**** BEGIN LOGGING AT Wed Dec 07 02:59:56 2005 Dec 07 03:03:37 RP: what are all the issues? Dec 07 03:05:42 ade|desk: good question. I'm not sure I can remember anymore... Dec 07 03:06:13 Something like whether to use include or override syntax. If using overrides, dots are a bad idea so the variable becomes much less readable. We then have the naming conflict Dec 07 03:06:30 if we make the var name more generic it would allow for future systems other than Zs and could be used for 2.7 etc kernels in the distant future Dec 07 03:08:39 add comment to # 513 Dec 07 03:08:47 CONF_KERNEL_VERSION Dec 07 03:08:48 s/add/added Dec 07 03:10:11 koen: How are the ipaqs handling this at the moment? Dec 07 03:10:21 I thought they used linux24 and linux26? Dec 07 03:14:25 RP: kernel 2.6.14+2.6.15-rc2-r4 Dec 07 03:15:08 RP: they add KERNEL to overrides and use kernel24 and kernel26 as the value for the override Dec 07 03:15:20 hrw|work: what I get is that I need to set wifi settings to any ssid and no wep, insert the card (sandisk connect) and then change the settings to my ssid and wep key Dec 07 03:15:26 we agreed on using overrides, but not on the name or the value Dec 07 03:15:45 koen: Is there an example of a machine still doing that in OE? Dec 07 03:16:24 hrw|work: like that I can access my home wifi (smc 2834wbr v.2) every time Dec 07 03:16:26 h3600.conf, h3900.conf and simpad.conf Dec 07 03:16:27 Bernardo: my card works fine except speed Dec 07 03:16:31 Bernardo: Ok, I'm only aware of one SD issue known in mainline and the moment and this doesn't sound like it. Someone's going to have to debug this I guess... Dec 07 03:17:00 ok, this is strange, because the card works well at first Dec 07 03:17:04 koen: They use kernel_kernel24 now which is why my grep didn't find it Dec 07 03:17:31 then (I'll have to see if it is after suspend) it suddenly locks every process trying to access it Dec 07 03:18:26 I do get i2c timeouts when resuming Dec 07 03:20:23 i2c shouldn't affect SD... Dec 07 03:20:25 and just now got the worst lock up ever on my akita - I never had to remove the battery to unlock it before Dec 07 03:21:01 It could be -rc2 - I know that release causes problems on desktops... Dec 07 03:21:10 Bernardo: i lose the bl when i resume my akita Dec 07 03:21:15 it's interesting seeing the mail led scintilating with every line on the console, though Dec 07 03:21:25 ade|desk: the battery led? Dec 07 03:21:33 back light Dec 07 03:22:12 ade|desk: here it takes a few secs (the i2c timeout?) but turns on eventually Dec 07 03:22:51 I had cacko on it for some time, and I had to use the intensity function keys from time to time, as it would turn the bl out sometimes Dec 07 03:23:14 Bernardo: just turned the device on from eariler, no back light still :( Dec 07 03:23:25 kernel 2.6? Dec 07 03:23:32 The backlight is controlled by i2c Dec 07 03:23:33 2.6.15-rc2 Dec 07 03:23:54 ade|desk: and if you plug it in, does it turn on? Dec 07 03:24:29 When you do get back into it, have a look at dmesg Dec 07 03:24:44 See if it says anything about fatal battery errors Dec 07 03:26:09 most of the dmesg log gets overwritten with "empty flash at ..." errors Dec 07 03:27:02 koen: Having looked over the Zaurus 2.6 and 2.4 config files, I think I'm convinced that includes are going to be a lot neater than overrides. I think we should move all the none machine files out the machine directory and into seperate one so the machine targets are clear and change the name of the variable Dec 07 03:27:09 Bernardo: That will only be at bootup Dec 07 03:27:11 sd gets detected properly - mmcblk0: mmc0:0001 SD 988160KiB Dec 07 03:27:56 Bernardo: I was just wondering if there were messages from over the suspend/resume Dec 07 03:28:03 ah, for ade Dec 07 03:28:31 I'll see now if the lockups I get using sd are after suspend Dec 07 03:28:33 RP: heh, we already agreed that includes should go Dec 07 03:28:51 koen: They look a lot neater than I remember Dec 07 03:29:41 too much duplication and useless clutter Dec 07 03:29:54 that was the reaction the last time Dec 07 03:29:59 koen: What do they duplicate exactly? Dec 07 03:31:04 morning @all Dec 07 03:31:08 Its not useless clutter as they have a use ;-) Dec 07 03:31:11 morning greentux Dec 07 03:31:30 hrw|work: recieved anything? Dec 07 03:31:54 greentux: postmail? not yet Dec 07 03:31:57 RP: handhelds-common for example Dec 07 03:32:22 hrw|work: ok, today or tom Dec 07 03:32:44 greentux: ok - will check mailbox Dec 07 03:32:58 snailbox Dec 07 03:33:12 koen: Most include handheld-common? Dec 07 03:33:14 greentux: that one I mean Dec 07 03:33:15 t Dec 07 03:33:41 RP: see the 2.6 clamshell file Dec 07 03:35:13 ade|desk: ok, I'm getting the same Dec 07 03:35:24 ade|desk: did you suspend using the power button? Dec 07 03:36:09 ade|desk: sorry, waking up now, but into a console Dec 07 03:36:14 koen: It includes handhelds-common so either this is unneeded or there is some problem with the code in hh-common for 2.6. That isn't a reason to prefer include or override syntax though Dec 07 03:36:16 ade|desk: with i2c output Dec 07 03:36:32 daughter woke up, have to run Dec 07 03:39:15 koen: Ther difference is hh-common provides an RRECOMMENDS whereas zaurus-clamshell provides DEPENDS Dec 07 03:44:19 basically we need one common for all PDA and basic machine confs which define features and kernel stuff Dec 07 03:44:53 USE_BT = 1/0, USE_WIFI, USE_PCMCIA, USE_ALSA etc or USE = "network/bt network/wifi sound/alsa" Dec 07 03:45:37 hrw|work: viewmtn 0.5 has a manifest browser: http://ewi546.ewi.utwente.nl/tmp/viewmtn/manifest.psp?id=79a998fb2cba55ee5a3451a77bcdc934c1ed2d8a&path=conf Dec 07 03:46:12 nice Dec 07 03:46:59 Anyone have a suggestion for the pathname for all the machine include files (to split them from the actual machine conf files)? Dec 07 03:47:06 koen: if your viewmtn will work then I will get one more reason to sigh about bk times Dec 07 03:47:30 conf/machine/includes/ Dec 07 03:47:37 I upgraded to monotone 0.24, so it should be more stable Dec 07 03:47:52 it will only block when someone is accessing the db Dec 07 03:49:11 koen: vi css, add "color:black" into "body" part Dec 07 03:49:58 done Dec 07 03:50:23 gracias Dec 07 03:51:14 The database spends a lot of time locked :-/ Dec 07 03:52:17 hmm.. greentux sent card so I should build new distro for using Dec 07 03:52:46 RP: you use gcc 3.4 or gcc 4.0 based rootfs? Dec 07 03:52:55 hrw|work: 3.4 Dec 07 03:52:57 do we want viewmtn to be realtime or semi realtime? Dec 07 03:53:27 sorry folks, had to pop out for a bit Dec 07 03:53:31 koen: If semi-realtime running off a separate db would resolve most of the locking issues, it might be an idea Dec 07 03:54:02 hrw|work: Specifically I still seem to use 3.4.3 Dec 07 03:55:25 is there a userland thing i can do to get the back light up again, i can't see what i am doing to give it network access so no ssh to it atm Dec 07 03:55:59 Bernardo|away: pluging in AC doesn't help Dec 07 03:56:02 ade|desk: echo 10 > /sys/class/backlight/corgi-bl/brightness Dec 07 03:56:12 RP: It's now using the snapshot db, which is updated via cron every hour Dec 07 03:56:32 (as long as vanille is up, that is Dec 07 03:56:42 koen: That sounds better Dec 07 03:57:57 RP: nope didn't work :( Dec 07 03:58:28 ade|desk: Can you dmesg > somefile before you reboot so we can see any error messages? Dec 07 03:59:42 lets see Dec 07 04:01:09 rebooting... Dec 07 04:01:24 RP: what id the empty flash crap about ? Dec 07 04:01:29 s/id/is Dec 07 04:02:03 ade|desk: Just ignore it Dec 07 04:02:21 known error ? Dec 07 04:02:44 ignorable error Dec 07 04:02:53 standard jffs2 verbosity Dec 07 04:03:28 ade|desk: Annoying warning Dec 07 04:07:13 bbl, training Dec 07 04:08:23 hi schurig Dec 07 04:08:27 hi Dec 07 04:10:28 hrw|work: do you know if gcc 4.0.2 already stable enought for PXA? Dec 07 04:10:49 schurig: RP told that his image atleast boots Dec 07 04:11:13 hehe, I guess I'll get a ton of errors when I compile Qt/E 3.x with gcc 4.0.2 ... Dec 07 04:11:24 hrw|work: and, it is noticable faster? Dec 07 04:11:28 schurig: how stable it is I will check on days - I plan to boot from SD to rootfs/gcc4 Dec 07 04:12:10 schurig: qt/e 'any version' and gcc newer then 2.95 is like calling for problems Dec 07 04:12:35 hrw|work: I'm using Qt/3 and gcc 3.x for more than a year Dec 07 04:12:52 schurig: I know - how many patches added to get it build? Dec 07 04:13:24 schurig: I did some benchmarking and gcc 4.0 performed better than/comparable to 3.4 Dec 07 04:13:34 RP: something for you: Dec 07 04:13:35 | In file included from arch/arm/mach-pxa/corgi.c:46: Dec 07 04:13:36 | arch/arm/mach-pxa/sharpsl.h:57: error: array type has incomplete element type Dec 07 04:13:36 | arch/arm/mach-pxa/sharpsl.h:58: error: array type has incomplete element type Dec 07 04:13:45 3.3 sometimes came out faster Dec 07 04:14:04 schurig: I was only using a basic bootstrap with an xserver Dec 07 04:14:31 RP: thats linux-openzaurus 2.6.15-rc2-r4 with gcc 4.0.2 Dec 07 04:14:42 hrw|work: This is with 2.6.15-rc2? Dec 07 04:14:55 ah, gcc 4.0.2 :-/ Dec 07 04:14:59 http://pastebin.com/452274 Dec 07 04:16:02 koen: isn't fakeroot working? Dec 07 04:16:23 zecke: no idea, I just get those error messages when trying to build fuse Dec 07 04:16:34 RP: any plan to fix it? or do we end with setting GCC_FOR_KERNEL or how it is called? Dec 07 04:18:19 hrw|work: You need to add an #include to corgi.c at a guess Dec 07 04:19:34 will check Dec 07 04:20:11 hrw|work: I'm using gcc-3.4.3 and no patch for QtE 3.3.4 Dec 07 04:20:41 atleast one version which TT did good Dec 07 04:20:41 RP: sounds good :-) Dec 07 04:21:07 hrw|work: however, there was once a patch somewhere, I have an uncommented line: Dec 07 04:21:10 # For gcc-3.4.2 environments: Dec 07 04:21:10 #SRC_URI += "file://gcc3.4.patch;patch=1" Dec 07 04:22:22 but for me it worked without this Dec 07 04:22:50 schurig: no aliasing problems in the gfx code? (no compile issues though) Dec 07 04:23:08 schurig: did QtE 3 improve in any way since QTE3.0? Dec 07 04:23:38 zecke: our apps all draw perfectly, not sure what you mean with "aliasing problems" Dec 07 04:23:59 greentux: I've now faxed the receipt for you. Did you get it ok ? Dec 07 04:24:33 zecke: Qt/E 3.x improved only the way normal Qt 3.x improved, no special improvements for embedded things, e.g. no /dev/input/eventX handling anywhere in Qt/E 3.x Dec 07 04:24:48 lrg: thanks :) did they wrote a date for sending back? Dec 07 04:25:09 greentux: end of Feb 2006 Dec 07 04:25:48 zecke: AFAIK Qt/E 2.x got some improvements because of Opie/QTopia. I don't really know what those improvements where, but for our apps Qt/E 3.x worked fine Dec 07 04:25:54 morning lrg Dec 07 04:26:01 lrg: ok, that was my info too Dec 07 04:26:07 morning RP Dec 07 04:26:59 schurig: some speed improvements (application start), some robustness changes Dec 07 04:27:19 RP: funny namespacing is in corgi.c... corgi_pm_machinfo use spitz_battery_levels_noac ;) Dec 07 04:27:22 schurig: but they used to ship QtE with a security flawed freetype, jpeg,png, zlib and didn't dare to update Dec 07 04:27:48 zecke: I'm using system JPEG & ZLIB and no freetype Dec 07 04:30:16 schurig: right, good choiche Dec 07 04:31:25 schurig: where you ever asked to implement suspend to flash? Dec 07 04:32:00 zecke: no, not. Would be impossible. We have 128 MB SDRAM, but only 32 MB Flash. Dec 07 04:32:46 zecke: but we have Suspend to RAM, where (with a full battery) we last more than three months. And even when you remove the battery (e.g. to switch it), we can keep the RAM for some days. Dec 07 04:33:01 s/switch/replace/ Dec 07 04:33:11 hrw|work: I know. Calling it corgi_battery would be equally confusing and various other names cause other issues so I gave up in the end :) Dec 07 04:33:49 zecke: also, even just writing 32 MB to the flash takes it's time. Could be the case that a reboot would be faster ... Dec 07 04:34:19 schurig: which cpu do you use nowadays? Dec 07 04:35:40 zecke: PXA255 and on some fork lift terminals i386 lookalikes Dec 07 04:35:49 RP: #include into corgi.c, corgi_ssp.c Dec 07 04:37:06 zecke: going from one cpu to the next (e.g. PXA27x) is quite costly. New schematics, new board, new tests, new drilling and placing instructions for the manufacturing company etc ... and the market doesn't need the newest CPU, they don't really care about the exact CPU type in it Dec 07 04:37:47 schurig: 'if it works, dont fix it' way? :) Dec 07 04:37:52 hrw|work: Thinking about it further, it probably should go into arch/arm/mach-pxa/sharpsl.h Dec 07 04:37:54 hrw|work: exactly Dec 07 04:38:27 RP: you are kernel master here Dec 07 04:39:01 schurig: true Dec 07 04:47:43 RP: kernel built Dec 07 04:47:56 while compiling gcc-cross-4.0.2-r1: checking for memcpy... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES Dec 07 04:48:30 RP: dont forget about that small fix:) Dec 07 04:51:03 hmm, that happens while it configures zlib. Didn't even know that there's a zlib in gcc ... Dec 07 04:51:54 koen: where can I turn this GNOME theme of viewmtn off? ;) Dec 07 04:52:22 hrw|work: I won't - need to make that before sending that patch to mainline :) Dec 07 04:52:59 koen: but viewmtn looks good, finally I can browse easily in the repo Dec 07 04:53:24 zecke: and look does 'pull/merge/update' is worth time Dec 07 04:53:54 zecke: I tried to change it to the tango theme, but the icon names don't map Dec 07 04:53:59 hmm, this is an output from some ./configure run, but it's in log.do_compile ... weird Dec 07 04:54:48 koen: 'Show current Head' fails here. Dec 07 04:55:07 someone know a tool (gui preferred: kde, gtk2) which allow me to select which mails to remove from pop3 mailbox? Dec 07 04:55:09 koen: I was scared a new spartial safari window would open once I entered the folder Dec 07 04:55:26 hrw|work: kmail? Dec 07 04:55:50 hrw|work: (leave on server, remove if locally removed) Dec 07 04:56:04 zecke: o... where it has it? Dec 07 04:56:20 zecke: there are some issues with the caching mod_python tries to do Dec 07 04:56:23 hrw|work: somewhere on the pop3 dialog ;) Dec 07 04:57:32 zecke: not in 3.5.0 then.. 'leave fetched on server', 'leave on server for X days', 'keep only the last X messages', 'keep only the last X megabytes', 'filter if greater then' Dec 07 04:58:31 hrw|work: wait a second, let me start XDarwin Dec 07 04:58:46 ok Dec 07 04:59:59 starting kmail from stable from my uni account Dec 07 05:01:38 hrw|work: weird Dec 07 05:01:49 hrw|work: I could have sworn I have seen that option Dec 07 05:02:04 * schurig things that Fastjar and zlib are only needed when I compile the gcc4 java stuff. I'll disable this and look what happens ... Dec 07 05:02:09 s/things/thinks/ Dec 07 05:02:46 zecke: happens Dec 07 05:03:08 zecke: btw - kitchensync from 3.5.0 is able to sync with opie? Dec 07 05:03:33 hrw|work: can you reach the same e-mail server via IMAP? Then use this. Dec 07 05:03:38 if that's using opensync you'll need to get the plugin from bluelightning Dec 07 05:03:57 schurig: if I would have imap there.... Dec 07 05:03:58 hrw|work: maybe they get removed anyway if you checked 'leave on server' Dec 07 05:03:58 hrw|work: unlikely. I need to catch bluelightning Dec 07 05:04:33 zecke: no - they stay (and I have 3M of mails there - 2.5M is spam probably) Dec 07 05:05:17 hrw|work: let me try evolution Dec 07 05:05:29 lol it crashed Dec 07 05:05:58 zecke: well it is a copy of Outlook Dec 07 05:07:16 ~lart konq/emb for being so old crap Dec 07 05:07:17 * ibot gives konq/emb a good seeing to for being so old crap Dec 07 05:08:56 hrw|work: "telnet 110", USER ...; PASS ...; LIST ...; DELE ... :-) Dec 07 05:10:38 schurig: http://www.inf.fu-berlin.de/inst/ag-nbi/lehre/0506/V_NP/folien/07-InternetDienste.pdf Dec 07 05:10:44 schurig: slide 16 and 17 Dec 07 05:11:12 schurig: actually look at 20 as well Dec 07 05:11:18 schurig: and you forgot about 'TOP number 30' to check subjects Dec 07 05:12:23 * france is back (gone 01:24:04) Dec 07 05:21:14 koen: could you add machine feed for tmp images? Dec 07 05:22:01 no, I deleted the feeds used for building Dec 07 05:22:06 ok Dec 07 05:37:13 RP: which glibc do you use with gcc 4.0.2? Dec 07 05:39:50 schurig: glibc-2.3.5+cvs20050627-r0 here Dec 07 05:40:30 hrw|work: this won't compile on for TARGET_ARCH = "i386". The lowlevellock.h and lowlevellock.c don't agree ... Dec 07 05:41:29 schurig: x86 targets are not well tested in OE.. Dec 07 05:50:17 koen: that are oz or familiar builds btw? Dec 07 05:50:37 which builds? Dec 07 05:50:53 tmp/OZ/ ones Dec 07 05:50:57 OZ Dec 07 05:51:02 ok Dec 07 05:51:26 on #openzaurus I have borzoi user which will test probably Dec 07 05:51:49 * koen has TMPDIR=/home/koen/OE/build/tmp/${DISTRO} in local.conf Dec 07 05:52:22 ok Dec 07 05:53:41 hi Dec 07 05:53:53 hey Mardy Dec 07 05:56:11 hi Mardy Dec 07 06:04:05 schurig: I've only used gcc 4.0.2 from OE on arm... Dec 07 06:05:12 RP: it seems that even in glibc's cvs the sysdeps/.../i386/lowlevellock.h doesn't match the generic/lowlevellock.c file. Maybe ntpl isn't for 386ers ... now i try with TARGET_ARCH="i586" ... Dec 07 06:05:39 one is defining __lll_lock_write() with return type int, the other with return type void Dec 07 06:06:28 schurig: I'm afraid I don't know anything about that problem... Dec 07 06:06:37 RP: no problem Dec 07 06:10:50 03pH5 07org.oe.dev * r635857df... 10/: module-init-tools: prefer 3.2.1 over 3.2-pre4 Dec 07 06:15:18 greentux: do you know if my borzoi was repartitioned, as I'm running out of space on the device when I reflash it. Dec 07 06:15:54 gints: pnong Dec 07 06:19:31 why is "sharp-aticore-oss" included in the package "sharp binary only"? Dec 07 06:19:44 it doesn't seem to be binary only Dec 07 06:21:16 Mardy: it was created by disassemble of sharp binary Dec 07 06:23:30 hrw|work: the guy who did it did a nice work Dec 07 06:24:08 I guess it's not easy to deduce the include file from the code Dec 07 06:24:50 lrg: you have to repartition... Dec 07 06:25:04 greentux: how do I do that ? Dec 07 06:25:16 greentux: its all in Japanese Dec 07 06:25:25 lrg: see oz page :) Dec 07 06:25:29 lrg: mom Dec 07 06:25:36 thanks. Dec 07 06:25:51 lrg: no it should be sharp english rom Dec 07 06:26:36 greentux: the bootloader is all in Japanese, there are a few roman symbols though Dec 07 06:27:11 lrg: ok. then boot into the sharp rom. open console and make a fdisk Dec 07 06:27:42 lrg: you need only the bootloader for flashing, but you need repartition before. Dec 07 06:28:07 greentux: ok i assume there is a sharp image on the cdrom that I can reflash Dec 07 06:28:46 lrg: i asked RP for repartioning in the updater script... but he declined Dec 07 06:29:13 why? the borzoi should boot the original sharp image... isnt it? Dec 07 06:29:48 lrg: why? the borzoi should boot the original sharp image... isnt it? Dec 07 06:30:13 lrg: best way would be to flash a cacko rom... -> console -> repartition -> flash OZ Dec 07 06:30:17 greentux: it did originally, until I reflashed with an oz image Dec 07 06:30:37 greentux: ah, ok Dec 07 06:30:47 lrg: just minute Dec 07 06:32:54 rather than repartiion in oz, I used to just change the argument sent to kernel in command line, does mtdblock stuff still obey those args? Dec 07 06:33:39 XorA: it doesn't even get that far as the updater runs out of space when unpacking the oz image Dec 07 06:34:06 lrg: ok Dec 07 06:34:55 lrg: it should be possible to build a flash version for borzoi... but i think repartitioning is also a way. Dec 07 06:35:22 lrg: i read somewhere that there is a build in emergency shell, perhaps its useful for partitioning Dec 07 06:35:50 ok, I'll search for that Dec 07 06:36:17 lrg: but the cacko way is also easy... ten minutes Dec 07 06:36:39 greentux: in that case I'll go with cacko. thanks Dec 07 06:37:49 lrg: cacko installes in flash, so you have the hdd free Dec 07 06:39:16 greentux: ok, I've found it Dec 07 06:45:59 RP: i got the dmesg output after suspend on battery Dec 07 06:48:39 RP: http://pastebin.ca/32785 Dec 07 07:00:05 morning Dec 07 07:04:10 hi chouimat Dec 07 07:07:05 greentux: now booting oz :-) thanks Dec 07 07:08:19 lrg: fine... Dec 07 07:10:06 ade|desk: Do you have sound modules installed or not? Dec 07 07:10:42 I'd guess not which makes me wonder what the timeout relates to... Dec 07 07:16:26 03koen 07org.oe.oz354fam083 * r4ff9f0d4... 10/: h2200.conf: adapt to changed module names Dec 07 07:16:30 03koen 07org.oe.oz354fam083 * r96ccf6b1... 10/: handhelds-pxa-2.6: add defconfig for h2200 Dec 07 07:20:42 RP: no sound modules installed Dec 07 07:21:19 ade|desk: and in that case, the backlight didn't come back? Dec 07 07:22:52 RP: indeed Dec 07 07:23:18 does echo 10 > /sys/class/backlight/corgi-bl/brightness help? Dec 07 07:23:32 or does that give a further tiimeout? Dec 07 07:30:55 RP: couldn't capture the dmesg that time, trying again Dec 07 07:31:22 i think i have a sticky shift key Dec 07 07:36:17 RP: echo 10 > /sys/class/backlight/corgi-bl/brightness Dec 07 07:36:21 RP: works Dec 07 07:37:13 ade|desk: Were you under gpe/opie at the time? Does it work any better if you're at a console with opie/gpe killed? Dec 07 07:37:28 RP: gpe Dec 07 07:42:48 lrg: ok, cu tom Dec 07 07:43:14 greentux_away: cu Dec 07 07:43:25 grr no path to /sbin make life so much harder Dec 07 07:44:24 RP: no difference with or without X Dec 07 07:45:16 ade|desk: ok, thanks Dec 07 07:45:39 RP: also I tested with and without opie, no difference here too Dec 07 07:46:26 * RP supects an i2c suspend/resume issue Dec 07 07:46:40 It was working... Dec 07 07:46:49 RP: do you want the i2c messages on resume? Dec 07 07:47:07 hi Dec 07 07:47:13 Bernardo: I think ade pastebinned them earlier Dec 07 07:47:30 ok, I'll check if mine are equal Dec 07 07:48:28 yep, identical Dec 07 07:48:56 how to regenerate all zImages kernels of openzaurus? Dec 07 07:49:16 bitbake them Dec 07 07:50:12 Dormi|OFF: rebuild kernel Dec 07 07:50:19 Dormi|OFF: which machine btw? Dec 07 07:50:29 just found out what is wrong with my wifi - a complaint by iwconfig that it doesn't know mode "infrastructure" - but does know mode "managed" and connects Dec 07 07:50:50 bitbake task-bootstrap but I need remove any files from tmp/stamps ? Dec 07 07:51:02 openzaurus collie Dec 07 07:51:47 rm tmp/stamps/openzaurus-sa-2.4.18-rmk7-pxa3-embedix-r22.do_* ? Dec 07 07:51:48 Dormi|OFF: then bitbake collie-kernels Dec 07 07:51:50 editing /etc/network/interfaces and replacing "infrastructure" with "managed" and now I get wifi always Dec 07 07:51:59 thanks hrw|work Dec 07 07:55:12 np Dec 07 07:56:22 <[lala]> hi guys Dec 07 07:56:46 <[lala]> http://bugs.openembedded.org/show_bug.cgi?id=498#c4 Dec 07 07:57:05 <[lala]> can anyone explain me the behavior of bitbake in that case? Dec 07 07:59:09 [lala]: ascii table: '.' is after '-' Dec 07 08:00:27 <[lala]> well, the question is if that behavior schould be changed in bitbake ... Dec 07 08:01:08 <[lala]> because it clearly misleads the user ;-) Dec 07 08:02:01 <[lala]> i wouldn't even been aware of the fact that i'm compiling an _old_ version if it hadn't been the case that the old version was not on the mirror anymore Dec 07 08:04:12 [lala]: probably PV parser need to be changed Dec 07 08:04:14 it shouldn't happen with out new versioning policy Dec 07 08:04:34 RP: removing sd card after suspend locks the akita Dec 07 08:05:39 hrw|work, collie-kernels it is a meta-package? Dec 07 08:05:55 Bernardo: It also locks c7x0 under 2.6.15-rc5... Dec 07 08:06:23 Dormi|OFF: yes Dec 07 08:06:40 At least its reproducable I guess... Dec 07 08:06:42 I dont have collie-kernels.bb file in package/meta Dec 07 08:07:12 Dormi|OFF: ops. its linux/collie-kernels-2.4-embedix.bb Dec 07 08:07:36 Dormi|OFF: basically you need to 'bitbake virtual/kernel' - let OE decide which to build ;) Dec 07 08:08:06 hrw|work thanks Dec 07 08:09:08 np Dec 07 08:09:26 RP: that might also be related to what I see accessing the sd after suspend / resume Dec 07 08:09:36 looks like the sd module dies Dec 07 08:10:46 Bernardo|away: Its highly likely Dec 07 08:15:28 * ade|desk forgot to erase /home partition .. wondered why i had qt crap all over the place Dec 07 08:55:39 heh.. I feel that gnumeric is not a tool for me.. more problems with chart then I had in oocalc Dec 07 09:02:33 hrw|work: whats wrong with gnuplot :-D Dec 07 09:02:49 gnuplot rocks Dec 07 09:04:12 but rotating 3d plots on an ipaq is sloooooooooooooow Dec 07 09:04:43 gnuplot is a nice candidate for iwmmx Dec 07 09:06:05 XorA: Idont want to run gnuplot to draw 12 point chart which I want to manipulate Dec 07 09:06:29 XorA: for me gnuplot is a tool when I want to make a chart from hundreds of entries Dec 07 09:41:28 Good evening. Dec 07 09:42:23 hi sirfred Dec 07 09:43:04 sirfred: good evening Dec 07 09:43:14 sirfred: meet Mardy Dec 07 09:43:26 Mardy: meet sirfred (Manuel) Dec 07 09:43:34 :) Dec 07 09:45:06 koen: Can you give a link to that mutli-machine build script again please? Dec 07 09:46:00 sure Dec 07 09:47:04 RP: http://ewi546.ewi.utwente.nl/tmp/scripts/ Dec 07 09:47:31 Mardy: Hello. Dec 07 09:47:40 the matching local.conf is also there Dec 07 09:47:41 hi sirfred :-) Dec 07 09:48:08 I still don't know what this microcode is Dec 07 09:48:15 koen: thanks - I've saved them this time. I'm about to try and set up some semi automated builds Dec 07 09:50:31 the script should guard against the most common OE pitfalls, but doesn't protect against PV downgrading and such Dec 07 10:06:22 hrw|work: Do you see a problem with putting some documentation about the w100 onto the oz wiki? Dec 07 10:07:29 RP: OE wiki? no problem with me Dec 07 10:07:58 hrw|work: Ah, OZ doesn't have a wiki does it :-/ Dec 07 10:11:51 RP: can't you create a new page on the wordpress Dec 07 10:13:12 lardman|work: I can but I'm not going to be the only person working on this - I'm recommending it goes on hh.org who hopefully won't mind :) Dec 07 10:15:43 RP: Ah, I see Dec 07 10:15:57 hh.org wiki is best probably as it already contain many misc PDA related docs Dec 07 10:16:05 agreed Dec 07 10:16:19 I was thinking OZ had a wiki but I'm just confused :) Dec 07 10:23:03 koen: With this script do you need to inherit multi-machine? Dec 07 10:23:36 yes, although the content is pasted into my local.conf Dec 07 10:32:52 When mt updating, I'm getting a bunch of these: "monotone: warning: discarding revision data packet 9beb55f69acc310fc2689a2b52611680e524a2e8 with unmet dependencies" Dec 07 10:32:55 should I be concerned? Dec 07 10:33:22 it's a side effect from a mistake in the nslu2 tree Dec 07 10:33:44 can I safely ignore them? Dec 07 10:33:47 a packet has ancestors in a branch that isn't in your db Dec 07 10:33:53 you can safely ignore them Dec 07 10:33:53 ah Dec 07 10:34:32 thanks koen Dec 07 10:45:54 cu Dec 07 10:59:09 any meta-package or script to organize all the ikg packages whith section dirs? Dec 07 11:09:33 meh Dec 07 11:09:34 ~today Dec 07 11:09:37 Wednesday sucks, because it is only half way through the week. Dec 07 11:11:12 What's up kergoth ? Dec 07 11:21:15 Goog evening all Dec 07 11:24:13 Anyone here ? Dec 07 11:26:00 heh. unemployment is becoming troublesome Dec 07 11:26:42 Quick question. Do we have already a package in oe, where I can find an initrd example ? Dec 07 11:26:59 kergoth: Want to work in germany ? Dec 07 11:29:25 Oh shit. Just did a pull and now have lots of parsing errors ... Dec 07 11:29:26 OTE: Handling BitBake files: \ (0888/2837) [31 %]ERROR: while parsing /home/sc/scdev/oe/org.openembedded.oz354fam083/packages/guile/guile-native_1.6.4.bb Dec 07 11:29:31 Any help ;-( Dec 07 11:45:19 kergoth: hey Dec 07 11:45:29 kergoth: do you like marvin? ;) Dec 07 11:58:23 kergoth: you left astrocorp? Dec 07 12:00:02 yeah, wasnt very happy there. plus i missed work a fair bitdue to depression before the medical benefits started and i got on the medication.. then 2-3 weeks for it to take effect... all good now, but a bit late Dec 07 12:00:03 heh Dec 07 12:01:21 kergoth: at least you have drugs now! Dec 07 12:01:31 kergoth: head up, you are not alone Dec 07 12:01:37 this is true, except that i may have trouble affording the prescription refill now :) Dec 07 12:01:39 * kergoth nods Dec 07 12:02:46 kergoth: how much are these pills? Dec 07 12:03:19 kergoth: did you have a look at my proposed changes to tslib? Dec 07 12:03:34 it was like $25 when i /had/ medical insurance. i expect around $100 or so Dec 07 12:03:43 Mardy: not yet, its in my mailbox, will do so Dec 07 12:03:56 kergoth: well ask for donations Dec 07 12:03:59 kergoth: looking forward to it :-) Dec 07 12:04:14 kergoth: I still have some money from treke on my paypal account Dec 07 12:04:38 kergoth: are you in other thearapy as well? Dec 07 12:04:54 no worries, i've still got one possible way to get out of the shithole i'm in Dec 07 12:05:00 going to try to sell my townhouse to my dad Dec 07 12:05:06 get some cash that way, since its worth more than it was when i bought it Dec 07 12:05:08 hmm Dec 07 12:05:17 can you rent it to him? Dec 07 12:05:45 although I can understand your need to pay off debt Dec 07 12:06:02 well, if i dojnt come up with cash within weeks, i'm losing my car, and will still owe $12k on the car loan Dec 07 12:06:05 :P Dec 07 12:06:40 I was afraid of that Dec 07 12:06:54 new cars suck Dec 07 12:06:56 hehe Dec 07 12:07:03 It can take weeks to arrange real estate transactions Dec 07 12:07:04 yeah Dec 07 12:07:12 thats true Dec 07 12:07:22 they are worth less than the loan the moment you drive them off the lot Dec 07 12:08:07 indeed. and some idiot hit my car like 2 months after buying it, which doesnt help the value any Dec 07 12:08:12 bastards Dec 07 12:08:15 ~lart car companies Dec 07 12:08:16 * ibot takes a rusty axe and swings it violently, taking car companies's head off Dec 07 12:08:32 kergoth: you don't get a new car in this case? Dec 07 12:08:56 kergoth: good luck! Dec 07 12:14:00 anyone have any advice on which pop server package I should use? mail at school hasgone to poop Dec 07 12:14:36 Crofton: pop1 or pop2? why not imap? Dec 07 12:15:10 no real preferece Dec 07 12:15:13 I was looking at adding dovecot to OE, but gave up after the third time configure wanted to run binaries Dec 07 12:15:28 heh Dec 07 12:54:29 hi pb_ ! Dec 07 12:56:21 hi gremlin[it] Dec 07 12:56:41 hi mickeyl, kergoth, all Dec 07 12:58:58 gruezi pb_ Dec 07 13:00:18 hail zecke Dec 07 13:02:51 hi pb_ Dec 07 13:04:28 hi rp Dec 07 13:05:01 * gremlin[it] start planning to change nick to "AG" ;) Dec 07 13:52:17 ~seen RP Dec 07 13:52:21 rp is currently on #gpe (4d 4h 27m 31s) #oe (4d 4h 27m 31s) #openzaurus (4d 4h 27m 31s) #handhelds.org (4d 4h 27m 31s). Has said a total of 203 messages. Is idling for 45m 37s Dec 07 13:52:33 evening lrg|home Dec 07 13:52:41 evening Richard Dec 07 13:52:59 RP: I have a theory on the resume Dec 07 13:53:28 this sounds promising Dec 07 13:53:56 RP: would you be able to check if the dpm_event is called in the 31 driver for state D0 after resume Dec 07 13:54:15 lrg|home: yes, I'll do it now Dec 07 13:54:27 RP: if so, then it looks like we have a cache coherency issue. Dec 07 13:54:42 RP: thanks Dec 07 13:55:16 RP: if not then there is a bug in the stream logic Dec 07 13:56:32 lrg|home: Did you see my comments on the ACTIVE register? Dec 07 13:57:39 RP: yes, the active bt was not being set Dec 07 13:58:16 RP: PM state D0 should do this Dec 07 13:59:40 lrg|home: Its in state D3hot Dec 07 13:59:46 ah Dec 07 14:00:18 RP: this means my stream logic is a litle wonky. I now know where to look. Thanks Dec 07 14:02:21 lrg|home: np. I'm pleased it gives a clue as to what's wrong :) Dec 07 14:03:29 RP: got it. I'll be sending a patch shortly :) Dec 07 14:03:48 That was quick :) Dec 07 14:05:05 RP: a stupid ommision :( I had made a change to the 53, but didn't apply it to the others. doh Dec 07 14:20:32 RP: patch in the mail.... Dec 07 14:47:49 koen: I just get many "monotone: warning: discarding revision cert/data packet ... with unmet dependencies" errors when trying to pull from ewi546 Dec 07 14:48:22 can packets from the org.openembedded.dev branch depend on packets in other branches? Dec 07 14:48:25 ignore those Dec 07 14:48:39 they can depend on other branches Dec 07 14:48:48 see jbowlers email to oe@ today Dec 07 14:48:50 03pH5 07org.oe.dev * r8183e5d9... 10/: gtk+-2.8.6: adapt patches from 2.6 Dec 07 14:49:04 ah right, pushing worked, too. Dec 07 15:01:44 lrg|home: Will test, thanks Dec 07 15:02:19 RP: thanks, cu tomorrow Dec 07 15:05:40 03koen 07org.oe.oz354fam083 * rbeebd164... 10/: handhelds-pxa-2.6 cvs: copy zImage into deploy/images, backported from .dev Dec 07 15:19:39 has anyone seen this error while building sdl-image? -> "Cannot find package sdl-base" Dec 07 15:25:23 In machine/*.conf, what exactly does the OLDEST_KERNEL mean? ie, in x86.conf, it's 2.6.5. What exactly does this effect? Dec 07 15:25:40 the oldest kernel that will run on that device Dec 07 15:26:33 mmkay, that's what I was thinking, but wasn't sure Dec 07 15:26:34 thanks Dec 07 16:03:14 <_law_> hey where can i find arm-toolchain (2.95.3) Dec 07 16:03:58 http://handhelds.org/download/projects/toolchain/archive/cross-2.95.3.tar.bz2 Dec 07 16:04:22 and read http://oe.handhelds.org/cgi-bin/moin.cgi/ZaurusKernels to see where the symlinks should point to Dec 07 16:05:08 <_law_> koen: thanks Dec 07 16:35:25 <_law_> monotone: 32.9 M | 1.3 M | 12,674 | 3,085 | 53 Dec 07 16:35:36 <_law_> :-( i think this will take hours Dec 07 16:35:55 grab the snapshot Dec 07 16:37:12 <_law_> or is this because i convert a dvd in the background ;-) ? Dec 08 00:42:35 hi all Dec 08 00:43:12 I've been talking with the guy who runs osnews about getting dillo to be recognised as a pda browser Dec 08 00:43:54 I have found where the useragent string is getting generetad in the code, however when I try to bitbake dillo2 it comes with complile errors Dec 08 00:44:21 I don't know why because dillo2 is part of the openzaurus dist Dec 08 00:45:01 dw_table.c:1181: error: invalid lvalue in assignment Dec 08 00:45:10 dw_table.c:1201: error: invalid lvalue in assignment Dec 08 00:45:41 line 1180 is EXTR_VALUE (sub_extremes[i]) = sub_extr_width; Dec 08 00:46:21 I notice that this file (dw_table.c) is included in the gtk2+ patch so perhaps it's not getting patched properly Dec 08 00:46:54 sorry the error should be dw_table.c:1180: error: invalid lvalue in assignment Dec 08 00:49:41 hi oyo, what embedded stuff are you working on? Dec 08 00:52:48 darmou_: your compiler can't find out what "EXTR_VALUE (sub_extremes[i])" is, an lvalue is the thing on the left side of an X=Y assigment Dec 08 00:53:20 darmou_: look at the definition of EXTR_VALUE() macro Dec 08 01:05:00 #define EXTR_VALUE(e) (max ? (e).max_width : (e).min_width) Dec 08 02:03:20 morning all Dec 08 02:09:09 morning greentux Dec 08 02:09:39 morning RP, you have the "job"? (i dont know the status of the contract...) Dec 08 02:10:18 greentux: I'd guess that paperwork would get delivered to my boss and he's away for a couple of days now so I honestly don't know Dec 08 02:11:15 RP: ok... the /tmp/oe images are 2.6 forborzoi? Dec 08 02:11:56 greentux: Which server are you referring to? Dec 08 02:12:23 RP: ewi... Dec 08 02:12:34 They'll be 2.4 ones Dec 08 02:13:06 The next version of OZ will be shipping with a 2.4 kernel, then I'll produce some 2.6 images for the new supported devices Dec 08 02:13:25 The problem is we'd need to make soo many changes to the branch to support recent 2.6 kernels Dec 08 02:13:33 s/soo/too/ Dec 08 02:13:35 ok, like a 3.5.4-new or 3.5.5 ? Dec 08 02:13:51 3.5.4.1 or something, yes Dec 08 02:14:55 I was testing my multi device build yesterday. I'll set it off to finish that build run and there might be some testing images... Dec 08 02:16:12 RP: we need not so much gpe or qt things, but the 2.6 kernel. so we can use your RCs... Dec 08 02:17:19 RP: ^^ testing images for what 3.5.4 or 3.5.4.1? Dec 08 02:17:36 3.5.4.1-rc :) Dec 08 02:18:27 RP: fine! is it in the monotone? Dec 08 02:18:41 morning Dec 08 02:18:51 It will be built from what's in monotone .dev Dec 08 02:19:37 RP: ok, but there is a bb file? ( i am not so good in developing, but "lala" does it for me ;) Dec 08 02:19:45 but the final release it likely to be a later kernel than what's there. I have some out of tree changes which I've yet to commit, mostly as they totally break cxx00 sound until I (or Liam) get it fixed Dec 08 02:19:50 morning XorA Dec 08 02:20:27 greentux: Just the usual .bb files with MACHINE=borzoi Dec 08 02:22:45 morning XorA Dec 08 02:23:26 The good news it that the next set of sound drivers should be the best yet as I've worked with Liam on a few issues and resolved them once and forall. c7x0 is working well and it should just be a matter of applying those changes to cxx00 Dec 08 02:24:17 so i gave liam a borzoi for testin g;) Dec 08 02:24:50 RP: did you get feedback on your LED code? Dec 08 02:25:13 koen: Not a lot. gregkh is happy with some minor changes though Dec 08 02:25:20 No word at all from rmk... Dec 08 02:25:41 I guess rmkl is too busy breaking serial to notice Dec 08 02:26:26 koen: He didn't actually break it - ttyS3 just refers to itself as ttyS0 now as our hack isn't working as well as it did... Dec 08 02:27:27 ah, ok Dec 08 02:28:06 ~lart fuse Dec 08 02:28:07 * ibot blames fuse for all the evil in the world Dec 08 02:28:16 ibot: botsnack Dec 08 02:28:17 aw, gee, koen Dec 08 02:28:37 morning koen RP Dec 08 02:28:45 hey gremlin[it] Dec 08 02:28:53 morning gremlin[it] Dec 08 02:31:01 morning all Dec 08 02:31:07 hey ade|desk Dec 08 02:32:28 I just realise that installing a dev node doesn't do much good if you use devfs or udev.... Dec 08 02:32:32 stinking fuse Dec 08 02:39:18 morning ade|desk Dec 08 02:48:11 koen: fuse as in the Filesystem in User ..... Dec 08 02:52:17 XorA: that one Dec 08 02:53:16 koen: cool Dec 08 02:53:44 I'm affraid i need to make a fuse-lib, fuse-utils and fuse-module .bb to get to work nicely Dec 08 02:58:16 good morning **** ENDING LOGGING AT Thu Dec 08 02:59:59 2005