**** BEGIN LOGGING AT Mon Apr 19 02:59:56 2010 Apr 19 05:57:22 ~nf Apr 19 06:16:46 TAsn: np Apr 19 06:17:13 TAsn: but i found another interesting scenario Apr 19 06:17:31 feitingen: i use mutt on my home pc via ssh Apr 19 06:18:17 feitingen: and if i didn't, i'd use mutt on my FR. Or probably Gnus. I'm starting to be more and more convinced that Emacs is the most reasonable and convenient general-purpose app for FR. Apr 19 06:19:03 I used it today to edit and read files, start stop scripts, it also has a nice irc client (and one use bitlbee or jabber.el for IM), etc etc. Apr 19 06:19:25 And when finally a usable telephony UI for emacs appears i'll switch completely i guess. Apr 19 06:22:01 TAsn: so imagine i have several calls from people not in my phonebook. Obviously they weren't resolved. Then i add/change contacts to add them. But my call list view is still open. I'm supposed to receive signals to update my view. For that you'd need on every contact list update (probably only those adding/changing numbers) to go through all active queries, re-resolve entries and if it changed, emit a signal. :) Apr 19 06:22:25 mrmoku|away: are we the only fsogsmd users on freerunner? Apr 19 06:23:53 mrmoku|away: if yes, can you please try activating and deactivating GPRS? Because for me pppd becomes zombie after that, and i want to rule out glib (as i'm using debian, hence old version). Apr 19 06:24:44 mrmoku|away: also do i understand it right that since network status signal contains strength info, it should be emitted on every strength change, is it the case for you? Apr 19 06:36:45 Guete Morge Apr 19 06:37:05 * spaetz grabs coffee Apr 19 06:44:17 PaulFertser: no GPRS over here... not yet. Ordered a new SIM with dialplan the other day... Apr 19 06:44:26 PaulFertser: but... IIRC vanous was doing GPRS with fsogsmd Apr 19 06:44:37 spaetz: pass me one... our machine is broken :/ Apr 19 06:44:54 *please* :) Apr 19 06:45:11 hehe, we have a rental machine in the office, as ours is broken too :) Apr 19 06:45:19 * spaetz slurps more coffee and smiles Apr 19 06:45:23 prost Apr 19 06:45:26 * spaetz is cruel Apr 19 06:45:54 * mrmoku hates spaetz ;) Apr 19 06:46:31 * mrmoku thinks about how he could sabotage the next testing release :P Apr 19 06:46:50 * spaetz wonders if he could manage to order a takeaway coffee to be sent to mrmoku over the internets Apr 19 06:46:57 :P Apr 19 06:52:49 mrmoku: gprs itself works if one uses "noauth", it's disabling that bothers me. Apr 19 06:54:46 ok, should we try to get iliwi into OE today? Apr 19 06:54:51 or is it in already? Apr 19 06:54:52 mrmoku: also one uncool observation: i didn't quite find the pattern yet but sometimes no matter how many times i restart fsogsmd it doesn't work. Though strace shows it does initial AT command configuration just fine. Apr 19 06:54:59 * spaetz updates git Apr 19 06:55:35 PaulFertser: :-( that's bad Apr 19 06:55:44 modem/timing issue? Apr 19 06:57:45 PaulFertser: hmm... do you have to set DEBUG in fsogsmd.conf to make it work? Apr 19 06:57:52 spaetz: no idea Apr 19 06:57:54 or does it work for you out of the box Apr 19 06:58:14 mrmoku: no, it works for me out of the box mostly. And it does so reliably after reboot. Apr 19 06:58:29 mrmoku: what about you? Apr 19 06:58:47 PaulFertser: interesting... because for me fsogsmd works _only_ if I set something to debug... like the libgsm0710 thing Apr 19 06:59:01 mrmoku: are you using a "fast" kernel? Apr 19 06:59:08 Because i'm not, my system is slooow :) Apr 19 06:59:16 ahh, iliwi recipe is in OE already. Apr 19 06:59:17 hmm... that might be it Apr 19 06:59:22 will pick for shr-t Apr 19 06:59:22 yes I'm using new kernel Apr 19 06:59:32 spaetz: iliwi is nice :) Apr 19 06:59:55 never used it, I need to pick in itno shr-t and then I can test it :) Apr 19 07:00:06 daniele_athome: I built it locally yesterday... and somehow broke my system :/ Apr 19 07:00:31 no idea what went wrong there... but while upgrading it hang... had to pull the battery... will reinstall freshly later and test (and bump then) Apr 19 07:00:57 daniele_athome: probably will bump without testing... as you tested it... and it build :) Apr 19 07:01:00 +s Apr 19 07:01:21 PaulFertser: will try later if it works fine with old kernel and no DEBUG later Apr 19 07:19:51 mrmoku: btw, since you all are now edj wizards, can someone prepare an E theme where terminal keyboard doesn't have that useless topbar? Apr 19 07:27:10 PaulFertser: me is far from being an edj wizard :/ and keyboard is special anyway Apr 19 07:28:33 mrmoku: do you understand the issue? DocScrutinizer was talking about for a year and it's still not fixed, that's just stupid :( Apr 19 07:41:24 never used it, I need to pick in itno shr-t and then I can test it :) Apr 19 07:41:26 +1 Apr 19 07:41:27 :) Apr 19 07:42:23 Weiss: hm why did you set the initial virtual framebuffer size to 480x640? the old fb code provided 480x1280 - the funny thing is that the old one did not provide panning while the new one does it. Apr 19 07:42:44 Is there a way to reconfigure the buffer for /dev/fb during runtime? Apr 19 07:49:18 graziano: trying to build it now on the buildhost Apr 19 07:50:40 PaulFertser, the scenario you described is already in the TODO :P Apr 19 07:51:05 spaetz: thanks, I LOVE this community :) Apr 19 07:51:44 hehe, I bet many others disagree. But yes, it's a nice community. Apr 19 07:51:49 PaulFertser, that's the whole idea about queries, that you won't need to do any thinking/polling on your own. The query will update and let you know it updated when something changed. Even if you changed a contact name (which resolved one of your numbers), the query should let you know. Apr 19 07:54:00 mrmoku, how much time does it take from cold boot to network registered with .32 and fsogsmd? Apr 19 07:54:18 TAsn: can't test now... my system is broken :/ Apr 19 07:54:26 :) Apr 19 07:54:29 So for now, slower :P Apr 19 07:54:29 something went wrong with upgrading last night Apr 19 07:54:44 well... very fast instead... until the segfault of enlightenment :P Apr 19 07:55:47 if efl were coded with help of the snake it wouldn't segfault :P Apr 19 07:55:55 mrmoku: fsodeviced segfault? Apr 19 07:56:25 mrmoku: it was PE bumped because of SRCPV merge.. so maybe you need to readjust to newer alsa screnarios Apr 19 07:56:33 JaMa: you are too fast for me. I just updated the iliwi.bb when git pulling failed because you had done the same already :) Apr 19 07:56:58 spaetz: you mean added? Apr 19 07:57:20 I had it in my queue since it was announced, but waited for SRCPV merge to save one conflict :) Apr 19 07:58:31 but now I have no access to my PC/neo home, because of crappy eth connector on my motherboard :/ Apr 19 08:00:45 :-( Apr 19 08:01:03 I updated the src rev, he has fixed a bug since you added the recipe. Apr 19 08:01:17 seen that, thanks Apr 19 08:01:35 Also, I am not extremely happy about the implicit assumption that we will have busybox installed with udhcpc, but I don't know any better way wither Apr 19 08:02:12 dependencies would need to be better specified... (I really miss Gentoo USE flag functionality in OE) Apr 19 08:02:58 RDPENDS+= "udhcp | busybox (USE +udhcpc)" Apr 19 08:03:24 yes, but adding RPROVIDES for all busybox links in busybox recipes seems a bit strange too Apr 19 08:04:05 it seems strange, but if busybox provides udhcpc, it should probably provide it.... Apr 19 08:04:20 not sure, just not happy with the overall dependency specifications. Apr 19 08:06:12 spaetz, there's a simple solution... Apr 19 08:06:16 spaetz, drop busybox :P Apr 19 08:07:38 hehe, that's what I had in RDEPENDS first :) Apr 19 08:07:47 normal udhcpc and psmisc Apr 19 08:08:42 JaMa: right, I wanted to add that too before I noticed your comment :) Apr 19 08:08:57 TAsn: if you drop it in SHR-u, shr-t will follow suit :) Apr 19 08:09:11 I am a mere lemming, following the hordes Apr 19 08:09:19 hehe ;) Apr 19 08:09:45 Just kidding, need to think on the implications, but I think we have enough memory/storage/whatever for the switch, maybe I'm wrong though. Apr 19 08:09:51 dropping is not that easy as it seems in OE Apr 19 08:10:06 Because the FR (and future phones) aren't really low-memory/cpu Apr 19 08:10:07 but there were couple of commits making it easier Apr 19 08:10:24 I mean, 600mhz + 128mb ram is as strong as my old computer Apr 19 08:10:42 as for storage, I don't use nand, so I have 8gb, which is as strong as my old computer :P Apr 19 08:10:46 (old laptop that is) Apr 19 08:10:49 TAsn: and our -lite image is still few MB bigger thant gta01 flash :/ Apr 19 08:11:12 TAsn: although *I* am happy with busybox. I seldomly had a reason to complain. Apr 19 08:11:13 with that in mind I removed those 2 and added comment Apr 19 08:11:28 and I don't really care what provides my killall as long as it works Apr 19 08:11:44 if iliwi happens to be our new default wlan manager it would be in -lite Apr 19 08:12:06 spaetz: I hate it everytime it doesn't know some option I'm using normaly on my desktop Apr 19 08:12:09 JaMa, well, it's terribly easy to "maintain" a gta01 image which is different, isn't it? just add different "dependencies" for image in gta01 and different in the rest... Apr 19 08:12:11 spaetz: ie grep -R Apr 19 08:12:34 TAsn: terribly easy? haha Apr 19 08:13:51 TAsn: it's easy to add different packages, for cost of making our task recipes PACKAGE_ARCH = MACHINE_ARCH Apr 19 08:14:11 NOOOO :) Apr 19 08:14:57 TAsn: but fixing in recipe RDEPENDS with machine overrides to make -lite smaller would be terrible mess Apr 19 08:15:44 spaetz: I'm not planing to do that :) Apr 19 08:16:31 spaetz: as no SHR dev is using/testing gta01 images I think it's better to make them as close to gta02 images as possible even for cost of mandatory uSD use there Apr 19 08:17:20 JaMa: no, first X segfaulted... fixed that by rebuilding/reinstalling X+mesa Apr 19 08:17:29 JaMa: now enlightenment segfaults Apr 19 08:17:53 stupid OE git, running make update for the 7th time now.... Apr 19 08:17:56 as while upgrading it hang completely somewhere in the middle I think that my system is just broken... will reinstall Apr 19 08:18:25 mrmoku: ah :/ Apr 19 08:27:29 JaMa, okie :( Apr 19 08:35:42 opkg install iliwi Apr 19 08:35:57 Collected errors: * opkg_install_cmd: Cannot install package iliwi. Apr 19 08:36:06 very enlightening, thank you opkg Apr 19 08:45:14 JaMa: btw. is there a recipe for lindi's om tool? Apr 19 08:46:10 no, I haven't tried it yet :/ Apr 19 08:46:20 k Apr 19 08:48:10 mrmoku: i'm not the only author and it's not only a command line tool, it's a library too :) Apr 19 08:48:25 lindi-: ok :) Apr 19 09:00:11 JaMa: iliwi configure fails here: configure.ac:8: error: possibly undefined macro: AM_PROG_VALAC Apr 19 09:00:29 do I need to update autotools or something? Apr 19 09:04:36 mmh, yes automake 1.10 in shr-t. too old for vala Apr 19 09:22:24 JaMa: any idea why I get the following building a shr-lite-image? Apr 19 09:22:26 * satisfy_dependencies_for: Cannot satisfy the following dependencies for task-shr-minimal-base: * kernel-module-snd-soc-s3c24xx-i2s * kernel-module-snd-soc-s3c24xx * kernel-module-snd-soc-wm8753 * kernel-module-snd-soc-core * kernel-module-snd-pcm * Apr 19 09:23:03 ohh... and I see the following in rootfs log Apr 19 09:23:05 mv: cannot move `/etc/gtk-2.0/gtkrc' to `/etc/gtk-2.0/gtkrc.old': Permission denied Apr 19 09:23:08 :P Apr 19 09:23:55 mrmoku: are you building .32 kernel only in your tmpdir? Apr 19 09:24:27 JaMa: hmmm... don't hope so... as I want this image to have .29 Apr 19 09:25:16 because I whould expect this message with newer .32 where all snd modules are now in kernel (and I don't see it only because I have also older .32 built in current tmpdir) Apr 19 09:26:20 but IIRC I dropped alsa from gta02 task-base few weeks ago.. but I did rebuild from scratch since than and haven't seen it too :/ Apr 19 09:26:54 http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=7fd9751f56bff7506f78388b2dfed57e118c1448 Apr 19 09:27:06 it was because of unwanted pull for alsa-state Apr 19 09:40:18 * spaetz cherry picks that Apr 19 09:41:09 spaetz: then you also need few other commits Apr 19 09:41:35 spaetz: ie this one http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=af847022ffd78df5ec3381d4c2a5fa878af48b84 Apr 19 09:41:38 maybe more Apr 19 09:42:13 ok, I'll revert that then. :) Apr 19 09:44:37 spaetz: mesa bump you cherry-picked is also not that easy... Apr 19 09:45:16 spaetz: you need ie http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=44998328268b1631894bf10b970108f8ba4cd685 Apr 19 09:45:56 spaetz: newer libdrm and all following changes to those 2 Apr 19 09:50:00 sigh, I just bumped that as I wanted to keep our preferred-shr-version synchronized... Apr 19 09:50:15 ok Apr 19 10:12:10 RDPENDS+= "udhcp | busybox (USE +udhcpc)"updating autoconf-native makes the build fail :-( Apr 19 10:12:12 JaMa: I think we better switch to fsogsmd independently of the new kernel Apr 19 10:12:16 oops Apr 19 10:12:25 updating autoconf-native makes the build fail :-( Apr 19 10:13:08 sh: autom4te: command not found Apr 19 10:13:37 while building autoconf-native: well yes, it's part of autoconf, stupid! Apr 19 10:14:27 me is tempted to simply revert all of the commits of today Apr 19 10:16:18 mrmoku: ok, if it suits you better Apr 19 10:16:47 JaMa: well... I think the new kernel is missing essential stuff Apr 19 10:16:56 the vibrator/missing leds for one Apr 19 10:17:05 I had no luck with GPS too... did you try that? Apr 19 10:17:55 spaetz: not sure if reverting that alsa commit was best (PR downgraded and IIRC this was in shr-t for few week, wasn't it?) Apr 19 10:18:14 JaMa: no, it was in shr-t for about 5 minutes Apr 19 10:18:42 I reverted it immediately after I cherry-picked it. Apr 19 10:19:14 mrmoku: IIRC GPS worked for me, but it's few weeks since I tried it Apr 19 10:19:49 spaetz: ah ok, not sure why in log sometimes I see cherry-picks with current time and sometimes with time of original commit (maybe depends if you had to resolve conflict or no) Apr 19 10:20:09 I guess that makes the time difference, yep Apr 19 10:20:26 trying a clean rebuild to see if it works. Apr 19 10:23:30 JaMa: NOTE: preferred version 2.24.0 of glib-2.0-native not available (for item glib-2.0-native) Apr 19 10:23:42 same for gtk+. It's not even in org.OE.dev yet Apr 19 10:23:57 but afaik, you bumped the preferred version to that Apr 19 10:24:42 ahh, I am wrong Apr 19 10:24:50 weird Apr 19 10:25:44 I'm sure it's in oe.dev :) Apr 19 10:26:31 doh, I was checking the wrong branch Apr 19 10:26:34 of course it is Apr 19 10:38:37 it seems a lot of what would be the hard work with mms on the moko has already been done: http://www.mbuni.org Apr 19 10:40:38 feitingen: looks like latest news is quite old though... Apr 19 10:44:55 still, it can be used, and it is enough to exite me Apr 19 10:45:37 :) Apr 19 10:45:39 spaetz: seems like you can merge whole oe.dev after todays commits :) Apr 19 10:45:49 didn't want to take away your excitement :) Apr 19 10:48:07 JaMa: seems like it would have been easier indeed :) Apr 19 10:48:35 mrmoku: the mailing lists are not that old, and i'm looking at the source, so none taken :) Apr 19 10:49:00 who needs mms, when you can have WSOD with colored stripes slowly making the whole screen black. Apr 19 10:49:13 much more psychedelic. Apr 19 10:50:00 JaMa: if it does not compile now, that is what I am going to do. Just merge the whole thing... Apr 19 10:50:18 * spaetz makes note to remove the AUTOREV from the beginning this time. Apr 19 10:53:57 spaetz: and it would be nice to have "snapshot" of shr-u before switching to fsogsmd :) Apr 19 10:54:29 spaetz: and as shr-t has the same fso revs etc.. I guess it's "quite safe" to merge all Apr 19 10:55:18 yeah, but "quite safe" is not enough. I'll merge if compile fails and keep it otherwise... Apr 19 10:55:43 we have git, so we can always take the snapshot later. Apr 19 10:56:00 ~wb Apr 19 10:56:01 thx Apr 19 10:56:13 ~uptime Apr 19 10:56:38 well if you didn't break it with partial Xorg/mesa/binutils/glib/gtk pick.. then I really think that merging all cannot break it more :) Apr 19 10:57:02 hehe Apr 19 10:57:26 I'll know soon. Building from scratch on a fast host now. Apr 19 10:58:15 and I still don't see libdrm/mesa-dri-glsl-native .. so it should fail to build for sure Apr 19 10:59:09 OK, I give up. I'll merge :) Apr 19 11:02:22 merged, rebuilding now. Apr 19 11:02:41 spaetz: failed already or just taken my advise? ;) Apr 19 11:02:53 taken your advice. Apr 19 11:03:01 I'm sure it would have worked :) Apr 19 11:03:39 now you can probably close all "fixed in shr-u" :) Apr 19 11:04:03 but after binutils bump, it would be nice to rebuild from scratch also on buildhost.. Apr 19 11:04:17 JaMa: right, either it is not fixed at all or fixed everywhere :) Apr 19 11:04:33 JaMa: already deleted 'tmp' (after saving the cache dir) Apr 19 11:04:41 and already rebuilding from scratch Apr 19 11:04:51 ah ok.. Apr 19 11:05:08 with our buildhost slowness now I guess it will take couple of days :) Apr 19 11:05:37 :( yep Apr 19 11:29:00 ~nf Apr 19 11:29:01 The #openmoko-cdevel Newsflash Bulletin Board. (continued at ~NF2. For help see ~NF-help) - - - Recommended URLs and channels(chanlogs etc):see ~RL - - - NEWSFLASH [2010-03-28 21:09 UTC]:buildhost got rebooted... new RAM will arrive beginning of next week Apr 19 12:31:22 Running task 4162 of 7519 Apr 19 12:31:31 so far so good on the new shr-t Apr 19 12:41:08 JaMa: I'm trying to package a lib... and have to notice once again that I'm an OE noob :P Apr 19 12:44:00 mrmoku, what lib? :P Apr 19 12:44:23 TAsn: curious? Apr 19 12:44:30 duh ;) Apr 19 12:44:34 pjproject Apr 19 12:44:36 lib-network-manager? :P Apr 19 12:44:40 pjproject? Apr 19 12:44:48 www.pjsip.org Apr 19 12:44:49 :) Apr 19 12:45:28 http://www.pjsip.org/release/0.5.0/pjproject-0.5.0.html Apr 19 12:45:30 yeah : Apr 19 12:45:32 :P Apr 19 12:45:33 Noticed. Apr 19 12:45:39 0.5.0 ? Apr 19 12:46:24 1.0.3 is what I want Apr 19 12:48:34 mrmoku: what's the difference between packaging lib and app from OE view? Apr 19 12:48:59 JaMa: none I guess... my defeciencies account for apps too :P Apr 19 12:49:06 hehe ;) Apr 19 12:49:09 JaMa: it installs to /usr/local Apr 19 12:49:19 and I don't know how to fix that correctly Apr 19 12:49:52 what is the correct class to inherit for a 'normal' configure/make/make install thing? Apr 19 12:49:58 isn't it hardcoded in configure.ac or something like that? Apr 19 12:50:15 autotools(_stage) Apr 19 12:50:45 there is no configure.ac... a aconfigure.ac instead Apr 19 12:50:51 and a configure script Apr 19 12:50:56 which has --prefix Apr 19 12:51:02 so I guess it's not hardcoded Apr 19 12:51:24 mrmoku: OE by default removes configure script and tries to regenerate that from configure.ac Apr 19 12:51:35 but never heard about aconfigure.ac :/ Apr 19 12:52:18 hehe and configure script as a wrapper for aconfigure... guess I just have to do it manually Apr 19 12:53:48 mrmoku: can you try mv aconfigure.ac configure.ac in do_configure_preppend? :) Apr 19 12:54:31 or ln -s aconfigure.ac configure.ac Apr 19 12:56:06 trying Apr 19 13:03:46 JaMa: guess it's some different configure thing... Apr 19 13:03:52 autoheader: warning: missing template: PJMEDIA_HAS_L16_CODEC Apr 19 13:03:55 get lots of those Apr 19 13:05:23 PaulFertser: looks like I need the DEBUG even with old kernel :/ Apr 19 13:07:59 mrmoku: isn't it defined in some m4 file in that archive? sometimes do_configre looks for attached m4 macros in different dir than where they are in upstream archive Apr 19 13:08:53 JaMa: trying with a sed in configure_prepend now Apr 19 13:08:58 mrmoku: at least I've seen few moves of those *.m4 to m4/ dir in some recipes IIRC (and I'm not autotools guru.. :/) Apr 19 13:09:02 to susbstitue /usr/local with /usr Apr 19 13:09:07 +t Apr 19 13:09:43 also try grep -R PJMEDIA_HAS_L16_CODEC . to check that definition Apr 19 13:10:20 and clearly, you're not OE noob, just picked unfriendly upstream this time :) Apr 19 13:10:37 ./aconfigure.ac: AC_DEFINE(PJMEDIA_HAS_L16_CODEC,0) Apr 19 13:10:37 ./aconfigure:#define PJMEDIA_HAS_L16_CODEC 0 Apr 19 13:18:02 JaMa: ok, seding the aconfigure _and_ the Makefile makes it go to /usr :) Apr 19 13:19:18 now to the next one... sflphone.org :) Apr 19 13:20:46 mrmoku: linphone is not enough? :) but sflphone looks also nice :) Apr 19 13:26:44 mrmoku: slfphone isn't easy Apr 19 13:27:12 i tried that already Apr 19 13:28:39 Heinervdm: really? Apr 19 13:29:13 I use that on my laptop as softphone... and it is nicely divided into the SIP/daemon part and the UI Apr 19 13:29:16 pjproject did compile Apr 19 13:29:18 using dbus Apr 19 13:29:44 it's some time ago i tried to compile it Apr 19 13:29:54 version 0.9.7 Apr 19 13:34:19 Heinervdm: did it fail with the gnome ui or with the common part? Apr 19 13:34:31 mrmoku: i don't remeber anymore Apr 19 13:34:53 i just have a recipe for the common part, so i think it failed there Apr 19 13:43:19 mrmoku: it looks like pjproject is the problem Apr 19 13:43:59 Heinervdm: sflphone has a copy included in it's source Apr 19 13:44:05 which version did you build? Apr 19 13:44:30 1.5 from pjsip.org Apr 19 13:45:04 Heinervdm: that probably can't work Apr 19 13:45:09 the one included is 1.0.3 Apr 19 13:45:23 can you mail me the recipe to try please? :) Apr 19 13:45:32 * mrmoku wants that sip stack :D Apr 19 13:47:48 mrmoku: http://heinervdm.dyndns.info/shr-beta-recipes/slfphone/ Apr 19 13:48:02 thanks Apr 19 14:15:05 One bombshell before I leave: lately I got two annoying bugs that I didn't have time to debug: 1. moko freezes completely. 2. gsm dies. Apr 19 14:15:15 If anyone else also seen something similar, please let me know. Apr 19 14:19:00 mrmoku: i got pjproject compile Apr 19 14:19:10 * JaMa has also some strange bug.. when typing on illume keyboard sometimes that hover effect on letter is not drawn small near touched letter, but big in middle of screen Apr 19 14:19:16 but it installes files to /usr/local Apr 19 14:19:28 Heinervdm: then I'm one step ahead :P Apr 19 14:19:42 Heinervdm: I sed aconfigure and Makefile in configure_prepend Apr 19 14:20:21 mrmoku: have you ever tried to debug that? Apr 19 14:20:38 PaulFertser: no... don't even know where to look at :/ Apr 19 14:22:04 mrmoku: can show me your modifications? Apr 19 14:23:38 do_configure_prepend () { sed -i -e 's#ac_default_prefix=/usr/local#ac_default_prefix=/usr#' aconfigure sed -i -e 's#^prefix = /usr/local#prefix = /usr#' Makefile Apr 19 14:23:48 mrmoku: ah, because it always works with DEBUG it's hard to get anything meaningful... Apr 19 14:23:56 mrmoku: nobody reverted my SRCPV patch in last > 12 hours, what about changing SHR makefile to start building from oe.dev? :) Apr 19 14:24:15 } PaulFertser yep :/ Apr 19 14:24:30 mrmoku: today I'll switch my build env to it and stop merging oe.dev to shr/unstable branch Apr 19 14:24:43 JaMa: hehe nice :) Apr 19 14:25:26 JaMa: if your buildhost builds as expected... no problems with me Apr 19 14:25:29 mrmoku: i think it'd be reasonable to use strace and to look at the sources trying to understand how that can happen, probably adding a single line of debug here or there, not an easy task :/ Apr 19 14:25:47 I'll check remaining oe.dev/shr/unstable diff, but I guess that all PV>PV should be OK Apr 19 14:26:13 PaulFertser: very tedious and time consuming task... the more for one not knowing the projects Apr 19 14:26:29 I think it is something which mickey|zzZZzz will have to do... unfortunatelly :/ Apr 19 14:26:47 JaMa: good Apr 19 14:27:00 JaMa: if you feel all is fine... just switch the Makefile Apr 19 14:27:37 current shr/unstable branch will stay in shr-t :) Apr 19 14:30:02 Heinervdm: hmm... no pkgconfig file for pjproject in staging :/ Apr 19 14:30:19 but there is a .pc in the source Apr 19 14:30:47 try inherit pkgconfig Apr 19 14:31:13 ok Apr 19 14:31:59 and i think we need pjproject-native Apr 19 14:39:51 * mrmoku still waiting for PA to finish building :/ Apr 19 14:40:33 why pa? Apr 19 14:40:39 checking for CCGNU2... configure: error: Package requirements (libccgnu2 >= 1.3.1) were not met: Apr 19 14:40:58 Heinervdm: PA is a build dependency of it... you can turn it off runtime, but I found no switch for building Apr 19 15:52:56 mrmoku: commoncpp2 if you haven't found it already Apr 19 16:09:36 -L/usr/lib isn't good when cross-compiling, is it? Apr 19 16:16:20 JaMa: do you have an idea how to remove the -L/usr/bin here?: http://shr.pastebin.com/0kvjRhXH Apr 19 16:19:22 Heinervdm: grep for where did it came from (some .la in ccrtp or in staging?), now leaving to gym.. Apr 19 16:19:48 ok Apr 19 16:37:46 Heinervdm: yep, found it... and added a recipe for libzrtcpp Apr 19 16:38:02 libzrtcpp? Apr 19 16:38:21 there is a ccrtp recipe Apr 19 16:38:22 needed too Apr 19 16:38:30 yep, is a different one... Apr 19 16:38:36 (which is needed too) Apr 19 16:38:48 it configures... but build fails Apr 19 16:38:50 mrmoku: have you fixed ccrtp? Apr 19 16:39:04 work/armv4t-oe-linux-gnueabi/sflphone-common-0.9.7-r0/sflphone-0.9.7/sflphone-common/libs/dbus-c++/tools/.libs/lt-dbusxx-xml2cpp: cannot execute binary file Apr 19 16:39:08 Heinervdm: fixed? Apr 19 16:39:39 mrmoku: it's not working for me, because it's searching libs in /usr/lib Apr 19 16:39:50 ohh... it built for me Apr 19 16:39:58 s/searching/tries to link against/ Apr 19 16:39:59 Heinervdm meant: mrmoku: it's not working for me, because it's tries to link against libs in /usr/lib Apr 19 16:40:43 daniele_athome: not yet :P Apr 19 16:40:50 but will do now :) Apr 19 16:40:58 hi mrmoku Apr 19 16:41:04 i tried yesterday, the focus issue remains Apr 19 16:41:10 ohh :/ Apr 19 16:41:12 :( Apr 19 16:41:26 but the E_CONTAINS is right, there is some problem with ecore_evas Apr 19 16:41:28 I guess Apr 19 16:41:32 ecore_x and so on... Apr 19 16:42:47 mrmoku, what about writing a new wm based on EFL? Apr 19 16:43:02 hehe ouch Apr 19 16:43:03 I mean, there are many useful functions in Ecore/Evas that can be used for X11 windows management Apr 19 16:43:20 and we could take something from E17 directly too Apr 19 16:43:41 daniele_athome: wouldn't it be better to help dh to improve illume2? Apr 19 16:43:52 mrmoku, ok one example: Apr 19 16:45:03 with illume2, when closing a window, the time before the window below is rendered is about half a second Apr 19 16:45:10 not talking about memory footprint Apr 19 16:45:16 sorry i'm against E wm :D Apr 19 16:45:25 it's really a waste of resources IMHO Apr 19 16:45:43 mrmoku, i think window management needs for the FR are really few Apr 19 16:46:08 that's why writing an ad-hoc wm should not be so hard Apr 19 16:48:43 mrmoku: can you send me your libzrtcpp recipe? Apr 19 17:00:42 daniele_athome: i'm all ears. while i don't have time to engage in yet another project, I would have uses for a lightweight window manager Apr 19 17:00:56 right now i don't even want X Apr 19 17:01:00 but that could change eventually Apr 19 17:01:17 as sooner or later devices without hardware keyboard need to be supported Apr 19 17:01:35 mickeyl, I could investigate the chance to write a new efl-based wm Apr 19 17:01:41 give me a few days :) Apr 19 17:01:49 cool Apr 19 17:02:08 and i'll see if it's worth a while Apr 19 17:02:17 (aaarrr bad english sentence) Apr 19 17:02:28 udnerstood Apr 19 17:10:55 Heinervdm: moment Apr 19 17:12:04 Heinervdm: scped it to tests/mrmoku Apr 19 17:13:00 ok, thx Apr 19 17:14:10 daniele_athome: hai troppo tempo :P Apr 19 17:14:39 daniele_athome: but you have my support Apr 19 17:14:47 mrmoku, ahah I'm telling you "give me a few days" but actually I meant "give me a few weeks" :P Apr 19 17:15:13 :) Apr 19 17:17:39 bbiab Apr 19 17:25:35 mrmoku: contacts yet crashes on addition Apr 19 17:44:41 mrmoku: sflphone-common/libs/dbus-c++/tools/.libs/lt-dbusxx-xml2cpp is a ARM binary... Apr 19 17:48:20 freesmartphone.org: 03mickey 07cornucopia * rd2194f2b9ee8 10/fsodeviced/ (6 files in 5 dirs): fsodeviced: add accelerometer plugin for Kionix KXSD9 Apr 19 17:50:32 Heinervdm: ohh Apr 19 17:50:45 Heinervdm: heh, quite logic :P Apr 19 17:51:10 freesmartphone.org: 03mickey 07cornucopia * r84fe5f8db00a 10/fsodeviced/conf/palm_pre/fsodeviced.conf: fsodeviced: conf/palm_pre: enable accelerometer_kxsd9 and backlight_omappanel Apr 19 17:52:00 mrmoku: so we need a native version of DBUSXX Apr 19 18:04:13 mrmoku: dbus-c++-native already exists Apr 19 18:07:02 but it doesn't compile because the host doesn't exists anymore Apr 19 18:54:50 mrmoku: sflphone uses relative paths in his tree to find pjproject instead of using data form pkgconfig Apr 19 18:55:07 mrmoku: i solved the dbus-c++ issue Apr 19 18:57:19 Heinervdm: great (the second part) Apr 19 18:57:26 * mrmoku was busy doing family stuff Apr 19 18:57:50 mrmoku: you can find the current recipe here: http://heinervdm.dyndns.info/shr-beta-recipes/slfphone/ Apr 19 18:58:06 i sent a patch dbus-c++ to oe-devel Apr 19 19:01:50 Heinervdm: just asked in #sflphone Apr 19 19:02:52 copying all those other project into theire sources makes it really hard for us... Apr 19 19:08:05 yeah... but I _think_ they should be interested to cooperate :) Apr 19 19:09:01 :) Apr 19 19:09:33 removing a line with sed was sed -e 'd/,,,/' fiel Apr 19 19:09:51 the other way round? Apr 19 19:09:54 /,,,/d Apr 19 19:10:38 i will try Apr 19 19:11:48 do_configure_prepend is growing and growing ;) Apr 19 19:15:36 hehe Apr 19 19:22:26 arm-oe-linux-gnueabi/bin/ld: cannot find -lpjnath-sfl-arm-oe-linux-gnueabi Apr 19 19:22:42 it's setting it's own name... Apr 19 19:23:01 TAsn: i know that's a stupid question but does opimd allow storage of blobs for photos? ;) And i guess one should at least "standardise" a field to store "ringtone name". Apr 19 19:23:32 PaulFertser, not a stupid question at all. Apr 19 19:23:44 blobs/photos: should. Apr 19 19:23:58 Haven't tested, but it should work. Apr 19 19:24:24 ringtone name: just call it Ringtone (of type "filename" iirc) Apr 19 19:24:30 We'll soon follow ;P Apr 19 19:25:44 Heinervdm: I really have high hopes to be able to use sflphone-common as our base for voip Apr 19 19:25:48 i should patch taht makefile to use pkg-config... Apr 19 19:26:31 and until then the gnome client should be pretty well usable on a small screen Apr 19 19:26:48 mrmoku: about using sip. best place for that should be ophoned Apr 19 19:27:12 Heinervdm: yep, though ophoned's future is quite undecided Apr 19 19:28:23 yeah but FSO is the right place to handle it Apr 19 19:28:41 mrmoku: it compiles Apr 19 19:28:52 ohh great Apr 19 19:30:05 mrmoku: (sip) what about telepathy? Apr 19 19:30:31 PaulFertser, quick API change Apr 19 19:30:47 PaulFertser, it's @contacts instead of @ContactId Apr 19 19:30:49 :) Apr 19 19:31:09 TAsn: np Apr 19 19:31:20 PaulFertser: no idea... at least on my (current, bleeding edge) fedora desktop it does not yet offer SIP telephony Apr 19 19:31:28 /usr/lib/sflphone/sflphoned looks like a executable Apr 19 19:31:36 why is it in $libdir? Apr 19 19:31:38 TAsn: but is it several contacts now? What type will "v" be? Apr 19 19:31:53 v? Apr 19 19:32:05 TAsn: btw, how to embed several phone numbers in a single field? Or should i rather create several fields with the same name? Apr 19 19:32:07 TAsn: a{sv} Apr 19 19:32:33 TAsn: so it'll be s for @contacts and v for result. Apr 19 19:32:44 PaulFertser, in order to add multiple "Phone" fields Apr 19 19:32:45 just Apr 19 19:32:54 Phone = [field1, field2, ...] Apr 19 19:33:02 that's the correct way of doing it. Apr 19 19:33:07 Heinervdm: it is the daemon part... dunno why they stuff it in lib Apr 19 19:33:13 didn't get the last part. Apr 19 19:33:19 Heinervdm: ahh... I know Apr 19 19:33:22 please provide context :P Apr 19 19:33:27 it get's started by dbus-activatio Apr 19 19:33:28 n Apr 19 19:33:33 ah Apr 19 19:33:35 though no need to put it in path Apr 19 19:33:44 last build now Apr 19 19:33:51 TAsn: i'm talking in dbus types semantics. The call returns a{sv} -- an array of {sv} that is string + variant. Apr 19 19:33:53 let's see if all packages are included now Apr 19 19:34:06 s/packages/files Apr 19 19:34:26 freesmartphone.org: 03tom 07framework * ra39929d3829f 10/framework/subsystems/opimd/db_handler.py: Apr 19 19:34:26 freesmartphone.org: opimd: Changed @ContactId to @contacts for future abstraction. Apr 19 19:34:26 freesmartphone.org: Added a bit more infrastructure for future extensions to this temporary method. Apr 19 19:34:46 PaulFertser, isn't a{sv} a hash table? Apr 19 19:34:56 Well anyhow Apr 19 19:35:05 you still get the id Apr 19 19:35:19 not a hash Apr 19 19:35:31 in the future, I'll provide an optional way of getting a hash table Apr 19 19:35:36 TAsn: so Phone can be either a single string or an array of strings, right? Apr 19 19:35:37 with values Apr 19 19:35:42 PaulFertser, yes. Apr 19 19:35:55 TAsn: why can't it be always an array of strings? Apr 19 19:36:00 PaulFertser, same for all fields, of course. Apr 19 19:36:04 PaulFertser, annoying overhead, nothing more. Apr 19 19:36:13 it's just annoying to handle (all this unpacking) Apr 19 19:36:22 and it's usually just 1 Apr 19 19:36:43 As a contact usually has 1 name, 1 photo, 1 surname, 1 cellphone, 1 work cellphone, etc etc Apr 19 19:36:54 I.e the usage of fields is encouraged instead :) Apr 19 19:37:27 TAsn: hm, sounds reasonable, i hope i won't have much difficulty parsing the return value converted to elisp types :) Apr 19 19:37:34 :) Apr 19 19:37:38 In c/python/whatever Apr 19 19:38:00 there's a simple glib (or in python, native because the wrapper already converts...) way of determining type. Apr 19 19:38:38 mrmoku: should we do all those things with sed or should i make a patch for this stuff Apr 19 19:38:56 TAsn: elist has "listp" function that tells if an object a list. Apr 19 19:39:12 Heinervdm: probably a patch would be nicer Apr 19 19:39:14 So should be no problem, just one more hurdle. Apr 19 19:39:23 PaulFertser, btw, sorry for this (and future?) API change, it's just that I preferred to provide you a working solution asap. Apr 19 19:39:47 mrmoku: but for that i need a .pc file for pjprojects, and i can't find it Apr 19 19:39:53 PaulFertser, one more hurdle, less overhead :) Apr 19 19:40:07 Heinervdm, mrmoku: sounds like they need a lesson in autotools :P Apr 19 19:40:12 TAsn: sorry for taking so much to actually make use of it :) I couldn't code today because i was at work, and couldn't implement all i wanted yesterday because i'm only a human and a very novice lisper. Apr 19 19:40:20 (and all the deving around it, like pkg-config) Apr 19 19:40:32 And changing programs in one's head is often so damn easy, so nevermind :) Apr 19 19:40:39 :P Apr 19 19:41:06 mrmoku: found the .pc but that file is wrong Apr 19 19:41:38 Heinervdm: well then let's go the sed way for now... and maybe talk them into fixing it upstream :P Apr 19 19:41:48 yes :) Apr 19 19:41:54 TAsn: i wonder if you can save some time by actually documenting opimd instead of answering all that questions :) Not everybody can easily and fast determine all needed info from the sources you know ;) Apr 19 19:42:20 PaulFertser, I think the same :P Apr 19 19:42:52 PaulFertser, Know what? Remind me *after* the 26/4 (I have a test that day). And I promise I'll make it a top priority. Apr 19 19:43:14 I'll write full documentation for the querying system. Apr 19 19:43:19 TAsn: please no promises, i got all the info i needed already ;) Apr 19 19:43:28 PaulFertser, I wanted to do it anyway. Apr 19 19:44:03 It's just that i'm really bad at reminding, i usually do not remember even my own tasks and i do not use org-mode. Apr 19 19:44:08 I'm just delegating remembering to you :P Apr 19 19:44:21 PaulFertser, that's great. Apr 19 19:44:24 TAsn: ok, i'll add it to my motd Apr 19 19:44:29 PaulFertser, I prefer that you'll forget ;P Apr 19 19:45:51 TAsn: about what will the test be? Apr 19 19:46:05 Probability. Apr 19 19:46:11 In which I will probably fail :P Apr 19 19:46:26 j/k, I won't fail, though I hate that. :| Apr 19 19:47:32 TAsn: do you describe probability in terms of the set theory? Apr 19 19:49:06 To be honest. Apr 19 19:49:13 I don't remember what they did in that course. Apr 19 19:49:27 I took it last semester, though I had no time to take the test Apr 19 19:49:44 so I registered to the second test. Apr 19 19:50:13 TAsn: oh, i see. I mean this definition: http://en.wikipedia.org/wiki/Probability_theory Kolmogorov and other soviet mathematicians were very strong. Apr 19 19:50:15 freesmartphone.org: 03morphis 07utilities * r27a15b624bc5 10/palmpre/hidduin/src/hidduin.c: hidduin: more debugging statements Apr 19 19:50:30 TAsn: were the waves between you and the test? :P Apr 19 19:50:54 mrmoku, actually, it was a dangerous mix of big waves and opimd_redesign IIRC. Apr 19 19:50:59 PaulFertser, I know what you meant. Apr 19 19:51:05 :P Apr 19 19:51:29 PaulFertser, and I truly don't remember anything we did there. Apr 19 19:51:34 It was a very bad semester ;( Apr 19 19:51:48 I was terribly busy and had no time to study Apr 19 19:51:56 especially not study courses I don't like :P Apr 19 19:52:45 http://mashable.com/2010/04/19/google-chrome-ditches-http/ Apr 19 19:52:48 douches ^ Apr 19 19:52:50 (gogle) Apr 19 19:52:52 (google) Apr 19 19:53:00 TAsn: ok, so now you have to study that course to pass it. I hope i won't be distracting you from now on (as i got all opimd-related info i wanted) :) Apr 19 19:53:16 PaulFertser, The test is in exactly a week. Apr 19 19:53:22 good thing we did not yet bump EFL Apr 19 19:53:24 I should really start studying, yeah :P Apr 19 19:55:34 TAsn: i think the probability theory is quite cool. I know nothing bad about it comparing to the other math courses. Apr 19 19:56:14 PaulFertser, I just find it less "pure" Apr 19 19:56:34 I mean, it feels more like physics than math ;P Apr 19 19:56:38 If you get what I mean. Apr 19 19:56:57 I just don't like it :P Apr 19 19:57:08 TAsn: that's probably because you do not study it deep enough. It has very sound grounds in fact. And the most important theorems in it are proven in a pure mathematical sense :) Apr 19 19:57:27 And yes, i think i get what you mean. Apr 19 19:57:55 PaulFertser, that's probably the reason. But I came to it with a lot of prejudice. Apr 19 19:58:05 TAsn: (impurity) but so are your favourite programming languages, come on, let's start coding in Haskell ;) Apr 19 19:58:32 PaulFertser, I love C, and C feels really pure :) Apr 19 19:58:42 On the other hand, c#/java/c++ don't feel pure to me. Apr 19 19:58:46 python just feels fun :P Apr 19 19:58:54 python is a game :P Apr 19 19:59:00 TAsn: ROTFL, in C almost every meaningful statement has a side effect. Apr 19 19:59:16 like? Apr 19 19:59:27 python is fun Apr 19 19:59:30 spaetz, :) Apr 19 19:59:35 TAsn: like changing variables or any state of anything else. Apr 19 19:59:45 changing variables? Apr 19 19:59:54 TAsn: yep, pure languages never change variables. Apr 19 19:59:57 TAsn: you will be happy to learn that shr-t was rebased to todays shr-u Apr 19 20:00:13 jama convinced me. So lots of bumps.... Apr 19 20:00:25 TAsn: it looks like you had no exposure to functional programming yet. So much fun ahead! Apr 19 20:00:26 spaetz, But there is a MAJOR issue Apr 19 20:00:35 there is always one Apr 19 20:00:43 what is it this time? :) Apr 19 20:00:44 PaulFertser, I don't get what you mean by changing variables. :P Apr 19 20:00:47 only one? Apr 19 20:00:52 hehe Apr 19 20:00:58 spaetz, j/k. Wanted to freak you out, but I guess you became fearless with time. Apr 19 20:01:19 TAsn: you call it fearless, I call it resignated :) Apr 19 20:01:26 TAsn: i guess i couldn't it explain any better, so sorry, here's a link: http://en.wikipedia.org/wiki/Purely_functional Apr 19 20:01:27 :) Apr 19 20:01:37 PaulFertser, thanks. Apr 19 20:01:38 three months testing maintainer... and all your fear is gone :P Apr 19 20:02:06 TAsn: if you like serious math, you'll like functional programming. Apr 19 20:02:07 mrmoku: no, all the fears have come true! Apr 19 20:02:13 hehe Apr 19 20:02:19 elisp! Apr 19 20:02:30 * spaetz shuts up now and drinks a beer Apr 19 20:03:09 PaulFertser, still reading + doing something else, sec :P Apr 19 20:04:35 spaetz: elisp is in fact a bad example Apr 19 20:05:11 not functional? Apr 19 20:05:37 spaetz: old weird dialect of lisp. And unfortunately it often makes people use imperative style. Apr 19 20:05:48 spaetz: it's functional but it's just a bad example. Apr 19 20:05:53 it makes me puke. but that's a different thing Apr 19 20:06:00 LOL Apr 19 20:06:22 spaetz: i spend the last 3 days thinking almost solely about related topics and i enjoyed it. Apr 19 20:06:38 it leads people to write undocumented code and that is bad in any language paradigm Apr 19 20:06:58 at least the lisp code I see is mostly undocumented Apr 19 20:07:33 spaetz: in all elisp documentation it's always suggested to add documentation strings to specific places and to document the code. Apr 19 20:07:53 Also, good code requires little documentation :) Apr 19 20:09:55 Funny thing, i waited for so long for someone at my uni to teach me Haskell and i feel like i finally grown up enough to do that myself and probably to teach someone else after that :) Apr 19 20:10:30 PaulFertser, purely_functional: nice. Apr 19 20:10:42 Nothing I encountered before, but nice :P Apr 19 20:13:00 right, that will eventually need integrating. Apr 19 20:13:03 oops Apr 19 20:23:34 freesmartphone.org: 03morphis 07utilities * r78bb6e8c0a10 10/palmpre/hidduin/src/hidduin.c: hidduin: next try ... Apr 19 20:26:55 freesmartphone.org: 03morphis 07utilities * r996eaa253df3 10/palmpre/hidduin/src/hidduin.c: hidduin: make it compile ... Apr 19 20:27:49 are there US people here? I need housing in Seattle for 3 weeks in May, Where would you recommend that I look. craigslist? what else? Apr 19 20:28:49 TAsn: fyi if you plan to learn some functional programming after your test: (from my uneducated point of view) if you learn lisp (or its dialect scheme), you'll find it beautiful because it has almost no syntax at all and yet is very powerful and expressive. And if you learn Haskell you'll be amazed by its complexity and power and purity and style. Apr 19 20:31:44 PaulFertser, list sounds nicer, but I'll look at both (when I have time, which will hopefully be soon.) Apr 19 20:31:56 deal Apr 19 20:32:41 :{ Apr 19 20:32:45 bah, meant. Apr 19 20:32:46 :> Apr 19 20:34:00 Night, folks, my bed is waiting for me so elisp will have to wait for tomorrow. Have a nice evening :) Apr 19 20:37:07 night. Apr 19 20:39:08 night PaulFertser Apr 19 20:40:39 ohh... bed waiting... yeees, mine too :) Apr 19 20:40:41 gnight all Apr 19 20:42:12 bed? ohh me tooo! night Apr 19 21:40:31 SHR: 03Martin.Jansa 07shr-makefile * rc7780dd0f5c1 10/Makefile: Makefile: use org.openembedded.org branch for shr-unstable Apr 19 21:42:50 opimd guys: you should really port opimd to berkeley dbs :) Apr 19 21:42:58 SHR: 03Martin.Jansa 07shr-makefile * r8943c22bf0dc 10/Makefile: Makefile: switch bitbake branch to 1.10 Apr 19 21:43:00 they are really fast Apr 19 21:45:05 as long as opimd is part of python frameworkd, I don't think the chce of db matters :-) Apr 19 21:45:09 SHR: 03Martin.Jansa 07shr-makefile * r2d1878ac3012 10/Makefile: Makefile: fix shr-unstable branch name Apr 19 21:46:42 daniele_athome, actually, sqlite is terribly fast Apr 19 21:46:48 The bottleneck is dbus. Apr 19 21:47:02 daniele_athome, are you using opimd in your app stack? Apr 19 21:47:08 *phone stack Apr 19 21:47:40 TAsn, no dbus is fast Apr 19 21:47:51 i'm doing right now some tests Apr 19 21:47:56 with bdb and dbus Apr 19 21:48:04 it's very fast Apr 19 21:48:12 TAsn, no i'm not using opimd, yet. Apr 19 21:48:46 hearing that "yet" made me happy, I was afraid you are going to implement another thing on your own (except for the phone stack and wm that is :P ) Apr 19 21:48:47 TAsn, anyway writing in a sqlite table, 7 columns, the majority are numbers, took almost 1sec Apr 19 21:49:12 daniele_athome, then you are doing something terribly wrong, sqlite is wicked fast here. Apr 19 21:49:31 omdex` Apr 19 21:49:32 daniele_athome, is there a recipe for berkley db? Apr 19 21:49:33 err Apr 19 21:49:34 index? Apr 19 21:49:39 TAsn, yes Apr 19 21:50:03 well it supports "sqlite SQL" so it should be just fine. Apr 19 21:50:05 mickeyl, index is the primary key (the builtin integer used as auto increment value) Apr 19 21:50:29 mickeyl, opimd is indexed, but I don't think he said opimd is slow, just praised bdb over sqlite Apr 19 21:50:33 you should add additional indices Apr 19 21:50:40 k Apr 19 21:50:57 mickeyl, why? they will slow down inserts Apr 19 21:50:57 mickeyl, actually, opimd is quite alright. Apr 19 21:51:06 and selects in my case is one-shot for all records Apr 19 21:51:18 mickeyl, getting 5000 calls + resolving contacts for all the numbers took here ~3 seconds. Apr 19 21:51:36 Including dbus overhead. Apr 19 21:51:47 that's speedy Apr 19 21:51:48 TAsn, so there is something I don't understand/do wrong :) Apr 19 21:52:04 daniele_athome: yes, but think about which options occur often and which not Apr 19 21:52:08 mickeyl, yeah, that's "good enough", and don't forget, it's python. Apr 19 21:52:16 i'd say in the typical scenario it's lookups which should be fast Apr 19 21:52:25 daniele_athome, I doubt it, as you are doing something really simple. Apr 19 21:52:36 daniele_athome, indexing is also important in your case. Apr 19 21:52:49 TAsn, indexing is for optimizing reads, not writes Apr 19 21:52:53 isn't it? Apr 19 21:53:01 unrelated to the issue you described Apr 19 21:53:04 I'm referring to Apr 19 21:53:19 and selects in my case is one-shot for all records Apr 19 21:53:29 yes Apr 19 21:53:33 anyhow, adding sorting indexes Apr 19 21:53:37 that already sort your data Apr 19 21:53:48 on insertion Apr 19 21:53:56 will be good for you as well. Apr 19 21:54:11 TAsn, of course, but still the overhead is on writing Apr 19 21:54:18 reading is pretty fast Apr 19 21:54:21 Of course. Apr 19 21:54:46 i wait almost 1 second between call end and call resume popup to be shown Apr 19 21:54:54 (time of sqlite write) Apr 19 21:55:26 daniele_athome, this won't change reading speed, will speed up things you would have probably done otherwise Apr 19 21:55:32 as for your write issue. Apr 19 21:55:47 is it possible that you are forcing an immediate flush Apr 19 21:55:53 i.e "no buffering" Apr 19 21:55:53 ? Apr 19 21:56:06 still, that's a lot for I/O, but just maybe, dunno. Apr 19 21:56:26 TAsn, actually I activated no options in sqlite Apr 19 21:56:30 normal open Apr 19 21:56:58 I wonder what's the default then. Apr 19 21:57:16 Anyhow, this sounds like something went wrong, rather than sqlite general performance. Apr 19 21:57:20 I really find it hard to believe... Apr 19 21:57:28 But maybe it's true, no idea. Apr 19 21:57:39 I do *a lot* of indexed writes Apr 19 21:57:50 *inserts Apr 19 21:57:54 in opimd Apr 19 21:58:34 <[Rui]> hey Apr 19 21:59:16 TAsn, in opimd code text_factory is set to OptimizedUnicode Apr 19 21:59:20 i don't set that Apr 19 21:59:27 <[Rui]> horizontal scrolling in contacts is s l oooooo o o wwww.... (> 200 contacts) Apr 19 21:59:27 could that be a difference? Apr 19 22:00:44 daniele_athome, you said you are inserting numbers, didn't you? Apr 19 22:01:06 Anyhow, if you are it doesn't matter, if you do, it shouldn't cause such a time penalty. Apr 19 22:01:14 <[Rui]> ah... seems to be the picture I added for my wife. for some reason it's causing some horizontal scrolling, slowly coming up from the left side. Apr 19 22:01:14 TAsn, one text field and a few numbers Apr 19 22:01:25 [Rui], That's mostly because of glamo I assume. Apr 19 22:01:37 [Rui], I don't think I quite get the last statement. Apr 19 22:01:42 * daniele_athome is already porting mokosuite to qvga :P Apr 19 22:01:43 <[Rui]> TAsn: I removed the picture and it was normal again... Apr 19 22:01:44 daniele_athome, shouldn't matter. Apr 19 22:01:58 [Rui], Oh, what's the natural resolution of the picture? (just wondering) Apr 19 22:02:14 <[Rui]> TAsn: I added a picture to my wife's contact, yesterday. so now I noticed that it's vewy slow! Apr 19 22:02:17 Does it happen "just before" and "just after" you get to your wife? Apr 19 22:02:25 (in the list) Apr 19 22:02:32 TAsn, anyway... nevermind for this... these bdb functions are really easy to use Apr 19 22:02:45 or all the time? (i.e including when she's not nearly visible, or when she's visible) Apr 19 22:02:50 daniele_athome, :P Apr 19 22:02:52 <[Rui]> TAsn: probaby 640x480 Apr 19 22:02:53 i find myself quite comfortable with it, so I guess I will continue to use it Apr 19 22:03:01 [Rui], that's probably it. Apr 19 22:03:02 i don't need sql queries anyway Apr 19 22:03:09 [Rui], scaling is expensive Apr 19 22:03:21 well at least not cheap Apr 19 22:03:43 my job is waiting for me!! had better to go to sleep Apr 19 22:03:54 [Rui], got any answers to my questions? :P Apr 19 22:04:01 thank you TAsn mickeyl Apr 19 22:04:04 night Apr 19 22:04:10 daniele_athome, night. Apr 19 22:04:59 <[Rui]> TAsn: suggestion: when specified an image, make two copies, scaled down. One for bigger display, other for list display. Apr 19 22:05:07 <[Rui]> TAsn: most people won't ever think of that. Apr 19 22:06:51 <[Rui]> TAsn: what questions? Apr 19 22:07:01 [Rui], I really think that's e's job. Apr 19 22:07:07 TAsn> Does it happen "just before" and "just after" you get to your wife? Apr 19 22:07:07 (in the list) Apr 19 22:07:11 or all the time? (i.e including when she's not nearly visible, or when she's visible) Apr 19 22:07:36 <[Rui]> TAsn: oh, sorry I didn't notice them... Apr 19 22:07:39 <[Rui]> all the time Apr 19 22:07:57 <[Rui]> as soon as I removed it, I don't even see any horizontal scrolling Apr 19 22:08:13 freesmartphone.org: 03mickey 07vala-dbus-binding-tool * rf86268e071d8 10/ (configure.ac src/vala-dbus-binding-tool.vala): Apr 19 22:08:13 freesmartphone.org: enable generating synchronized versions of dbus interfaces in addition Apr 19 22:08:13 freesmartphone.org: synchronized versions will have the name Sync, which Apr 19 22:08:13 freesmartphone.org: allows for synchronized calls, which may come in handy at times. Apr 19 22:08:14 freesmartphone.org: If you do not want to have that, pass the argument --no-synced Apr 19 22:08:34 [Rui], all the time? i.e even in the other end of the list? Apr 19 22:08:36 <[Rui]> TAsn: well, I use plenty images, albeit smaller ones in source, in elmdentica Apr 19 22:08:38 <[Rui]> TAsn: yes Apr 19 22:08:47 ok, thanks. Apr 19 22:09:08 Sounds like an e issue to me. Apr 19 22:09:24 It's e's job to scale down and keep a scaled down copy, and not worry about it all the time Apr 19 22:09:35 Or at least, e should provide a solution. Apr 19 22:10:16 <[Rui]> TAsn: ah, nice Apr 19 22:11:56 I bet they didn't do it because they don't want to waste memory, but they should at least let us force them to do it in some cases, or decide a scaling factor for which they do it. Apr 19 22:12:41 It's unreasonable to maintain (well it is possible, just not fun) Apr 19 22:12:58 Damn "Pretty Woman" is *ALWAYS* on tv. Apr 19 22:13:06 They really like this movie here, I think. Apr 19 22:13:07 <[Rui]> TAsn: there too? :) Apr 19 22:13:32 [Rui], it's the same with your cable company?! Apr 19 22:14:23 <[Rui]> TAsn: I don't really have dedicated channels here, but Pretty Woman does go on quite frequently Apr 19 22:14:57 What the hell is wrong with this world?! Catching LOTR/Matrix/Monty Python/ is almost impossible, but all the Julia Roberts movies air all the time. Apr 19 22:15:10 [Rui], I see. :P Apr 19 22:15:29 [Rui], btw, asked in e, but it seems that that channel is dead, we should wait for raster, I guess. Apr 19 22:15:40 <[Rui]> TAsn: and it's not as if LOTR doesn't have enough of drollicious women :) Apr 19 22:16:02 <[Rui]> TAsn: raster is the man for those things Apr 19 22:16:48 [Rui], if by those you meant everything, so yeah, he is :P Apr 19 22:17:03 Maybe with the small exception of illume2, where devilhorns is the man :P Apr 19 22:17:13 actually *THE* man. Apr 19 22:17:15 :) Apr 19 22:28:25 [Rui], ok, must be a julia roberts "marathon", just looked in the guide and the previous two movies were "Mystic Pizza" and "Notting Hill", and the next movie is "Everyone Says I Love You". Apr 19 22:28:41 <[Rui]> lol! Apr 19 22:29:28 <[Rui]> well, I'm starting to cache CURL* useragents in elmdentica. Fetching timelines and posting seems to work again, but I borked user subscription :) Apr 19 22:29:51 [Rui], are you sure you broke it? Maybe they just changed ayi. Apr 19 22:29:52 api* Apr 19 22:31:48 <[Rui]> easy to check, just need to see if what's on the phone works or not :) Apr 19 22:33:33 <[Rui]> TAsn: did you read that they might have changed recently, somewhere? what's on the phone also crashed Apr 19 22:33:47 <[Rui]> and I was testing on my yet non-commited/packaged code Apr 19 22:33:56 [Rui], no. Apr 19 22:34:04 [Rui], I don't do "Social networking" Apr 19 22:34:09 I don't even have a facebook account :P Apr 19 22:34:19 <[Rui]> TAsn: face-what? :) Apr 19 22:34:20 I just assumed, as this is usually the case with those kind of websites. Apr 19 22:34:32 <[Rui]> I started with Twitter, but for me it's mostly identi.ca now, it's where all the cool guys are :) Apr 19 22:34:53 If they don't offer a dedicated API they will probably break whatever API you are using. Apr 19 22:35:01 * [Rui] . o O ( well, those that do social networking, I mean ) Apr 19 22:35:21 As they make money/plan to make money out of ads Apr 19 22:35:31 So they really rely on entries. Apr 19 22:35:53 *visits Apr 19 22:36:36 <[Rui]> well, no, it's working.... Apr 19 22:36:45 <[Rui]> so it must've been my changes... Apr 19 22:38:00 <[Rui]> I used to create new CURL* everytime (of course that's not efficient, but previously I had only one and when I moved to multiple-account support... I never tackled this issue again... Apr 19 22:42:12 <[Rui]> damn, ssl connection also broken :( Apr 19 22:42:53 <[Rui]> sleep.... tomorrow's another day Apr 19 22:42:56 <[Rui]> ta! Apr 20 02:28:40 alphaone: (dfu && ! & conundrum) congratulations, i think you got it right ! ;-) Apr 20 02:36:27 DocScrutinizer: ping **** ENDING LOGGING AT Tue Apr 20 03:00:03 2010