**** BEGIN LOGGING AT Fri Jul 22 23:59:56 2005 Jul 23 00:31:03 could monotone be any slower? Jul 23 00:58:58 I understand there were some problems with the monotone servers yesterday. Today I cannot update monotone but I get an error "fatal: std::runtime_error: failure extracting schema from sqlite_master". Are the servers up and running fine? Jul 23 02:43:03 I can compile nunome by hand but via bitbake it fails. More details at http://rafb.net/paste/results/ulEZbA47.html Jul 23 02:43:47 I fail to see what is going wrong here. The error message seems to suggest there is no rule for make install but as I verified by compiling by hand there really is. Jul 23 02:46:31 morning Jul 23 02:48:32 lamikr: morning. How are you? Jul 23 02:52:18 Fine, actually trying to make gpe-imape for ipaq h6340. I had some problems yesterday so I have deleted whole tmp and have now second attempt. sofar gross-gcc compiled and now I will start gpe-image. Jul 23 02:53:01 nice. Jul 23 02:53:09 Good luck. Jul 23 02:53:43 lamikr: If you have some spare time, would you mind taking a look at why http://rafb.net/paste/results/ulEZbA47.html fails? Jul 23 02:55:11 laibsch: I can try this after I get gpe-image done from familiar-0.83 branch. Jul 23 02:55:24 Cool. Jul 23 02:56:24 btw: If I execute gpe-image, should it also add kernel image to /boot and kernel images to lib/modules directory? Jul 23 02:57:04 I am not much of an expert. Let's hope they wake up soon from their delirium ;-) Jul 23 02:58:02 I have not yet compiled a complete image but I would expect that bitbake produces a flashable image. Everything else you install from ipk packages. Jul 23 02:58:37 That being said, currently I still use the stock Sharp ROM on my Zaurus because unless I got *some* kind of Japanese input, I cannot switch. Jul 23 03:00:32 For some reason that did not happened for me when I build it yesterday. That was propably because I deleted part of the tmp/work directory during the build node because I had hd space problems. Jul 23 03:01:51 ;-) Jul 23 03:02:00 You should not do that, I guess. Jul 23 03:07:23 Hi all. Jul 23 03:09:58 Yes, I made now more space for the oe build. Jul 23 03:10:14 Laibsch: So you are from japan? Jul 23 03:10:30 no Jul 23 03:10:33 guess again Jul 23 03:10:37 ;-) Jul 23 03:10:44 Hi dmx_ Jul 23 03:14:28 laubsch: Laibch sounds almost like Leibzig so I guess Germany :-) Jul 23 03:15:21 Wow, good guess! Jul 23 03:15:59 Did you really have the association Laibsch - Leipzig? Or did you cheat and make a tracert? Jul 23 03:16:38 The sound of Leipzig and Laibsch are quite dissimilar, BTW. Jul 23 03:17:19 Just the locals with their Sächsisch dialect slur it to be close to Laibsch. Jul 23 03:17:39 * Laibsch like lots of things about Leipzig but dislikes Sächsisch Jul 23 03:20:18 laibsch: Yes!, well that was just a guess without cheating with tracert. Sorry btw, so slow for writing. I have 2 year old daugter and she needs a lot of attention. Jul 23 03:20:50 ** pads daughter of lamikr: Be a good girl Jul 23 03:20:58 'mornin koen. Jul 23 03:21:15 lamikr: That is alright. Jul 23 03:21:37 lamikr: Better slow in writing than slow in thinking ;-) Jul 23 03:21:57 laibsch: Well not always sure even from that ... :-) Jul 23 03:22:15 ;-) Jul 23 03:22:39 Laibsch: I just have visised once in the germany and spend some time in Leipzig and Dresden. So in which area of germany you are? Jul 23 03:22:54 zecke_sad: what is the problem? Did your favourite gadget die on you? Jul 23 03:23:38 lamikr: You really are slow in thinking, aren't you? ;-) No, it is not Dresden althought I live there for half a year. Jul 23 03:24:12 s/live/have lived/ Jul 23 03:25:24 Laibsch: So where is Sächsisch? Jul 23 03:25:34 zecke_sad: I hope it is nothing really serious in which case I would have shown my ignorance with the previous comment. I hope everything is alright with you. Jul 23 03:25:55 * koen guesses in Saches Jul 23 03:25:57 good morning all Jul 23 03:26:10 Sachsen* Jul 23 03:26:16 There he is, koen, another good guesser ;-) Jul 23 03:27:10 Morning Koen! Where do you live, let me guess, also in Germany. Jul 23 03:27:18 nope Jul 23 03:27:21 koen: morning Jul 23 03:27:23 lamikr: you only 5km wrong :) Jul 23 03:27:35 morning all Jul 23 03:27:46 koen: Netherland? Jul 23 03:27:48 lamikr: Enschede, The Netherlands Jul 23 03:28:17 see http://ewi546.ewi.utwente.nl/maps/OE.php Jul 23 03:28:24 lamikr: Tip, do you /whois $nick and very often you can find out for yourself. Jul 23 03:28:31 lamikr: Tip, do a /whois $nick and very often you can find out for yourself. Jul 23 03:30:39 koen: nice map, do you know who is living in the south australia? I thought that place would be almost like antarctic. Jul 23 03:31:08 lamikr: you can click on the flags. Jul 23 03:31:54 laibsch: Wow, almost like a game! Well being today really slow... Jul 23 03:32:02 lamikr: I believe the south of australia is pretty warm. It's rwhitby that lives in adelaide Jul 23 03:33:09 koen: I'd really appreciate if you could spare a minute and take a look at http://rafb.net/paste/results/ulEZbA47.html. nunome still does not compile. Almost same error as yesterday. Jul 23 03:34:03 koen: I wonder why mickeyl is not in the map. Jul 23 03:34:11 Does he not want to be listed? Jul 23 03:34:40 Laibsch: he wasn't around when I asked people on IRC Jul 23 03:34:49 I see. Jul 23 03:34:53 Laibsch: I suspect S is set wrong Jul 23 03:35:03 S? Jul 23 03:35:13 could you do an 'ls /tmp/work/nunome-1.0-r0/' ? Jul 23 03:35:22 koen: I did yesterday my first attempt successfull with bitbake gpe-image for h6340 but for some reason the /boot and /lib/modules directories where empy in the result tar.gz. Should they be included there? Jul 23 03:36:06 koen: If yes, the problem was propably caused because I needed to "rm -rf" some directories in tmp/work because I had space problems. Jul 23 03:36:23 koen: AUTHORS COPYING dicman image Makefile Nnmlib nunome-1.0 nunome_uni.dic server temp ui Jul 23 03:36:52 * Laibsch studies wiki for S Jul 23 03:36:59 Laibsch: aha, OE expects the tar.gz to contain a dir called 'nunome-1.0' Jul 23 03:37:35 Laibsch: http://ewi546.ewi.utwente.nl/tmp/d/keyS.html Jul 23 03:37:43 Thanks. Jul 23 03:37:53 S = "${WORKDIR}" should fix it Jul 23 03:38:14 or S = "${WORKDIR}/" Jul 23 03:38:33 So you are working on docu? Very nice. Jul 23 03:39:16 Laibsch: that is the output of ' bitdoc -c ../monotone/openembedded/conf/documentation.conf -o/var/www/tmp/d' Jul 23 03:39:38 masterhacker zecke created the documentation hook Jul 23 03:39:39 I did try a while ago in the wiki and am still motivated to write things up. I guess my strengths are in structuring, my weakness is knowledge of OE which might either be an advantage or a disadvantage, depending on how you see it. Jul 23 03:39:49 I see. Jul 23 03:40:42 Even though it is in tmp, I guess it will be up long enough to warrant a link from the wiki? Jul 23 03:40:52 yeah Jul 23 03:43:21 Done. Jul 23 03:47:51 koen: Also made a link to your dev map from the front page. Hope you do not mind. Jul 23 03:48:07 no, it's googles bandwith :) Jul 23 03:49:35 I am sure they have plenty to spare ;-) Jul 23 03:49:42 DaKa2: I removed the armfix patch from libusb, your original patch works fine Jul 23 03:50:21 :) Jul 23 03:50:31 I guessed so :-) Jul 23 03:51:06 I can transfer pictures from my ixus to my ipaq, but the snapshot function doesn't work Jul 23 03:51:15 but I'm very happy with this result Jul 23 03:51:58 yeah, never got the snapshot to work on my ixus either, but thats not what I needed anyway Jul 23 03:53:01 ah, ok Jul 23 03:53:16 I was blaming my usb card: http://handhelds.org/scap/port.17733.png Jul 23 03:53:31 the drivers for that don't do full usb 1.1 Jul 23 03:55:56 it is a bit strange since gphoto states that it should work Jul 23 03:59:38 DaKa2: it did work two years ago on my workstation Jul 23 03:59:50 hm.. thats not good.. Jul 23 03:59:56 I have to try again.. Jul 23 04:03:47 *** Error (-1: 'Unspecified error') *** :-) Jul 23 04:04:05 that's what I see too on my ipaq Jul 23 04:07:31 "gphoto2-port(2): Writing message (request=0x4 value=0x10 index=0x0 size=88=0x58)... canon/usb.c(2): canon_usb_dialogue: read 1 of 0x40 bytes failed! (Error reading from the port)", I might look at that later, could be sending the wrong data to the camera and it doesn't reply Jul 23 04:13:04 koen: grepping through the packages dir I did not find another example of using S= in a bb file. What I did is 'do_install() {\n oe_runmake TARGET_DIR=${D} USR_DIR=${D}/usr S="${WORKDIR}"\n }' but that does not work. Where do I need to put that S-key? Jul 23 04:14:09 OE should do 'cd ${S}' before running commands Jul 23 04:14:22 but you could but that in to be sure Jul 23 04:14:58 Laibsch: you could look at the vlan package, it uses S= Jul 23 04:16:00 hey mickeyl Jul 23 04:16:06 morning Jul 23 04:16:18 DaKa2: Indeed. The reason I did not find it is that is has "S =" instead of "S=" for which I grepped. Jul 23 04:17:08 koen: hey, add me to your developer map, will you? I'm @ E: 8,69°; N: 50,08° Jul 23 04:17:34 Laibsch: ahh, yes :-) My package so I knew it was there, and Im probably the only user of it :-) Jul 23 04:17:37 mickeyl: hehe. We were just talking about that a minute ago. Jul 23 04:17:47 heh Jul 23 04:18:07 DaKa2: I suspected that it was your package. Thanks for pointing it out. Jul 23 04:18:21 :) Jul 23 04:19:32 could anyone tell me why we didn't think of using keymaps also for the Z models running 2.6 ??? Jul 23 04:19:39 ~hail CoreDump|afk for the idea to do that Jul 23 04:19:40 * ibot bows down to CoreDump|afk for the idea to do that and chants, "I'M NOT WORTHY!!" Jul 23 04:19:43 mickeyl: done, see http://ewi546.ewi.utwente.nl/maps/OE.php Jul 23 04:19:50 s/2.6./2.4/ Jul 23 04:20:41 koen: thanks. hey quite a croud in germany Jul 23 04:20:45 2.4 has keymap support? Jul 23 04:20:49 mickeyl: because you thought it was easier to patch QT/e to death Jul 23 04:21:11 koen: keymap patch has more lines changed than the keytable ;) Jul 23 04:21:43 * koen points people to http://article.gmane.org/gmane.comp.version-control.monotone.devel/4161 Jul 23 04:24:12 thanks Jul 23 04:24:52 zecke: yeah, i've tried it - and it just worked :) Jul 23 04:25:09 zecke: now we have nearly all special buttons working on the Zs Jul 23 04:25:21 ~lart us fscking with qte Jul 23 04:25:21 * ibot puts on a hockey mask and jumps out at us fscking with qte Jul 23 04:25:47 mickeyl: keymaps are fluff and lower compatibility with closed source programs Jul 23 04:26:18 ah righto Jul 23 04:26:25 that was the reason why we didn't do it earlier :D Jul 23 04:27:39 | install: cannot stat `/home/koen/OE/build/tmp/familiar/work/h3900-linux/initscripts-1.0-r51/alignment.sh': No such file or directory Jul 23 04:28:21 did I miss something? I thought alignment.sh was in the arm/ subir Jul 23 04:28:22 alignment issues break compability ARM Ltd. should introduce a better ABI Jul 23 04:28:41 koen: that looks like a problem in the fetcher code... it should pick it up in the arm dir Jul 23 04:29:07 ar_: give us your GPS cords and be added into http://ewi546.ewi.utwente.nl/maps/OE.php Jul 23 04:29:45 koen: we will set up tinderbox tonight Jul 23 04:29:51 zecke_sad: cool Jul 23 04:29:52 mickey|bnb: uff. gps code? i seach ;) Jul 23 04:30:24 now I've to learn the rule of streets ;) Jul 23 04:36:10 hi zecke_sad Jul 23 04:53:39 OE core devs: a new version of the update-alternatives files: http://david.thg.se/saker/openslug/update-alternatives ready for your review .... Jul 23 04:54:50 thx for the critical review yesterday - DaKa2 has made all the changes requested and completed the outstanding files. We haven't rewritten the bbclass (we're not at that level of mastery yet), but pledge to retrofit these files in the future if someone else does upgrade the update-alternatives bbclass. Jul 23 04:55:30 I hope I made all the changes requested :) Jul 23 04:57:01 well, DaKa2 even did console-tools as requested :-) Jul 23 04:57:36 DaKa2: if you are really motivate you could do something like update-alternatives --install ${bindir}/chvt chvt chvt.${PN} 100 Jul 23 04:58:38 Well, that is possible, is it something you want me to do? Jul 23 04:59:28 DaKa2: only if you have time for it, it's a (personal) style preference Jul 23 04:59:56 wow, coreutils really got expanded :) Jul 23 05:00:08 style and consistency are always important :-) Jul 23 05:00:29 koen: yes, well, I tried to find a better solution on coreutils, but that was the cleanest I could come up with Jul 23 05:04:45 DaKa2: you could mail the diff to the oe mailinglist for good measure Jul 23 05:05:01 DaKa2: if there are no issues found in this round, are we good to check this all in soon? Jul 23 05:06:03 Well, I can commit anytime, but the ML might be a good idea Jul 23 05:06:14 hi all Jul 23 05:06:37 I've found a small bug in coreutils, but that is just two lines that shouldnt be there Jul 23 05:06:47 hey CoreDump|home Jul 23 05:07:00 nod on the ML, but we want to push this through so we can complete the OpenSlug native build environment ... Jul 23 05:08:25 rwhitby: you can always put it in the nslu2-branch and don't propagate till you get an ACK Jul 23 05:11:23 * CoreDump|home is flashing a shiny new Akita image and hopes for the best Jul 23 05:11:26 koen: true, but we like to stay close on general OE files ... don't want to invoke the merge hydra ... Jul 23 05:12:41 morning Jul 23 05:12:46 reenoo_: http://david.thg.se/saker/openslug/update-alternatives - updated Jul 23 05:12:48 hi reenoo_ Jul 23 05:13:43 hi rwhitby Jul 23 05:13:46 hey pb_ Jul 23 05:17:00 hi reenoo_ Jul 23 05:18:06 ok guys, the packages in update-alternatives are looking pretty good. Feel free to check those in. Based on that ground work, adding an update-alternatives.bbclass will be a breeze Jul 23 05:22:38 rwhitby: AFAICT, there are still a couple of do_install_append's left. also, I guess I'd replace those mv $foo $foo.$PN with for loops to make things easier to read. other than that, like mickeyl said, good job Jul 23 05:22:57 reenoo_: thank DaKa2 not me :-) Jul 23 05:23:08 ~hail DaKa2 Jul 23 05:23:09 * ibot bows down to DaKa2 and chants, "I'M NOT WORTHY!!" Jul 23 05:23:35 I've just changed into ${PN} Jul 23 05:24:06 Should I replace the do_install_appends where there are no -native packages too? Jul 23 05:24:21 :) Jul 23 05:24:38 might as well for consistency, and as a template for future copies of your work to be robust Jul 23 05:25:37 ok, Ill do that, do a last round of tests, commit, and send a mail the the ML Jul 23 05:25:38 DaKa2: yeah, please replace all do_install_appends. just to avoid future problems Jul 23 05:25:53 DaKa2: Good Plan Jul 23 05:28:54 DaKa2: as for the large mv and update-alternatives blocks I was thinking of something along the lines of: Jul 23 05:29:01 bindir_progs = "[ basename cksum comm" Jul 23 05:29:01 for i in $bindir_progs do Jul 23 05:29:01 mv ${D}${bindir}/$i ${D}${bindir}/$i.$PN Jul 23 05:29:01 done Jul 23 05:29:51 (just an example from coreutils) Jul 23 05:30:10 yes, that would be the preferred solution Jul 23 05:30:38 that way the eventual transition to a bbclass would be even more straight Jul 23 05:33:16 Ok, that should work.. Jul 23 05:33:36 good solution for coreutils atleast Jul 23 05:35:35 hey CoreDump|home (sorry, didn't see your "hi" in the first place) Jul 23 05:36:16 :) np Jul 23 05:38:12 DaKa2: well.. that approach should work wherever you run virtually the same command multiple times Jul 23 05:38:27 later all Jul 23 06:21:51 03koen 07org.oe.dev * r7d0136e5... 10/packages/gpsdrive/files/gpsdrive-pda-fix.patch: Fix gpsdrive patch again Jul 23 06:25:12 morning Jul 23 06:31:21 morning XorA Jul 23 06:31:24 nunome still does not build for reasons I fail to fathom. I have set TARGET_DIR=${D} USR_DIR=${D}/usr for oe_runmake. "bitbake nunome" fails alternately with "no rule to make target install" or trying to install into /plugins thus ignoring $TARGET_DIR. Jul 23 06:32:09 The current setup is at http://rafb.net/paste/results/ml8t9P42.html, the one from this morning at http://rafb.net/paste/results/ulEZbA47.html Jul 23 06:32:27 Why does bitbake ignore my request to pass $TARGET_DIR? Jul 23 06:32:53 mickey|breakfast: http://zecke.blogspot.com/ comments? Jul 23 06:33:00 ignore spelling errors and such... Jul 23 06:34:00 * mickey|breakfast reads Jul 23 06:34:13 zecke_sad: zigbee will be big, networking wise too Jul 23 06:34:32 CP|Packing: yeah forgot zigbee Jul 23 06:34:41 and don't drop the pda in the jacuzzi Jul 23 06:34:42 CP|Packing: If we've zigbee I will get a criminal... Jul 23 06:34:42 Laibsch: probably a recursive make problem Jul 23 06:34:50 CP|Packing: a it is waterproof ;) Jul 23 06:35:03 CP|Packing: just scan how much toys a house hold has Jul 23 06:35:06 Laibsch: some hardcoded makefiles don't use MAKE, but use make directly, hence our handcrafted environment variables get lost Jul 23 06:35:12 CP|Packing: and if two plasma monitors answer you can rob them ;) Jul 23 06:35:23 CP|Packing: and hopefully PDAs will be lead less Jul 23 06:35:41 mickey|breakfast: Guten Appetit ;-) Jul 23 06:35:46 zu spät .) Jul 23 06:35:50 mickey|breakfast: I'll inspect the makefile Jul 23 06:37:39 zecke: well writte! Although i'd change "provide smaller libs" into something like "Provide libraries with a fine granular feature based configuration system, so that they can be smaller on mobile clients" Jul 23 06:37:42 zecke_sad: nice Jul 23 06:37:57 koen: sadly I won't go to akademy... Jul 23 06:38:08 koen: maybe I will apply for next GUADEC Jul 23 06:38:23 mickeyl: I still think the "one huge lib, but with a *build time* config system" is stupid for embedded devices Jul 23 06:38:53 mickeyl: BTW, when I call "TARGET_DIR=./image/ USR_DIR=./image/usr/ make install" inside /tmp/work/nunome-1.0-r0/ the build goes through as expected. Jul 23 06:38:54 mickeyl: provide small fine tunable, pluggable libraries ;) Jul 23 06:38:55 koen: it may be stupid, but it may be the only thing we can get by the KDE team Jul 23 06:39:21 mickeyl: KDE4 will have smaller libraries Jul 23 06:39:24 mickeyl: I wonder what happens when libkde.so is 400MB big in a few years Jul 23 06:39:31 koen: Of course I'd like them to write small libraries, however I'm afraid that large libraries with a configuration based system will be the best we can get from the team Jul 23 06:39:34 CosmicPenguin: I missed DVB-T as well Jul 23 06:39:34 zecke: oh that'd be nice Jul 23 06:39:58 Laibsch: yeah, that supports my recursive make theory Jul 23 06:40:02 Laibsch: workaround: Jul 23 06:40:04 mickeyl: yeah, I was affraid the KDE people are attached to the ONE HUGE lib dogma Jul 23 06:40:16 Laibsch: do_compile() { for i in dir; oe_runmake ...; done ) Jul 23 06:40:21 (something like that) Jul 23 06:40:26 koen: KDE devels are humans as well Jul 23 06:40:28 mickeyl: Thanks, I'll try that. Jul 23 06:41:07 zecke_sad: not wanting to admit the errors of their ways? Jul 23 06:41:15 *cough* spatial *cough* Jul 23 06:41:57 koen: Errors? What is that? my version of wikipedia does not have that word Jul 23 06:41:58 best would be small libraries + a configuration system for the larger ones Jul 23 06:42:16 koen: is Ewi now 32bit by default? Jul 23 06:42:36 mickeyl: yes for example exchange XMLGUI with something better for PDAs Jul 23 06:43:58 zecke_sad: 32bit kernel, 32 bit userspace, since iptables didn't work with a 64bit kernel Jul 23 06:44:15 koen: netfilter is overrated Jul 23 06:48:30 koen: seems to work here Jul 23 06:48:46 mickeyl: http://marccramdal.blogspot.com/ Jul 23 06:49:58 nice launcher idea Jul 23 06:50:52 btw., did you ever see tomtomgo or the tomtom software for pocketpc? Jul 23 06:51:03 it has a nice and clean interface. nothing fancy, but up to the task Jul 23 06:51:06 i like it very much Jul 23 06:51:36 mickeyl: yes on CeBit Jul 23 06:52:43 all that makes me think that the #1 thing for better usability is: less options, less features, but make the ones which are present to just work (tm) Jul 23 06:52:57 yes Jul 23 06:54:58 mickeyl: what no emacs? Jul 23 06:55:40 mickeyl: why are you speaking gnome mantra's? Jul 23 06:56:22 koen: because recently I've tried to use some of our apps :D Jul 23 06:56:43 * CosmicPenguin laughs Jul 23 06:57:12 and i'm clearly separating desktop and pda here. I love rich desktop apps, don't get me wrong Jul 23 06:57:22 me too Jul 23 06:57:22 03daka 07org.oe.nslu2-linux * r613d3067... 10/packages/ (17 files in 14 dirs): Updated to use update-alternatives and fix some FHS bugs Jul 23 06:57:32 i just question whether it's wise to give the people the full power of desktop apps to their mobile devices Jul 23 06:57:39 I got pissed off when they butchered all the functions I used in galeon Jul 23 06:57:57 "reload all tabs" "reload tab every 10 minutes", stuff like that Jul 23 06:58:03 ya, that was handy Jul 23 06:58:27 mickeyl: I've always wondered that Jul 23 06:58:55 This is why palm was successful, and the others were... not as successful Jul 23 06:59:32 *nod* I know that a lot of people are really fancy about the pdaXrom and debian workstation stuff... but... honestly. How many of them are typical users? Jul 23 06:59:56 mickeyl: the biggest complaint on zug seems to be, "where are the PDA apps" Jul 23 07:00:00 simplicity!!! Jul 23 07:00:11 and usability!!! Jul 23 07:00:19 let us invent a new Itsy with more sensors Jul 23 07:00:23 to create a new feel Jul 23 07:00:49 its our advantage being linux based, we can cater to both crowds, though the more advanced crowd should start learning how to cater for themselves Jul 23 07:00:56 XorA: yep. our devices do a lot of fancy things that isn't possible with wince or palm, however we really suck at the "standard" apps Jul 23 07:01:09 XorA: that should be the featured quote on the OE home page Jul 23 07:01:57 mickeyl: its because no-one here is a typical user Jul 23 07:02:15 mickeyl: I know I support the apps I use, which happen to be mplayer and firefox mostly Jul 23 07:02:40 XorA: right. I guess we need some fundings to buy every Opie and GPE developer a Tungsten or something like that Jul 23 07:02:46 just for use Jul 23 07:02:49 NOT for tweaking! Jul 23 07:02:49 a Treo650 Jul 23 07:02:50 :D Jul 23 07:02:56 Treo? Jul 23 07:02:57 mickeyl: reflashing disabled :-D Jul 23 07:02:57 * mickeyl googles Jul 23 07:03:01 XorA: hehe, righto Jul 23 07:03:14 oh sweet Jul 23 07:03:32 Ive actually only ever used an Ipaq 5550 and I hated it, it was 10x more complex than opie Jul 23 07:03:42 Treo is my PDA/phone of choice, and Treo has been for the past four years through three generations Jul 23 07:04:06 (Treo270, Treo600, Treo650) Jul 23 07:04:07 If I ever used a PDA, it would be a palm - but I don't even find those useful Jul 23 07:04:44 http://www.linuxdevices.com/news/NS8201733582.html <- MV 75m$... Jul 23 07:05:07 hmm Jul 23 07:05:22 not bad Jul 23 07:05:35 I hate MV... Jul 23 07:05:40 i love VC :) Jul 23 07:05:47 * koen thinks OE could raise some VC as well Jul 23 07:05:48 Just today I was at the park, and ssh'd into our nslu2-linux build machine over GPRS to kick off a build. Then listened to an eBook, which was automatically put on pause when an incoming phone call came in, and automatically started again when I hung up (after switching mid call to my calendar to enter a new entry). Jul 23 07:05:53 * koen starts seeing M$ Jul 23 07:05:53 VC for OE would rock Jul 23 07:06:01 (M as in mega) Jul 23 07:06:15 rwhitby: wow, that's a pretty cool scenario. Jul 23 07:07:03 I could have been here as rwhitby-treo too, but it demands to be the foreground app, so the phone call would have broken that connection. Jul 23 07:07:25 (the IRC client, that is) Jul 23 07:09:35 XorA: nslu2-linux raised USD$2K in the last two months :-) Jul 23 07:09:53 rwhitby: cool Jul 23 07:10:39 we spent most of it on a new server for the wiki/ipkgs/monotone. we're hoping to have enough left to buy the core devs a 770 each (joking). Jul 23 07:10:56 hehe Jul 23 07:10:59 heh Jul 23 07:11:04 that 770 is slick Jul 23 07:11:05 can't wait for the 770. Jul 23 07:11:22 did I see something about a developer discount in koen's blog? Jul 23 07:11:32 rwhitby: yes Jul 23 07:11:40 how do you qualify for that? Jul 23 07:11:51 rwhitby: mail nokia... Jul 23 07:12:14 Nokia brought a entire regiment to OLS Jul 23 07:12:23 rwhitby: http://maemo.org/maemowiki/FrequentlyAskedQuestions Jul 23 07:12:25 I was trying to get my work to buy some for tradeshow demos, but they arent foresited enough Jul 23 07:12:25 yes, but what criteria will they use to sift my gem of an email out from the dross that everyone else sends them ... Jul 23 07:12:59 no idea :) Jul 23 07:13:29 hmm Jul 23 07:13:36 is base-passwd just failing on my end? Jul 23 07:13:54 /home/repvik/unstable/openslug/tmp/staging/i686-linux/bin/install-sh: ./-p does not exist. Jul 23 07:14:41 Hmm - that maemo.org URL is not working for me either. Seems that there is some weird problem with the internet in the last day or so which is causing some sites to be unroutable for me. Jul 23 07:15:32 I currently cannot ping www.sf.net - from here I can ping ping 4.68.102.17, but cannot ping ping 4.68.127.190 (which is the next hop on a traceroute from an nslu2-linux host in the states to www.sf.net). Jul 23 07:16:00 s/ping ping/ping/ Jul 23 07:16:17 that's odd indeed Jul 23 07:16:21 NAiL: jbowler fixed that Jul 23 07:17:09 mickeyl: yes, it's very annoying Jul 23 07:21:03 * koen sees his name in "Most active authors today" at http://cia.navi.cx/ Jul 23 07:21:15 congrats Jul 23 07:21:37 ~lart Sharp for an inaccessible USB host connector when the PDA is in the cradle Jul 23 07:21:38 * ibot takes large quantities of Krispy Kream donuts and stuffs them one after another down Sharp's throat until Sharp puts on 150lbs for an inaccessible USB host connector when the PDA is in the cradle Jul 23 07:21:51 pfff, what a mild ~lart Jul 23 07:22:32 bah koen you're average is only 18.45 hours between messages. Jul 23 07:22:42 s/you're/your/ Jul 23 07:23:42 http://cia.navi.cx/stats/author/rwhitby <- 5.25 hours between messages over 0.87 years Jul 23 07:24:12 but congrats on the top 10 anyway :-) Jul 23 07:24:23 koen: time? got milk? Jul 23 07:25:35 zecke_sad: I got time, but I'm out of milk atm Jul 23 07:26:08 do you've enough sno(o)(o)pje? Jul 23 07:26:28 * koen searches for candy Jul 23 07:26:53 zecke_sad: yes, I have enough snoepje :) Jul 23 07:27:23 okay let us setup the first tinderbox then Jul 23 07:28:55 zecke_sad: ok, what need to go into httpd.conf? Jul 23 07:29:22 can we enter Unslung and OpenSlug distro build results into your tinderbox? Jul 23 07:29:56 rwhitby: we're not there yet ;) Jul 23 07:30:13 sorry, when it's up and running, I mean :-) Jul 23 07:30:15 rwhitby: in theory do builds via a cron job put INHERIT += "tinder" in your local.conf Jul 23 07:30:45 zecke_sad: cool. we *will* do that. Jul 23 07:31:23 rwhitby: every oe build can be used as tinderclient ;) Jul 23 07:31:53 zecke_sad: we will put it in our master makefile ;-) Jul 23 07:32:20 koen: we need one email address and some procmail magic to work on the subject Jul 23 07:32:40 Okay here is my plan: Jul 23 07:32:58 We setup one tinderbox for BitBake and one for OpenEmbedded Jul 23 07:33:12 ok Jul 23 07:33:18 we have one tinderbox at ewi email address to receive logs Jul 23 07:33:48 and distribute them doing subject recognizing to the right box Jul 23 07:33:57 that email address will be a bit of a problem, since all ports are closed, except ssh, http and monotone Jul 23 07:34:12 s/problem/delay/ Jul 23 07:34:26 we could try http post... Jul 23 07:35:20 The BitBake will only run the regression tests Jul 23 07:35:32 The OE TinderBox will have multiple 'trees' Jul 23 07:36:05 We will use bittest to report missing source and not applying patches as one tree Jul 23 07:36:31 and then as the other tree we can use the output of builds Jul 23 07:36:34 what is missing Jul 23 07:36:56 we need a VC for monotone and we need to interface with viewcvs and viewmtn Jul 23 07:37:13 zecke_sad: you're going to do world builds, or pick a handful of meta targets? Jul 23 07:37:51 CosmicPenguin: that is up to the tinderclients Jul 23 07:38:04 CosmicPenguin: I assume to run some tests on ewi Jul 23 07:38:56 koen: http://www.nslu2-linux.org/developers.php Jul 23 07:39:09 (note the new "G_HYBRID_TYPE" map type) Jul 23 07:39:42 the hybrid is nice Jul 23 07:39:50 koen: okay let us tackle apache now Jul 23 07:39:55 it solves some of the projection problems Jul 23 07:40:18 koen: /data/tinderbox,2/html_out needs to be served Jul 23 07:40:33 koen: /data/tinderbox,2/cgi-bin needs to be executable Jul 23 07:40:39 rwhitby: with LargeMap control, you will get a slider for zooming in and out Jul 23 07:41:19 koen: we've restricted that to between 10 and 15 anyway Jul 23 07:41:49 and the largemap control impinged on dyoung's marker :-) Jul 23 07:41:58 koen: html_out needs to be owned by apache user (httpd?) Jul 23 07:42:04 koen: same applies for the log Jul 23 07:43:34 zecke_sad: what runes do I need to type into httpd.conf? Jul 23 07:44:46 koen: sites-available and sites-enabled need to be tuned... Jul 23 07:46:03 koen: we create two new sites in this directory Jul 23 07:47:19 * koen in etc/apache2/sites-available Jul 23 07:47:33 koen: yes Jul 23 07:47:44 * koen vi default Jul 23 07:47:59 cp default tinderbox1 Jul 23 07:48:05 edit... Jul 23 07:48:10 cp tinderbox1 tinderbox2 Jul 23 07:50:08 hi all Jul 23 07:50:15 hi florian Jul 23 07:50:38 Is there a merge tool for monotone which doesn't need X? Jul 23 07:51:29 florian: the lightest one I could find was xxdiff (which does need X, but doesn't require thirty thousand KDE or Gnome libs like all the other merge tools). Jul 23 07:52:01 I guess you could run xemacs in tty mode ... Jul 23 07:52:22 florian: vim or emacs? Jul 23 07:52:48 florian: look at the default hooks of monotone Jul 23 07:52:54 florian: and make it use vim only Jul 23 07:53:01 in fact i have no idea why mt would need it here... but it requests it. Jul 23 07:53:30 zecke_sad: okay thanks... Jul 23 07:54:43 zecke_sad: /me vi tinderbox1 Jul 23 07:58:33 rwhitby: xemacs in tty is just plain emacs, you know... :) Jul 23 07:59:38 CosmicPenguin: um, no. It's completely different. Jul 23 07:59:40 * CosmicPenguin wanders off in search of gooey OLS knowledge Jul 23 08:00:36 CosmicPenguin: ask me one day about how I tried to get vc-clearcase.el added to both GNU Emacs and XEmacs, and the very different response I got from Stallman vs Zawinski. Jul 23 08:00:51 They are very much *not* the same. Jul 23 08:01:01 Hi! I would need some advices for creating rootfs with oe. Can somebody help? Jul 23 08:01:25 koen: /data/tinderbox1/html_out should be the DocumentRoot? Jul 23 08:01:46 zecke_sad: ok Jul 23 08:03:07 ok, changed Jul 23 08:03:26 mv /data/tinderbox/ /data/tinderbox2 Jul 23 08:04:07 zecke_sad: pssst: Jul 23 08:04:08 drwxrwxrwx 13 zecke users 4096 Jul 22 21:34 tinderbox Jul 23 08:04:08 drwxr-xr-x 2 zecke users 6 Jul 22 21:42 tinderbox2 Jul 23 08:04:50 zecke@bitbake:/data$ mv tinderbox tinderbox1 Jul 23 08:04:50 mv: cannot move `tinderbox' to `tinderbox1': Permission denied Jul 23 08:06:06 ok, done Jul 23 08:06:15 (and adapted apache for it) Jul 23 08:07:24 koen: heh? you already had /dat/tinderbox1 already? Jul 23 08:07:35 lamikr, bitbake boostrap-image ? Jul 23 08:07:40 Did I get that right? Jul 23 08:08:29 03florian 07org.oe.dev * rbc3daa84... 10/packages/meta/meta-sdk.bb: Jul 23 08:08:29 Add some bits to fixup some rough edges in the SDK filesystem: Jul 23 08:08:29 - Remove broken *.la files. Jul 23 08:08:29 - Fix pathes in *.pc files. Jul 23 08:08:47 koen: now we need two symlinks into sites-enabled/ Jul 23 08:08:57 Crofton: I have just build full gpe-image tar-ball but for some reason it does not include my kernel libs. The kernel is however build and modules are in dir tmp/work/linux-h6300-hh-2.6.12-rc5-r0/install but Jul 23 08:08:58 modules are not in dir tmp/work/linux-h6300-hh-2.6.12-rc5-r0/ Jul 23 08:09:31 lamikr: your image needs to rdepend on virtual/kernel and the modules Jul 23 08:09:45 koen: let us remove the /doc layout Jul 23 08:10:02 [Sat Jul 23 17:09:21 2005] [warn] NameVirtualHost *:0 has no VirtualHosts Jul 23 08:10:08 koen: How should I do that? Jul 23 08:10:24 lamikr: look at how h3900.conf does that Jul 23 08:10:27 * zecke_sad does a local setup as well Jul 23 08:16:19 koen: I have following h6300.conf Jul 23 08:16:22 #@TYPE: Machine Jul 23 08:16:24 #@NAME: OMAP1510 cpu on an H2 board Jul 23 08:16:26 #@DESCRIPTION: Machine configuration for the OMAP H2 development board with a 1510 processor Jul 23 08:16:27 TARGET_ARCH = "arm" Jul 23 08:16:29 IPKG_ARCHS = "arm9tdmi ${MACHINE}" Jul 23 08:16:31 PREFERRED_PROVIDER_xserver = "xserver-kdrive" Jul 23 08:16:32 PREFERRED_PROVIDER_virtual/kernel = "linux-h6300-hh" Jul 23 08:16:33 PREFERRED_PROVIDERS += "virtual/${TARGET_PREFIX}depmod:module-init-tools-cross" Jul 23 08:16:35 BOOTSTRAP_EXTRA_DEPENDS += "virtual/kernel modutils-collateral" Jul 23 08:16:37 BOOTSTRAP_EXTRA_RDEPENDS += "modutils-collateral" Jul 23 08:16:38 SERIAL_CONSOLE ?= "115200 ttyS0" Jul 23 08:16:39 lamikr: pastebin Jul 23 08:16:39 EXTRA_IMAGECMD_jffs2 = "--pad --little-endian --eraseblock=0x40000" Jul 23 08:16:41 PACKAGE_ARCH = "arm9tdmi" Jul 23 08:16:55 http://pastebin.ca Jul 23 08:17:16 florian: heads merged Jul 23 08:17:20 * koen hugs meld Jul 23 08:17:32 * koen stabs remote X to XDarwin Jul 23 08:18:46 koen and croft: Thanks for the hint: http://pastebin.ca/18335 Jul 23 08:20:50 koen: yeah... i wonder what it tried to merge before, i removed meld and tried again and it didn't try to merge anything. Jul 23 08:22:47 is the h6300 an omap machine? Jul 23 08:23:01 lamikr: arm9tdmi isn't correct Jul 23 08:23:17 lamikr: see how nokia770.conf specifies it (omap1710) Jul 23 08:23:31 Florian: Yes, it is based on omap1510 Jul 23 08:25:18 florian: it has arm9tdmi and atleast gpe itself build with these settings. I have also get things working arm9tdmi when I have build crosscompiler and gpe apps earlier by hand. But is it just too painfull in that way to try to figure out all required configs and dependencies... Jul 23 08:26:41 lamikr: but arm9 isn't a real arch, armv5 is Jul 23 08:26:43 * zecke_sad gets some food Jul 23 08:28:38 fwiw, omap1510 doesn't actually use arm9tdmi anyway Jul 23 08:31:04 oh, actually, maybe it does. I thought all those TI chips were arm926, but apparently not. Jul 23 08:31:20 1510 is arm925 Jul 23 08:31:28 Koen: You are correct: The processor h6300 has is arm925t Jul 23 08:32:03 hmm Jul 23 08:32:23 koen,florian: i don't see bc3daa844870742c024624bb588b3a94af824c6d in the head branch Jul 23 08:32:24 koen: hmm how do we put tinderbox1/ in the '/' Jul 23 08:32:27 pb_: I think 926 is used on omap1610 or 1710 Jul 23 08:33:18 it looks like 7d0136e54bdf205173cc235d2b4a5d747c2bd8bc&id2=88f48c5aa876635c2ee02c93af2b2f3e66b797d1 reverted the changes Jul 23 08:33:18 lamikr: yah, so it seems. I wonder why they used arm925 in the 1510. Jul 23 08:33:20 was that intentional ? Jul 23 08:33:30 i.e. meta-sdk is back at PR = r2 Jul 23 08:33:32 http://people.apache.org/~rbowen/presentations/apacheconEU2005/hate_apache.pdf Jul 23 08:34:50 mickeyl: hmmm, I remember merging it to R4 Jul 23 08:35:19 did you merge to LEFT (or CENTER when you used 3way) ? Jul 23 08:35:30 aha Jul 23 08:35:34 left on 3 way Jul 23 08:35:37 oops Jul 23 08:35:39 that's wrong Jul 23 08:35:40 pb: I can use -mcpu=arm9tdmi when compiling for h6300 for -march there is not that kind of option. Jul 23 08:35:46 lamikr: anyway, "armv4t" would probably be the appropriate PACKAGE_ARCH for arm9tdmi. Jul 23 08:36:12 indeed, since arm9tdmi is a CPU and not an architecture Jul 23 08:36:37 mickeyl: it did a 2 way merge followed by a 3 way Jul 23 08:36:39 How about armv4, I have thought that armv4t would force using 16 bit thump instruction set by default. Jul 23 08:36:49 mickeyl: the first time it said bla bla *LEFT* bla Jul 23 08:36:57 A small question, the package cpp-symlinks (gcc/gcc-package.inc) never gets any files, because the link ${bindir}/cpp never gets created. Is there any reason why? Jul 23 08:37:01 koen: the second time it usually says bla bla *CENTER* bla :) Jul 23 08:37:16 mickeyl: I see that after scrolling up :( Jul 23 08:37:26 DaKa2: what should the symlink link to ? Jul 23 08:37:29 kdiff3 seems superior to meld Jul 23 08:37:31 i'd use that Jul 23 08:37:44 So I can not specify cpu options to machine conf, only architecture? Jul 23 08:37:45 it looks like it's automatically prefered when you have it installed Jul 23 08:38:00 mickeyl: that would require install 300+ mb of kde, just for merging Jul 23 08:38:01 ln -sf ${bindir}/${TARGET_SYS}-cpp ${D}${bindir}/cpp should probably be added, or? Jul 23 08:38:33 koen: ah, so you have a kde free workstation Jul 23 08:38:37 mickeyl: on a box without a monitor attached, and I allready had gtk+gnome for kaffee Jul 23 08:38:42 i see Jul 23 08:39:31 * koen tries to monotone disapprove 8d875e3e0f79f3c0d11d7ee0371624b84c02e603 Jul 23 08:40:05 mickeyl: the big update-alternatives patch is about to hit your org.openembedded.dev branch. The only outstanding issue we see is the one that DaKa2 just reported about a missing cpp symlink. Jul 23 08:40:47 florian: could you secure the correct meta-sdk.bb, pull, patch it and sync ? Jul 23 08:41:05 who uses c++ anyway? Jul 23 08:41:07 * koen runs Jul 23 08:41:34 lamikr: I think you are ascribing powers to PACKAGE_ARCH that it does not possess. Jul 23 08:42:11 the only function of that variable is to control what ends up in the Architecture field of the .ipk files. It does not have any influence on compiler options, code generation, instruction set choice, or anything like that. Jul 23 08:42:13 koen: Yeah the preprocessor is deprecated anyway Jul 23 08:42:15 rwhitby: hmm cpp symlink? Jul 23 08:42:27 re Jul 23 08:42:33 rwhitby: will deal with it when it occurs Jul 23 08:42:58 * koen gets annoyed by monotone: misuse: revision 8d875e3e0f79f3c0d11d7ee0371624b84c02e603 has 2 changesets, cannot invert Jul 23 08:43:03 lamikr: ah cool... yes right, if you intend to build a whole system oe is way better than compiling by hand. Jul 23 08:43:06 yes. the gcc-symlinks package symlinks all target tools into bindir, with the notable exception of cpp. Jul 23 08:43:25 pb: Ok. So what is you opinion for the preferred name used: armv4 or armv4t? Jul 23 08:43:39 * koen goes does the dishes to make some food Jul 23 08:43:53 lamikr: personally, I would use armv4t, but either would be fine. Jul 23 08:43:54 rwhitby: ah. that shouldn't be hard to fix Jul 23 08:44:15 mickeyl: DaKa2 knows how to fix it - we're just asking whether it might be intentional for some reason ... Jul 23 08:44:35 lamikr: if you use armv4, you should avoid emitting any BX instructions, but that shouldn't be hard to arrange. Jul 23 08:44:45 ah. i don't think it was, but then again, i'm not a toolchain expert. pb ? Jul 23 08:45:02 I don't think it was intentional Jul 23 08:45:06 mickeyl: http://monotone.vanille.de/viewmtn/revision.psp?id=cbe32b306c5e7609bbe24ffbbb6c1f179f6bede5 is not in your repo with the update-alternatives patch. Jul 23 08:45:11 DaKa2: please go ahead and fix it Jul 23 08:45:14 s/not/now/ Jul 23 08:45:15 03rwhitby 07org.oe.nslu2-linux * r76260c1a... 10/packages/meta/ (openslug-native.bb openslug-packages.bb): Added openslug-native - set of packages required to be installed for the OpenSlug native build environment Jul 23 08:45:16 pb: What are BX instructions? Jul 23 08:45:40 lamikr: "BX" is the only instruction that is in armv4t (and hence supported by your arm9tdmi) but not in plain armv4. Jul 23 08:45:48 rwhitby: ok, good so far. Jul 23 08:45:52 * rwhitby and DaKa2 bunker down waiting for any update-alternatives fallout ... Jul 23 08:45:56 hehe Jul 23 08:46:04 i will launch a fresh built of openzaurus and will see Jul 23 08:46:14 * mickeyl waits for florian to repair meta-sdk Jul 23 08:46:17 lamikr: if you choose to select "armv4" as your PACKAGE_ARCH, you should avoid generating any instructions that will not run on a real armv4 cpu Jul 23 08:47:38 pb: Ok. Is it the TARGET_ARCH that defines the architecture option passed when OE starts building cross-gcc? Jul 23 08:48:13 lamikr: no, it's the TARGET_CC_ARCH Jul 23 08:49:11 any apache guru around? Jul 23 08:49:17 TARGET_ARCH should be set to the cpu family, usually "arm". Jul 23 08:49:30 or "armeb", maybe, for those crazy big endian guys Jul 23 08:49:51 pb: Where it is defined, I tried to check from h3900 and nokia770.conf files Jul 23 08:49:55 mickeyl: that doesn't seem to be that easy... mt just destroyed my local version. Jul 23 08:50:03 lamikr: typically in tune-.conf Jul 23 08:50:35 florian: oh dear Jul 23 08:50:46 * florian wonders how this could happen Jul 23 08:51:10 did you type "mt annihilate" or any such command? Jul 23 08:54:08 no... it wanted to merge using melt which failed because of lacking X, after this it didn't try to do this again and the next sync+update seems to have destroyed my local copy. Jul 23 08:55:19 no problem, just check out the PR4 revision Jul 23 08:55:30 we have a SCM on our hands, remember? :) Jul 23 08:55:43 yep, i found it... Jul 23 08:55:48 good Jul 23 08:56:43 pb or koen: Should following make the kernel to be included with gpe-image? BOOTSTRAP_EXTRA_RDEPENDS += "virtual/kernel sysfsutils console-tools modutils-collateral" Jul 23 08:57:46 pb or koen: I took idea from nokia770.conf. h3900.conf seems to define also a lot of other things like BOOTSTRAP_EXTRA_RDEPENDS_append_kernel26 = " udev module-init-tools" is that required? Jul 23 08:58:23 it's required if you require it Jul 23 08:58:33 look at the packages and think about what they do Jul 23 08:58:38 then decide if you need them to be in your image or not Jul 23 08:58:45 Here we go... Jul 23 08:59:51 Did anyone here try that SDK package i built? Jul 23 09:00:10 koen: can we do Alias /bitbake_qa /data/tinderbox1/html_output? Jul 23 09:00:15 mickey: I mostly wanted to know is the difference between keywords BOOTSTRAP_EXTRA_RDEPENDS_append_kernel26 and BOOTSTRAP_EXTRA_RDEPENDS Jul 23 09:00:31 ah. the difference is that one is a conditional append Jul 23 09:00:34 03florian 07org.oe.dev * r86f9faea... 10/packages/meta/meta-sdk.bb: One more try to establish r4 version. Jul 23 09:00:39 there other one (+=) is an unconditional append Jul 23 09:00:42 s/there/the/ Jul 23 09:00:58 read in the bitbake manual about OVERRIDES Jul 23 09:01:11 lamikr: the h3900 does manipilate OVERRIDES as well Jul 23 09:01:45 zecke: Ok, I will now try to make another build... Jul 23 09:01:47 florian: ok, then I can push my change now Jul 23 09:01:52 thanks for all. Jul 23 09:02:27 mickey|break: okay... Jul 23 09:05:31 zecke_sad: see http://ewi546.ewi.utwente.nl/bitbake_qa/ Jul 23 09:11:33 is there any daemon in oe that can turn off a computer after a certain amount of inactivity? Jul 23 09:12:25 micropal: when you say computer you mean.... Jul 23 09:12:35 a amd computer Jul 23 09:12:51 geode? Jul 23 09:13:20 asus amd duron atx motherboard. I want to make a htpc distro which turn off the computer when it's inactive. Jul 23 09:14:10 power off or suspend? You're being too vague Jul 23 09:14:36 Power off Jul 23 09:15:04 03daka 07org.oe.nslu2-linux * r494c257a... 10/packages/gcc/ (gcc-package.inc gcc_3.4.4.bb): Add symlink to cpp, so package cpp-symlinks actually gets created Jul 23 09:15:50 The bios is featureless, so I need a software solution. Jul 23 09:16:04 you don't have ACPI or APM at your disposal? Jul 23 09:16:09 mickey|break: update-alternatives patch complete. Jul 23 09:16:39 Yeah, I have acpi. Jul 23 09:16:48 But I have no bios feature to turn it off. Jul 23 09:17:14 (to turn the computer off) Jul 23 09:17:20 So what you're saying is that you have ACPI, but it doesn't expose power states to you? Jul 23 09:18:09 I want to turn the computer automaticly off after a certain amount of inactivity. 5 min without any hid event and mplayer not running for instance. Jul 23 09:19:09 ok - so whats your problem then? Jul 23 09:19:34 I wonder if there is any such program in oe Jul 23 09:19:38 probably not Jul 23 09:20:20 but its pretty basic functionality - stuff like that usually doesn't get packaged up because people don't want to bother with a release Jul 23 09:20:45 Either write a script for a screensaver, or write your own - that will take care of the HID events, and you can do a script to check for mplayer Jul 23 09:21:01 Then you can shut it down Jul 23 09:21:37 yeah, I think I'll do that or something similar. Jul 23 09:22:15 micropal: I do think you would be more pleased if you did a suspend to disk - ten you wouldn't have to deal with BIOS post Jul 23 09:22:22 s/ten/then/ Jul 23 09:23:15 http://ewi546.ewi.utwente.nl/bitbake_qa/cgi-bin/tinder.cgi 404 Jul 23 09:23:18 We need more people playing with suspend2 and swsusp Jul 23 09:27:01 Yeah, the biosdelay is an issue, but I want to be able to just unplug the power without doing a proper shutdown. I believe suspend2 has a mode where it doesn't update the suspended-imagefile. It might be what I'm looking for, as I prefer disk-on-module storage. Jul 23 09:30:21 koen: I would have found the sites-available/ thing more nice Jul 23 09:30:35 koen: but we can put everything in httpd.conf as well Jul 23 09:38:03 tinderbox1 and tinderbox2 should be set up correctly now Jul 23 09:42:54 zecke_sad: cgi-bin is in place Jul 23 09:43:28 koen: now we need to fix file permissions Jul 23 09:43:51 /OE_qa points to tinderbox2 Jul 23 09:46:40 zecke_sad: which files need what permissions? Jul 23 09:47:13 koen: check error.log Jul 23 09:47:23 koen: I think html_output needs to be owned by www-data Jul 23 09:48:14 zecke_sad: [Sat Jul 23 18:48:06 2005] [error] [client 83.169.171.16] Could not open logfile: /tmp/tinderbox2/tinderbox2.log Jul 23 09:49:49 now? Jul 23 09:51:08 zecke_sad: [Sat Jul 23 18:50:47 2005] [error] [client 83.169.171.16] Could not open logfile: /tmp/tinderbox2/tinderbox2.log[Sat Jul 23 18:50:47 2005] [error] [client 83.169.171.16] Premature end of script headers: tinder.cgi[Sat Jul 23 18:50:47 2005] [error] [client 83.169.171.16] File does not exist: /var/www/favicon.ico Jul 23 09:51:31 zecke_sad: you should be able to read /var/log/apache2/error.log now Jul 23 09:52:32 koen: apt-get install libsvn-core-perl Jul 23 09:52:48 done Jul 23 09:53:09 anyone know why bitbake isn't in monotone, btw? Jul 23 09:53:39 NAiL: http://oe.handhelds.org/, 4th item Jul 23 09:54:20 NAiL: because svn suffices for bitbake, and doesn't suffer from monotone's drawbacks Jul 23 09:54:29 koen: could you allow me to get www-data? Jul 23 09:54:36 zecke_sad: ? Jul 23 09:54:43 koen: or run /data/tinderbox1/cgi-bin/tinder.cgi as www-data? Jul 23 09:56:03 Thats the first time I've heard somebody complain about bitbake *not* being monotone Jul 23 09:56:12 heh Jul 23 09:56:26 not complaining, just wondering Jul 23 09:56:36 especially since I can't get bitbake right now :P Jul 23 09:58:26 zecke_sad: how do I convince linux/apache to do that? Jul 23 10:00:26 koen: it does it automatically Jul 23 10:00:44 koen: you specify which user should run the cgi scripts Jul 23 10:00:55 koen: and if you try to run it as another user it complains and does not work Jul 23 10:02:36 it just keeps on saying [Sat Jul 23 19:02:27 2005] [error] [client 83.169.171.16] Could not open logfile: /tmp/tinderbox2/tinderbox2.log[Sat Jul 23 19:02:27 2005] [error] [client 83.169.171.16] Premature end of script headers: tinder.cgi Jul 23 10:02:58 koen: I see that but it shouldn't do that ;) Jul 23 10:03:05 this is why I want to run tinder.cgi by hand Jul 23 10:03:13 ah, I see Jul 23 10:03:31 do you have screen running? Jul 23 10:03:44 yes Jul 23 10:05:09 koen: For some reasons kernel modules where still not included to either rootfs or gpe-images. Does it help if I add H6300_MODULES_virtual/kernel = " omapts" to h6300.conf in http://pastebin.ca/18340 Jul 23 10:05:34 koen: We have kernel module called omapts for touschreen... Jul 23 10:05:41 are the h6300 modules in BOOTSTRAP_RDPEPENDS? Jul 23 10:05:46 local_conf/TinderConfig.pm:137:$ERROR_LOG = "/data/tinderbox2/tinderbox_oe.log"; Jul 23 10:06:39 aha Jul 23 10:08:31 koen: So I should define them in this way: BOOTSTRAP_EXTRA_RDEPENDS += "kernel-module-omapts" and then they would be included in? Jul 23 10:09:15 that should work Jul 23 10:17:07 http://ewi546.ewi.utwente.nl/bitbake_qa/ Jul 23 10:17:21 http://ewi546.ewi.utwente.nl/bitbake_qa/Bitbake/status.html Jul 23 10:18:45 woohoo! Jul 23 10:19:16 koen: now only OE_qa Jul 23 10:19:26 koen: some more permissions are wrong Jul 23 10:19:45 NAiL: you can not connect svn.berlios.de? Jul 23 10:21:18 koen: we could create tinderbox-oe AT gmx.de Jul 23 10:21:30 zecke_sad: hotmail! Jul 23 10:21:33 * koen hides Jul 23 10:21:36 koen: and then let getmail/fetchmail run Jul 23 10:21:42 koen: can you pop3 hotmail? Jul 23 10:21:58 * florian wants a gpe-image for his indy Jul 23 10:22:51 heh - nobody has a gmail account they don't use any more? Jul 23 10:23:06 * koen has a zillion invites left Jul 23 10:23:32 * zecke_sad lost once again his login + password Jul 23 10:23:36 do you see a pattern? Jul 23 10:23:50 zecke_sad: what's up with the Sat Jul 23 19:23:46 2005] tinder.cgi: Could not rename: '/data/tinderbox2/html_output/Tmp.OpenEmbedded.1122139426.29673 to /data/tinderbox2/html_output/OpenEmbedded. Is a directory ? Jul 23 10:24:08 koen: some kind of permission error?! Jul 23 10:24:43 zecke_sad: it does seem to be moving a *file* to a *dir* Jul 23 10:24:43 koen: the directory is owned by users? Jul 23 10:24:55 oh right Jul 23 10:25:10 if you rm both, they get recreated with the same error Jul 23 10:25:36 oh yeah Jul 23 10:25:44 tilde lart me Jul 23 10:26:05 koen: that is due having spaces in the tree name Jul 23 10:26:43 heh Jul 23 10:27:13 koen: I can not write that file anymore Jul 23 10:28:14 try now Jul 23 10:28:52 http://ewi546.ewi.utwente.nl/OE_qa/ Jul 23 10:29:11 koen: now I would love to have three cron jobs Jul 23 10:29:43 zecke_sad: svn: Can't connect to host 'svn.berlios.de': Connection refused Jul 23 10:29:43 make: *** [bitbake/bin/bitbake] Error 1 Jul 23 10:30:21 Nail: I put binaries to here: http://aragorn.kortex.jyu.fi:8080/h6300/bitbake-1.3.1.tar.gz Jul 23 10:30:41 zecke_sad: put the scripts in a dir and I'll dump then in /etc/cron.d Jul 23 10:31:23 koen: Could you the jobs? I would like to setup clients now Jul 23 10:31:52 zecke_sad: what should the jobs be? Jul 23 10:32:29 koen: call tinder.cgi Jul 23 10:35:42 koen: I'm going to setup a gmx account now Jul 23 10:45:27 http://ewi546.ewi.utwente.nl/bitbake_qa/Bitbake/status.html Jul 23 10:50:49 rwhitby-asleep: ping Jul 23 10:54:52 Crofton__|laptop: can you run a tinderclient at your facility? Jul 23 10:55:07 hmmmm Jul 23 10:55:11 what is it ? Jul 23 10:55:41 Crofton__|laptop: do builds for your board regulary Jul 23 10:55:48 Crofton__|laptop: (automatically) Jul 23 10:55:57 probably Jul 23 10:56:02 do you have an URL? Jul 23 10:56:33 Crofton__|laptop: http://ewi546.ewi.utwente.nl/OE_qa/OpenEmbeddedBuild/status.html Jul 23 10:56:48 Crofton__|laptop: your machine will be in one column Jul 23 10:57:30 Do you a link to some instructions on what I need to do? Jul 23 10:57:38 put INHERIT += "tinderclient" in your local.conf Jul 23 10:57:57 and then build from cron? Jul 23 10:58:22 TINDER_ADMIN, TINDER_BUILD, TINDER_ERROR, TINDER_TREE Jul 23 10:58:48 ? Jul 23 10:58:49 TINDER_MAILTO, TINDER_FROM Jul 23 10:59:00 these values need to be filled out in your conf/local.conf as well Jul 23 10:59:02 you are confusing me now Jul 23 10:59:05 ah Jul 23 11:00:04 I haven't been following the tinderbox stuff stuff much Jul 23 11:00:24 But you are looking for people to rebuild various systems on a regualr basis? Jul 23 11:00:25 http://tinderbox.mozilla.org/showbuilds.cgi?tree=XULRunner Jul 23 11:00:31 Crofton__|laptop: there wasn't much to follow :) Jul 23 11:00:45 maybe thats why I didn't follow it :) Jul 23 11:00:52 Crofton__|laptop: Yes we want people to build oe for different configurations regulary Jul 23 11:01:38 zecke: each time from scratch? ouch :) Jul 23 11:01:45 I'll look at setting up something to keep a repo updated and do a build Jul 23 11:02:03 I probably shouldn't use the repo I work in since it id frequently broken Jul 23 11:02:10 * koen going to watch SG: atlantis 2x02, bbl Jul 23 11:02:27 send it my way ;) Jul 23 11:02:36 zecke, Bug me again late next week Jul 23 11:02:48 I need to write a paper this week Jul 23 11:03:04 and I would much rather set up tinderbox than do real work:) Jul 23 11:03:37 I want to do this for our project also now Jul 23 11:03:49 morning Jul 23 11:04:12 svn.berlios.de seems down. Jul 23 11:04:31 Crofton__|laptop: we can send you our configuration... Jul 23 11:05:01 Cool Jul 23 11:05:08 Do you have my email? Jul 23 11:05:19 philip@balister.org Jul 23 11:05:49 Crofton__|laptop: I might ask again for your address Jul 23 11:07:00 heh Jul 23 11:07:07 My mind is goign also :) Jul 23 11:09:30 anyone here know if you can hang a usb wireless adapter on a nslu2 box? Jul 23 11:09:52 don't see why not. There's nothing special about the nslu2's USB. Jul 23 11:10:16 I want to make a portable webcam :) Jul 23 11:10:22 heh Jul 23 11:10:35 hmm, how much on board flash is there? Jul 23 11:10:40 I know I can look that up .... Jul 23 11:10:55 8mbyte if I recall correctly Jul 23 11:10:56 spyonmycats.com Jul 23 11:11:05 not much. a few megs, maybe 4 or 8 Jul 23 11:11:18 oe can build images now, roght? Jul 23 11:11:29 indeed, building images is one of oe's core competencies Jul 23 11:11:35 What about getting a wireless router with usb, then you can have a wireless webcam :) Jul 23 11:11:37 heh Jul 23 11:11:49 hmmm Jul 23 11:12:00 know of one that I can build images for? Jul 23 11:12:17 micropal: heh. or, of course, you could just go out and buy a wireless webcam. Jul 23 11:12:32 I need to get my advisor to buy me more disk space for all these projects :) Jul 23 11:12:32 yeah, easier ;) Jul 23 11:13:26 but with the nslu2 I have more flexibility Jul 23 11:13:35 after the wireless webcam gets boring Jul 23 11:24:15 hold tight guy I'm syncing :\ Jul 23 11:28:34 enter passphrase for key ID [coredump@openembedded.org]: Jul 23 11:28:34 monotone: bytes in | bytes out | revs in | revs out | revs written Jul 23 11:28:34 monotone: 18.5k | 94.3k | 0 | 4 | 0 Jul 23 11:28:45 the damn thing is just sitting there Jul 23 11:30:23 And it will Jul 23 11:31:26 nm, monotone: successful exchange with ewi546.ewi.utwente.nl Jul 23 11:32:48 * chouimat is away: food shopping Jul 23 11:50:13 hmm Jul 23 11:50:14 * mickeyl surprised Jul 23 11:50:24 I plugged a USB keyboard into my PDA and Jul 23 11:50:25 ... Jul 23 11:50:27 it just worked! Jul 23 11:50:33 a shocking experience Jul 23 11:50:41 even with opie Jul 23 11:51:49 * mickeyl hugs the usb subsystem Jul 23 11:51:53 lol, horrifying Jul 23 11:52:06 * koen looks at his camera connected to his h5550 Jul 23 11:52:29 well, camera is usb-storage... that's fairly common. usb-hidp is much more exciting :) Jul 23 11:52:44 mickeyl: ptp camere :) Jul 23 11:52:48 libgphoto2 Jul 23 11:52:52 03coredump 07org.oe.dev * r98bc20b1... 10/packages/libsdl/ (libsdl-qpe-1.2.7/SDL-Akita.patch libsdl-qpe_1.2.7.bb): Akita: Add machine definition for SDL Jul 23 11:52:56 03coredump 07org.oe.dev * rce7180a4... 10/packages/ipkg/ (3 files in 2 dirs): Akita: Add SD, CF and home entries to ipkg.conf and don't move status to /var Jul 23 11:53:14 hmm Jul 23 11:53:22 CoreDump|home: why ? Jul 23 11:53:25 hmm well... not for me, my workstations only have usb for hid devices Jul 23 11:53:36 CoreDump|home: why not use ODevice? Jul 23 11:53:47 aehem huh? Jul 23 11:54:15 CoreDump|home: look at the ipaq patch Jul 23 11:54:16 SDL has definition for all other machines, not akita. Where the problem when I add it? Jul 23 11:54:40 CoreDump|home: use ODevice where possible but I will look at your patch now Jul 23 11:54:50 thx Jul 23 11:55:01 it's like um 5 lines of code Jul 23 11:55:38 "and don't move status to /var" Jul 23 11:55:40 why that? Jul 23 11:56:02 mickeyl: ipkg status in /var? Jul 23 11:56:02 because Akita has a ton of flash :) no need to loose the data on each reboot Jul 23 11:56:07 that's asking for troubles Jul 23 11:56:11 eh not status Jul 23 11:56:13 my bad Jul 23 11:56:20 server data only Jul 23 11:56:21 hmm Jul 23 11:56:23 typo Jul 23 11:56:34 that will make reboots *incredibily* slow Jul 23 11:57:14 since ipkg parses the lists on boot Jul 23 11:57:31 why would that be? It doesn't matter where the server data is IMO. On-flash has the advantage that it isn't lost during a reboot Jul 23 11:57:33 and ipkg will be a lot slower since it has to read the lists from flash instead of ram Jul 23 11:57:50 20:57 < koen> since ipkg parses the lists on boot Jul 23 11:58:22 well, play with it, you have the machine. if you think it's better on flash, then go for it. Jul 23 11:58:28 koen: I'll check, never was a notable problem on collie tough Jul 23 11:58:30 CoreDump|home: the patch itself is okay (SDL) Jul 23 11:58:39 zecke: thx :) Jul 23 11:59:01 CoreDump|home: try adding a big feed Jul 23 11:59:07 boot time + 5 minutes Jul 23 11:59:15 I'll pull all OZ feeds Jul 23 11:59:28 should be big enough =) Jul 23 11:59:36 ipkg reads lists on *boot*? Jul 23 11:59:41 why would ipkg run on boot? Jul 23 11:59:51 to see if its needs to configure packages Jul 23 11:59:59 .....*what*?? Jul 23 12:00:03 yes Jul 23 12:00:08 but wouldn't it read "status" for that? Jul 23 12:00:11 isn't that something that needs to happen only on the first boot? Jul 23 12:00:24 CoreDump|home: whatever you do with ipkg, it always parses the lists Jul 23 12:00:30 ah i c Jul 23 12:00:35 CoreDump|home: not that I like that we parse that info twice (once in libopie and once in libsdl) Jul 23 12:01:15 zecke: yep, noticed that, too. Not efficient. Jul 23 12:01:50 not efficient, but faster and more compatible Jul 23 12:01:54 * koen hides Jul 23 12:02:12 * mickeyl chuckles Jul 23 12:02:15 heh Jul 23 12:02:23 efficiency is "fluff" :D Jul 23 12:02:32 koen: again, I ask, isn't this something that should happen only on first boot? I can think of no reason whatsoever that ipkg needs to configure packages on boot except first boot. Jul 23 12:02:39 * koen should stop repeating trolltech marketing mantras Jul 23 12:02:43 CoreDump|home: know what's pretty cool? Jul 23 12:02:45 JustinP: no idea Jul 23 12:02:52 mickeyl: what? Jul 23 12:03:03 CoreDump|home: we will perhaps have a localized keymap for your hentges 3.5.4 without the damn keyz hack Jul 23 12:03:14 koen: then it shouldn't be running on every boot. There should be an rc entry which runs ipkg on first boot then removes itself Jul 23 12:03:22 mickeyl: ahem, kewl! Jul 23 12:03:26 03mickeyl 07org.oe.dev * r694e3e88... 10/packages/libusb/libusb_0.1.10a.bb: libusb: disable parallel make to fix compiling Jul 23 12:03:29 CoreDump|home: finally all our SL5500G will behave Jul 23 12:03:45 * JustinP checks his Zaurus to see if his zaurus is running ipkg on boot Jul 23 12:04:02 should have thought earlier of that hehe. It's a kinda obvious hack. Damn lol Jul 23 12:04:13 mickeyl: I guess you run more -j's than I do ;) Jul 23 12:04:17 i never thought that the loadkeys thing would also work on our retarded kernel Jul 23 12:04:22 koen: *nod* :)) Jul 23 12:04:36 now that you came with the akita keymap everything fell into place Jul 23 12:04:44 JustinP: I don't know, you would have to talk to the people who added that Jul 23 12:04:51 and thanks to Überhacker Zecke's qte keymap driver of course Jul 23 12:04:54 ~praise zecke Jul 23 12:04:55 All hail zecke! Jul 23 12:05:01 well, the guy at oesf.org found out that it worked :) Jul 23 12:05:20 you could have asked me as well ;) Jul 23 12:05:32 it's a bit ashaming that we shipped so many releases without a proper keyboard when it would've been _so_ easy Jul 23 12:05:37 any linux kernel >= 0.23 probably would have worked Jul 23 12:06:03 =) Jul 23 12:06:39 next step will be to attach a mouse Jul 23 12:06:49 CoreDump|home: now if we could educate the oesf people to use bugzilla instead of expecting devs to to inspect every forum in the world..... Jul 23 12:07:01 koen: ummm..... /etc/rcS.d/S98configure has an exit 0; before the ipkg-cl Jul 23 12:07:04 koen: right. Jul 23 12:07:06 this has already be done, but i'm looking to do it in the Opie core, not with YAA (yet another applet) Jul 23 12:07:06 koen: we should say bugzilla is a forum Jul 23 12:07:21 hehe Jul 23 12:07:27 bugzilla _is_ a form Jul 23 12:07:29 forum Jul 23 12:07:32 kind of Jul 23 12:07:45 JustinP: ok, so somebody changed that when I wasn't paying attention Jul 23 12:07:50 mickeyl: yeah most people would not recognize the difference Jul 23 12:08:11 mickeyl: can we have pictures of those japanese cartoons in bugzilla? Jul 23 12:08:12 mickeyl: I think I will release the next version of the keymap driver soon Jul 23 12:08:20 koen: just letting you know Jul 23 12:08:25 mickeyl: if not, I'm affraid bugzilla isn't geeky enough Jul 23 12:08:28 koen: hehe Jul 23 12:08:51 koen: regarding changes... I think I want to include diffs in the commits Jul 23 12:08:56 mickeyl: me too Jul 23 12:09:01 we have twice the people with commit access now Jul 23 12:09:07 mickeyl: although I use the rss feed Jul 23 12:09:09 ~lart gentoo for making peole think forums are acceptable Jul 23 12:09:10 * ibot brandishes Excalibur! "With this sword, I vanquish thee, gentoo!" and lops off gentoo's head for making peole think forums are acceptable Jul 23 12:09:17 feed://cia.navi.cx/stats/project/openembedded/.rss Jul 23 12:09:24 there's too many chances people change things in places they're not supposed to be Jul 23 12:09:26 wow, ibot is doing some l33t parsing Jul 23 12:09:39 at least not without discussing the consequences Jul 23 12:10:18 mickeyl: we could split it up in 'core' and 'stuff' Jul 23 12:10:20 * JustinP doesn't understand rss....what kind of reader makes it worthwhile? Jul 23 12:10:37 koen: do you've the complete Stargate Atlantis episodes? Jul 23 12:10:48 safari rss is doing a pretty good job Jul 23 12:10:50 is there a commit mailing list for OE? Jul 23 12:10:51 cute Jul 23 12:10:59 JustinP: oe-commits@handhelds.org Jul 23 12:11:01 * JustinP doesn't use OS X unless he has to Jul 23 12:11:01 zecke_sad: yes Jul 23 12:11:14 koen: can one rsync the files somewhere? Jul 23 12:11:34 zecke_sad: as a 'personal backup' ? Jul 23 12:11:39 yes Jul 23 12:11:49 and what would be more secure than my localhorst Jul 23 12:11:50 hehe Jul 23 12:13:20 koen: I'm not sure. I don't want people to get the feeling some have more rights than others. I'd rather alarm people to think about how many different platforms and distributions OE is used, to broaden their focus and to make them discuss changes. Although that may slow down development a bit I think it's really necessary. Our QA system will help as well, but we shouldn't rely on technique, we should rely on responbility and care. Jul 23 12:13:39 true Jul 23 12:14:09 guess I should write something on the wiki about that Jul 23 12:14:15 trust is good, control better Jul 23 12:14:25 that's my credo Jul 23 12:14:49 after seeing fellow project members blow up more than one core IC Jul 23 12:15:05 hehe Jul 23 12:15:06 "I'm sure it can take 5A" *poof* Jul 23 12:15:13 :D Jul 23 12:15:21 ~lart CMOS Jul 23 12:15:21 * ibot acting on orders from an unspecified client drags CMOS into court suing for $200 million Jul 23 12:16:00 I agree with mickey; I don't think it would be very desirable to establish some kind of eleet "oe-cabal" group, and banish the proletariat to some kind of "contrib" ghetto. Jul 23 12:16:10 CoreDump|home: or forgetting to put resistors between and lcd and the driver IC Jul 23 12:16:27 if people can't be trusted to make changes in a responsible fashion, they shouldn't have write access to the tree at all. Jul 23 12:16:28 * zecke_sad wishes he could do dpkg-reconfigure exim4 on his BSD box Jul 23 12:16:30 oh a cabal is forming :) Jul 23 12:16:35 wow - mr. dictionary has arrived.. :) Jul 23 12:16:59 I wouldn't spell proletariat the same way twice in my entire life Jul 23 12:17:17 koen: ouch Jul 23 12:17:42 hell, I wouldn't *pronounce* it the same way twice Jul 23 12:17:43 CoreDump|home: I only said it once every ten minutes: "put resistors in between" Jul 23 12:18:00 people can't be _that_ stupid Jul 23 12:18:15 CoreDump|home: trust is good, control is better Jul 23 12:18:22 indeed Jul 23 12:19:26 * mickeyl suddenly gets remembered about what he liked about british literature Jul 23 12:19:28 ummm....*** glibc detected *** double free or corruption (!prev): 0x0000000000612c90 *** Jul 23 12:19:31 the wording rules Jul 23 12:19:33 should I be worried about this? Jul 23 12:19:45 (during gpe-image do_rootfs) Jul 23 12:20:22 which glibc is that ? your local ? Jul 23 12:20:27 then i'd be worried Jul 23 12:21:20 I'm not sure. It just is ouputting this during the do_rootfs Jul 23 12:22:23 it depends - is it a double free, or is it corruption? Jul 23 12:22:26 double free is bad Jul 23 12:22:34 /is/isn't that/ Jul 23 12:23:38 zecke_sad: could you push a tinderbox.inc containing the needed vars? Jul 23 12:23:39 * JustinP doesn't know....that's all it says Jul 23 12:23:48 mickeyl: is vanille down again Jul 23 12:24:05 koen: look at $HOME/build/simpad/conf/local.conf (at the bottom) Jul 23 12:25:21 bbl, food Jul 23 12:25:56 zecke_sad: restarted Jul 23 12:26:00 03freyther 07org.oe.dev * r50481fac... 10/classes/tinderclient.bbclass: Jul 23 12:26:00 openembedded/classes/tinderclient.bbclass: Jul 23 12:26:00 -Fix spacing and imports now we can test the tinderclient Jul 23 12:27:14 mickeyl: how can I get just one file out of the db? Jul 23 12:27:56 zecke@bitbake:~/org.openembedded.dev$ monotone update Jul 23 12:27:56 monotone: warning: missing classes/tinderclient.bbclass Jul 23 12:28:47 if it's in the db (added *and* committed) montone updte will get it Jul 23 12:29:45 koen: it is already there Jul 23 12:29:59 and it should be updatet Jul 23 12:30:34 sadly, i have no idea. I'm still a monotone newbie as well Jul 23 12:30:53 mickeyl: has any1 been working on the OE for motorola phones? Jul 23 12:31:04 no Jul 23 12:31:13 yes and no Jul 23 12:31:28 actually I've invited the folks who got the kernel source Jul 23 12:31:34 * JustinP would like to point out that there's a very interesting discussion going on in the monotone IRC channel about the merging algorithm Jul 23 12:31:45 looks like they're going to eb changing the way that it works Jul 23 12:31:51 excellent Jul 23 12:32:05 seems like not much is happening.... Jul 23 12:32:12 mickeyl: argh vanille is broken? Jul 23 12:32:23 zecke: only if you broke it again Jul 23 12:32:25 mickeyl: I synced, updatet on one machine Jul 23 12:32:34 * koen points zecke to a local netsync server Jul 23 12:32:35 koen: of course you've missed a bunch of stuff that just happened ;-) Jul 23 12:32:35 mickeyl: now viemtn shows two heads Jul 23 12:32:48 zecke: sounds like you didn't update your copies previously Jul 23 12:32:55 mickeyl: and on ewi I want my changes to tinderclient.bbclass Jul 23 12:32:55 this happens when you're working with two machines Jul 23 12:33:06 I always use pull and push, not sync, btw. Jul 23 12:33:21 monotone sync monotone.vanille.de org.openmebedded.* Jul 23 12:33:26 what is wrong here? Jul 23 12:33:47 zecke: you misspelt openembedded Jul 23 12:34:05 pb_: :( Jul 23 12:34:07 * zecke_sad hides Jul 23 12:34:38 * koen already merging Jul 23 12:34:42 how long is a typical initial monotone pull? Jul 23 12:34:50 very long Jul 23 12:34:50 ive been waiting for a while... Jul 23 12:34:52 cyph_: initial? 4 hours or so Jul 23 12:34:55 use the snapshot instead Jul 23 12:34:58 cyph_: better grab the snap and update it Jul 23 12:35:09 4!! Jul 23 12:35:14 ok good idea Jul 23 12:35:36 GettingStarted doesnt say anything abt it though Jul 23 12:35:50 cyph_: there are a lot of small embedded groups working based on OpenEmbedded but not contributing back. We need to explicitly invite and contact those people if we want to keep the OE tree as the definite source Jul 23 12:36:07 koen: it is not merged here? Jul 23 12:36:15 cyph_: maybe you'd like to update the GettingStarted page to talk about the snapshot Jul 23 12:36:31 pb_: i've to find the snapshot first :> Jul 23 12:36:48 03koen 07org.oe.dev * r8b425961... 10/conf/tinder.conf: Add tinder.conf containing the vars needed for a tinder client Jul 23 12:36:49 http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 Jul 23 12:37:10 koen: might be helpful if you would put a date into it? Jul 23 12:37:27 ah, in the snapshot Jul 23 12:37:46 poitn the browser to http://ewi546.ewi.utwente.nl/OE/ and see ;) Jul 23 12:37:52 koen: what about using documentaton.conf Jul 23 12:37:57 * koen set that up as a test Jul 23 12:38:08 zecke_sad: I was a bit too lazy for that, sorry Jul 23 12:38:29 zecke_sad: could you make bitdoc sort the keys by alfabet? Jul 23 12:38:47 koen: like it is for one week? Jul 23 12:39:11 mickeyl: will see what i can do about the moto phones, but we are not even close to building the whole system from source Jul 23 12:39:23 zecke_sad: ah, my bitdoc must be out of date Jul 23 12:39:51 mickeyl: you want the groups to do something like the nslu2 people? Jul 23 12:39:58 mickeyl: lots of proprietary stuff.. the gsm drivers, SD card support etc Jul 23 12:41:23 koen: yes and no. I'd like people to contribute stuff into the main OE repository if it doesn't conflict with our stuff, i.e.more machine types, more kernels, etc. Branches are great for groups which need more customizing than OVERRIDES, filenames, and stuff Jul 23 12:41:50 mickeyl: does OE have a blanket license? Jul 23 12:42:14 i think so Jul 23 12:42:16 well Jul 23 12:42:19 MIT IIRC Jul 23 12:42:25 MIT Jul 23 12:42:30 which should be perfectly ok for non-commercial and commercial stuff Jul 23 12:42:30 which implies? Jul 23 12:42:40 koen: We can sue them for paten violation Jul 23 12:43:02 koen: but we can't revoke their rights to use it Jul 23 12:43:04 what I absolutely don't like is that people like the motorola guys are hacking opie to support some cool things (i.e. cooperation with the ezx framework) without telling us :/ Jul 23 12:43:16 that's really sad Jul 23 12:43:23 mickeyl: moto is not using opie Jul 23 12:43:36 motorola guys = motorola user Jul 23 12:43:50 ah ic Jul 23 12:43:54 that'll be me Jul 23 12:43:54 yeah i need to be more precise in speech Jul 23 12:43:57 righto :) Jul 23 12:44:32 made opie's qws run with ezx's qws Jul 23 12:44:37 congrats Jul 23 12:44:40 that's pretty cool stuff Jul 23 12:44:46 but that build is really awkward Jul 23 12:44:53 * koen gets slightly excited by running X on a phone (and rendering it pretty useless) Jul 23 12:44:55 part of it depends on leaked header files Jul 23 12:45:01 uh oh. that bad? :D Jul 23 12:45:07 which i believe are illegal trade secrets.... Jul 23 12:45:18 i see. then you were right about not telling us :) Jul 23 12:45:22 pb_: The Getting Started page does talk about the Snapshot. I added that yesterday. Jul 23 12:45:30 ah. heh. Jul 23 12:45:39 I guess cyph_'s browser is malfunctioning. Jul 23 12:46:17 cyph_: you have an E or an A ? Jul 23 12:46:33 i'm considering the A780 with the GPS which is supposed to hit germany in august. Jul 23 12:46:35 mickeyl: i dont get u.... Jul 23 12:46:48 mickeyl: u mean E??? or A??? Jul 23 12:46:52 yep Jul 23 12:47:00 e680 hardware Jul 23 12:47:05 flashed to e680i Jul 23 12:47:36 koen: does ewi deliver mails? Jul 23 12:47:40 nice. how would you rate the possiblity to have an own environment talking to the closed source drivers ? Jul 23 12:48:30 the guys at MFC has been making neat progress... Jul 23 12:48:42 we need a GSM module Jul 23 12:48:49 should be simple AT commands Jul 23 12:49:06 strace shows us AT commands going to /dev/mux2 Jul 23 12:49:08 zecke_sad: ehm, I think you can send mails from there Jul 23 12:49:22 need to figure out how gprs works Jul 23 12:49:37 sounds cool. I'm really thinking about buying one, but I want _one_ device that I actually can continue to use after a lobotomy ;) Jul 23 12:49:38 gprs is just ppp over an uart Jul 23 12:49:51 mickeyl: your shoes Jul 23 12:50:00 koen: not quite.... we are talking abt gprs on the phone itself Jul 23 12:50:28 cyph_: yes, I know Jul 23 12:50:37 gprs is used through a kernel module gprsv.o with net interface gprsv0 Jul 23 12:50:42 hmmm Jul 23 12:50:47 no source code :< Jul 23 12:50:56 ok, it's a little more lowlevel as a cf gprs card Jul 23 12:51:57 cyph_: welcome to the family. we had similar problems with some sharp stuff Jul 23 12:52:25 haha... well part of me is hoping they will be leaked.... just like the ezx headers Jul 23 12:52:26 :> Jul 23 12:52:29 (damn sharp stuff....at least the Spitz has more open hardware) Jul 23 12:52:41 hehe /dev/mux2... Jul 23 12:52:44 while pulling from monotone.vanille.de: Jul 23 12:52:45 monotone: warning: discarding revision cert packet 470f2685dfe5cb3786786067c1edf1282dd1c536 with unmet dependencies Jul 23 12:52:47 sounds like GSM 11.07 Jul 23 12:52:48 monotone: warning: discarding revision cert packet 50481fac31a3c89de7959f87001b670c05ea5e7a with unmet dependencies Jul 23 12:52:58 hey Jul 23 12:52:58 JustinP: vanille disconnected you Jul 23 12:53:11 ok...that's interesting Jul 23 12:53:20 worked this time Jul 23 12:53:21 that's harmless. you will get the revs the next time Jul 23 12:53:25 zecke_sad: we also have orphan certs in our db Jul 23 12:53:48 I'm willing to bet people will have at least 7 serious errors in db check Jul 23 12:53:53 koen: could reconfigure ewi to deliver mail? Jul 23 12:54:07 yay, multiple heads.... Jul 23 12:54:13 zecke_sad: whats that? Jul 23 12:54:16 the monotone server seems a little flakey combined with our half-dead router :/ Jul 23 12:54:26 <-- GSM n00b Jul 23 12:54:29 cyph_: ETSI multiplexer specification Jul 23 12:54:37 koen: almost all errors are from yours and zeckes branch Jul 23 12:54:39 i checked last nite Jul 23 12:54:57 mickeyl: heh my branch had only one commit Jul 23 12:54:59 we killed the revs, but the certs are still there Jul 23 12:55:15 this is why I'm using my own local branch now and propagating org.oe.dev changes....merging multiple heads locally causes issues Jul 23 12:55:34 hmm Jul 23 12:55:37 i'm not sure if this is good Jul 23 12:55:54 we had some revs living in 'unknown' branches Jul 23 12:56:08 zecke_sad: thx... hope that helps.... Jul 23 12:56:20 cyph_: do ATI on /dev/mux0 Jul 23 12:56:21 mickeyl: do those orphaned certs show up in synced/pulled copies? Jul 23 12:56:28 JustinP: yes Jul 23 12:56:33 cyph_: what vendor is that modem from? Jul 23 12:56:36 that's why it's nearly impossible to kill 'em Jul 23 12:56:42 we would get them back all the time Jul 23 12:56:51 we can block a branch, but not orphaned revs Jul 23 12:56:55 certs, even Jul 23 12:56:55 mickeyl: does a db rebuild fix them? Jul 23 12:56:56 cyph_: i have no idea..... and im not opening up my phone :> Jul 23 12:57:00 JustinP: no idea Jul 23 12:57:12 of course a db rebuild re-hashes all of the revs Jul 23 12:57:23 cyph_: ATI will tell you Jul 23 12:57:26 might be worth trying. I'm not sure what happens when a rebuilt db syncs to an old one though Jul 23 12:57:36 zecke: what u mean ATI? Jul 23 12:57:43 RP: you there? Jul 23 12:57:46 I think everyone has to get a new copy.... Jul 23 12:57:56 zecke_sad: how do I force debian to reconfigure exim4 ? Jul 23 12:57:56 cyph_: ATI is a AT command to show you the vendor and version Jul 23 12:58:05 cyph_: send "ATI" to the modem and see what it says Jul 23 12:58:09 oic... Jul 23 12:58:18 dpkg-reconfigure exim4 ? Jul 23 12:58:19 koen: dpkg-reconfigure -p low exim4 Jul 23 12:58:19 zecke_sad: dpkg-reconfigure --force exim4 does nothing Jul 23 12:58:38 still nothing Jul 23 12:58:39 you might need to add a number as well; try ATI1 or ATI2 Jul 23 12:58:59 on some modems, regular ATI just gives you a meaningless response Jul 23 12:59:06 OK Jul 23 12:59:06 ATI Jul 23 12:59:06 960 Jul 23 12:59:09 means anything? Jul 23 12:59:10 for other stuff I would have to check the my data sheets... Jul 23 12:59:13 dpkg-reconfigure exim4-config Jul 23 12:59:14 sadly not Jul 23 12:59:20 no, that's no good Jul 23 12:59:21 try ATI1 Jul 23 12:59:34 micropal: thanks Jul 23 12:59:50 I'm pulling from vanille and getting about a dozen " warning: discarding revision data packet ... with unmet dependencies" Jul 23 12:59:52 ATI1 Jul 23 12:59:52 000 Jul 23 12:59:58 heh, still no good. ATI2? Jul 23 13:00:24 ATI3 Jul 23 13:00:24 OK Jul 23 13:00:24 ATI2 Jul 23 13:00:24 OK Jul 23 13:00:24 ATI3 Jul 23 13:00:27 OK Jul 23 13:00:29 ATI4 Jul 23 13:00:31 L Series Jul 23 13:00:33 OK Jul 23 13:00:35 ATI5 Jul 23 13:00:37 010 Jul 23 13:00:39 OK Jul 23 13:01:00 doh, evidently no help to be had there. Jul 23 13:01:01 hmm not too good as well Jul 23 13:01:44 im wondering if the AT command interpreter is in mux_cli.o or in some co-processor on the board Jul 23 13:03:00 yeah, could be either Jul 23 13:05:49 zecke_sad: I think mail delivery should work now Jul 23 13:05:54 will look more into it.. Jul 23 13:05:57 koen: hmm could you check if I just sent a mail? Jul 23 13:06:01 koen: I got yours Jul 23 13:06:04 zecke_sad: you should be able to use {fetch,get}mail Jul 23 13:06:35 * koen waits for the broken uniserver Jul 23 13:08:26 zecke_sad: I didn't get your testmail yet Jul 23 13:10:01 koen: I didn't send one I guess Jul 23 13:10:11 my tinderclient should send one Jul 23 13:11:45 2005-07-23 22:12:18 1DwQM6-0002iV-0z <= tinderbox@ewi546.ewi.utwente.nl H=localhost (bitbake.utwente.nl) [127.0.0.1] U=zecke P=esmtp S=773 Jul 23 13:11:48 2005-07-23 22:12:18 1DwQM6-0002iR-04 => tinderbox-oe@gmx.net R=smarthost T=remote_smtp_smarthost H=smtp.utwente.nl [192.87.17.19] Jul 23 13:11:51 about twice a minute Jul 23 13:13:44 koen: it looks it didn't like freyther at hh.org as sender Jul 23 13:14:10 kewl now we need to process the mail Jul 23 13:16:59 akita looks very good now. One bug is left however. Screen-tap and key-click sounds are b0rked since Akita was added to libopie Jul 23 13:18:06 do you hear sound at all? Jul 23 13:18:08 is qss started ? Jul 23 13:19:28 mickeyl: normal sound (ie from games) work Jul 23 13:20:06 qss isn't running Jul 23 13:20:10 ah Jul 23 13:20:13 that would be a problem Jul 23 13:20:36 heh Jul 23 13:20:42 * CoreDump|home installs opie-qss Jul 23 13:25:16 * koen gets annoyed at the /opt person at familiar@ Jul 23 13:26:03 mickey|tv: thx, that worked Jul 23 13:26:11 http://article.gmane.org/gmane.comp.handhelds.linux.familiar/28725 <- updating 2500+ bbs, idiot Jul 23 13:26:39 koen: I need help with procmail and fetchmail Jul 23 13:28:20 zecke_sad: see /tmp/f for a fetchmail rc Jul 23 13:28:48 03freyther 07org.oe.dev * rb71bc754... 10/ (classes/tinderclient.bbclass conf/tinder.conf): Jul 23 13:28:48 TinderClient.bbclass: Jul 23 13:28:48 -Add a TINDER_SUBJECT field, mails will carry that subject Jul 23 13:28:48 tinder.conf: Jul 23 13:28:48 -Mention TINDER_TREE and TINDER_SUBJECT. Without TINDER_TREE the Jul 23 13:28:49 tinderbox will not know how to associate your report. Jul 23 13:29:11 koen: you've some serious mental issues (refering to your password) Jul 23 13:29:17 heh Jul 23 13:29:25 that isn't my password ;) Jul 23 13:29:37 koen: I don't know why people use /opt for such things...certain things in /opt are ok, such as, say, the Sun JVM, but making all package sinstall there?? What the hell are they smoking? Jul 23 13:29:47 also note I used donkey instead of ass ;) Jul 23 13:36:55 koen|tv: any idea? Jul 23 13:49:40 is it possible to use distcc with bitbake? Jul 23 13:50:24 yeah, I think jacques did some work on that a while back Jul 23 13:51:35 micropal: better you can use it with icecream Jul 23 13:51:51 micropal: INHERIT += "icecc" ICECC_PATH="/usr/bin/icecc" Jul 23 13:52:03 micropal: your toolchain will get distributed automatically to the slaves Jul 23 13:52:09 ahh nice :) Jul 23 13:52:25 mickey|tv: make.cfg can be reached somehow? Jul 23 13:52:50 from a package? no. Jul 23 13:57:42 now we have an issue.. Jul 23 13:57:57 I used to save the TINDER_STARTED in make.cfg... Jul 23 13:58:40 pb_: where could we store it? Jul 23 13:58:50 zecke: what does TINDER_STARTED do? Jul 23 13:59:41 sounds liek a PID lock Jul 23 13:59:53 pb_: it safes the start date Jul 23 14:00:18 return time.strftime('%m/%d/%Y %H:%M:%S', time.localtime()) Jul 23 14:00:46 CoreDump|home, for your akita builds, DISTRO = "openzaurus-3.5.4"? any cvsdate workarounds? Jul 23 14:01:01 (getting ready to start a build) Jul 23 14:01:33 jh_: I'm using CVSDATE 20050722 Jul 23 14:01:36 zecke_sad: ah, hm, tricky. Jul 23 14:01:40 * chouimat is back. Jul 23 14:01:46 could save it in a file keyed from the PID, I guess. Jul 23 14:02:30 jh_: be sure to use a tree from today, might want to force rebuild of initscripts and opie-taskbar Jul 23 14:04:23 cool, thanks Jul 23 14:05:01 np Jul 23 14:06:18 hmm I could let bitbake add that variable to configuration automatically Jul 23 14:06:57 yeah, that'd work Jul 23 14:07:33 BUILD_START or such... Jul 23 14:08:29 wow, my build is proceeding at a glacial pace Jul 23 14:08:33 * pb_ stabs nfs Jul 23 14:11:52 ah nfs is a good idea... Jul 23 14:17:00 zecke_sad: I'm back again Jul 23 14:19:27 okay Jul 23 14:19:58 zecke_sad: www-data does have a shell Jul 23 14:21:35 why am I getting merge conflict propagating to my own branch...all I've ever done is add files... Jul 23 14:22:43 monotone has some serious issues I think... Jul 23 14:23:23 JustinP: can't judge without seing your error Jul 23 14:24:06 03freyther 07org.oe.dev * r657428f7... 10/ (classes/tinderclient.bbclass conf/tinder.conf): Jul 23 14:24:06 classes/tinderclient.bbclass: Jul 23 14:24:06 -setting and reading make.cfg can not work Jul 23 14:24:06 conf/tinder.conf: Jul 23 14:24:06 -set TINDER_START there for now... Jul 23 14:24:37 zecke_sad: can you cook up a procmailrc? Jul 23 14:25:06 lol Jul 23 14:25:10 should it work? Jul 23 14:25:13 my procmail and regexp powers are to low Jul 23 14:25:44 I have to reboot my FreeBSD machine Jul 23 14:26:00 as it can not mount dirty ext2 (ext3) volumes Jul 23 14:27:43 fetchmailrc is written and .forward points to procmail Jul 23 14:27:59 that should do the trick. *crosses fingers* Jul 23 14:28:25 zecke_sad: it's not an error, it just makes me manually merge a bunch of files. Usually openslug-related. I have to go the the org.oe.dev tree and look up the newest head's file to know which one to merge Jul 23 14:28:39 weird Jul 23 14:28:49 koen: I'm ripping an example procmail apart Jul 23 14:30:57 I get this nearly every time I merge my changes back in. I think it has something to do with using a local copy like this....but this should not be happening Jul 23 14:34:06 koen: /home/zecke/procmailrc Jul 23 14:36:19 fetchmail: Authorization failure on tinderbox-oe@pop.gmx.net Jul 23 14:36:48 tinderbox-oeATgmx.net is the user Jul 23 14:38:27 I can't login to gmx.net via web with the password Jul 23 14:39:07 replace AT with the right sign (my xkb is broken) Jul 23 14:39:17 yeah, I know Jul 23 14:42:35 is anyone able to reproduce the problem with pcre (#33)? Jul 23 14:43:13 bug #33 that is Jul 23 14:43:18 haven't look to close? is it about libtool and .la files? Jul 23 14:44:03 sort of Jul 23 14:44:21 something (maybe libtool) is picking up host library paths Jul 23 14:44:32 I can't reproduce it here Jul 23 14:45:46 i.e. pcre builds fine with or without pcre installed on the host system (koen suggested it might be related to pcre headers being present on the host) Jul 23 14:47:06 you used to have pcre-dev installed to get pcre to work Jul 23 14:50:02 koen: doesn't seem to matter here Jul 23 14:50:43 JustinP: are you able to reproduce bug #33? Jul 23 14:51:35 yes Jul 23 14:51:52 I've been getting it on one system but not another Jul 23 14:52:07 both gentoo and recently updated Jul 23 14:52:16 different programs installed, though Jul 23 14:54:03 both have libpcre 5.0 installed Jul 23 14:54:16 I've gotten similar errors with other packages Jul 23 14:55:05 pb_: in absence of kergoth could you take a look at our procmailrc? Jul 23 14:55:14 sure Jul 23 14:55:45 not that I've used procmail for about a decade, but I'll see what I can do Jul 23 14:55:55 heh Jul 23 14:56:10 http://ewi546.ewi.utwente.nl/.procmailrc Jul 23 14:56:49 is anything in particular the matter with it? Jul 23 14:57:21 yes if Subject is BitBake* mail should get passed to /data/tb1 Jul 23 14:57:34 if Subject is OpenEmbedded mail should get passed to /data/tb2 Jul 23 14:58:10 currently only mails/build_db gets written even with "OpenEmbedded Tinder Log" as subject Jul 23 14:58:29 me looks for miss spelling again Jul 23 14:58:50 koen: I can not read from Jul 23 14:59:08 your browser can Jul 23 14:59:50 try * ^Subject:.*BitBake.* Jul 23 15:00:20 . == any Jul 23 15:00:37 CoreDump|home: that doesn't matter Jul 23 15:00:39 same for OE? Jul 23 15:01:08 ah, the subject really is just "Bitbake"? Jul 23 15:01:16 nm then Jul 23 15:01:27 CoreDump|home: no it would be "BitBake Tinder Log" Jul 23 15:01:45 then try * ^Subject:.*BitBake.* ;) Jul 23 15:02:16 CoreDump|home: again. that doesn't matter. it will match regardles.. Jul 23 15:02:23 regardless even Jul 23 15:02:42 hmm Jul 23 15:02:59 maybe up the loglevel to see what goes wrong Jul 23 15:04:14 http://ewi546.ewi.utwente.nl/mail/from Jul 23 15:04:15 ahhh Jul 23 15:04:24 it is not case sensitive Jul 23 15:04:37 koen: it does not seem to use fetchmail... Jul 23 15:05:33 JustinP: I guess you could inspect what pcre's build system is doing. but it certainly isn't "a general problem either with how bitbake or openembedded handled linking paths" Jul 23 15:06:49 JustinP: pcre is performing a couple of stunts that may very well blow up (I can't think of any particular reason, but who knows when libtool is involved and used in non-standard ways) Jul 23 15:07:46 hmmmmm....of course I have no idea where to look Jul 23 15:08:50 www-data@bitbake:~$ fetchmail -c Jul 23 15:08:50 94 messages for tinderbox-oe@gmx.net at pop.gmx.net (6483464 octets). Jul 23 15:09:22 zecke_sad: ready to start fetching? Jul 23 15:10:04 yes Jul 23 15:10:31 "Zugriff durch POP3, IMAP oder ein anderes Browser-Fenster derzeit gesperrt." Jul 23 15:10:54 procmail: Skipped "|/data/tinderbox2/bin/processmail_bugs" Jul 23 15:10:54 From tinderbox@ewi546.ewi.utwente.nl Sun Jul 24 00:11:06 2005 Jul 23 15:10:54 Subject: Tinder-Client build log Jul 23 15:11:17 yeah that is the old subject Jul 23 15:11:35 I wonder how it matches BUG... there Jul 23 15:13:03 hmm Jul 23 15:13:08 tinderbox blows up Jul 23 15:13:16 /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..Tinder.Client.Test/EWI.1122154932.1122156816.16096: No such file or directory at /data/tinderbox2/lib/Persistence/Storable.pm line 86 Jul 23 15:13:46 ah no Jul 23 15:13:57 I had no Build tree specified... Jul 23 15:14:08 ok Jul 23 15:15:21 the procmailrc seems to be working Jul 23 15:17:44 hmm Jul 23 15:20:09 yo Twiun Jul 23 15:21:39 heh Jul 23 15:21:41 zecke_sad: 21 neue Mails in Ihrem Spamverdacht-Ordner Jul 23 15:21:49 hey reenoo_ Jul 23 15:22:27 koen: the first thing I did when the gmx people introduced their spam filter was to turn it off >:) Jul 23 15:22:40 :) Jul 23 15:24:03 zecke_sad: it keeps giving the same error, are you sure that's ok? Jul 23 15:24:04 damn Jul 23 15:24:07 no Jul 23 15:25:05 koen: could you turn off fetchmail for now? Jul 23 15:25:18 done Jul 23 15:25:50 koen: how can we generate a maildir? instead of mbox? Jul 23 15:26:15 add a trailing / in the procmailrc Jul 23 15:27:01 zecke_sad: I do something like this: http://pastebin.com/319352 Jul 23 15:27:09 which is an ugly hack Jul 23 15:27:22 ah I can imagine the error Jul 23 15:28:35 koen: just $MAILDIR/.spam/ would do. procmail will interpret it as a maildir folder Jul 23 15:29:06 I recall that that didn't work for some reason Jul 23 15:29:10 but I might be wrong Jul 23 15:29:19 'if it aint broke, don't fix it' Jul 23 15:31:47 well. I have >20 of those in my procmailrc and it works Jul 23 15:32:24 the dir may have to exist beforehand though. I don't remeber Jul 23 15:32:32 koen: let us turn fetchmail on again Jul 23 15:32:51 done Jul 23 15:34:24 is the tinder.cgi cron job still on? Jul 23 15:34:32 zecke_sad: did you release the spam out of the spamsuspect folder? Jul 23 15:34:36 zecke_sad: it is still running Jul 23 15:34:51 Jul 24 00:35:01 bitbake /USR/SBIN/CRON[4337]: (www-data) CMD (/data/tinderbox1/cgi-bin/tinder.cgi --daemon-mode) Jul 23 15:35:18 hmm not good Jul 23 15:35:35 [Sun Jul 24 00:33:45 2005] Error in: Persistence::save_structure writing file: /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..Tinder-ClientTest/EWI.1122158006 Jul 23 15:35:38 can't create /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..Tinder-ClientTest/EWI.1122158006.1122158025.16494: No such file or directory at /data/tinderbox2/lib/Persistence/Storable.pm line 86 Jul 23 15:35:53 * koen turns of fetchmail Jul 23 15:36:50 * chouimat is away: dinner Jul 23 15:43:09 koen: let us clean the html_output directory Jul 23 15:43:51 done Jul 23 15:44:22 Test/EWI was the issue '/' is not allowed Jul 23 15:44:33 could it be the '/' in the filename? Jul 23 15:44:36 LOL Jul 23 15:45:31 should I turn fetchmail back on and delete html_output again? Jul 23 15:45:36 monotone on vanille seems down again Jul 23 15:45:53 keturn: ewi546.ewi.utwente.nl Jul 23 15:46:21 data_dir was important but I still was www-data ;) Jul 23 15:46:28 yeah turn on fetchmail again please Jul 23 15:46:39 done Jul 23 15:48:59 Subject: OpenEmbedded Tinder Log Jul 23 15:49:01 Folder: mail/builds_oe 2082 Jul 23 15:49:08 zecke_sad: seems to be working Jul 23 15:49:21 http://ewi546.ewi.utwente.nl/OE_qa/OpenEmbeddedBuild/status.html Jul 23 15:52:14 zecke_sad: cheers Jul 23 15:53:31 http://ewi546.ewi.utwente.nl/OE_qa/OpenEmbeddedBuild/all_notices.html#1122159240 Jul 23 15:53:34 zecke_sad: build: done, blog : WIP Jul 23 15:53:46 same applies here Jul 23 15:53:49 heh Jul 23 15:54:14 koen: I get an access denied error for branch koen. is that to be expected? Jul 23 15:54:21 yes Jul 23 15:54:27 it's not serving that branch Jul 23 15:54:53 zecke_sad: it's missing http://ewi546.ewi.utwente.nl/star1.gif Jul 23 15:55:08 it describes it as a "protocol error", which is unfortunate Jul 23 15:58:19 keturn: what version of monotone do you've? Jul 23 15:58:28 koen: I will fix that soon Jul 23 15:58:30 0.20 Jul 23 15:58:40 koen: oh you did Jul 23 16:01:10 zecke_sad: OE does seem to use svn.berlios.de, but I guess you don't have a monotone tinderbox yet Jul 23 16:05:42 not yet Jul 23 16:05:49 we need a VC_mnt.pm Jul 23 16:06:10 and we need a VCDisplay (viewcvs and viewmtn) Jul 23 16:06:32 we should integrate the bug tracker first Jul 23 16:07:55 zecke_sad: I fixed tinder.conf to have no slashes Jul 23 16:08:13 thanks Jul 23 16:08:19 cia will follow when mickeyl restarts vanille.de Jul 23 16:08:20 treke|home: ping Jul 23 16:08:58 http://ewi546.ewi.utwente.nl/tmp/viewmtn/getdiff.py?id1=b71bc754c5facba821ce7681d2298bd6fbb584d5&id2=7a1a1d1dc2d072ab7829bb7c3314d4604d4445d6&fname=conf/tinder.conf Jul 23 16:09:13 koen: will you mail oe? I'm too tired Jul 23 16:09:20 OK Jul 23 16:09:44 we should not make the tinderbox-oe address public? Jul 23 16:10:10 process-builds is a bit picky and a malformed email can... Jul 23 16:11:20 well, you can't have it both ways. if you want people to be able to run tinderbox clients, they will have to know the address to mail results to. Jul 23 16:11:43 pb_: sure Jul 23 16:11:47 zecke_sad: mail sent Jul 23 16:12:09 pb_: hmm right, I would have demanded people asking us for the data Jul 23 16:12:33 koen: hehe Jul 23 16:12:56 for example mailing tinderbox: starttime: 0 Jul 23 16:13:07 to mozilla could make their tinderbox useless Jul 23 16:13:51 really? iirc, it ignores mails dated in the past Jul 23 16:14:20 Jul 23 16:14:20 [Sun Jul 24 00:46:10 2005] tinder.cgi: Error in updatefile: /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..None.1122150354, starttime: 0, is not a valid time. at /data/tinderbox2/lib/TinderDB/Build.pm line 998. Jul 23 16:14:24 Jul 23 16:14:28 and it didn't update the status ;) Jul 23 16:14:34 so it recognizes the error Jul 23 16:14:48 but stops processing Jul 23 16:15:03 ah, that does suck Jul 23 16:15:11 koen: I only have five serious errors in my mt db Jul 23 16:15:26 lucky you Jul 23 16:15:53 I won't sync with you guys anymore Jul 23 16:16:48 good plan Jul 23 16:18:13 I'm starting a meta-opie for h3900 now as well.. Jul 23 16:18:44 oh, my meta-sdk build is finally building gcc-cross-sdk Jul 23 16:18:50 I wonder if mickey's error will occur for me as well Jul 23 16:19:24 I hope it does ;) Jul 23 16:19:43 heh Jul 23 16:19:49 should I hope that your meta-opie build fails as well? Jul 23 16:20:02 pb_: OT: have you used a non GNU/Linux like distribution recently? Jul 23 16:20:24 pb_: hehe, if you look deep deep deep inside of you... Jul 23 16:20:38 zecke_sad: I'll try to do a world build on dominion tomorrow Jul 23 16:20:56 no seriously I trust mickeyl as a bugreporter and if you see the bug as well Jul 23 16:20:59 it helps to get it fixed Jul 23 16:21:31 zecke: not recently, no Jul 23 16:21:40 * chouimat is back. Jul 23 16:22:00 unless you count Windows 2000, which I use fairly often Jul 23 16:22:05 pb_: It is unbelievable how much faster FreeBSD boots and stops Jul 23 16:22:23 zecke_sad: try zeta Jul 23 16:22:28 and how much less noise is printed... Jul 23 16:22:38 koen: actually I'm tempted to buy it Jul 23 16:22:48 yeah, linux is quite poor in that respect Jul 23 16:23:25 given that my workstation need 2minutes for an ECC check on boot, boottime isn't a big thing for me Jul 23 16:23:36 the boot time has gotten a lot worse with recent distributions. my desktop machine spends ages doing hotplug stuff now that it never used to. Jul 23 16:24:21 I do not know bsd's init system but I think it is faster because it does less than debians' Jul 23 16:24:28 yeah, almost certainly Jul 23 16:24:47 if that's try suse should boot faster too Jul 23 16:24:48 debian's does seem to do a lot of stuff that I suspect could be removed Jul 23 16:25:06 s/try/true/ Jul 23 16:26:05 and freebsd sends sigterm and waits for buffers to be flushed and then quits Jul 23 16:26:07 hehe Jul 23 16:26:32 an hour ago I said "only 20 minutes till I go to bed" Jul 23 16:26:56 drat Jul 23 16:26:57 at 20h I said only 10 minutes ans you will learn the theory for your drivers license.. Jul 23 16:27:05 my gcc-cross-sdk build failed, but I have a different error than mickey Jul 23 16:27:26 Uh... did i break it? Jul 23 16:27:36 * france is back (gone 21:12:07) Jul 23 16:27:43 no, I think it's some environment thing Jul 23 16:27:55 | /nexus/home/pb/oe/build-oz/tmp/work/gcc-cross-sdk-3.3.4-r2/gcc-3.3.4/gcc/f/com.c:11080: error: conflicting types for 'ffecom_gfrt_basictype' Jul 23 16:27:55 | /nexus/home/pb/oe/build-oz/tmp/work/gcc-cross-sdk-3.3.4-r2/gcc-3.3.4/gcc/f/com.h:236: error: previous declaration of 'ffecom_gfrt_basictype' was here Jul 23 16:28:26 it looks like it might be trying to use gcc 4 or something, although I thought I had changed the default to 3.4 in my host system Jul 23 16:28:41 koen: we need an admin Jul 23 16:29:25 what does that do? Jul 23 16:30:04 koen: he sets the message of the day ;) Jul 23 16:30:07 If you don't mind stupid question every now and then you could fill in my name Jul 23 16:30:45 only localhost can create an admin password for now Jul 23 16:33:39 for reference: some lines of my error is @ http://pastebin.ca/18363 Jul 23 16:33:41 03koen 07org.oe.dev * r7a1a1d1d... 10/conf/tinder.conf: remove / and add a warning about it. Any volunteers to add the keys to documentation.conf? Jul 23 16:35:27 * koen decides slipknot is not very pleasant at 1:35 am Jul 23 16:35:54 the Corrs it is Jul 23 16:38:27 alright.. time to read a couple of papers and go to bed Jul 23 16:38:29 'night all Jul 23 16:38:35 'night reenoo_ Jul 23 16:38:51 03jbowler 07org.oe.nslu2-linux * r822ad5fc... 10/packages/coreutils/coreutils_5.1.3.bb: Jul 23 16:38:51 Use of [.coreutils caused problems because the coreutils [ link ended up Jul 23 16:38:51 being called \[ (update-alternatives cannot handle an unescaped [). Fixed Jul 23 16:38:51 by using lbracket.coreutils (based on the actual source name of the .c file Jul 23 16:38:51 used to build this.) Jul 23 16:39:32 jbowler: ping Jul 23 16:40:50 zecke_sad: pong Jul 23 16:41:08 jbowler: do you want to try a test build for NSLU? Jul 23 16:41:22 testing what ;-) Jul 23 16:41:44 jbowler: http://ewi546.ewi.utwente.nl/OE_qa/OpenEmbeddedBuild/status.html Jul 23 16:42:16 well...it looks like the pcre Makefile is using LIBDIR for linking, which is the lib install dir (/usr/lib). Should I try to patch the Makefile after configure runs? Jul 23 16:42:29 zecke_sad: how do I do it? Jul 23 16:43:13 jbowler: http://monotone.vanille.de/viewmtn/getfile.py?id=646b15ef6bd216e5256116d981f25623b19ea029&path=conf/tinder.conf Jul 23 16:43:28 JustinP: isn't there something you can do prior to running configure, i.e. a configure.in or a test in .m4 ? Jul 23 16:43:45 s/a/a patch to/ Jul 23 16:44:37 jbowler: you must be able to deliver emails Jul 23 16:44:50 jbowler: put INHERIT += "tinderclient" in your local.conf Jul 23 16:45:05 Ah, ok, that was the clue I was missing... Jul 23 16:45:12 mickeyl: I suppose so...unfortunately I know nothing about autoconf as yet Jul 23 16:45:23 I believe low level mail delivery works on my build machine, but I need to check. Jul 23 16:45:41 koen: ??? Jul 23 16:45:44 JustinP: me neither, but I think if we want to do it the right way (tm) we need to do it like that Jul 23 16:45:50 koen: how does your TINDER_START look? Jul 23 16:47:18 okay anyway I need to sleep Jul 23 16:47:55 zecke_sad: no idae Jul 23 16:48:07 koen: rm /data/tinderbox2/data_dir/OpenEmbeddedBuild/db/Build.Update..None.1122150354,' Jul 23 16:48:07 zecke_sad: I will try it, but it will take some time as I have to configure this machine for sending mail. Jul 23 16:48:15 koen: and run tinder.cgi please Jul 23 16:48:30 * mickeyl gladly notices some new patches in rmk patch tracker Jul 23 16:48:32 good night Jul 23 16:48:34 the configure.in has a note about cross-compiling.... Jul 23 16:48:35 g'night florian Jul 23 16:48:43 florian: good night Jul 23 16:48:46 'night florian Jul 23 16:49:00 koen: let us finish it tomorrow ;) Jul 23 16:49:05 g'night zecke Jul 23 16:49:15 g'night all Jul 23 16:49:18 'night all Jul 23 16:49:21 http://zecke.blogspot.com/ Jul 23 16:49:31 two denial of services against tinderbox... Jul 23 16:50:04 http://permalink.gmane.org/gmane.comp.handhelds.openembedded/5264 Jul 23 16:56:51 you're right, of course, it's pcre's Makefile.in that's breaking the linking Jul 23 16:57:39 anyone in here know the value that *should* be used for the linking libdir? or how I can use STAGING_LIBDIR in it? Jul 23 17:07:54 I think it's just -L${STAGING_LIBDIR}, and I think it is there by default. Jul 23 17:08:43 (The tools in cross, at least, know where the cross library directory is.) Jul 23 17:18:54 well pcre isn't linking right Jul 23 17:20:19 zecke: pong Jul 23 17:20:58 can someone send me the tinderbox configuration example that I will need to set up the clients for nslu2-linux's Unslung and OpenSlug distros ? Jul 23 17:21:32 Crofton__|laptop: dyoung, NAiL and I all have brand new zd1211-based wireless usb dongles, and we're going to get them working on OpenSlug for nslu2 Jul 23 17:41:58 well I see why this is happening...the log shows -L/usr/lib but I don't know where to fix it Jul 23 17:43:51 hum. Where to get quick help on linux arm kernel hacking? Jul 23 17:45:25 ah...libpcreposix.la Jul 23 17:45:39 rwhitby: zecke gave me this URL: jbowler: http://monotone.vanille.de/viewmtn/getfile.py?id=646b15ef6bd216e5256116d981f25623b19ea029&path=conf/tinder.conf Jul 23 17:45:52 and that URL is just the checked in tinder.conf Jul 23 17:46:05 for some reason, modules are unstripped on this build... is that due to some option I checked, and can it serve any useful purpose wrt ksymoops? Jul 23 17:47:23 or not...argh Jul 23 17:48:38 JustinP: bad autoconf or a bad makefile, assuming this isn't a 'native' build. Jul 23 17:49:34 I'm sure I've seen that problem, or something very like it, before with pcre Jul 23 17:53:32 jbowler: well I've looked through the Makefile.in, configure.in, libpcreposix.la, and I don't see the offending compile line Jul 23 17:54:34 It's automagically generating it - because that would be the default setting. Jul 23 17:55:09 Oh, ok - I thought I'd fixed it. What does your do_compile say? Jul 23 17:56:24 the error: /usr/lib/libc.a: could not read symbols: File format not recognized Jul 23 17:56:51 No, the do_compile in the .bb file Jul 23 17:57:02 oh Jul 23 17:57:04 one sec Jul 23 17:57:13 (The error is because it is using the wrong cc) Jul 23 17:58:11 oe_runmake CC_FOR_BUILD="${BUILD_CC}" CFLAGS_FOR_BUILD="-DLINK_SIZE=2 -I${S}/in Jul 23 17:58:12 clude" LINK_FOR_BUILD="${BUILD_CC}" Jul 23 17:58:29 pcre says it needs a host cc as well, though... Jul 23 17:58:52 Hum, and the failure is compiling dftables - right? Jul 23 17:59:02 (well, linking dftables) Jul 23 17:59:07 in the configure.in it mentions that cross-compiling will need to alter the Makefil or pass the right thing to configure Jul 23 17:59:11 no Jul 23 17:59:16 libpcreposix Jul 23 17:59:35 ccache arm-linux-gcc -march=armv5te -mtune=xscale -shared .libs/pcreposix.o -L/home/papercrane/oe/build/tmp/work/pcre-4.4-r1/image/usr/lib -L/usr/lib -lpcre -march=armv5te -mtune=xscale -Wl,-soname -Wl,libpcreposix.so.0 -o .libs/libpcreposix.so.0.0.0 Jul 23 18:00:33 Curious, let me find that line in my build log... Jul 23 18:00:56 ok Jul 23 18:01:22 thanks for looking...I've been trying to fix this for a few days now Jul 23 18:01:30 ccache armeb-linux-gcc -march=armv5te -mtune=xscale -shared .libs/pcreposix.o -Wl,--rpath -Wl,/home/work-tmp/jbowler/nslu2/openslug/work/pcre-4.4-r1/pcre-4.4/.libs ./.libs/libpcre.so -march=armv5te -mtune=xscale -Wl,-soname -Wl,libpcreposix.so.0 -o .libs/libpcreposix.so.0.0.0 Jul 23 18:01:53 Note that there are no -L's Jul 23 18:02:04 interesting Jul 23 18:03:44 so you don't know what it might be either? Jul 23 18:03:55 could it be the machine or distro? Jul 23 18:04:25 JustinP: libpcreposix's configure script is doing bad things. its searching for libpcre in your /usr, finding it, and using a -L there. Jul 23 18:04:39 JustinP: likely theres a configure option to specify the location of it, or a variable Jul 23 18:05:04 oh is *that* it....I'll look Jul 23 18:05:46 Maybe, but I have libpcre in /usr/lib and it works for me... Jul 23 18:06:22 perhaps it's the version of gcc? Jul 23 18:06:31 JustinP: check the libtool line which generates the gcc line: /bin/sh ./armeb-linux-libtool --mode=link ccache armeb-linux-gcc -march=armv5te -mtune=xscale -I/home/work-tmp/jbowler/nslu2/openslug/staging/armeb-linux/include -fomit-frame-pointer -frename-registers -fweb -funit-at-a-time -Os -D_REENTRANT -I. -I/home/work-tmp/jbowler/nslu2/openslug/work/pcre-4.4-r1/pcre-4.4 -rpath /usr/lib libpcre.la -version-info \ Jul 23 18:06:31 '0:0:0' -o libpcreposix.la pcreposix.lo Jul 23 18:06:31 I notice yours is different... Jul 23 18:07:49 /bin/sh ./arm-linux-libt Jul 23 18:07:49 ool --mode=relink ccache arm-linux-gcc -march=armv5te -mtune=xscale -I/home/papercrane Jul 23 18:07:52 /oe/build/tmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer Jul 23 18:07:55 -frename-registers -O2 -D_REENTRANT -I. -I/home/papercrane/oe/build/tmp/work/pcre-4.4-r Jul 23 18:07:58 1/pcre-4.4 -rpath /usr/lib libpcre.la -version-info 0:0:0 -o libpcreposix.la pcreposix. Jul 23 18:08:01 lo -inst-prefix-dir /home/papercrane/oe/build/tmp/work/pcre-4.4-r1/image/bin/sh ./arm-linux-libt Jul 23 18:08:04 ool --mode=relink ccache arm-linux-gcc -march=armv5te -mtune=xscale -I/home/papercrane Jul 23 18:08:07 /oe/build/tmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer Jul 23 18:08:11 -frename-registers -O2 -D_REENTRANT -I. -I/home/papercrane/oe/build/tmp/work/pcre-4.4-r Jul 23 18:08:13 1/pcre-4.4 -rpath /usr/lib libpcre.la -version-info 0:0:0 -o libpcreposix.la pcreposix. Jul 23 18:08:16 that didn't paste quite right... Jul 23 18:09:15 I'll trye again Jul 23 18:09:19 /bin/sh ./arm-linux-libtool --mode=relink ccache arm-linux-gcc -march=armv5te -mtune=xscale -I/home/papercrane/oe/build/tmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -D_REENTRANT -I. -I/home/papercrane/oe/build/tmp/work/pcre-4.4-r1/pcre-4.4 -rpath /usr/lib libpcre.la -version-info 0:0:0 -o libpcreposix.la pcreposix.lo -inst-prefix-dir /home/papercrane/oe/build/tmp/work/pcre-4.4-r1/i Jul 23 18:09:25 there we go Jul 23 18:10:22 link vs relink... Jul 23 18:10:27 mine says relink Jul 23 18:11:01 And there is an extra -inst-prefix-dir on the end, and that is all. Jul 23 18:12:47 just cleaned and tried again and it's the same Jul 23 18:14:24 http://dioxide.randomvoids.com/~papercrane/pcre Jul 23 18:14:45 my logs Jul 23 18:14:46 There is definately some bad stuff with the pcre config - the dftables build was attempted with the target cc originally. The conf stuff wasn't fixed - kergoth made it compile dftables outside the autoconf Makefile, then I fixed it up so the internal compile would work. Jul 23 18:15:28 hehe. thats my least favorite hack to fix such things. should really patch it up to use the build compiler when appropriate itself :\ Jul 23 18:16:44 ... I think it must have the inverse problem somewhere (use build for target), but I don't see anything obvious. Jul 23 18:17:47 if you look at the configure.in it mentions cross-compiling Jul 23 18:18:50 I assume you've seen that before, though... Jul 23 18:21:07 it says you should be able to override the host cc with a configure option Jul 23 19:26:34 how can I set env vars for the configure step? Jul 23 19:26:44 with inherit autotools Jul 23 20:27:25 this has to be a problem with the libtool script in here... Jul 23 20:28:21 * JustinP wishes he knew what magic sed replace would fix this Jul 23 20:28:36 * JustinP goes off to bug the pcre author Jul 23 22:53:11 * france is away: Away **** ENDING LOGGING AT Sat Jul 23 23:59:56 2005