**** BEGIN LOGGING AT Tue Nov 15 02:59:57 2011 Nov 15 04:35:08 * pabs3 wonders if it is time for a new SHR blog post Nov 15 08:31:35 hi Nov 15 08:31:50 moin Nov 15 08:32:41 i tried shr-core-testing on my gta02 , seems contacts are not working, i can not edit/add contactfields, any idea ? Nov 15 08:33:07 what's shr-core-testing ? Nov 15 08:33:19 . i dont like editing the sqlite db by hand .. Nov 15 08:34:21 shr-core the testing image , /etc/issue show "shr-2010527 " Nov 15 08:35:57 hmm still doesn't make sense.. Nov 15 08:36:08 maybe it's just shr-testing ? Nov 15 08:36:29 shr-core says at least 2011 Nov 15 08:37:42 i downloaded it from /build.shr-project.org/shr-testing/images/om-gta02/ Nov 15 08:38:15 then it's shr-testing Nov 15 08:38:31 and there should be sqlite db upgrade script in it imho Nov 15 08:39:27 should i better use the unstable image ? Nov 15 08:41:03 you still have to convert old sqlite all 3 shr versions Nov 15 08:41:30 and how ? Nov 15 08:41:51 09:37:12 < JaMa> and there should be sqlite db upgrade script in it imho Nov 15 08:41:59 check files shiped with frameworkd Nov 15 08:42:20 frameworkd postinst was also showing what to do Nov 15 08:42:29 so you can read it Nov 15 08:42:47 if you really have that ancient db Nov 15 08:42:58 maybe you don't have contacts in db only in sim? Nov 15 08:43:43 the phoneui-contacts doesnt allow me to enter text, Nov 15 08:44:33 its a new/clean installation , maybe i upgrade to some image from shr-core/images/om-gta02 Nov 15 08:45:09 you have to reflash, not opkg upgrade.. just in case Nov 15 08:46:26 .. grab out the sim card put it into my netwalker, reformat an untarring, i have the android image on the nand ... Nov 15 08:46:52 s/sim/sd/ Nov 15 08:46:53 pwgen meant: .. grab out the sd card put it into my netwalker, reformat an untarring, i have the android image on the nand ... Nov 15 08:54:42 any idea how i can get a "new" gta02 battery , or at least the charging electronic. Nov 15 08:57:27 .. shr-core.. booting .. Nov 15 09:07:38 hmm shr-core seems not working Nov 15 09:13:10 pwgen: i think Pulster and GoldenDelicious still have that in stock. Are you sure a Nokia one doesn't fit your needs? Nov 15 09:14:01 i am running wit an nokia battery, but i can not get charging and power information Nov 15 09:18:42 did anyone got the shr-core image running ? mine is allways suspending( and not awakening ) after some seconds. Nov 15 09:22:48 pwgen: using the dumb battery driver? Nov 15 09:22:58 pwgen: what exactly are you missing, isn't the approximation "good enough"? Nov 15 09:26:51 . actually i am missing an image, that is running on the GTA02. shr-core makes a nice brick, .on shr-testing i can not edit contacts, and shr-unstable runs "unstable" Nov 15 09:27:47 qtmoko 35 works Nov 15 09:28:46 i had never seen battery information except i use a gta02 battery , ( actually i have a nokia bl5c in ) Nov 15 09:29:18 pwgen: just use the platform_battery driver Nov 15 09:29:30 pwgen: you'll get an approximation that is imho acceptable. Nov 15 09:29:30 and all my gta batteryss ( 2 for the gta01 and one for the gta02 ) are blown up Nov 15 09:30:25 gta01 is a worse case: basically you can't easily get it to charge a non-gta01 battery with a reasonable current. Nov 15 09:30:33 So an external charger might be an option. Nov 15 09:31:05 Alternatively, you can solder ~8k in parallel to the ID pin right at the battery connector, so it should start charging nokia battery. Nov 15 09:31:17 pwgen: yes I have shr-core image running Nov 15 09:33:03 JaMa: when i install it, it booting, asking for configuration, and after some time its goining to sleep and will never awake again Nov 15 09:33:32 there is some problem with display brightness going down Nov 15 09:33:39 are you sure it's suspended? Nov 15 09:34:19 I had the same on last n900 image.. so I guess it's caused by some change in fso Nov 15 09:35:06 hmm no raction on putton press and inserting it to a usb to a host did not shoiw Nov 15 09:35:47 plugging it to a notebook does not show up a usb client, i would say .. sleeping , very deep Nov 15 09:38:02 i am running from sd card. can this be the problem ? Nov 15 09:41:26 I am also running from sd card Nov 15 09:42:17 SHR: 03Martin.Jansa 07shr-chroot * r8ce1a1afe5d3 10/ (923 files in 55 dirs): system upgrade Nov 15 09:43:36 how can i tune the apm/sleep timeouts ? Nov 15 09:44:24 /etc/phonefsod.conf Nov 15 09:45:54 lars_: hi, what's the status of glamo mainlining? ;) Nov 15 09:49:11 lindi-: I got an email from someone about it last week, but I've been really busy Nov 15 09:49:21 Weiss: ooh Nov 15 09:50:20 Weiss: how would you feel if I added a wiki page that has all the 2.6.39 openmoko patches and then we could write comments about their mainlining status? or is there already something like that? Nov 15 09:55:19 lindi-: that'd probably be good, I don't know if one exists any more Nov 15 09:55:47 upstreaming Glamo (DRM kernel bits, libdrm bits and Mesa bits) would probably be straightforward, but the question is: would it help anyone? Nov 15 09:56:18 Weiss: well we could get openmoko kernel support to debian at least Nov 15 09:56:51 Weiss: mesa bits as in userland stuff? Nov 15 09:57:37 Mesa bits as in the 3D stuff (about enough to do a glClear()) Nov 15 09:57:59 Weiss: ok but I'd be happy with 2D Nov 15 09:58:17 Weiss: that only needs the kernel bits surely Nov 15 09:58:31 Weiss: this https://gitorious.org/pkg-fso/linux-2-6-gta02/blobs/gta02/debian/patches/features/gta02/0015-Add-glamo-driver.patch Nov 15 09:59:23 ah, that's the non-DRM stuff Nov 15 09:59:32 I was thinking about the DRM stuff Nov 15 10:02:18 Weiss: ah, I don't think I use it at all Nov 15 10:02:24 Weiss: and I'm still happy :) Nov 15 10:02:38 Weiss: so how's the non-drm stuff? Nov 15 10:03:26 what ist aurora-image ? Nov 15 10:03:46 I don't really know anything about non-DRM... Nov 15 10:04:08 does the GTA04 have any kind of graphics acceleration? Nov 15 10:04:39 Weiss: we're using 2.6.39 now and last drm patches exists for 2.6.34 Nov 15 10:08:32 JaMa: it should be possible to rebase it without too much hassle, though Nov 15 10:08:48 perhaps a little bit to use the mainlined platform-DRM stuff Nov 15 10:11:00 it was causing some issues even with 2.6.34 so I've dropped it from our shr kernel patchset for 2.6.37.. Nov 15 10:11:43 and I'm completely happy with 2D in 2.6.39 :) Nov 15 10:11:58 I just want all the stuff mainlined Nov 15 10:15:06 To be able to install plain mainline Debian, yay. When there'll be not more that 100 users of the gta02 left :( Nov 15 10:21:36 PaulFertser: we can get more users :) Nov 15 10:23:04 Weiss: JaMa: seems there's "git notes" that could be used Nov 15 10:29:17 lindi-: applied your patch to shr-3.6.39-nodrm https://gitorious.org/shr/linux/commit/eaedb6e64687ad5b08dac80953718fbd9cf80aa2 looks better then my old https://gitorious.org/shr/linux/commit/7d0d327464eeee664e6aa26538d4990caea13ed7 Nov 15 10:30:09 JaMa: thanks :) Nov 15 10:30:17 what gcc do you use? Nov 15 10:30:32 zub, ping Nov 15 10:30:42 seems to work fine here with gcc-4.6 Nov 15 10:31:00 JaMa: 4.4.6-11 Nov 15 10:31:16 JaMa: wifi driver also needs some noinline I think Nov 15 10:31:31 JaMa: uses huge amounts of stack and I think that might be partially due to inlining Nov 15 10:32:04 JaMa: "/home/lindi/debian/debian-linux-2.6/linux-2.6-2.6.39/debian/build/source_armel_none/drivers/ar6000/ar6000/ioctl.c:2531: warning: the frame size of 1896 bytes is larger than 1024 bytes" Nov 15 10:32:07 and have you tried http://markmail.org/message/xm5syil3vzhhjwzz#query:+page:1+mid:l7ni2cog7zxlmbwe+state:results instead of reverting all Per's patches? Nov 15 10:32:36 JaMa: nope, I just followed your SHR work happily :) Nov 15 10:33:44 yay efl "1.1" alpha Nov 15 10:42:21 lindi-: I've pushed his suggestion to gitorious in shr-2.6.39-nodrm-mci-wip, but cannot test it on device so it will probably just end there for reference Nov 15 10:58:39 JaMA: no chance to get actual shr-core running. Nov 15 11:11:42 JaMa: can you consider pulling http://docs.openmoko.org/trac/ticket/2478 too? Nov 15 11:30:45 lindi-: I would like at least mrmoku's comment on that one Nov 15 11:31:12 mrmoku: can you have a look http://docs.openmoko.org/trac/ticket/2478 ? Nov 15 11:32:02 JaMa: yes sure Nov 15 11:32:34 JaMa: I also found some usb bugs but I think I'm going to send them to mainline outside any openmoko circles since they are not critical or openmoko specific Nov 15 12:11:24 pespin: pong... I've had a look at the code and you were right, there's no this Nov 15 12:14:18 zub, do you think could we remove the has_target=false and use the delegates as class methods? so we could have access to class members Nov 15 12:14:22 JaMa: PaulFertser: Weiss: http://wiki.openmoko.org/wiki/Kernel/Upstreaming Nov 15 12:16:29 pespin: as usual, I'm at work now and don't have the code here :-( ... and I don't know exactly how is "this" passed, but in general it must work Nov 15 12:16:48 what all data does the callback receive? we still need a way to identify which actual item is the callback invoked for Nov 15 12:17:02 also, what all code are we breaking by libeflvala changes? I guess not much Nov 15 12:17:41 zub, basically, examples, your apps, my apps and Iliwi (which we can patch in OE) Nov 15 12:17:41 :P Nov 15 12:17:41 and ffalarms which we can also patch in OE Nov 15 12:18:36 that's a huge user base :D Nov 15 12:18:38 zub, anyway, I think genlist is unusuable for complex utilities if we don't improve the way they are used in vala Nov 15 12:18:54 zub, yeah hahaha Nov 15 12:19:25 I agree the api is suboptimal Nov 15 12:19:28 fix it :D Nov 15 12:19:49 I can be more cooperaive in teh evening, but now I'm at work :( Nov 15 12:19:55 cooperative Nov 15 12:22:02 I need to look more into it as I never used it before Nov 15 12:22:32 zub, I got another problem too when trying to set content with for example a bubble with a Label inside it Nov 15 12:22:49 as genlist can free that content whenever it wants Nov 15 12:23:44 I'm currently, still, stuck on glib and some main loop fun Nov 15 12:23:50 (w.r.t. my app) Nov 15 12:24:12 I don't know how to tell vala not to free those objects (I can pass the ownership of the parent widget when returning the object, but I don't know how to do that with children) Nov 15 12:24:38 zub, I'll dig more into it and show you this night or so if you have time :) Nov 15 12:24:56 I think i'll be here Nov 15 12:25:28 btw there's #vala on irc.gimp.net Nov 15 12:37:12 * pespin lunch, bbl Nov 15 12:50:19 lindi-: serial console and gsm are on the same uart in gta01 :( Nov 15 12:50:32 PaulFertser: aha so I can ignore that patch too? Nov 15 12:50:41 lindi-: i think so Nov 15 12:51:03 PaulFertser: I updated my comment on that Nov 15 12:52:22 lindi-: jbt is not the backlight, it's the IC inside LCM. Nov 15 12:52:44 iirc Nov 15 12:53:40 PaulFertser: ah ok, makes sense Nov 15 12:53:50 PaulFertser: could that be responsible for the odd colors in 320x240 mode? Nov 15 12:54:25 PaulFertser: I've recently started watching TV with my phone and especially in dark scenes the colors are pretty weird Nov 15 12:55:10 lindi-: "washed out" colours seem to be somehow related with the "pixelclock" of the qvga mode, if you set it higher, they're getting better. Nov 15 12:55:21 lindi-: i remember one needs to tweak jbt for landscape mode. Nov 15 12:55:40 PaulFertser: tweak as in poke /sys or patch kernel? Nov 15 12:56:41 PaulFertser: my TV solution is: http://paste.debian.net/145543/ Nov 15 12:56:44 lindi-: kernel should do that on its own iirc. Nov 15 12:56:59 PaulFertser: ok Nov 15 12:57:21 gitorious sucks, doesn't show me commits without JS :( Nov 15 12:57:50 but github is even more non-free :( Nov 15 12:58:10 let me see if there's a workaround Nov 15 12:59:21 PaulFertser: append "?format=patch" Nov 15 13:00:14 lindi-: (accelerometer) there's another driver for those devices, and it's upstream but it sucks a lot. Nov 15 13:00:21 I'll apt-get install xul-ext-itsalltext to be able to do search-and-replace in the wiki Nov 15 13:04:00 lindi-: (jitterless touchscreen) iirc there was a cool hack to read out the TS values during the vblank, otherwise it was jittery. Nov 15 13:04:35 PaulFertser: should work without javascript now Nov 15 13:05:08 PaulFertser: but why does screen refresh have such an effect? isn't ts supposed to be independent? Nov 15 13:06:17 or something like that Nov 15 13:06:30 right, updated again :) Nov 15 13:06:52 lindi-: well, they're near and they affect each other over the air i think :) Nov 15 13:08:40 lindi-: the last link on the wiki seems to be a branch and not a patch Nov 15 13:09:17 PaulFertser: it's a patch but it's just not in SHR branch yet Nov 15 13:10:21 PaulFertser: added link to http://docs.openmoko.org/trac/ticket/2478 Nov 15 13:11:39 lindi-: btw, that systemtap stuff you did is cool. Nov 15 13:13:06 PaulFertser: knowing systemtap four years ago would have helped a lot more :) Nov 15 13:23:07 lindi-: (tslib pressure) hm, worth investigating more: is not returning pressure at all a valid thing to do for a TS driver? Nov 15 13:24:20 PaulFertser: can I add that under "Paul: " there? ;) Nov 15 13:24:40 lindi-: http://lwn.net/Articles/304326/ Nov 15 13:24:45 lindi-: sure Nov 15 13:26:05 PaulFertser: I've seen my laptop TS do that (events with zero pressure) Nov 15 13:26:18 so tslib can just be outdated? Nov 15 13:29:06 I think it worth investigating, just that, the patch doesn't seem outright wrong (i.e. non-suitable for upstream). Nov 15 13:29:31 hi, is there a browser that support ssl for the freerunner under shr-core? Nov 15 13:58:43 JaMa: seems to be a backlight prongnis runtiblem i have connected the debug board and can see Nov 15 13:59:05 how can i switch the backlight on ? Nov 15 14:01:11 freesmartphone.org: 03angelo 07aurora * r38b7ba5f9527 10/aurora-daemon/ (10 files in 4 dirs): Nov 15 14:01:11 freesmartphone.org: aurora-daemon: include Aoura browser v0.2 Nov 15 14:01:11 freesmartphone.org: * Imported from angelox/testing branch Nov 15 14:01:11 freesmartphone.org: * Edited to fit in current aurora components changes Nov 15 14:01:25 /sys/devices/platform/s3c2440-i2c/i2c-0/0-0073/pcf50633-backlight/backlight/pcf50633-backlight/brightness Nov 15 14:01:28 pwgen: ^ Nov 15 14:01:45 wich value ? Nov 15 14:03:16 lindi-: (gps power-on/off fix) i wouldn't be so sure it makes sense only with u-boot, quick look doesn't confirm it. Nov 15 14:04:12 lindi-: looks more like a mandatory bugfix to me. Nov 15 14:05:08 pwgen: bigger then current and smaller than /sys/devices/platform/s3c2440-i2c/i2c-0/0-0073/pcf50633-backlight/backlight/pcf50633-backlight/max_brightness Nov 15 14:07:05 pwgen: in general, /sys/class/backlight is there for all sane backlight devices. Nov 15 14:07:44 bl_power contains 0 .. Nov 15 14:08:29 echo "63" > brightness voila i can see something ... Nov 15 14:10:54 when i start the xservernodm the backlight goes off .. Nov 15 14:27:15 how can i activate the 2 buttons ...? Nov 15 14:35:59 pwgen: elm_shr_softkey Nov 15 14:39:22 ahh i get closer ... all services are per default disabled Nov 15 14:51:12 but no network no gsm , seems there is nothing working Nov 15 14:55:19 pwgen: what image did you install? Nov 15 14:57:12 the oe-core image from bearstech Nov 15 14:57:21 om-full. Nov 15 14:59:11 http://shr.bearstech.com/shr-core/images/om-gta02/full-om-gta02.tar.gz. but it seems this is a pain in the ..... Nov 15 15:00:41 pwgen: "opkg info systemd' says what? Nov 15 15:00:57 and /etc/shr-version Nov 15 15:02:40 SHR 20111109 Nov 15 15:02:40 Built from branch: shr Nov 15 15:02:40 Revision: b94b4fe3cda6034194212fb8ddf7767972df0913 Nov 15 15:02:47 and no systemd installed Nov 15 19:39:23 JaMa, lindi-: (alsa/suspend) interesting ... Nov 15 19:41:20 heyho Nov 15 19:41:22 angelox|laptop: ping Nov 15 19:41:58 angelox|laptop: sorry I don't have enought time to write a mail to you today Nov 15 19:42:04 angelox|laptop: but as a first step to ease development Nov 15 19:42:48 angelox|laptop: we should adjust our development cycle that way that we develop features in own branches Nov 15 19:43:03 angelox|laptop: I already started with development of call management in an own branch I will push later Nov 15 19:43:22 so we can merge features into master branch when they are ready and stable Nov 15 19:43:29 so we keep master branch all time stable Nov 15 19:43:44 which is something we should have when targeting a release this year Nov 15 19:43:59 so we can even decide to not include some features if we don't have enough time Nov 15 19:46:40 morphis: pong; sorry for delay Nov 15 19:46:44 * angelox|laptop reads Nov 15 19:48:13 morphis: ok; that's what i do at every change i do,i re-compile whole aurora and check all features to maintain it stable Nov 15 19:49:17 thats the minimum Nov 15 19:49:25 but we should work with branches Nov 15 19:50:04 hmm Nov 15 19:50:21 one for each feature? Nov 15 19:50:26 that we want to work Nov 15 19:50:54 yes Nov 15 19:52:13 which features we have to work on we need to discuss per mail Nov 15 19:53:25 ok Nov 15 19:53:55 there's a while i don't see the user stories to check which features we need to work Nov 15 19:57:48 yes, we need to define them too ... Nov 15 20:04:08 are you going to use fso mailing list to talk? Nov 15 20:04:12 yes Nov 15 20:04:14 fso-userland Nov 15 20:04:42 angelox|laptop: I will define my user stories and add them to the wiki Nov 15 20:04:47 hmm,i think i'm not subscribed in that one Nov 15 20:05:38 morphis: ok; i'll define my ones too,but i'm a bit busy right now Nov 15 20:07:12 no problem I am too Nov 15 20:43:31 morphis: (AlmostOT) Do you remember the first 'logo'(icon) of aurora? Nov 15 21:08:09 JaMa: next systemd problem is syslog... guess we would want a newer busybox or use some other syslog Nov 15 21:10:12 JaMa: there are patches for syslogd socket passing from syslogd but only in unstable 1.19 git Nov 15 21:12:42 hi pespin Nov 15 21:27:37 pespin: hola! looking at DOC and generated code, I believe the has_target=false is simply wrong for the genlist callbacks. it currently works by accident, and it also explains why did I have to do an ugly cast to get it working... Nov 15 21:28:19 pespin: currently wit the VAPI as it is, valac uses one-arg-less function as the callback Nov 15 21:28:51 this works bc of c calling convention, but it also means what we get as "Evas_Object *obj" is in fact the data field (void* data) Nov 15 21:29:27 ~ I had to cast this Evas_Object* to my own, completely unrelated class... which is ugly and I was surprised why they expect Evas_Object - now it's explained Nov 15 21:29:28 You had to cast this Evas_Object* to my own, completely unrelated class... which is ugly and I was surprised why they expect Evas_Object - now it's explained? Nov 15 21:29:50 pespin: good you noticed :) Nov 15 21:30:23 buuut... it we change it to has_target=true, this way the prototype will be ok, and we get delegates working Nov 15 21:30:43 but, can one still associate custom data just with the particular item? Nov 15 21:30:58 possibly via this Evas_Object Nov 15 21:31:13 because we need a way to tell wich item it was that the notification is about Nov 15 21:31:33 I've been using the data field (incorrectly passed as *obj) to store this information Nov 15 21:32:06 to keep it that way, we'd add void* arg first and heep has_target=false, but then no delegates... so I think has_target=true is the way to go Nov 15 21:32:14 just - how to figure out which item was selected? Nov 15 21:32:22 additional data per item would be cool Nov 15 21:33:32 I'll ask in #e Nov 15 21:34:01 zub: genlist items have data IIRC Nov 15 21:34:43 yup, they do - genlist_item.data_get() - but can I get to the genlist item structure in the callback? Nov 15 21:34:52 http://docs.enlightenment.org/auto/elementary/group__Genlist.html#ga5084a6406b0f996633b9afd1e74a8bc6 Nov 15 21:35:20 zub: callback for which event? Nov 15 21:35:46 basically all the genlist ones, e.g. Elm_Genlist_Item_Label_Get_Cb Nov 15 21:35:56 it takes: (void *data, Evas_Object *obj, const char *part) Nov 15 21:36:01 what is the obj? Nov 15 21:36:18 if that doesn't help, then we're ...ed and we have to keep has_target=false I think Nov 15 21:36:31 (but fix the prototype then) Nov 15 21:36:39 I *think* item related callbacks should get the item passed... Nov 15 21:37:01 lemme ask the #e gurus Nov 15 21:38:34 and a obj parameter that points to the genlist object itself Nov 15 21:38:35 hmm Nov 15 21:38:48 zub: yeah, asking e freaks is the best :) Nov 15 21:39:40 zub: All of them receive a data parameter that will point to the same data passed to elm_genlist_item_append() and related item creation functions, and a obj parameter that points to the genlist object itself. Nov 15 21:40:03 tells me that obj is the genlist itself Nov 15 21:40:43 yes, that's why I've been using the data for passing my own stuff Nov 15 21:40:56 but pespin would (not unreasonably I have to say) like to use it for this Nov 15 21:41:04 because now the callback has to be static Nov 15 21:41:33 but if we use it for this, we have a way to figure out the actual item Nov 15 21:41:43 if there's no such way, then it just has to stay static Nov 15 21:42:34 (but then it has to be fixed anyway because the VAPI now declares it static, but with 2 args... so what is names obj is in fact data, and what is named part is obj, and part is not accessible at all) Nov 15 21:42:39 * pespin_ back, needs min to read backlog Nov 15 21:46:26 zub: question is what you need the item for in the func callbacks? Nov 15 21:47:27 mrmoku: if elm asks me for item description, I need to teel which item it's asking for :-P Nov 15 21:47:28 my guess is you should _not_ need it Nov 15 21:47:31 methinks :) Nov 15 21:47:32 ? Nov 15 21:48:01 but the description does not depend on the item itself.. .more on the data you gave it when adding it, no? Nov 15 21:48:10 and that one you get Nov 15 21:48:21 mokuuuuuuuuu! :) Nov 15 21:48:29 :P Nov 15 21:48:37 yes, unless you use that to pass "this" so that you could use real delegates in vala, not only static Nov 15 21:48:56 pespin: would like to have real delegates... the idea sounds nice... point is: can we do it? Nov 15 21:49:05 ahh, now I start to understand :) Nov 15 21:49:36 zub: you're doomed ;) Nov 15 21:49:51 we're dooooooooooomed, na ja :) Nov 15 21:52:53 ok I'll explain a posibility I have in mind Nov 15 21:53:35 remove has_target=false Nov 15 21:54:02 then, that means we have a target, so an object from which the func is called Nov 15 21:54:32 so remove "static" from delegates Nov 15 21:54:47 nah, now I realized the semantics is wrong for such usage anyway :-P Nov 15 21:54:48 one way would be Nov 15 21:54:59 but I'm listening first... go on Nov 15 21:55:20 creating 1 object which has get_content, get_label, etc. methods Nov 15 21:55:26 1 object per item Nov 15 21:55:44 ok Nov 15 21:55:45 then we can feed the object with whatever we need. Nov 15 21:56:36 this could work Nov 15 21:56:39 so we have 1 main class/object which holds tobjectshe genlist + the Item Nov 15 21:56:56 so we have 1 main class/object which holds the genlist + the Item objects Nov 15 21:57:01 but then you're forcing this one-instance-of-cb-handler per item view Nov 15 21:57:06 sorry, netbook touchpad ;) Nov 15 21:57:20 yeah Nov 15 21:57:29 thats 1 approach Nov 15 21:57:32 other approach Nov 15 21:57:32 you can do this with static as well, just with a bit more typing :-P and you keep the other "pass any data you want" available too Nov 15 21:58:37 other approach would be manually creating a struct and feed it with whatever data you need, and make delegates static Nov 15 21:58:45 but that would be ugly for vala imho Nov 15 21:58:59 on the other hand Nov 15 21:59:23 having 1 object instance per item, is overkilling Nov 15 21:59:51 because we are somehow going against genlist philosphy of freeing content when it's not needed Nov 15 22:00:04 anothe thing: how would you, neatly in vala, make the association between the class that handles the callbacks and the genlist? Nov 15 22:00:16 the callbacks are set in item class Nov 15 22:00:20 cause the glib object would need to still be there when get_content is called to recreate the content Nov 15 22:00:49 zub, just pass a reference to the main class as param of the constructor of the Item class Nov 15 22:01:10 and store it as attribute Nov 15 22:01:16 for later use when delegates are called Nov 15 22:01:23 not what I asked (or meant to ask): how do you set the actual delegates Nov 15 22:01:29 so that genlist calls them Nov 15 22:01:54 e.g: item_class.func.label_get = on_ap_list_get_label; ... Nov 15 22:02:04 so how would you change this to look nice? Nov 15 22:02:23 var object_Item = new ObjectItem(this); genlist.append(object_item, blabla); Nov 15 22:03:09 hmmm Nov 15 22:03:20 no, pregunto sobre algo differente :) Nov 15 22:03:24 yeah :) Nov 15 22:03:43 you can have more item classes, by the way Nov 15 22:04:05 one should make sure it's difficult to mix up different classes and their this pointers :) Nov 15 22:04:31 yeah that item_class.func.label_get = on_ap_list_get_label; Nov 15 22:04:34 is a problem Nov 15 22:04:42 you would have to create some wrappers for the item classes, and that wrapper would set all from, say an interface (abstract class) Nov 15 22:04:57 ah, vala has interfaces... so, ignore the abstrac class note Nov 15 22:05:04 yep Nov 15 22:05:29 I guess we'd need to go to the 2nd approach Nov 15 22:05:36 creating struct with personal content Nov 15 22:06:06 like { main class reference, key for item } Nov 15 22:06:59 and in get_content() { return data.main_class_reference.get_content(data.key)); } Nov 15 22:07:30 i.e. keep the delegates static and let the user handle it Nov 15 22:07:32 we could use generics for that I guess Nov 15 22:07:34 yes Nov 15 22:07:38 yup Nov 15 22:08:13 I've never played with generics in Vala before (I used them, but nver created a generics class) so if you want to give it a try you're welcome Nov 15 22:08:18 if not, I'll look into it :) Nov 15 22:08:42 JaMa, mrmoku hi ffalarms segfault when adding an alarm Nov 15 22:08:54 hi pespin_ Nov 15 22:09:00 GNUtoo, hi :) Nov 15 22:09:06 pespin_: I'm tempted to do it, but dunno when I find the time Nov 15 22:09:19 ok, let's share info when someone does it Nov 15 22:09:20 any news on better bluetooth audio support? Nov 15 22:09:23 pespin_: nevertheless, let me propose this as the first fix: http://linux.fjfi.cvut.cz/~zub/SHR/libeflvala/delegates-arg-count-fix.patch Nov 15 22:09:31 the arg count is, for the static case, wrong Nov 15 22:09:36 damn Nov 15 22:09:37 :) Nov 15 22:09:44 wrong patch... uno momento Nov 15 22:09:49 GNUtoo, I ordered bluetooth headsets, they should arrive in 1-2 weeks Nov 15 22:09:53 ok Nov 15 22:10:07 zub, un momento :) Nov 15 22:10:16 currently playing music is a pain Nov 15 22:10:31 between the setup and the restart of intone.... Nov 15 22:10:37 you spent too much time Nov 15 22:11:00 and mplayer2 is not good enough ;) Nov 15 22:11:05 ah? Nov 15 22:11:10 pespin_: it's been too long since I used any spanish... teach me! :) Nov 15 22:11:12 too slow in my gta02 Nov 15 22:11:24 yes but did you try before or after mad and such things Nov 15 22:11:42 GNUtoo, mad is still too slow. It's mostly ok but still stops sometimes Nov 15 22:11:50 ok Nov 15 22:11:52 every 3-4 secs Nov 15 22:11:54 I use mainly ogg vorbis Nov 15 22:11:55 ouch Nov 15 22:12:10 I think the way to go is gstreamer Nov 15 22:12:11 we must bother JaMa about that Nov 15 22:12:13 enjoy Nov 15 22:12:17 ah? Nov 15 22:12:19 really? Nov 15 22:12:22 let me retry Nov 15 22:12:33 pespin_: http://linux.fjfi.cvut.cz/~zub/SHR/libeflvala/0001-eflvala-fix-prototypes-of-GenList-callbacks.patch Nov 15 22:12:47 neeee Nov 15 22:12:54 again wrong :-/ Nov 15 22:13:01 wtf is wrong with me :( Nov 15 22:13:11 but I have to check enjoy yet, as people in #e are playing with emotion lately (some commits) which broke emotion, and so enjoy (I think it's now fixed) Nov 15 22:13:20 zub, lol Nov 15 22:13:20 I have some local improvements, but that should be separate, if at all Nov 15 22:15:27 zub, btw, I asked some questions about genlist callbacks because the docs where not pretty enough and in response they added info to the docs, so if you have some question fetch latest elementary, build it and look at Elementary.h Nov 15 22:17:45 zub, oh I now see you just asked about obj in #edevelop. I asked it this morning and they told me is the genlist object (not the item, but the list itself) Nov 15 22:18:22 http://linux.fjfi.cvut.cz/~zub/SHR/libeflvala/0001-eflvala-fix-prototypes-of-GenList-callbacks.patch - finally Nov 15 22:19:16 yeah Nov 15 22:21:02 there is also the microsd issues to fix.... Nov 15 22:21:11 I'll look at the other genlist changes I have here Nov 15 22:21:24 also, there are some deprecations/renamings, these should be reflected Nov 15 22:21:36 zub, I got other problem while playing with get_content and genlist Nov 15 22:22:01 related to genlist freeing objects Nov 15 22:22:35 GNUtoo, and SD is not being mounted by default Nov 15 22:22:41 also may I ask... http://linux.fjfi.cvut.cz/~zub/SHR/libeflvala/0003-eflvala-get-rid-of-a-warning.patch - it's trivial, but kill off a warning :) (one of many, but still...) Nov 15 22:23:18 zub, yeah, but that looks ugly. imho it looks uglier than getting a warning :P Nov 15 22:23:42 warnings are wrong :-/ Nov 15 22:23:50 but you're the boss :) Nov 15 22:24:04 I'll think about it :) Nov 15 22:24:43 zub, the problem I got, is the following: Nov 15 22:25:25 pespin_, that's very easy to fix Nov 15 22:25:29 vim /etc/fstab Nov 15 22:25:46 imagine I want to set a bubble which contains a label as content Nov 15 22:25:53 scp root@192.168.7.2:/etc/fstab ./meta-smartphone... Nov 15 22:26:01 and commit and push Nov 15 22:26:09 zub, but I should declare both as unowned because genlist will free them whenever it wants... Nov 15 22:26:11 pespin_: listening, but I'm quite confused about elm/e* lifetime, I mostly rely on hope in this issues :) Nov 15 22:27:07 zub, afaik I can pass reference of the top object using: return (owned) bubble; Nov 15 22:27:14 so bubble is not freed Nov 15 22:27:27 when the delegate ends Nov 15 22:28:08 yup Nov 15 22:28:13 but no idea on how to force vala to not to free the label under bubble. Nov 15 22:28:18 bye Nov 15 22:28:43 ah, interesting issue Nov 15 22:29:17 would it be auto-freed by E somehow (if vala didn't free it)? Nov 15 22:29:49 sure, genlist frees it (that's why the get_content exist, to recreate the evas objects when they are needed) Nov 15 22:30:39 so you need to make vala forget about the instance w/o deleting it? Nov 15 22:30:56 yeah Nov 15 22:31:18 I guess I'll have to ask in #vala for some solution (maybe some ccode?) Nov 15 22:32:36 if no solution appears, I guess I can try using edje so only parent object is created by vala, but would be nice to create it on code for now Nov 15 22:34:19 well, I think you could use some naaaasty hack, like creating a c function that doesn't delete and passing ownership to it (?), but that's seriously ugly Nov 15 22:35:11 zub, I tried something like bubble.pack_end( (owned) label); Nov 15 22:35:14 btw. (a lame question, I know) the trouble all stems form the fact that Evas_Object is not refcounted? Nov 15 22:35:15 but it seems it's not working Nov 15 22:35:23 maybe the method needs to be declared as owned? Nov 15 22:35:40 I'd think, but I'm far from sure... just ask in #vala Nov 15 22:35:49 zub, yeah Nov 15 22:36:25 zub, I guess we could think on addig owned to pack_end(). that would solve lots of problems with elementary Nov 15 22:36:49 if that works the way... Nov 15 22:37:13 afaik that would be correct in the sense that when you pack something inside another thing, the first gets reparented, and when the parent is freed, childs are freed too Nov 15 22:37:28 yes, it makes sense Nov 15 22:37:47 but I'm not sure what does vala think w.r.t. passing ownership that way Nov 15 22:38:07 * zub sees how much reading he has to do Nov 15 22:38:13 yeah hehe Nov 15 22:38:31 I have no trouble with refcounting in c++, but with vala, because it's so automatic and invisible, I'm always unsure :-( Nov 15 22:39:12 I think parents should get ownership of children Nov 15 22:39:29 yes, I agree with that Nov 15 22:39:53 when they kill the children on their own will, they do have the ownership, so libeflvala should reflect that Nov 15 22:39:59 yeah Nov 15 22:40:45 playya__, mickey|babybusy if you are still alive over there and have time to read all this, please opinions are welcome :) Nov 15 22:43:57 zub, while programming using eflvala, I feel all the time like sweating to maintain all efl widgets alive, which is harder in that way than programming in C+efl Nov 15 22:44:37 hm, true... I've been storing them in containers and that's PITA Nov 15 22:44:43 yeah Nov 15 22:45:08 maybe adding "owned" if it's posible would be nice. I'll try to look further into it during next days Nov 15 22:57:24 hm, I'm still fighting with my event loop issue :( Nov 15 22:58:03 zub, which problem do you have? Nov 15 22:58:50 zub, btw I lost the link to you project repo, so if you could give it to me would be nice :) Nov 15 22:59:57 a multitude... when writing to a pipe via giochannel, I get "g_io_channel_write_chars: code should not be reached", so I'm doing something wrong... but then I see I get a call wring messed up this (-1), so there's probably more fundamental issue Nov 15 23:00:17 but I've been too tired recently (damned daily job) that I'n again not really progressing Nov 15 23:02:21 pespin_: git://zub.lamer.la/shr_supplicant.git but no recent commit there, my working copy is guite ... unstable, at the moment :) Nov 15 23:03:06 ok :) Nov 15 23:03:31 I had the main look nicely working, at least so I thought... but as I was using a polling event source (well, I knew it's polling, and I wanted to do it w/o polling, but thought the polling freq will be ok anyway) Nov 15 23:03:45 ~ I discovered it completely burns the CPU even if idle :) Nov 15 23:03:46 You discovered it completely burns the CPU even if idle :)? Nov 15 23:03:58 apt: nö! Nov 15 23:04:24 so now I'm changing it to use a pipe + giochannel Nov 15 23:04:43 because this way I can wake the main loop w/o the main loop having to poll Nov 15 23:04:53 but I messed it up somehow Nov 15 23:05:13 anyway, off to bed :( Nov 15 23:05:16 'night Nov 15 23:07:41 zub, wasn't this a wifi client using wpa_supplicant? Nov 15 23:07:55 doesn't wpa_supplicant have a dbus interface? Nov 15 23:26:25 pespin: it is, it does, it uses it Nov 15 23:26:54 zub, then all those pipe / gio_channel? Nov 15 23:27:00 are them related to dbus? Nov 15 23:27:41 no, it's related to me wanting to run the app in 2 threads Nov 15 23:28:36 ok ;) Nov 15 23:28:42 time to go to sleep, gnight! Nov 15 23:28:46 I could mostly do with 1 thread and async calls, but I wanted sync lib... and generally wanted ui w/o any freezes whatsoever Nov 15 23:29:24 a bit mean/superfluous thing, but hey, it's my crazy app :) Nov 15 23:29:39 buenas noches Nov 15 23:29:46 yeah for sure :) Nov 15 23:49:11 good night people **** ENDING LOGGING AT Wed Nov 16 02:59:57 2011