**** BEGIN LOGGING AT Wed Sep 10 02:59:57 2008 Sep 10 09:37:10 raster: know how you feel... at work we have tons of ancient Fortran which likes to provoke bugs in all the compilers Sep 10 10:35:27 Weiss: oooh this is a brand new compiler.. new chip. new architecture Sep 10 10:35:56 i just happen to have a knack of proviking the bugs in anything i touch.. except stuff i write :) Sep 10 10:53:24 i have created a sample deb package abcd.1.0.0_i386.deb. is there any build environment to cross-compile for Arm?? (because i have installed debian on my free-runner) Sep 10 11:09:30 maddy: take a look at the section "Develop Applications for and on Debian" on http://wiki.debian.org/DebianOnFreeRunner Sep 10 11:36:40 shr.bearstech.com is back and building again Sep 10 11:37:02 nice Sep 10 11:37:10 we could check the shr images? Sep 10 11:37:44 It's churning through the changes since Sep 1, so it might take some time before fresh images are ready Sep 10 11:37:56 (assuming everything still builds) Sep 10 11:38:16 shr image is on 99 of 5136 Sep 10 11:45:54 it doesn't work the shar-unstable image (with makefile) Sep 10 11:46:02 i'll checked yesteday and it fails Sep 10 11:46:12 let's see what happens today :D Sep 10 11:46:18 yeah, I'm testing a fix for that now Sep 10 11:47:27 freesmartphone.org: 03rod 07fso-makefile * r0c3a1c36b18e 10/setup-env: setup-env: Extend BBPATH for shr overlay Sep 10 11:47:35 try that Sep 10 11:47:43 (make update will pick it up) Sep 10 11:48:11 that should get the build started, dunno whether the shr build will finish yet Sep 10 12:00:00 let's see Sep 10 12:00:24 i'm trying it in my company server (they assign me one for oe compilations) :D Sep 10 12:32:13 hi all Sep 10 12:32:39 I have a question: does anyone here know what's the equivalent of xorg.conf in Om2007.2 Sep 10 12:32:51 ? Sep 10 12:32:52 hi all, I need a keyboard for a software... how could I do? Sep 10 12:34:32 anybody here? Sep 10 12:44:17 ok, another question: whenever I try to boot a kernel image from sd card, I get an " invalid magic number" error. Can anybody tell me why that could be? I made sure I had good copies. Sep 10 12:45:03 "Please stay on topic and join #openmoko for non-developer discussion" Sep 10 12:46:24 I AM TRYING to develop, but if I can't get the blasted basics to work, how am I supposed to do any useful work? Sep 10 12:46:33 so throw me a bone instead of criticising me. Sep 10 12:47:12 Q4eq8rsa: the question you last asked on #openmoko is answered the wiki. get your own bone ;-) Sep 10 12:47:45 searching now... Sep 10 12:48:36 first google hit for "openmoko usb host mode" Sep 10 12:48:43 Q4eq8rsa: I'm not criticising you... I'm just reminding the rules. Sep 10 12:48:50 (/topic) Sep 10 12:50:06 Q4eq8rsa: for a start, you didn't say whether you were booting kernel from fat or ext2 or ext3 partition on sd card, you didn't say which kernel you were booting, you didn't give a pastebin of your u-boot environment variables, ... Sep 10 12:52:23 @OdyX: if any of that had solved my problems, I wouldn't be here asking questions Sep 10 12:52:46 @rwhitby: this is irc, not a forum. You could have asked. Sep 10 12:53:11 Q4eq8rsa: I shouldn't have to play 20 questions. you're asking for my time, remember. Sep 10 12:53:22 Q4eq8rsa: no bashing from me, but this is not a user channel ... Sep 10 12:54:15 guess he didn't really want answers ... Sep 10 13:18:21 Ainulindale: did you run some nslu2-linux makefile stuff in /home/shr? Sep 10 13:29:35 mwester: r.e. my questions last night, my wifi issues could be a hardware issue. ticket at https://docs.openmoko.org/trac/ticket/1983 Sep 10 13:29:58 I saw that on the list; that would be unfortunate. :( Sep 10 13:32:51 mickeyl: we should talk about feeds for fso milestone3 Sep 10 13:34:36 rwhitby: i took the liberty and added shr.bearstech.com/fso-unstable Sep 10 13:34:39 is that ok? Sep 10 13:34:46 here's the thing. Sep 10 13:35:08 it's been offline for a week (disk filled, then they mounted nfs, but then there was an inode problem, and they didn't copy the .ssh dir) Sep 10 13:35:27 then today I found someone had splatted the /home/shr directory with some other makefile Sep 10 13:35:35 and I don't have root access on that machine Sep 10 13:35:59 So, I've started a parallel build infrastructure on our nslu2-linux server at OSUOSL. Sep 10 13:36:07 hmm Sep 10 13:36:15 (which we can set up to respond to a freesmartphone.org cname if you wish) Sep 10 13:36:17 i have already built the gold build Sep 10 13:36:31 shr is back online Sep 10 13:36:34 and is building Sep 10 13:36:40 sounds like we need to ship an update to distro-feed configs then Sep 10 13:36:46 (which should be ok imo) Sep 10 13:37:01 so it's working now, but I don't have the same response time for problems that I do with nslu2-linux machines Sep 10 13:37:45 I think for fso4 we should set up a cname and use that - then it can be pointed elsewhere in an emergency Sep 10 13:37:50 agreed Sep 10 13:38:13 fingers crossed there will be no more problems with shr.bearstech.com and no-one will notice Sep 10 13:38:38 but I see it as a temporary location, not a permanent solution Sep 10 13:39:17 they have been gracious with the donation, so I don't want to speak badly of them either. Sep 10 13:39:25 sure thing Sep 10 13:40:22 is the gold build what's in fso-testing now? Sep 10 13:40:28 (no autorev) Sep 10 13:41:18 * rwhitby hopes you haven't made the cardinal mistake of a gold build that isn't done from a stable public configuration tag ;-) Sep 10 13:41:19 no, i still need to adjust the srcrevs Sep 10 13:41:26 doh! Sep 10 13:41:42 the gold build has to happen *after* that! Sep 10 13:41:59 why? i can see all revs Sep 10 13:42:02 Unless you run ClearCase ;) Sep 10 13:42:23 mickeyl: ah, but you're human. Sep 10 13:42:28 heh Sep 10 13:42:37 mwester: I did 10 years of ClearCase :-) Sep 10 13:42:51 Me too, actually. Sep 10 13:44:05 mwester: we need to teach mickeyl about the 3 external builds rule Sep 10 13:44:12 rwhitby, you are still only Human Sep 10 13:44:20 :p Sep 10 13:45:07 Three external builds won't be a problem for FSO; there are folks eagerly waiting to do that. Sep 10 13:45:34 In nslu2-linux, we have this rule: "A binary release can only be done once three external parties have successfully built and booted the image from the golden public CM tag" Sep 10 13:52:25 I consider it an affront to the open source community that someone would release a binary image before the source code configuration is publicly available ... Sep 10 13:53:34 and why 3 ? Sep 10 13:55:28 2 could be a fluke Sep 10 13:55:49 * rwhitby hopes that mickey|bbl knows I'm just teasing him Sep 10 13:58:40 mickey|bbl: ogpsd: gta01 support Sep 10 13:58:44 (for http://wiki.openmoko.org/index.php?title=OpenmokoFramework/Status_Update_4) Sep 10 14:00:15 Anybody done any work with displaying images on the gta-02? Sep 10 14:00:44 Graphics appears to be incredibly slow on my machine Sep 10 14:03:44 mickey|bbl: just saw the OE commits - thanks. Sep 10 14:06:12 * rwhitby remembers http://wiki.openmoko.org/index.php?title=OpenmokoFramework/Status_Update_4 is a wiki page ... Sep 10 14:06:33 oops, someone beat me too it. Sep 10 14:16:54 anyone know if FSO - MS3 has been released? Sep 10 14:17:11 Sup3rkiddo: officially, no. Sep 10 14:17:29 rwhitby, wokay, thanks Sep 10 14:17:33 unofficially, the source configuration tag has been published, and there are files on downloads ... Sep 10 14:18:28 Sup3rkiddo: once http://downloads.openmoko.org/framework/milestone3/NOT_YET_ANNOUNCED disappears I expect it's officially released :-) Sep 10 14:18:56 rwhitby, ah, ok.. Sep 10 14:39:12 hi, anybody knows the scancode of aux button? Sep 10 15:06:36 rwhitby, Sup3rkiddo: also note that milestone3 was officialy tagged 93 minutes ago ;] Sep 10 15:07:13 (from now) Sep 10 15:07:31 torindel, cool, thanks Sep 10 15:07:33 Woohoo! Sep 10 15:08:58 get it when its hot, http://downloads.openmoko.org/framework/milestone3/ Sep 10 15:09:32 that's what i'm doing :) Sep 10 15:09:33 Sup3rkiddo: well, there is still a file UPLOAD_IN_PROGRESS in that directory... Sep 10 15:09:45 File contents: "please have patience..." Sep 10 15:09:48 that will be for the repo Sep 10 15:10:02 because the images seems of proper size Sep 10 15:10:09 i will report ;) Sep 10 15:10:15 Excellent :-) Sep 10 15:10:25 I'm very interested in hearing what the address book is like Sep 10 15:14:07 note to myself: first mount the partition and only them write to it Sep 10 15:28:02 it's starting Sep 10 15:31:25 jOERG_zzZZzz: Talking on irc in your sleep? Sep 10 15:31:50 Sleeptyping... Sep 10 15:32:48 are the screws in GTA02 really Torx T6? my T6 screwdriver doesn't appear to want to open it... Sep 10 15:33:12 also, does the sticker over the left hand screw indicate that bad things will happen to me (warranty wise) if i open it? Sep 10 15:33:26 Both my GTA01 and 02 open up just fine with a Torx T6 here. Sep 10 15:34:03 hmm, i wonder if the problem is that i didn't pay enough for my screwdriver Sep 10 15:34:15 And I suspect that the sticker is indeed to mark if you've "tampered" with it, but Om rather expects that you tamper with it -- I've not heard any issues with warantee just because of screw removal. Sep 10 15:34:48 that depends on his importer ;] Sep 10 15:34:54 i vaguely remember reading somewhere that warranty issues are different when a reseller is involved. i might email Pulster to check first... Sep 10 15:35:07 * mwester examines his cheapo torx screw driver, and tries to remember if it was less than $2 or less than $3 Sep 10 15:35:30 Weiss: I didn't think about the resellers; that's a good idea. Sep 10 15:35:43 * torindel brought his for less than 3$ long time ago Sep 10 15:36:28 pulster is a friendly guy, so will probably agree that it's better to save shipping things back and forth if it's really just a case of reseating the wifi module, and i can show that raised a ticket to get "official" advice first... Sep 10 15:38:40 Weiss: in Germany, computers are supposed to be opened by customer (there's been a court decision on this). Neo is a computer, and "user manual" exactly specifies how to open it to connect debug board. So nor Pulser neither any other reseller might consider warranty void just because of opening he case Sep 10 15:42:25 Weiss: pulster is fine with opening the case Sep 10 15:42:40 I mailed him three weeks ago Sep 10 15:42:58 first impression about fso m3: if you call it, it don't ring. if you call it for the second time it's ring, but you cannot pick up. if you try to phone from you freerunner to another phone, and the other pick up, the call get cancelled Sep 10 15:43:10 DocScrutinizer / bumbl: thanks :) Sep 10 15:43:37 I'm going back to 2008.8 ;) Sep 10 15:46:22 dolf1074: :-( Sep 10 15:46:27 mwester: did you find the screws quite tight the first time? (sorry, tried to make that not sound dodgy) Sep 10 15:47:42 Yes, they are quite snug. Sep 10 15:47:54 One advantage: the power led glows instead of the aux if it's charching Sep 10 15:48:35 dolf1074: The argument about that with Om is what convinced me that Om's marketing team is totally insane. Sep 10 15:49:20 yeah Sep 10 15:49:46 they are just not ready yet for making a totally opensource image Sep 10 15:49:55 It's just a little to much Sep 10 15:50:43 Well, at least we should be happy that they are only doing cell phones; if the Om marketing folks worked for Toyota, we would have autos where the headlights would turn on when the brake pedal was depressed. Sep 10 15:51:05 If they would double the persons that are working in Om and hire a good manager, everything would be better Sep 10 15:51:34 but I don't think they have enough money for that Sep 10 15:51:35 Yep. Instead they drive away some of their best engineers, and have fewer. Sep 10 15:52:17 yes. Sep 10 15:52:33 dolf1074: well ringing worked for me Sep 10 15:52:34 that they are spread over the world is also a disadvantagement Sep 10 15:52:35 SHR build fails on Sep 10 15:52:36 NOTE: Tasks Summary: Attempted 1363 tasks of which 1362 didn't need to be rerun and 1 failed. Sep 10 15:52:38 ERROR: '/home/didier/IT/Périphériques/OpenMoko/Distros/SHR/shr-unstable/openembedded/packages/python/python_2.5.2.bb' failed Sep 10 15:52:39 make: *** [image] Erreur 1 Sep 10 15:53:04 dolf1074: you sure you didnt play with sound profiles? ;] Sep 10 15:53:04 torindel: it rings the second time, but I cannot pick up Sep 10 15:53:08 OdyX: Is there a reference to a .log file? There should be a more specific error message. Sep 10 15:53:23 for me it ring first and second time and i can pickup Sep 10 15:53:48 calling 3rd time too Sep 10 15:53:53 torindel: I see the 'hangup' or 'pickup' dialog, but i'm unable to press pickup Sep 10 15:54:18 maybe I have a bad provider, but it's working on all the other images Sep 10 15:55:38 wurp2|working: http://debian.pastebin.com/m2a891874 Sep 10 15:59:36 Hrm, I wonder if it's somehow related to the LANG on your PC Sep 10 15:59:46 Have you looked at build/temp.linux-x86_64-2.5/libffi/fficonfig.py on line 41? Sep 10 16:01:22 or with the 64bit arch of my PC ? Sep 10 16:03:58 ah ha... Sep 10 16:03:59 Hmm, maybe Sep 10 16:04:16 I have accents in my directory name Sep 10 16:04:20 "Périphériques" Sep 10 16:04:22 => UTF-8 Sep 10 16:26:22 hi Sep 10 16:31:43 hi aleix Sep 10 16:33:52 wurp2|working: hey! how is it going? Sep 10 16:34:22 Same as always :-) Sep 10 16:35:03 How about you? Sep 10 16:35:46 wurp2|working: a bit busy last two weeks Sep 10 16:36:22 wurp2|working: and still waiting for my FR.. cannot help more without it... Sep 10 16:36:46 aleix: Ack, I didn't know you didn't have it yet Sep 10 16:37:51 wurp2|working: you know now... I just was playing with the dev environment, to gain some time Sep 10 16:38:16 aleix: Well, it was very helpful to us :-) Sep 10 16:39:12 wurp2|working: thanks for the compliment :-) Sep 10 16:40:07 wurp2|working: is the omgta01 emulator working? i mean, could i try things there until i get the phone? Sep 10 16:42:26 aleix: Yep Sep 10 16:42:29 I never used it much Sep 10 16:42:37 It was astoundingly slow on my machine Sep 10 16:43:44 wurp2|working: what do you have? Sep 10 16:46:11 gta01 Sep 10 16:46:18 Oh, my home machine Sep 10 16:46:40 yes Sep 10 16:46:51 hmm, it was an old AMD, underclocked to 750 mhz since it didn't have any headers on the clock speed jumper when I got it (!) Sep 10 16:47:16 I have another machine now, but I still haven't tried, especially since I have the phone, now Sep 10 16:47:38 And I haven't found really any time at all to work on sw for the phone for a couple of months :-( Sep 10 16:48:24 :-( Sep 10 16:48:47 ok, i'll see what I can do with the emulator... Sep 10 16:57:12 wurp2|working: i'll try to merge the qemu stuff from MokoMakefile into FSO Makefile and see if we can load SHR images Sep 10 17:15:37 aleix: Cool! Sep 10 18:21:01 Ainulindale, is daria forty two your repository? Sep 10 18:27:32 hi, i'm trying to use qemu to load Openmoko-openmoko-qtopia-image-glibc-ipk-P1-Snapshot-20080721-om-gta01.rootfs.jffs2 Sep 10 18:28:39 the main screen shows, but nothing reacts... there is a Qtopia icon which does nothing, the same for the Star and Phone icon Sep 10 18:31:16 anyone has tried it? Sep 10 18:32:17 wurp2|working, here? Sep 10 18:51:34 quickdev: Sort of Sep 10 18:51:48 wurp2|working, yesterday you talked of the call status bug, ri ght? Sep 10 18:52:12 Hmm, did I? That phrase doesn't ring any bells Sep 10 18:52:24 But then again, I am old, so that may be the problem. Sep 10 18:53:07 wurp2|working, sorry, just that that it was AndreasD Sep 10 18:53:15 *just noticed Sep 10 18:53:25 AndreasD, available? :) Sep 10 18:53:36 Cool, I prefer that to thinking I've gone crazier. Sep 10 18:54:54 hehe ;) Sep 10 18:57:06 quickdev: yes Sep 10 18:57:21 AndreasD, you said you fixed an issue with callStatus handler? Sep 10 18:57:44 yes it should work now Sep 10 18:58:44 AndreasD, you submitted it to daria's svn? I just checked that out and compiled it and no change - what did you fix exactly? :) Sep 10 19:02:01 hey :) Sep 10 19:02:19 can anyone tell me the correct settings for CARCH, CHOST, CFLAGS and CXXFLAGS when cross-compiling for the FR? Sep 10 19:05:54 quickdev: yes, I changed the marshalling from UINT to INT Sep 10 19:07:05 AndreasD, this line: dbus_g_proxy_add_signal (callBus, "CallStatus", G_TYPE_UINT, G_TYPE_STRING, dbus_get_type_g_string_variant_hashtable(), G_TYPE_INVALID); ? Sep 10 19:07:39 yes Sep 10 19:09:06 AndreasD, it's working now. Expected g_printf() to output something but I shouldn't have ;) Sep 10 19:11:55 okay great, actually this was the last thing to make the talking gui pop up and disappear on calls :-) Sep 10 19:15:39 quickdev: I have to go now, if there is any problems just drop me a mail Sep 10 19:15:52 AndreasD, what's your mail? Sep 10 19:16:15 andreas.dalsgaard at gmail dot com Sep 10 19:16:28 AndreasD, ok, thanks, bye Sep 10 19:16:30 see you Sep 10 19:16:33 bye Sep 10 19:18:42 quickdev: Which dialer are you guys talking about? Sep 10 19:19:11 wurp2|working, I'm playing around with a simple efl dialer Sep 10 19:22:23 Oh, so the change was to lfg? Sep 10 19:22:33 mwester: Just wondering - what "impedance mismatch" are you referring to? Sep 10 19:23:46 It doesn't interface well. Sep 10 19:23:49 At many levels. Sep 10 19:24:56 One example: you don't have communications with the chip until very late in the boot process, and the current Linux structure is such that comms with the chip is lost during suspend/shutdown -- so that the 02 has to "coast" on capacitors during the final power-down because it cannot control the chip during the final portion of shutdown. Sep 10 19:27:10 mwester: Do you think a a smart PMU (e.g. a second CPU) could remedy that? Sep 10 19:27:33 We could fix it by restructuring the device initialization in Linux. Sep 10 19:28:30 I'm betting that the various serial busses are brought up very early in Windows CE or Windows Mobile, or that perhaps the driver start/stop order can be more easily controlled there. Sep 10 19:28:44 mwester: So the chip isn't to blame per se, it seems Sep 10 19:28:53 But if we want to keep Linux the way it is, a smart PMU will allow us to code the PMU to suit Linux. Sep 10 19:29:33 No, the chip is fine. Hence I used the term "impedance mismatch" -- there's no problem with either Linux, nor with the PMU -- they just don't fit well together, is all. Sep 10 19:31:27 Ah, gotcha Sep 10 19:31:32 Thanks for clearing it up Sep 10 19:59:19 * Weiss finds that a Nectar card makes an excellent substitute for a guitar plectrum for taking his FR to bits Sep 10 20:10:27 hi! Sep 10 20:28:34 oo, possible progress on my WLAN troubles Sep 10 20:32:47 maybe not :( Sep 10 20:37:35 someone skilled in python/e_dbus/zhone ? Sep 10 20:39:25 write_erase, what's the problem? Sep 10 20:40:25 quickdev, no problem. Just wondering if I could run zhone on my computer, and connect to FR'dbus trough network ? Sep 10 20:40:54 What's the difference between dbus and edbus ? Sep 10 20:41:34 Is there a way to use regular dbus in zhone (for developement purposes)? Sep 10 20:42:22 e_dbus is afaik the interface to the efl things, there is a difference, yes Sep 10 20:42:41 zhone uses regular dbus? Sep 10 20:44:20 nope.... it imports e_dbus.... which I don't have on my computer.. Sep 10 20:44:53 what's "regular dbus"? Sep 10 20:44:54 dbus is dbus Sep 10 20:44:57 it's a protocl Sep 10 20:45:02 and there are bindings Sep 10 20:45:20 like... pydbus, edbus, rubydbus, dbus-glib Sep 10 20:48:21 mickeyl, In fact I planned to run zhone on my PC. I get an error while "import e_dbus". Sep 10 20:48:33 sounds like you need to build python-edbus Sep 10 20:48:48 that's in the E svn Sep 10 20:49:19 The only package I have at the moment is python-dbus. I don't know what edbus is compared to dbus. Sep 10 20:49:56 ecore mainloop integration Sep 10 20:50:05 dbus is inherently async Sep 10 20:50:21 if you use a library that has a mainloop, you need a binding Sep 10 20:50:35 edbus glues dbus to ecore Sep 10 20:54:18 mickeyl, alsa says no soundcards found on my fso image, any suggestions? kernel issue, alsa issue? Sep 10 20:54:46 mickeyl, thx... Sep 10 21:05:10 mickeyl, If I compile ecore + edbus + python-edbus, may I be able to run zhone on my laptop (connected to the phone througe dbus-TCP) ? Sep 10 21:30:05 dbus-TCP is an adventure, good luck :) Sep 10 21:30:13 you need some more python stuff Sep 10 21:30:20 python-evas e.g. Sep 10 21:30:30 since zhone is using the EFL Sep 10 21:40:37 freesmartphone.org: 03stefan 07artwork * rb1b0f6adc59e 10/illume/illume-config/e.src: Higher the shutdown time on power button press. 2 seconds are to short. We go for 5. Sep 10 21:42:19 how large (GB) micro-sd cards can freerunner handle? Sep 10 21:46:13 avtobiff: theoretically 32GB microSDHC but I know of people using 8 Gig cards. Sep 10 21:46:31 Kensan, thanks! Sep 10 21:51:02 is the battery performance still only one day with the current OM (2008.8?)? Sep 10 21:51:07 or has it improved?! Sep 10 21:56:26 avtobiff: http://wiki.openmoko.org/wiki/Neo_FreeRunner_Battery#Notes_about_expected_battery_life Sep 10 21:56:36 avtobiff: there's a LOT of information in the wiki. Sep 10 21:56:54 Kensan, i am digging through it :) Sep 10 21:57:55 Kensan, i am looking at the OM 2008.8 Keyboard Guide atm :) Sep 10 21:58:13 very neat Sep 10 22:32:56 mickeyl: I'm getting a build failure on cairo_1.6.4 for fso-testing Sep 10 22:33:42 oh, nm - good be a local issue Sep 10 22:34:20 yep, false alarm Sep 10 22:35:18 FYI: i had to drawback the release Sep 10 22:35:25 two showstoppers *sigh* Sep 10 22:35:33 andy kernel no resume Sep 10 22:35:42 power button shutdown timeout too small Sep 10 22:35:54 both fixed, rebuilding and testing now/tomorrow Sep 10 22:36:04 using the stable kernel for the time being Sep 10 22:36:06 *sigh* Sep 10 22:36:16 although andy has some important accel fixes Sep 10 22:36:48 mickeyl: updated public configuration tag? Sep 10 22:37:18 the other reason why you want to do that before the release is so I have time to make sure the feeds are up to date before someone tries to opkg upgrade from them Sep 10 22:37:40 i didn't move the milestone3, but yes, we adjusted the SRCREV Sep 10 22:37:53 we have a problem anyways with the feed Sep 10 22:38:04 ok, so building fso-testing now will get what you're building? Sep 10 22:38:09 ya Sep 10 22:38:21 it's pretty likely we will downgrade git revs Sep 10 22:38:21 what's the feed problem? a package gone backwards? Sep 10 22:38:28 since gitr is local Sep 10 22:38:41 so there's no warrante that your build tree has smaller than mine Sep 10 22:38:58 which means opkg could think your packages are newer Sep 10 22:39:04 right - maybe Koen's idea about PR first isn't such a bad one ;-) Sep 10 22:39:14 nah, that's too much manual work Sep 10 22:39:21 we need to accept zecke's SRCPV fix Sep 10 22:39:26 i don't see another way Sep 10 22:39:30 or move away from git Sep 10 22:39:33 and again use svn Sep 10 22:39:36 which I'm very tempted atm. Sep 10 22:40:00 * stefan_schmidt has an eye here Sep 10 22:40:08 hash-based SCM's kind of really suck for development Sep 10 22:40:23 public development, that is Sep 10 22:40:25 mickeyl: I see at least three other people in your team that are not tempted... Sep 10 22:41:01 mickeyl: Well, that's what releases are for. Sep 10 22:41:12 mickeyl: could you give some detail on why you think that way? Sep 10 22:41:19 Not the whole world like to go with SCM versions all the time Sep 10 22:41:42 Kensan: hash-based revisions are not orderable Sep 10 22:41:45 Anyway, I think you should do your slides now, or you will not sleep this night... Sep 10 22:41:56 there's no way to tell whether sodghodsghoauhguoag is newer than aofosdfhgoudgodagh Sep 10 22:41:58 indeed Sep 10 22:42:00 * mickeyl back to his slides Sep 10 22:43:42 mickeyl: well, every hash-id (revision) has an ancestor etc. They might not be "human-readable" as in revision numbers but there's many tools that help imho. Sep 10 22:44:46 Kensan: Problem for git is that you need a local checkout for this to check. Not possible via git remote. For svn this is easy. Sep 10 22:45:29 Kensan: I see mickeys point, still I will not go back to svn just for version numbers. Sep 10 22:46:14 stefan_schmidt: Ah I see your point. Sep 10 22:46:27 stefan_schmidt: or mickeyl's :) Sep 10 22:47:20 Kensan: In a buildsystem like OE you want to know a lot scr revs but normally don't want a local checkout for this. Sep 10 22:50:45 stefan_schmidt: i see. but what's the plus when you can use integers to correlate to revisions besides the chronological order you can easily setup? You will still need to know which revision has what changes etc? Sep 10 22:52:35 Kensan: The problem are package updates. You don't know if foo-gitrHASH is newer the foo-gitr-HASH2 Sep 10 22:53:01 Kensan: Hard to decide if to uprade or not Sep 10 22:53:04 upgrade Sep 10 22:54:20 stefan_schmidt: my point is exactly this: you might know that one package is "newer" than the other, but to decide whether you would want to switch depends on the differences between the two packages. Sep 10 22:54:36 stefan_schmidt: of course for most software one can assume that newer also means improved. Sep 10 22:55:13 Kensan: Well, that's a bit out of scope of a package manager, right? Sep 10 22:55:28 Kensan: Asking the user technical stuff is no solution. Sep 10 22:55:54 Kensan: It's the guy who cares about the packaging who decides this Sep 10 22:56:10 stefan_schmidt: ok there's the misunderstanding on my part :) Sep 10 22:56:22 Kensan: But this guy don't want to set the version correctly. To much manual work, error prone, etc Sep 10 22:56:39 Kensan: No misunderstanding. Valid questions. Sep 10 22:57:29 Kensan: But if we ever like to become a "consumer product" updates must be really simple for the user. Means a lot more work for the technical guys upfront. :) Sep 10 22:58:06 stefan_schmidt: indeed. Less of the "oh, that looks fun, let's try that" :) Sep 10 22:58:46 Kensan: We can still do this in our playgrounds, just not ship it. It's a long way anyway. Sep 10 22:58:49 hi thos Sep 10 22:58:57 thos: How is intel? Sep 10 22:58:59 stefan_schmidt: oh ok, I think now I understood the "real" problem. Sep 10 22:59:27 stefan_schmidt: too tired to really think... Sep 10 22:59:28 thos: You guys still have the o-hand office as workplace? Sep 10 22:59:45 Kensan: Same here. 1am. Need to sleep soon Sep 10 23:00:05 stefan_schmidt, we have new "temporary" offices in the centre of london Sep 10 23:00:44 stefan_schmidt: totally different question: do you know who's coming to Openexpo next week? Sep 10 23:00:54 thos: ah, bigger company -> better office ;) Sep 10 23:01:18 stefan_schmidt, I guess so, soon we get our own new office though, that will be even better :-) Sep 10 23:02:21 Kensan: I only know about Harald, giving a keynote. No one from the active people sadly. I'm on vacation in .au and the rest of the team will fly to .tw Sep 10 23:02:23 thos: heh Sep 10 23:02:49 stefan_schmidt: you going anywhere near Adelaide in .au? Sep 10 23:02:52 stefan_schmidt: oh alright. That's too bad. Sep 10 23:02:54 thos: So you are still working as one group together. No split of into different teams in the open source lab? Sep 10 23:03:43 rwhitby: No. Starting in Sydney, raster visit, over Brisbane to cairns Sep 10 23:03:58 cool - catch you next time perhaps Sep 10 23:04:03 rwhitby: You're in Adelaide? Sep 10 23:04:16 yep Sep 10 23:04:27 rwhitby: heh, hopefully there will be a next time. :) But I guess so. Sep 10 23:04:50 rwhitby: Really looking forward to 4 nice weeks in your country. Sep 10 23:04:56 stefan_schmidt: just quickly revisiting the git-rev issue: I would assume when you put software into your consumers hands, that they have some release numbers attached to them... Sep 10 23:05:08 We have a tight schedule. 12 cities. Sep 10 23:05:19 Scuba diving at the reef, etc. Sep 10 23:05:38 stefan_schmidt, yeah, the whole reason this deal went though is because they wanted to hire the whole team to work on moblin! Sep 10 23:06:19 Kensan: That was what I was refering to in the beginning to mickey. Once we are getting stable we really should make releases and bugfixing releases on top, etc.. Sep 10 23:06:49 Kensan: It's just pretty hard to handle all this different development tress, integration issues, buildsystem, etc. Hard but fun Sep 10 23:07:29 thos: cool, hope all o-hand workers were ok with this. Sep 10 23:07:50 thos: It's sometimes tricky when a small company gets bought by a big one Sep 10 23:08:12 stefan_schmidt, yeah I think it was mostly ok, one or two people a bit suprised Sep 10 23:08:20 we shall see how it pans out of course :-) Sep 10 23:08:42 thos: yeah Sep 10 23:09:27 thos: One question that come into my mind was how poky fit into the ubunto^Wfedora strategy of moblin Sep 10 23:09:54 thos: I know that they tried hard to bring OM to the point giving up on OE... :) Sep 10 23:10:07 stefan_schmidt, that is currently under lots of discussion, but they definitely want to continue Poky Sep 10 23:10:14 stefan_schmidt: yes, there's multiple sides for everything :) Sep 10 23:10:28 stefan_schmidt, the question is how to get all the required features aligned Sep 10 23:10:30 thos: ok, interesting Sep 10 23:10:38 thos: As always. :) Sep 10 23:11:00 stefan_schmidt: well thanks for your explanations Sep 10 23:11:13 Kensan: np, youre welcome Sep 10 23:11:47 stefan_schmidt, so have you finished studies for good yet? Sep 10 23:12:38 thos: At least this semester. Still some to go. :) Sep 10 23:12:57 thos: No is vacation time and I will use it. Sep 10 23:13:01 stefan_schmidt, ah, when do you restart? Sep 10 23:13:12 thos: End of october Sep 10 23:13:22 ooh, that's very late Sep 10 23:13:39 most UK universities will all have started by end of september Sep 10 23:13:55 stefan_schmidt: what are you studying? Sep 10 23:14:27 thos: Guess we start later with the vacation the. :) Sep 10 23:14:35 Kensan: computer science Sep 10 23:14:47 guess so! Sep 10 23:14:54 anyway, time to get some shut eye for me Sep 10 23:14:57 * thos zzz Sep 10 23:15:16 thos: It's actually called "lecture-free time" here. I'm writing almost all my exams in this time. So spending most of it with learning Sep 10 23:15:22 night thos Sep 10 23:21:21 stefan_schmidt: what kind of major if I might ask? Sep 10 23:21:55 stefan_schmidt: I used to study CS myself not so long ago :) Sep 10 23:22:32 Kensan: operating system and networks Sep 10 23:22:49 stefan_schmidt: cool, sounds interesting. Sep 10 23:23:49