**** BEGIN LOGGING AT Wed Dec 22 02:59:57 2010 Dec 22 07:53:44 mickeyl: ping Dec 22 07:57:30 moin Dec 22 08:02:46 moin JaMa Dec 22 08:06:15 is it worth to update dbus to 1.4.1? Fix for CVE-2010-4352: sending messages with excessively-nested variants can crash the bus. Dec 22 08:08:05 no idea... if it comes without regressions :P Dec 22 08:11:12 JaMa: there is the g_ether init script which is quite gta02 specific, but installed on n900 too Dec 22 08:11:33 we probably should make it arch=gta02 Dec 22 08:11:44 dunno if it applies to gta01 too though Dec 22 08:15:10 mrmoku: we can move it to machine specs as palmpre has mountdevtmpfs.sh Dec 22 08:15:36 * JaMa building dbus-1.4.1 for test Dec 22 08:18:10 JaMa: yeah, even better Dec 22 08:30:19 morphis: maybe you know that too... for n900 we have C parts that use g_debug for logging... any idea how to get that to fsogsmd.log? Dec 22 08:31:04 g_debug uses stderr? Dec 22 08:31:25 to do that you need the logger class from libfsobasics Dec 22 08:31:41 then you can let stderr point to fsogsmd.log Dec 22 08:31:51 but why? Dec 22 08:31:51 some logger callback to use with g_log_set_default_handler would be nice :) Dec 22 08:32:23 because I want to see what the C parts taken from ofono have to say :) Dec 22 08:33:58 ah ok Dec 22 08:34:53 then the best would be exchange all g_debug against some function which uses a log function supplied from fsogsmd Dec 22 08:37:30 mrmoku: http://library.gnome.org/devel/glib/unstable/glib-Message-Logging.html#g-log-set-handler Dec 22 08:37:45 09:30 < mrmoku> some logger callback to use with g_log_set_default_handler would be nice :) Dec 22 08:37:52 that's what I want to do Dec 22 08:37:54 ah Dec 22 08:38:03 don't saw that line :) Dec 22 08:38:08 just I don't know how theLogger from fso translates into C :) Dec 22 08:39:22 you can use it from C, look at fsobasics.h Dec 22 08:39:32 then you can even supply the logger to libisi Dec 22 08:39:56 so I am off to university Dec 22 08:39:57 bye Dec 22 08:40:07 ahh Dec 22 08:40:29 fso_framework_logger_debug :) Dec 22 09:44:47 mickey|office: morning Dec 22 09:45:47 mickey|office: what is the most elegant way to get g_debug output from the C parts to fsogsmd.log? Dec 22 09:46:05 there is fso_framework_logger_debug in fsobasics Dec 22 09:46:30 but that lacks (in vala non existing IIRC) optional arguments Dec 22 09:47:07 I wonder if it would make sense to have something in libfsobasics that could be used as glib log callback for those C parts Dec 22 09:47:28 or should I do that locally for now? Dec 22 09:47:42 I need it in the lowlevel_nokia900 plugin for now Dec 22 10:27:48 GNUtoo|laptop: i'm convinced it's only the netlink stuff missing in nokia-gpio Dec 22 10:27:59 GNUtoo|laptop: and good morning :) Dec 22 10:28:01 ah? Dec 22 10:28:03 hi Dec 22 10:33:34 GNUtoo|laptop: porting nokia-gpio from ofono you dropped that Dec 22 10:33:37 or it was not there yet, dunno Dec 22 10:33:45 I don't remember Dec 22 10:33:56 I wonder if we should rewrite that thing in vala Dec 22 10:34:06 if you want to you can Dec 22 10:34:20 ok, will take a look then how difficult it is Dec 22 10:36:54 mrmoku: i agree, we should allow this from C. i need to think about a system similar to VA_ARGS Dec 22 10:37:18 mickey|office: ok :) Dec 22 10:37:18 mickey|office: i don Dec 22 10:37:22 ups Dec 22 10:38:01 mickey|office: not urgent though... I will try to rewrite the thing in vala now Dec 22 10:38:33 mickey|office: i don't want to switch to gdbus in shr-testing but there are some fso bugs in the current version, so would it be possible to make a legacy branch? Dec 22 10:39:18 * HeinervdmWork lunch Dec 22 10:39:32 mickeyl,the on/off button in shr-settings doesn't work anymore with fsogsmd, no dbus calls are made, strange isn't it Dec 22 10:39:41 I meant gsm on/off Dec 22 10:39:45 HeinervdmWork: yes, a legacy branch makes sense Dec 22 10:40:35 mickey|office: that Dec 22 10:40:37 s nice Dec 22 11:16:03 how come intone has a pink icon in SHR? I can't find any patches in OE about that Dec 22 11:21:26 pabs3: it's our icon theme Dec 22 11:54:59 mickey|office: hmm... is vala that nifty that it expands an enum value automatically to text when logging? Dec 22 12:00:44 whats a good app to look at for learning about vala & EFL? Dec 22 12:30:33 mrmoku: these days, yes Dec 22 12:38:33 mickey|office: nice :) Dec 22 12:43:25 pabs3: there aren Dec 22 12:43:40 t much apps with vala & efl Dec 22 12:44:03 ffalarms is the only one i know Dec 22 12:44:30 * HeinervdmWork hates using us keyboard with de layout Dec 22 12:44:59 mickey|office: how does something like this translate to vala: unsigned have_gpio_switch:1; ? Dec 22 12:45:12 with have_gpio_switch being a private class member Dec 22 12:45:38 HeinervdmWork: ffphonelog too Dec 22 12:46:29 JaMa: yes, but i never used it :) Dec 22 12:47:54 iliwi seems to be another one Dec 22 12:53:08 mickey|office: private bool restartRequested { get; set; default = false; } Dec 22 12:53:20 works without get and set too I assume :) Dec 22 13:07:05 #define GPIO_WRITE(line, value) \ (self.have_ ## line ? gpio_write(#line, value) : 0) Dec 22 13:07:08 is interesting too :P Dec 22 13:53:09 mrmoku: unsigned have_gpio_switch is a boolean Dec 22 13:53:24 restartRequested does not necessarily need a property Dec 22 13:53:50 the macro translates perfectly into a function Dec 22 14:02:33 mickey|office: yeah, made that a boolean Dec 22 14:03:27 mickey|office: what I'm trying to do instead of that macro is have an array of bools with an enum as index Dec 22 14:03:44 as the enum nicely translates into a string when constructing the filename Dec 22 14:07:39 right Dec 22 14:27:51 mickey|office, mickeyl: ping Dec 22 14:32:26 at your service Dec 22 14:37:27 mickeyl: we should talk in next year about combining our different software knowledge system (trac, mediawiki, gitweb) Dec 22 14:37:52 to have a better overview an better start to deal with all kind of information with not switching between all systems Dec 22 14:38:14 s/an better/and better/ Dec 22 14:38:15 morphis meant: to have a better overview and better start to deal with all kind of information with not switching between all systems Dec 22 14:38:39 in detail: switching with our wiki to trac or take a new system like redmine Dec 22 14:38:52 morphis: yes. ideally, we should also move git.fso Dec 22 14:38:59 jepp Dec 22 14:38:59 git.fso is atm. hosted at university of Braunschweig Dec 22 14:39:02 redmine has git support Dec 22 14:39:04 ok Dec 22 14:39:07 and the rest of FSO? Dec 22 14:39:12 but with the Braunschweigers no longer working for/with me, this is not optimal Dec 22 14:39:12 amethyst? Dec 22 14:39:35 hmm, let me think... no, everything is hosted at the same server atm. Dec 22 14:39:40 I will have some hours next year free next year Dec 22 14:39:44 so if that one has a problem, we have a problem Dec 22 14:39:50 sounds good. if you could do some admin duties Dec 22 14:39:54 this is always so time consuming... Dec 22 14:39:55 jepp Dec 22 14:40:05 I know ... have running my own server Dec 22 14:40:23 but we need a better starting point for people who are interested in FSO Dec 22 14:40:40 will write down my ideas on a mail to the FSO ml after christmas Dec 22 14:40:48 so we can talk about this more in detail Dec 22 14:41:18 great Dec 22 14:42:54 hmmm Dec 22 14:43:00 yo GNUtoo|laptop Dec 22 14:43:04 hi mickeyl Dec 22 14:43:36 I've unknown symbols.... Dec 22 14:43:42 with a clean kenrel build Dec 22 14:43:44 strange Dec 22 14:43:45 isn't it Dec 22 14:43:51 everything clean Dec 22 14:43:55 modprobe g_nokia Dec 22 14:44:20 say unknow symbol usb_gadget_probe_driver Dec 22 14:44:29 mrmoku, do you have g_nokia working? Dec 22 14:45:23 yup Dec 22 14:45:32 GNUtoo|laptop: ahh... modprobe twl4030_usb before Dec 22 14:45:53 wanted to tell you that, but you already left last night :-) Dec 22 14:46:08 took me quite some time to find out Dec 22 14:46:16 and no idea why it isn't loaded automatically Dec 22 14:49:11 ok Dec 22 14:49:13 thanks a lot Dec 22 14:49:18 yw Dec 22 14:53:32 btw bad news, your sys nodes were in fact /sys/class/net/phonet0 like I suspected Dec 22 14:53:59 sure, I thought that was clear Dec 22 14:54:08 mickey|office: just another idea, you know the german podcast Chaosradio Express? Dec 22 14:54:39 I thought about asking to make a podcast about FSO/SHR Dec 22 14:56:41 mickey|office: just another idea, you know the german podcast Chaosradio Express? Dec 22 14:56:43 I thought about asking to make a podcast about FSO/SH Dec 22 14:57:45 yes, i know chaosradio Dec 22 14:58:27 would like to write tim a mail about this idea Dec 22 14:58:38 sounds good Dec 22 14:59:52 maybe we could do that at cebit next year Dec 22 15:05:34 tim should be at 27c3, too Dec 22 15:08:18 playya_: jepp but I think mickey is not Dec 22 15:08:25 mickey|office: or you are? Dec 22 15:08:50 afaik he has no ticket. Dec 22 15:09:36 ok than it should be impossible to get one ... Dec 22 15:09:41 anyway Dec 22 15:09:47 probably not this year. i don't have anything to show, it collides with my birthday (as every year), and the ticketing is pretty braindead this year Dec 22 15:09:54 :) Dec 22 15:10:01 we should do that next year Dec 22 15:10:14 we should do many things next year :) Dec 22 15:10:51 mickey|office: btw. do you looked at the flattr thing? Dec 22 15:15:24 btw. because we have a lot of binary protocols these days, I'm writing a BinBuilder class Dec 22 15:15:35 playya_: oh yeah! Dec 22 15:18:11 where do you place it, in libfsobasics? Dec 22 15:18:32 as I would really like to use something like this in msmcommd for building the binary data of a frame Dec 22 15:21:09 yes Dec 22 15:21:45 btw. i think msmcommd should read some data before writing Dec 22 15:22:40 TIL read 5120 bytes and receives a lot of sync and config packets before sending data to the modem Dec 22 15:22:47 what is the vala/fso way for setsockopt ? Dec 22 15:23:06 these packets seem to confuse msmcommd Dec 22 15:23:29 mrmoku, what socket type do you use? Dec 22 15:23:35 AF_PHONET Dec 22 15:23:40 well... no Dec 22 15:23:47 it's a Netlink.Socket Dec 22 15:24:15 that's a simple posix socket? Dec 22 15:24:43 Netlink.Socket from libnl.vapi Dec 22 15:24:47 netlink???? Dec 22 15:24:48 dunno what that translates into Dec 22 15:24:50 probably yes Dec 22 15:24:55 we use a different netlink Dec 22 15:25:04 ? Dec 22 15:25:05 it's not the one from libnl Dec 22 15:25:10 for the isi modem Dec 22 15:25:22 brb. spaghetti Dec 22 15:25:45 fd = socket(PF_NETLINK, SOCK_DGRAM, NETLINK_ROUTE) Dec 22 15:25:46 hmm Dec 22 15:26:07 ? Dec 22 15:26:20 ouch oh noes Dec 22 15:26:25 I lost ssh Dec 22 15:26:31 and no touchscreen Dec 22 15:26:32 etc... Dec 22 15:26:36 noooo not again Dec 22 15:26:40 I just added netlink Dec 22 15:26:46 s/netlink/bridge Dec 22 15:28:56 mrmoku: hmm, i'm not sure whether i have bound setsockopt yet Dec 22 15:29:24 mickey|office: if GNUtoo|laptop is right and it's not a libnl netlink socket... Dec 22 15:29:33 where do I find posix.vapi? Can't find it nowhere Dec 22 15:29:53 $prefix/share/vala/vapi-/posix.vapi Dec 22 15:30:03 err Dec 22 15:30:09 basically it's just a name collision between ofono and standard stuff Dec 22 15:30:10 $prefix/share/vala-/vapi/posix.vapi Dec 22 15:30:21 another name colision: libisi is a proprietary nokia lib Dec 22 15:30:31 ahh... I just looked int vala not vala- Dec 22 15:31:04 mrmoku: we can add it to vapi/PosixExtra until I have submitted it upstream Dec 22 15:31:07 * GNUtoo|laptop wanted to try bridge-utils Dec 22 15:31:14 bridge-utils? Dec 22 15:31:35 yes for binding phonet0 and upnlink0 Dec 22 15:31:36 mickey|office: ok Dec 22 15:31:40 GNUtoo|laptop: oh yeah! Dec 22 15:31:50 yes but I've modules issues Dec 22 15:32:03 It's really not my day Dec 22 15:32:14 basically I added BRIDGE in the kernel Dec 22 15:32:20 and now no : Dec 22 15:32:23 *touchscreen Dec 22 15:32:27 *usbnet Dec 22 15:32:30 playya_: jepp you are right there is some optimization needed Dec 22 15:32:32 so I can't do anything Dec 22 15:32:38 I can't control the phone Dec 22 15:32:59 and of course I added the right modules Dec 22 15:34:50 sigh Dec 22 15:34:57 no framebuffer console => no console without X Dec 22 15:35:19 playya_: I will investigate some more time in msmcommd after christmas Dec 22 15:35:26 aha usbnet worked Dec 22 15:36:46 mickey|office: [CCode (cheader_filename = "sys/socket.h")] Dec 22 15:36:52 public int setsockopt (int sfd, int level, int option, const void* option_value, socklen_t option_len); Dec 22 15:36:55 does that look right? Dec 22 15:37:38 remove the const Dec 22 15:37:42 ahh, ok Dec 22 15:37:44 and you need to define the socklen_t upfront Dec 22 15:37:53 socklen_t is not defined anywhere iirc Dec 22 15:37:55 it is already there in posixextra :) Dec 22 15:37:59 really? Dec 22 15:38:00 d'oh Dec 22 15:38:04 excellent Dec 22 15:38:11 while you're there... Dec 22 15:38:14 fancy adding getsockopt? Dec 22 15:38:14 :) Dec 22 15:38:24 hehe, ok Dec 22 15:38:42 thanks Dec 22 15:38:58 just make the socklen_t an out param for the get case Dec 22 15:40:28 ok Dec 22 15:40:52 public int getsockopt (int sfd, int level, int option, void* option_value, out socklen_t option_len); Dec 22 15:41:07 looks good Dec 22 15:41:28 can't add phonet0 to bridge br0: Invalid argument Dec 22 15:41:29 since we don't use misc/vapi atm. we should copy it to fsodeviced for now Dec 22 15:42:13 GNUtoo|laptop: btw, there is something strange ongoing with lowlevel_nokia900 as is Dec 22 15:42:19 let me paste the log Dec 22 15:42:22 ok Dec 22 15:43:20 http://shr.pastebin.com/RbJA9FM6 Dec 22 15:43:28 that is repeating for quite some time Dec 22 15:43:30 phonet up Dec 22 15:43:31 phonet down Dec 22 15:43:55 yes it's strange Dec 22 15:44:00 it worked fine before Dec 22 15:56:02 hmm... there is more stuff missing from posix Dec 22 15:57:06 what else do you need? Dec 22 15:57:16 note that the non-posix additions are in linux.vapi Dec 22 15:57:49 SOL_NETLINK Dec 22 15:57:58 but that is not even in sockets.h Dec 22 15:58:40 and SIZE_NLMSG Dec 22 16:00:25 GNUtoo|laptop: are you sure it's not a netlink socket like in libnl? Dec 22 16:00:31 SOL_NETLINK is defined there Dec 22 16:00:49 mrmoku, doubble check Dec 22 16:00:54 but it seem a bit different Dec 22 16:01:11 at least it builds without an external libnl Dec 22 16:01:24 maybe it's a modified netlink .c and .h? Dec 22 16:02:15 gisi/netlink.c:#ifndef SOL_NETLINK Dec 22 16:02:15 gisi/netlink.c:#define SOL_NETLINK 270 /* libc!? */ Dec 22 16:03:01 include/netlink-local.h:#define SOL_NETLINK 270 Dec 22 16:03:04 (from libnl Dec 22 16:03:07 ) Dec 22 16:03:44 ofono might not use libnl... but do the same with plain posix socket Dec 22 16:03:54 but as we have libnl :-) Dec 22 16:04:00 yeah Dec 22 16:04:17 better use libnl2 for that, if it's clear _what_ they are doing Dec 22 16:05:17 fd = socket(PF_NETLINK, SOCK_DGRAM, NETLINK_ROUTE); Dec 22 16:05:21 is what they are doing Dec 22 16:05:31 * mrmoku checks libnl code Dec 22 16:07:12 hmm Dec 22 16:07:17 no PF_NETLINK in there :/ Dec 22 16:07:27 and NETLINK_ROUTE only used when connecting a socket Dec 22 16:07:37 duh Dec 22 16:07:45 grr brb Dec 22 16:17:51 isn't AF_NETLINK the same as PF_NETLINK? Dec 22 16:17:59 brb Dec 22 16:21:10 hmm.. good question Dec 22 16:23:23 indeed it is Dec 22 16:23:28 /usr/include/bits/socket.h:#define AF_NETLINK PF_NETLINK Dec 22 17:07:51 mmm... dinner. bbl Dec 22 17:14:38 k Dec 22 17:32:20 freesmartphone.org: 03mickey 07gdbus * r3e86288f3a91 10cornucopia/libfsobasics/ (configure.ac fsobasics/utilities.vala): libfsobasics: check whether the C library supports backtrace facilities, uclibc does not Dec 22 17:40:40 freesmartphone.org: 03mickey 07gdbus * rcf9cf7f97929 10cornucopia/libfsobasics/fsobasics/Makefile.am: libfsobasics: if backtrace is not available, fill result with dummy string Dec 22 17:41:05 freesmartphone.org: 03mickey 07gdbus * re53544c1d4f4 10cornucopia/tests/ (org.fresmartphone.Usage.py testserver.py): tests: chmod a+rx Dec 22 17:53:55 freesmartphone.org: 03mickey 07gdbus * r4ffcddf453a1 10cornucopia/fsodatad/tests/ (Makefile.am testworld.vala): fsodatad: don't link tests against libdbus Dec 22 19:39:27 hmm... does g_io_channel_unix_new(fd) connect if fd is a socket? Dec 22 19:50:19 wolfspraul: thank you, will see what Joachim replies... Dec 22 21:01:12 mickeyl: hmm... the vapis in misc... which recipe builds them? Dec 22 21:11:26 duh, can't find them :-/ Dec 22 21:11:43 misc-vapi? Dec 22 21:12:28 dos1: do you have a misc-vapi bb ? Dec 22 21:13:12 i think it was there somewhere Dec 22 21:13:20 I hmm Dec 22 21:14:04 morphis: do you have some clue? Dec 22 21:14:29 freesmartphone: misc-vapi is gone; remove traces Dec 22 21:15:01 (that's some commit description) Dec 22 21:15:08 heh :/ Dec 22 21:15:11 so it *was* there, but it is not now Dec 22 21:15:21 maybe everything from misc-vapi was integrated in plain vala? Dec 22 21:16:21 mrmoku: misc-vapi? Dec 22 21:16:33 morphis: cornucopia/misc Dec 22 21:16:37 the vapis in there Dec 22 21:16:40 like posixextra Dec 22 21:17:12 dos1: for sure not the stuff I added today :P Dec 22 21:17:15 ah ok Dec 22 21:17:39 * mrmoku is in urgent need of setsockopt Dec 22 21:18:07 and it is not build in OE? Dec 22 21:18:26 or whats the problem? Dec 22 21:18:40 there is no recipe in OE Dec 22 21:18:50 but somehow once was as dos1 discovered Dec 22 21:19:21 but it was removed? Dec 22 21:19:25 yup Dec 22 21:19:29 -require cornucopia.inc Dec 22 21:19:29 -DEPENDS = "alsa-lib" Dec 22 21:19:29 -PV = "0.2.0+git${SRCREV}" Dec 22 21:19:29 -PR = "${INC_PR}.0" Dec 22 21:19:34 this was misc-vapi.bb :P Dec 22 21:19:38 mrmoku: where do you need setsockopt? Dec 22 21:19:42 fsogsmd? Dec 22 21:19:54 morphis: in the lowlevel_nokia900 plugin Dec 22 21:20:07 so in fsogsmd Dec 22 21:20:14 yup Dec 22 21:20:15 and setsockopt is not in upstream vala? Dec 22 21:20:24 no Dec 22 21:20:41 ok Dec 22 21:20:47 and mickeyl told me to add it to posixextra ;) Dec 22 21:20:52 hm Dec 22 21:21:06 but then we need the depencency in fsogsmd on misc again Dec 22 21:21:20 which was removed quite some time ago as upstream vala provided all we need Dec 22 21:21:24 until now Dec 22 21:21:43 yeah, more then one year ago Dec 22 21:21:45 it should be definitly in posix.vapi or linux.vapi of upstream vala Dec 22 21:21:47 ok Dec 22 21:21:52 then I would do the following Dec 22 21:22:02 1. add a linux.vapi/posix.vapi to fsogsmd/vapi Dec 22 21:22:09 and prepare a patch for upstream vala Dec 22 21:22:27 hmm... a copy of posix.vapi then? Dec 22 21:22:34 and then remove the vapi from fsogsmd when the patches was commited upstream Dec 22 21:22:38 mrmoku: no Dec 22 21:22:45 it's additional? Dec 22 21:22:55 a fsogsmd/vapi/posixextra.vapi Dec 22 21:23:04 is setsockopt posix? Dec 22 21:23:06 or linux? Dec 22 21:23:09 posix Dec 22 21:23:30 ok Dec 22 21:23:45 then a fsogsmd/vapi/posixextra.vapi and a patch for upstream Dec 22 21:23:50 should be best solution Dec 22 21:23:59 ok Dec 22 21:24:09 the patch for upstream is mickeys job :) Dec 22 21:24:22 morphis: thanks Dec 22 21:24:27 and don't have the rat tail to create a recipe for misc-vapi again and add it as dependency ... Dec 22 21:24:30 no problem :) Dec 22 21:57:41 morphis: I have another problem... Dec 22 21:57:41 | testasync.o: In function `test_async_reactorchannel': Dec 22 21:57:41 | testasync.c:(.text+0xe4): undefined reference to `fso_framework_async_reactor_channel_unref' Dec 22 21:57:45 | collect2: ld returned 1 exit status Dec 22 21:57:49 when building current libfsobasics Dec 22 21:59:17 hmm Dec 22 21:59:35 * mrmoku puts fso on autorev Dec 22 22:01:02 hmm... is on autorev Dec 22 22:05:54 mrmoku: wait Dec 22 22:06:10 with OE? Dec 22 22:07:39 locally I can build latest version Dec 22 22:08:36 morphis: I'm trying a local build of libfsobasics Dec 22 22:08:56 do you have any modifications? Dec 22 22:08:58 INHERIT_append_pn-libfsobasics = "srctree gitpkgv" Dec 22 22:08:58 SRCREV_pn-libfsobasics = "${AUTOREV}" Dec 22 22:08:58 S_pn-libfsobasics = "/OE/src/cornucopia/libfsobasics" Dec 22 22:09:06 small ones Dec 22 22:09:33 in your local.conf? Dec 22 22:10:32 no, in linux.vapi Dec 22 22:10:37 ahh, that in my local.conf, yes Dec 22 22:10:58 + [CCode (cheader_filename = "linux/netlink.h")] Dec 22 22:10:58 + public const int NETLINK_ADD_MEMBERSHIP; Dec 22 22:11:09 just added some missing consts Dec 22 22:11:40 hm Dec 22 22:11:45 hmm... maybe I should just commit them Dec 22 22:11:50 they should be harmless Dec 22 22:11:55 do that Dec 22 22:12:00 ok Dec 22 22:12:07 but for the locally build it should be irrelevant Dec 22 22:13:27 freesmartphone.org: 03mok 07cornucopia * r8efafdb88dfa 10/libfsobasics/vapi/linux.vapi: Dec 22 22:13:27 freesmartphone.org: libfsobasics: add some netlink / socket consts in linux.vapi Dec 22 22:13:27 freesmartphone.org: Signed-off-by: Klaus Kurzmann Dec 22 22:16:51 so I am off Dec 22 22:16:55 gn8 Dec 22 22:18:03 gnight Dec 22 22:18:29 morphis: it built :-) Dec 22 22:27:52 where did you put the setsockopt now? Dec 22 22:30:36 mickeyl: misc-vapi is no more Dec 22 22:30:41 I added it there :P Dec 22 22:30:55 now I copied the posixextra.vapi to fsogsmd/vapi Dec 22 22:31:04 did not commit though Dec 22 22:31:57 ok Dec 22 22:32:14 * mrmoku wants to test that damn thing if it finally builds ;) Dec 22 22:32:15 for the gdbus switch we need a bleeding upstream version of vala anyways, so it's a good chance to sync our vapis Dec 22 22:32:40 good :) Dec 22 22:37:29 mickeyl: still no c++ bindings possible with vala? Dec 22 22:37:35 I'm starting to like it :-) Dec 22 22:38:12 :)) Dec 22 22:38:38 | plugin.vala:610.60-610.72: error: Argument 5: Cannot convert from `ulong' to `PosixExtra.socklen_t' Dec 22 22:38:42 if it would only build :/ Dec 22 22:39:11 unfortunately no, but we can think about writing our UI in C++, writing the communication agents and backend in Vala and gluing this to C++ via a gobject-based C lib which vala is emitting Dec 22 22:39:27 how's line 610 looking? Dec 22 22:39:36 socklen_t is the same size as an int Dec 22 22:39:49 so ulong will fail unless you explicitly truncate Dec 22 22:39:53 (i.e. cast) Dec 22 22:39:53 yeah, but sizeof( int ) seems to give ulong Dec 22 22:40:16 ah, right Dec 22 22:40:18 just cast it then Dec 22 22:40:23 it'll work Dec 22 22:40:25 brb Dec 22 22:40:30 ok Dec 22 22:40:35 * mrmoku rb too Dec 22 22:54:23 hmm Dec 22 22:54:32 AF_LOCAL will rather go into posix.vapi Dec 22 22:55:13 ahh, ok Dec 22 22:55:27 it's synonym to AF_UNIX Dec 22 22:55:31 so you could as well use AF_UNIX Dec 22 22:55:34 which already is in posix.vapi Dec 22 22:55:42 ok Dec 22 22:57:54 mickeyl: there are two more missing I can't find... SOL_SOCKET and SOL_NETLINK Dec 22 22:58:11 ahh and SO_RCVBUF Dec 22 22:58:19 let me check Dec 22 22:58:41 /usr/include/asm-generic/socket.h:#define SO_RCVBUF 8 Dec 22 22:58:52 SOL_SOCKET is there as well Dec 22 22:58:54 1 Dec 22 22:59:53 SOL_NETLINK is in some variants of socket.h Dec 22 23:00:03 so all of those go into linux.vapi Dec 22 23:01:00 hmm Dec 22 23:01:03 i'll add the whole bunch Dec 22 23:01:26 ok, thanks Dec 22 23:01:26 | plugin.c:766:18: error: incompatible types when assigning to type 'enum LowLevelNokia900Gpio[1]' from type 'enum LowLevelNokia900Gpio *' Dec 22 23:01:39 does not work what I wanted to do :/ Dec 22 23:01:51 private bool have_gpio[5]; Dec 22 23:01:53 can you pastebin the code? Dec 22 23:01:58 ok Dec 22 23:02:32 mickeyl: http://paste.pocoo.org/show/308655/ Dec 22 23:05:14 which line is triggering the error? Dec 22 23:05:18 (which vala line) Dec 22 23:08:23 mickeyl: I guess it's 550 Dec 22 23:08:34 have_gpio[l] = gpio_line_probe( l ); Dec 22 23:09:11 l is not an int Dec 22 23:09:18 no Dec 22 23:09:31 * mrmoku is used to enums being ints in same way in C++ :P Dec 22 23:09:32 you might need to drag a counter with you Dec 22 23:09:49 for the left side of the assignment Dec 22 23:09:58 hmm Dec 22 23:10:24 or just iterate over the int Dec 22 23:10:26 and cast it to the enum Dec 22 23:10:33 for the right side Dec 22 23:11:13 * mrmoku tries Dec 22 23:13:05 for ( int f = 0; f < GPIO_COUNT; f++ ) have_gpio[f] = gpio_line_probe( (Gpio)f ); Dec 22 23:13:10 something like that? Dec 22 23:13:31 yes Dec 22 23:14:09 but that won't work later on then... in gpio_write Dec 22 23:14:17 if ( !have_gpio[line] ) Dec 22 23:14:29 with line being enum Gpio Dec 22 23:14:44 hmm... will have to use int there too then Dec 22 23:14:52 and just use the enum to convert to string :P Dec 22 23:16:32 well Dec 22 23:16:38 enum and int are still the same types in vala Dec 22 23:16:49 but in contrast to C-derived languages there is not implicit cast Dec 22 23:16:54 you have to cast it explicitly Dec 22 23:16:58 which is ok and doesn't hurt :) Dec 22 23:17:06 but I can cast from enum to int? Dec 22 23:17:11 sure Dec 22 23:17:13 it's the same type Dec 22 23:17:17 good :-) Dec 22 23:17:19 just to tell the compiler you know what you're doing Dec 22 23:17:42 I hope so :P Dec 22 23:17:54 btw. what is your coding style for arrays? Dec 22 23:18:06 have_gpio[line] is fine? Dec 22 23:18:13 or have_gpio[ line ] ? Dec 22 23:18:58 have_gpio is fine Dec 22 23:19:02 err Dec 22 23:19:03 heh Dec 22 23:19:06 have_gpio[line] Dec 22 23:19:24 even with cast to int? Dec 22 23:19:31 have_gpio[(int)line] ? Dec 22 23:21:10 this is where it gets unreadable a bit Dec 22 23:21:17 so i guess i would add whitespace there Dec 22 23:21:27 ok :-) Dec 22 23:21:34 common sense then Dec 22 23:21:55 yep Dec 22 23:22:37 hmm, those sockopt levels are fun Dec 22 23:22:53 they're hiding ;) Dec 22 23:23:04 hehe... that's why I did not find them :P Dec 22 23:23:23 i guess they're pretty much undocumented, but at least i got a list now Dec 22 23:23:36 it's just that you have to have custom headers Dec 22 23:23:40 gisi/netlink.c:#ifndef SOL_NETLINK Dec 22 23:23:40 gisi/netlink.c:#define SOL_NETLINK 270 /* libc!? */ Dec 22 23:23:42 else it won't compile everywhere Dec 22 23:23:43 is what ofono does Dec 22 23:23:52 yep, that's in line with what i found here: Dec 22 23:24:17 http://www.takatan.net/lxr/source/include/linux/socket.h#L268 Dec 22 23:26:44 http://git.gnome.org/browse/vala/log Dec 22 23:27:43 mickeyl: great :-) Dec 22 23:27:56 if anything more is missing, just holler Dec 22 23:28:07 ok, do we use vala from git? Dec 22 23:28:11 don't think so... Dec 22 23:28:16 no, i cut a new snap Dec 22 23:28:20 and add it to OE tomorrow Dec 22 23:28:21 ahhh. it finally build :-) Dec 22 23:28:27 cool Dec 22 23:28:30 ok, I'm going to bed now anyway Dec 22 23:28:38 me too, gnight Dec 22 23:28:45 gnight Dec 22 23:47:34 mrmoku: if you read this tomorrow... strings are immutable. you need to Dec 22 23:47:38 filename = filename.printf( ... ) Dec 22 23:47:46 filename.printf() won't alter the filename Dec 22 23:47:58 you're doing this in a number of places in that code Dec 23 00:15:18 <[Rui]> hey, is the repo already gdbusified or can I still make an update? Dec 23 00:22:34 nothing new in feed today Dec 23 00:23:12 and I won't sync gdbus to public feed before we're satisfied with it Dec 23 00:26:55 <[Rui]> ok :) Dec 23 00:27:27 <[Rui]> I've been having weird lock ups every now and then, hope it's not a *recent* sd already dying Dec 23 00:28:17 <[Rui]> - Option "EmulateRightButton" "1" Dec 23 00:28:17 <[Rui]> + Option "EmulateRightButton" "True" Dec 23 00:28:44 <[Rui]> I'm not sure this works, in fact i think input-tslib expects it to be a number *really* Dec 23 00:35:52 <[Rui]> zZzZz **** ENDING LOGGING AT Thu Dec 23 02:59:58 2010