31 21:33:55 no, nothing in reach Mar 31 21:34:02 ggilbert: hw Mar 31 21:34:08 ggilbert: how are you? Mar 31 21:34:13 I have a package of Girl Scout cookies? Mar 31 21:34:46 zecke: Heya Mar 31 21:34:49 Doing pretty good Mar 31 21:35:55 nice :) Mar 31 21:38:26 ggilbert: You mean a tmp/deploy/ipk directory? Mar 31 21:39:28 RP: the patch looked sane, I only wonedered about the != None. I see it was copy and pasted from below Mar 31 21:39:35 RP: actually I'm just wondering what the default OE_MIRROR setting is, since I'm still seeing them hitting the treke.net domain :) Mar 31 21:39:46 which I thought I'd changed years ago Mar 31 21:39:53 zecke: I wondered about that too but as you note, it was copied Mar 31 21:40:04 RP: but this way if we append to a non existing flag we will get the initial value from the original var Mar 31 21:40:18 zecke: I think I'll commit and see if we get any complains ;-) Mar 31 21:40:25 also, since I'm here. Anything in particular needed bugzilla wise? Mar 31 21:40:31 RP: and this is something we might not want to have :} Mar 31 21:40:48 zecke: ah, good point Mar 31 21:41:31 ggilbert: it seems to work and I think the bits I want to have tweaeked can be changed by koen Mar 31 21:42:07 ggilbert: sometimes people manage to make bugs BitBake specific and this way bugs can't be publically viewed Mar 31 21:42:08 ggilbert: treke.net or oesources.org? Mar 31 21:42:16 RP: I'm seeing them hitting treke.net Mar 31 21:42:27 I wonder if someone is just distributing an ancient packages directory Mar 31 21:42:44 Not a big deal, just amuses me whenever I hit the logs Mar 31 21:43:02 hehe Mar 31 21:43:04 ggilbert: A quick grep of the OE tree shows no reference to treke.net Mar 31 21:43:14 Must be something ancient Mar 31 21:43:23 someone still building 3.3 or something :) Mar 31 21:43:27 When did this change occur? Mar 31 21:43:44 mwester: Uh if I had to guess I'd say 2004 or 2005 Mar 31 21:44:05 Ah ok, that would predate the tarball for Openslug 3.10 Mar 31 21:44:11 probably Mar 31 21:44:43 September of 2004 sounds right Mar 31 21:47:32 zecke: Reading this code, I'm not sure you would append to the standard variable if the flag didn't exist? Mar 31 21:48:21 RP: I have not read this part (it wasn't in your context). I assume we will set the var flag Mar 31 21:48:40 RP: but start with the initial value of the var for the flag Mar 31 21:49:10 RP: e.g. I would remove the != None and return VarFlag or "" as initial value, but I don't know :) Mar 31 21:50:32 zecke: http://www.rpsys.net/openzaurus/temp/bb-flag-appends.context Mar 31 21:51:03 zecke: See if the context changes your mind or not :} Mar 31 21:52:21 ah right groupd is from our regex Mar 31 21:52:46 it is just the flag name.... Mar 31 21:52:54 F[] = f Mar 31 21:53:03 will assign f to Var F Mar 31 21:53:47 but that is arguably correct? Mar 31 21:54:22 historically correct Mar 31 21:54:31 :} Mar 31 21:58:21 commit and we might see what is going to break Mar 31 21:59:39 zecke: ok :) Mar 31 22:00:31 I just hired at M$ :} Mar 31 22:01:08 03rpurdie * r803 10bitbake/ (ChangeLog lib/bb/parse/parse_py/ConfHandler.py): ConfHandler.py: Allow operations other than assignment on flag variables Mar 31 22:01:14 * zecke tries to not get fooled this year Mar 31 22:01:49 zecke: You're working for M$? Mar 31 22:02:16 RP: oh well, It is 1st of april and I felt like... Mar 31 22:02:44 zecke: Still 11pm here ;-) Mar 31 22:03:19 Nothing would surprise me atm :/ Mar 31 22:05:24 RP: oh? Mar 31 22:56:16 zecke: http://www.rpsys.net/openzaurus/temp/oe-depends-fixes.patch is how I'm thinking of resolving the current dependency issues... Mar 31 22:56:38 zecke: An upgrade path could prove tricky :/ Mar 31 22:58:51 RP: why tricky? Mar 31 22:58:57 RP: as we will require 1.8? Mar 31 22:59:22 zecke: True. I guess we just wait until we require 1.8, then check this in... Mar 31 23:03:40 :) Mar 31 23:03:40 zecke: Why not just have the URI in the checksums ini file? Mar 31 23:04:04 RP: good thinking :) Mar 31 23:04:26 you made me realize that the scoping is not needed at all and the same URI should have the same sum :) Mar 31 23:04:37 right :) Mar 31 23:14:39 RP: please bitch ross and say sound juicer is awfully broken in feisty :) Mar 31 23:14:58 (actually I think this is something gnome specific though) Mar 31 23:15:35 zecke: Hes offline at the moment but next time I see him... :) Mar 31 23:16:18 RP: there is no way to rip to OGG and if you want to edit audio profiles the dialog is blocked (no user input is possible) Mar 31 23:22:51 * florian shoots libgps for needing libstdc++ Mar 31 23:24:02 ~lart xlocale Mar 31 23:24:02 * ibot throws xlocale's poor little doggy off a cliff Mar 31 23:24:18 good night Mar 31 23:25:25 'night zecke Mar 31 23:31:09 03rpurdie * r804 10bitbake/ (ChangeLog lib/bb/runqueue.py lib/bb/taskdata.py): taskdata/runqueue.py: Add code to handle inter-task dependencies Mar 31 23:44:07 'night all Mar 31 23:53:06 * rwhitby tries an angstrom build for ixp4xxbe Mar 31 23:56:34 :-O Mar 31 23:58:17 jacques: koen|away said that there is untested EABI BE support. So I'm testing to see if it at least compiles (it didn't last time I tried). Apr 01 00:00:20 rwhitby, did the toolchain change? because I hacked in be support a few months ago for my build of it (which didn't boot) Apr 01 00:01:01 which toolchain components is it using now? Apr 01 00:01:09 (component versions) Apr 01 00:01:13 dunno, building to find out Apr 01 00:01:38 binutils-cross-2.17.50.0.5 Apr 01 00:02:03 I should fire up the macbook and try a build myself Apr 01 00:02:27 gcc-cross-initial-4.1.1-r12 Apr 01 00:02:37 hmm Apr 01 00:02:50 no fancy 4.2.0 ? :-\ Apr 01 00:03:10 maybe there are eabi be patches now which weren't there before Apr 01 00:03:50 800-arm-bigendian.patch Apr 01 00:03:57 801-arm-bigendian-eabi.patch Apr 01 00:05:04 aha Apr 01 00:13:11 NOTE: DO PACKAGE QA Apr 01 00:13:17 NOTE: Checking Package: linux-libc-headers-dev Apr 01 00:13:30 ERROR: Exception:exceptions.KeyError Message:'armeb' Apr 01 00:14:26 insane.bbclass doesn't support linux-gnueabi/armeb :-) Apr 01 00:15:42 well, he *did* say "untested" ;-) Apr 01 00:18:12 03florian 07org.oe.dev * r9b9f036b... 10/ (1 packages/gpe-aerial/gpe-aerial_0.3.0.bb): gpe-aerial: add 0.3.0 Apr 01 00:18:16 03florian 07org.oe.dev * rdf91b721... 10/ (1 packages/prismstumbler/prismstumbler_0.7.4pre1.bb): prismstumbler: add 0.7.4pre1 Apr 01 00:18:21 that problem should be not-too-hard to fix though Apr 01 00:18:43 either by fixing insane.bbclass, or just not inheriting it :-P Apr 01 00:18:49 there are still new releases of prismstumbler? Apr 01 00:19:14 I wish I hadn't inherited insanity :-( Apr 01 00:19:19 haha Apr 01 00:27:09 linux-libc-headers-2.6.15.99 Apr 01 00:27:16 glibc-intermediate-2.5 Apr 01 00:27:33 jacques: that enough info about the toolchain? Apr 01 00:29:18 rwhitby, that be-eabi patch is encouraging Apr 01 00:29:33 thanks for the info, I'm going to try a build Apr 01 00:29:47 let me know how you solve the insanity issue :-) Apr 01 00:30:53 * rwhitby waits for CIA-11 Apr 01 00:31:43 03rwhitby 07org.oe.dev * rcd22877f... 10/ (1 classes/insane.bbclass): insane.bbclass: Support armeb EABI. Apr 01 00:31:54 nice Apr 01 00:32:09 you resisted making an insanity pun Apr 01 00:32:55 it's Apr 1 - a day for jokes, not puns. Apr 01 00:33:13 uh-oh Apr 01 00:33:36 if you're joking about the be-eabi patch I'm gonna be upset Apr 01 00:33:39 oh, crap Apr 01 01:03:46 * rwhitby goes bowling with the kids ... Apr 01 02:00:18 is it possible to run fluxbox with the openzaurus rom Apr 01 02:23:14 where does shasum come from? I have sha1sum sha224sum sha256sum sha384sum sha512sum but no shasum Apr 01 02:24:02 which causes this error: Apr 01 02:24:03 sh: shasum: command not found Apr 01 02:24:03 NOTE: The SHA256 Sums do not match. Wanted: 'c0810fb3ddb6cb73c9ff045965e542af6e3eaa7f2995b3037181766d26d5e6e7' Got: '' Apr 01 02:24:22 ln -s sha256sum shasum Apr 01 02:24:26 just a thought Apr 01 02:25:27 I would prefer a less hackish solution Apr 01 02:26:51 would you prefer to get past this issue now? Apr 01 02:26:56 file a bug, hack it, move on Apr 01 02:29:21 is it a bug? I thought I just needed to install another package Apr 01 02:30:25 if you are calculating a sha256 sum, and you have sha256sum installed, it looks like a bug to me. Apr 01 02:30:44 ok I tried the symlink Apr 01 02:30:47 shasum: invalid option -- a Apr 01 02:30:47 Try `shasum --help' for more information. Apr 01 02:30:52 and then same as before Apr 01 02:30:58 oh well, it was a thought Apr 01 02:35:55 seems like there exists somewhere a program called shasum which is called something like 'shasum -a 256 foo' Apr 01 02:47:24 OK I found the answer Apr 01 02:47:38 ironically, shasum is part of a perl package Apr 01 02:47:49 nice Apr 01 02:48:12 now why one would want to start a perl interpreter for every checksum when one could just call sha256sum directly is beyond me Apr 01 02:48:21 but the package is Digest-SHA **** ENDING LOGGING AT Sun Apr 01 02:59:57 2007