IRC log of #harmattan for Saturday, 2012-03-17

*** NIN101 is now known as NIN101|afk00:01
*** lmoura has quit IRC00:11
*** M4rtinK has joined #harmattan00:18
*** aleksander_m has joined #harmattan00:23
pai have a problem with qtmobility.. has anybody used it?00:33
pabasically i am trying to query all the available qcontactmanagers  for contacts, but for each one of them (including memory and tracker) i get 0 contacts00:33
*** pasko has joined #harmattan00:42
*** pasko has quit IRC00:43
paok i actually get a failed to open store error00:44
pai think i know whats the problem00:47
pai probably dont have tracker read or write access00:47
pacan it be?00:47
paim just running the script from the terminal00:48
pahavent packaged it yet00:48
*** diorahman has joined #harmattan00:54
diorahmanhey guys what do you think about Nokia's Adam Weinrich post: ?00:57
pai mean, if i run a python script from terminal, how do i give aegis credentials to it?00:58
befordaegis-exec maybe? I have no idea maybe itsnotabigtruck knows00:58
paah thanks, let me check01:01
*** niqt has quit IRC01:02
*** [lutunen] has joined #harmattan01:03
*** luke-jr has quit IRC01:03
*** luke-jr has joined #harmattan01:03
paok i think i found it01:04
paapparently it is possible to give credentials to a script, if ran with #!/bin/python01:04
pawhich brings  me to the next problem: how can i add some credential file (aegis file) on the fly, without having to package it?01:05
RzRdiorahman: newlines are missing  ...01:05
*** Natunen has quit IRC01:06
*** lardman has quit IRC01:08
diorahmanRzR: here you are,, from Adam , Quim and Adam01:08
*** aleksander_m has quit IRC01:14
RzRi managed to read it01:16
RzRwell i donno01:16
*** e-yes has quit IRC01:16
RzRi was looking at what is freelancer.com01:16
pawhere do i find aegis-su?01:17
*** RzR is now known as rZr01:20
paseemingly in aegis-sdd-.tools01:20
pabut i dont have it01:20
paand apparently it is not in the repos01:20
pawell ok i know, its friday night, not many are around :)01:23
*** diorahman has quit IRC01:25
*** James_Blond has joined #harmattan01:28
paok, even aegis-manifest-dev is not available01:32
pai start freaking out..01:33
pabloody nokia01:33
*** jaywink has quit IRC01:49
*** Leeks912 has quit IRC01:49
panow aegis is rejecting the package i have to install01:53
*** heymaste_ has joined #harmattan01:55
*** jluisn has quit IRC01:55
pai guess rzr repos are unmantained, arent they?01:56
*** heymaster has quit IRC01:58
*** NIN101|afk has quit IRC02:03
*** hardaker has quit IRC02:04
*** tbf_ has quit IRC02:04
*** oberling_ has quit IRC02:04
*** Enforcer has quit IRC02:04
*** hardaker has joined #harmattan02:06
*** tbf_ has joined #harmattan02:06
*** oberling_ has joined #harmattan02:06
*** Enforcer has joined #harmattan02:06
*** heymaste_ has quit IRC02:09
*** [lutunen] has quit IRC02:10
*** Necrosporus has quit IRC02:19
pait seems this bloody package was available in beta302:21
Sazpaimon_have you given freerdp 1.0 a try?02:27
parZr, are your repos abandoned?02:28
paanyway, its so annoying that aegis-su is even mentioned in the nokia docs02:31
paand it's not fucking available..02:31
*** arcean has quit IRC02:33
*** diorahman has joined #harmattan02:46
rZrpa: not really i was busy on some other projects02:48
rZrbut now there are many apps in a4m repo02:49
rZrSazpaimon_: i did start to look at freerdp  but nothing happend so far02:49
rZrfreerdp_0.8.2-2 depends on cups02:51
parZr, but midnight commander is still broken02:53
pait has still broken dependencies02:53
rZrreport bugs :)02:53
pai think someone did already02:53
rZror better fix them02:53
parZr, by the way, do you know where to find aegis-dss-tools?02:54
Sazpaimon_rZr, freerdp 1.0 is stable02:54
Sazpaimon_in fact theyre at 1.0.102:54
Sazpaimon_and iirc it has pulseaudio support02:55
Sazpaimon_so that means audio redirection should work02:55
rZrSazpaimon_: i imported it but it depends on cups02:55
rZrwanna try to fix it ?02:55
Sazpaimon_rZr, i think that's a module02:56
rZrpa: rebuilding mc02:56
Sazpaimon_ie, it can be built without it02:56
rZrSazpaimon_: that's what i thought02:56
rZrSazpaimon_: i can try if you want02:56
Sazpaimon_at least on gentoo, it has a cups use flag02:56
Sazpaimon_keep in mind freerdp 1.0 and 0.8 are completely different02:57
Sazpaimon_1.0 is much more modular02:57
Sazpaimon_rZr, I can try it myself if youre too busy02:57
parZr, the problem is not mc, it is some libsomething (that i dont remember) that cannot be installed for some reasons02:57
Sazpaimon_I'd be using sbox and not obs though02:57
paSazpaimon_, does it happen you know where to find aegis-su, by chance? :)02:58
Sazpaimon_pa, what do you need it for02:58
Sazpaimon_iirc aegis-su doesnt disable aegis02:58
pai know02:58
rZrpa: no idea that's itsnotabigtruck baby02:58
Sazpaimon_i believe its only on the n950 firmware02:58
pai am just starting with a little python script02:59
Sazpaimon_i dont know that for sure though02:59
paand i dont want to package it every time i change it02:59
pabut i need tracker read access02:59
pai mean, it is even mentioned in nokia sdk docs02:59
pabut the package is not available anywhere02:59
pabof, whatever, i'll go to sleep now03:01
*** M4rtinK has quit IRC03:02
parZr, by the way.. how do i find a4m store?03:05
paapparently google cannot find it03:05
rZrthen you can install apps like03:06
rZrapps://1943 (0.4.0)03:06
rZrthis is the most stable repo to use03:07
paand the app list is in apps.formeego?03:08
paah right03:08
pai clicked the wrong icon03:08
pabasically a free market, i guess?03:10
djszapinpm: how to solve this issue then ?03:10
*** diorahman has quit IRC03:10
pabut there+s not much.. about 15 apps in total03:10
djszapito initialize the selectedIndex property with that return value ?03:10
*** lamikr has quit IRC03:27
*** benares_98 has joined #harmattan03:29
rZrSazpaimon_: i wont have time to hack those cmakefiles03:35
rZrif you wanna fork this package03:35
Sazpaimon_rZr, hacking cmake isnt required03:36
Sazpaimon_just modify the debian/rules03:36
Sazpaimon_but yeah, once I figure out obs I'll do it03:36
Sazpaimon_looks like freerdp wont build without xcursor though, no big deal im pretty sure harmattan has it available03:36
Sazpaimon_still a bug since they allow building without it03:37
*** benares_98 has left #harmattan03:38
Sazpaimon_actually I guess xcursor isnt includes03:39
*** lamikr has joined #harmattan03:42
*** Sazpaimon_ has quit IRC03:51
*** Sazpaimon has joined #harmattan03:51
*** Sazpaimon has quit IRC03:51
*** Sazpaimon has joined #harmattan03:52
*** Sazpaimon has quit IRC03:52
*** Sazpaimon has joined #harmattan03:53
*** beford has quit IRC04:03
*** dymaxion has joined #harmattan04:27
djszapipa stop this nonsense :)04:34
Sazpaimongot freerdp working04:34
djszapiyou need to request for creds as they work, aka from manifest04:34
Sazpaimonaudio redirection is working too04:34
Sazpaimonwhich is good04:34
SazpaimonI'll try to push it to obs in a bit04:35
*** oberling_ has quit IRC05:05
*** oberling has joined #harmattan05:07
*** Necrosporus has joined #harmattan05:14
*** pinheiro_ has joined #harmattan05:15
*** pinheiro has quit IRC05:16
*** risca has quit IRC05:44
*** oberling_ has joined #harmattan05:59
*** oberling has quit IRC06:02
*** heymaster has joined #harmattan06:37
Sazpaimoni cant get osc build working06:43
*** heymaster has quit IRC07:09
*** Natunen has joined #harmattan07:19
itsnotabigtruckpa: did you get your stuff sorted07:33
itsnotabigtrucki'm not sure if aegis-su will do what you want anyway07:34
itsnotabigtruckaegis-exec lets you run your script with arbitrary credentials07:34
itsnotabigtrucktry running a shell with precisely the credentials you want:07:35
itsnotabigtruckopensh -c 'aegis-exec -c -a cred1 -a cred2 -a cred3 ... sh'07:35
itsnotabigtruckthen run your script from there07:35
djszapiwhy would he do that crap ?07:35
itsnotabigtruckyou might have to do aegisctl -s from opensh, so that it'll let you run your script with high privileges without packaging07:35
itsnotabigtruckor just run it like: python foo.py07:35
itsnotabigtruckdjszapi: for validating what aegis creds are required, for one07:36
itsnotabigtruckif you're not sure what creds you need, tweak/repackage/copy/reinstall/test/repeat is slow07:36
itsnotabigtruckaegis-exec/tweak/repeat is quick07:36
djszapithe whole point (and how the security system works) is that you need to request credentials from the manifest.07:37
djszapiplease read the security introdction07:37
itsnotabigtruckmaybe for a normal scenario, but this isn't a normal scenario, and aegis-exec exists for a reason07:37
itsnotabigtruck(to filter credentials to exactly what you want)07:37
djszapiit was never meant for adding credentials from third party developer point of view07:38
djszapibut I am sure you know better how we designed, as usual...07:38
itsnotabigtruckwell, i'm not suggesting going live with this, i'm suggesting using this while hacking on the script07:38
djszapiagain, please listen to others07:39
djszapiwhat you think "hack" is cool, it is not cool.07:39
djszapisee the last accident about the wasted hours for yourself, and the guy whose time you wasted.07:39
itsnotabigtruckdjszapi: i'm not suggesting i know better, i'm just thinking maybe you're being a bit too...purist07:39
djszapibecause you wasted others' and your time last time ?07:40
itsnotabigtruckthe manifest approach is clearly not the way to go for quick testing purposes07:40
djszapieven if there was an initial warning it made no sense ?07:40
itsnotabigtruckit's for real stuff you're really distributing07:40
djszapiwell, sure go for it....07:40
djszapiwho cares then07:40
itsnotabigtruckwell, you were right about the force-remove-fail thing :p07:40
djszapiand many other things07:40
djszapibut the point is that you do not still listen07:40
itsnotabigtruckbut i don't think you had the solution either07:41
djszapihe needs to have something working and he asked whether he can request on the fly: answer was a big no-go for several reasons.07:41
itsnotabigtruck(for the record, the answer was: move the scripts out of the way, then do the remove)07:41
itsnotabigtruck(then install a fixed version of the package that doesn't fail, then remove that)07:41
djszapiI did have the solution07:41
djszapi*sigh* heck I even mentioned07:41
djszapiI do not have time for this ohhh big itsnotabigtruck07:41
itsnotabigtruckif i remember correctly, what you said was to install a new version of the package over the existing one, without any other tweaks07:41
itsnotabigtruckwhich would fail, due to invoking the broken script07:42
djszapiwaste others' time then, and shut down then people's suggestion who made the platform07:42
*** djszapi has left #harmattan07:42
itsnotabigtruckugh, why does it always get like this07:42
*** risca has joined #harmattan07:42
Sazpaimonokay, I cannot get osc build to work07:47
Sazpaimonevery time I run it, it tries to reinstall everything07:47
Sazpaimoneven though the build environment should already *be* installed07:47
*** hardaker has quit IRC07:48
itsnotabigtruckSazpaimon: yup, that's how obs works :(07:48
itsnotabigtruckjust try to minimize obs builds, though unfortunately that's unavoidable when you're doing trial and error because obs is hard to predict07:48
Sazpaimonso i should just commit every little change?07:50
itsnotabigtruckSazpaimon: no, don't commit until you're ready to release07:50
itsnotabigtruckand try to get your buildeps right the first time, that caused no end of trouble for me07:50
itsnotabigtrucksince obs doesn't install something if you don't ask for it, period07:50
itsnotabigtruckmodulo some core packages07:50
itsnotabigtrucka lot of upstream debian packages have broken buildeps under those conditions07:51
Sazpaimonso I need to rebuild the build evirionment whenever i make a little change and want to obs build?07:52
*** djszapi has joined #harmattan07:52
djszapiSazpaimon: that is the power of obs07:52
djszapiif we drop that feature, we do not need obs07:52
djszapithat is mostly the main idea behind obs07:52
djszapiand it is not "hard" to predict07:53
djszapiSazpaimon: also, with osc, you have an option to keep stuff07:53
djszapiso it is not even true "that is obs works".07:53
Sazpaimonalso the "No AEGIS_HASH_FDS environment" is a bit annoying07:54
Sazpaimonjust a bit07:54
djszapiSazpaimon: why would it ?07:54
djszapiyou can simply skip it07:54
Sazpaimonbecause its filling up my build log07:54
djszapiand ... ?07:54
djszapiyou do not have 5k space ?07:54
djszapion your disk ?07:55
Sazpaimoni didn't say its a concern07:55
Sazpaimoni said its annoying07:55
djszapiagain, why ?07:55
djszapiyou do not need to deal with those lines at all07:55
Sazpaimonbecause i have to look at 5000 lines of "No AEGIS_HASH_FDS environment" when I want to look at the actual backlog07:55
itsnotabigtruckSazpaimon: it is pretty ugly...just minimize the number of OBS commits you do07:56
itsnotabigtruckone per release, plus any extras to fix unanticipated bugs07:56
itsnotabigtruckwell, FTBFS events07:56
*** icequeen has joined #harmattan07:57
icequeenyou do not need07:57
*** icequeen is now known as djszapi_07:57
djszapi_that is the whole point07:57
djszapi_it is in the beginning of the whole log.07:57
* djszapi has never done one per release commit07:58
djszapiSazpaimon: also, I highly recommend using the webinterface07:59
*** djszapi_ has quit IRC07:59
djszapiwhich never does one commit per release08:00
*** heymaster has joined #harmattan08:00
djszapiand yes, please commit often for each logical change, and not a big bunch08:00
djszapithat is a basic attitude with version control systems for sanity08:00
itsnotabigtruckdjszapi: but afaict you aren't supposed to put every single code change into OBS08:01
itsnotabigtruckthat'll just clog up the build workers and stuff08:01
itsnotabigtruckcode changes go into your own repository, *releases* go into obs08:01
itsnotabigtruckso an obs commit would be once every time you bump the version number, in theory08:01
djszapiexcept that you do not just commit tarball into the osc repository08:01
djszapithat is the whole point behind it08:02
itsnotabigtruckexcept that's exactly what you least that's what everyone seems to do08:02
itsnotabigtruckwell, tarball + dsc file08:02
djszapiyou do not even know how to use it basically,  but spread the ideas ?08:02
djszapisorry, I need to ignore you08:02
djszapiseems you cannot stop it08:03
djszapiand I am not interested, and since you speak up the whole channel, pretty hard to ignore it otherwise.08:03
djszapiand that clearly shows it does not just  store the project tarball08:04
djszapiit has the debian packing is the damn thing you work on08:04
djszapiand every change can be committed for trying....08:04
djszapiand what most of the people do08:04
djszapithat is *08:05
djszapianyway, ignore here, you just waste my weekend leisure time08:05
itsnotabigtrucksheesh, not everything has to be turned into an argument08:05
Sazpaimonalright looks like it worked08:07
itsnotabigtruckSazpaimon: yay08:08
itsnotabigtruckwhich app is this, if you don't mind me asking?08:08
itsnotabigtruckfreerdp i guess08:08
itsnotabigtruckwhoops, too late :p08:08
djszapiSazpaimon: it should even ask you to make a clean build08:09
djszapiso it really is possible to not start from scratch, though I do not recommend it08:09
djszapiI always make a clean build since I would like to make sure it builds nicely from scratch after every single change.08:10
djszapiwhat you wanna use is scratchbox for fast-prototyping.08:10
Sazpaimonthats what I figured08:10
djszapito catch silly typos, and similar issues for which you do not need to have clean enviroment.08:10
djszapiso just fix your workflow :)08:10
Sazpaimonexcept a scratchbox environment can get polluted08:10
Sazpaimonas has mine08:10
Sazpaimoni had to dump it and start a new scratchbox environent08:11
Sazpaimoninstalled too many dependency packages from other places08:11
djszapiexcept that it does not matter08:11
djszapisince a typo issue in a C++ code will be the same gcc warning with any environment08:11
djszapiregardless to whether it is clean or not08:11
Sazpaimonyeah for typos sure08:11
djszapiand that is exactly what you want.08:11
djszapito fix your workflow to catch silly mistakes like that without rebuilding everything08:12
djszapiand once you have a working package, use obs, done08:12
djszapigcc error*08:12
djszapithat is how people work.08:13
djszapialterego: ping08:14
Sazpaimonpackage seems to work08:15
Sazpaimonxfreerdp -f --plugin rdpsnd -u Sazpaimon --sec nla
Sazpaimonand i get sound over pulseaudio08:15
djszapiyou are welcome...08:16
Sazpaimonindeed, thanks08:16
SazpaimonI guess I can submit this back to rzr08:17
djszapiyou can, and we will review.08:17
djszapiwe did not get any requests so far.08:19
djszapiwill try later.08:19
Sazpaimonsays its request 454308:20
itsnotabigtruckSazpaimon: you sent it to home:rzr08:21
itsnotabigtruckinstead of home:rzr:harmattan08:21
itsnotabigtruckbetter revoke it and resend08:21
djszapinot here.08:21
djszapiI will review later today08:22
Sazpaimonits not actually letting me do that08:22
Sazpaimonobs thinks the package is forked from home:rzr/freerdp08:22
Sazpaimonand there is no freerdp on home:rzr:harmattan08:23
Sazpaimonit's in home:rzr08:23
djszapithat is a different thing08:24
djszapijust submit against home:rzr:harmattan08:24
djszapiif it does not allow that, please open a bug report against upstream obs08:25
Sazpaimoneven though the package was forked from home:rzr?08:25
djszapiyes, why not ?08:25
djszapithey are for entirely different goals.08:25
Sazpaimongot it08:25
Sazpaimonokay, re-submitted request08:26
itsnotabigtruckdjszapi: home:rzr is rzr's personal zone, i think08:28
itsnotabigtruckerm, that was @sazpaimon08:28
itsnotabigtruckhome:rzr:harmattan otoh would be the community repo for anything08:28
Sazpaimonfreerdp didnt exist on the community repo, so i branched it from his personal one08:29
Sazpaimonthough I assume that was because it wouldnt build08:29
djszapiSazpaimon: please just use incremental numbers for the package version08:31
djszapiwe do not need the X.Y.Z schema twice08:31
djszapiwe do not need the "harmattan1" suffix either08:31
djszapiplus I would personally prefer bz208:31
djszapiif you are at it to change.08:31
Sazpaimondjszapi, sorry, I was basing it off of rzr's naming for it08:32
Sazpaimonhe named it 1.0.1-1.0~rzr108:32
itsnotabigtruckis it based off of an upstream debian package?08:32
itsnotabigtrucktake whatever version it uses and stick a maemo1 after it08:32
djszapiyeah, but I would not like it this way in the community repository because of complication and inconsistency.08:32
itsnotabigtruckdebian seems to have 1.0.1-108:32
itsnotabigtruckso make it 1.0.1-1maemo108:33
itsnotabigtruckor maybe harmattan1 is better but i think everybody's been using maemo, and iirc rzr himself advised using it08:33
itsnotabigtruckso go for that08:33
djszapiit is a simple fix after all08:33
Sazpaimonyeah just rename and edit the dsc, changes, and debian tarbsll08:34
itsnotabigtruckSazpaimon: erm, edit debian/changelog and regen the source package08:34
itsnotabigtruckdon't mutate the files directly08:34
Sazpaimoneh alright08:34
itsnotabigtrucki mean, you *can* but you shouldn't08:34
Sazpaimonwas too lazy to extract the file but its easier in the end08:34
djszapialso, please do not use debian repository urls in the control file.08:35
djszapiplease mention the debian authors as the original ones.08:35
djszapiplease use uesr/* sectionb08:35
djszapiso that it can be installed by the pkgmgr if someone wanna do that08:36
itsnotabigtruckoh yeah, harmattanized control checklist:08:36
Sazpaimonso rename Maintainer to Changed-By and restore the old Maintainer>?08:36
djszapiplease remove the debian history from the changelog08:36
itsnotabigtrucka) XSBC-Maemo-Display-Name08:36
itsnotabigtruckb) XB-Maemo-Flags08:36
djszapicheck out other packages how it is done08:36
itsnotabigtruckc) XB-MeeGo-Desktop-Entry-Filename (if installing a desktop file)08:37
itsnotabigtruckd) XB-Maemo-Icon-2608:37
djszapianybody can take a look a look at the debian stuff in the debian repository08:37
Sazpaimon[02:36] <djszapi> please remove the debian history from the changelog08:37
djszapiwe do not need to ship debian life history in harmattan history08:37
Sazpaimonyou mean the ubuntu specific entries?08:37
itsnotabigtrucke) use a section from the list on this wiki page starting with user/08:37
Sazpaimonleave the regular versions right? like 1.0.1-108:38
*** heymaster has quit IRC08:38
itsnotabigtruckSazpaimon: nah, always change the version if you changed anything08:38
itsnotabigtruckthat's what the maemox stuff is all about08:38
Sazpaimonitsnotabigtruck, i mean the old log entries08:38
itsnotabigtruckdjszapi: (if you aren't still ignoring me) it seems most proper to keep the debian changelog08:39
itsnotabigtruckto preserve continuity08:39
itsnotabigtruckbut changing the maintainer in the control file makes sense08:39
Sazpaimonthis is what I have08:39
itsnotabigtruckSazpaimon: perfect08:39
itsnotabigtrucknot sure why the section is stable all of a sudden :p08:39
Sazpaimoni can change it back to unstable08:40
Sazpaimondjszapi, is this a good control?08:41
*** heymaster has joined #harmattan08:42
itsnotabigtruckSazpaimon: look at the 5 items above08:42
djszapiSazpaimon: nah08:42
itsnotabigtruckmaybe skip the icon for now if you don't have a 64x64 icon ready to go08:42
djszapiyou didd not fix the things I said.08:42
itsnotabigtruckbut you should at least hit A/B/C/E08:42
djszapiactually you did not fix any of them :p08:43
itsnotabigtruckalso, if you put in an icon, please pngcrush it before base64ing it, icon bloat is the spawn of the devil :p08:44
Sazpaimonthere's no icon, its a terminal program08:44
djszapiyou do not need icon08:45
Sazpaimon[02:34] <djszapi> also, please do not use debian repository urls in the control file.08:45
Sazpaimonsorry didnt see this line08:45
Sazpaimonshould i just get rid of thos vcs lines altogether?08:45
djszapiyes pls.08:46
Sazpaimon<itsnotabigtruck> e) use a section from the list on this wiki page starting with user/08:47
Sazpaimonbroken link08:47
djszapiSazpaimon: just follow what I said pls, if you wanna get it accepted :)08:47
itsnotabigtruckSazpaimon: link works on my end08:47
Sazpaimonoh that isnt a whole link08:48
itsnotabigtruckah yeah, lol08:48
itsnotabigtruckjust up through Packaging08:48
SazpaimonSection: user/utilities08:49
Sazpaimonshould be accurate enough08:49
itsnotabigtruckSazpaimon: since there's no desktop icon then skip C08:49
itsnotabigtruckbut if you can find a good 64x64 icon you should still do D so the uninstall list looks pretty08:49
Sazpaimon[02:35] <djszapi> please mention the debian authors as the original ones.08:51
Sazpaimondidn't I do that in my pastebin?08:51
djszapi08:36 < djszapi> check out other packages how it is done08:52
SazpaimonI checked rzr's freerdp package for reference08:52
djszapithat is not the community repository.08:53
djszapiXSBC-Original-Maintainer: -> you can use this for instance08:53
itsnotabigtruckdammit, djszapi beat me to the punch08:53
itsnotabigtruckMaintainer -> XSBC-Original-Maintainer08:53
itsnotabigtruckChanged-By -> Maintainer08:54
Sazpaimonah ok08:54
itsnotabigtruckso drop the 2 Vcs- fields08:54
itsnotabigtruckdo the maintainer stuff08:54
itsnotabigtruckand add all the harmattan stuff08:54
Sazpaimoni checked some random packages in the community repo, and didnt see those08:54
Sazpaimonjust regular Maintainer lines08:54
Sazpaimon better?08:55
Sazpaimonitsnotabigtruck, i couldnt find any documentation on XB-Maemo-Flags, but what i found on google told me "visible" was what to use08:56
djszapia little bit08:56
djszapisections are still not okay.08:56
djszapiSazpaimon: please do /not/ use XB-Maemo-Flags08:56
Sazpaimondjszapi, make it user/* instead?08:56
itsnotabigtruckSazpaimon: yeah, visible is the only one you use on normal packages08:56
itsnotabigtruckdjszapi: why not?08:56
djszapiwell you have two sections for freedrp at the top08:56
itsnotabigtruckSazpaimon: the other ones are undocumented and have to do with rebooting and stuff08:57
djszapidrop 1108:57
Sazpaimonsorry i didnt notice that08:57
djszapichange 208:57
djszapiand change all the other sections too08:57
Sazpaimondjszapi, should I not bother with categories and just use user/*?08:57
itsnotabigtruckSazpaimon: i think * is a placeholder08:58
itsnotabigtrucknot a literal value you should put in08:58
itsnotabigtruckalso maybe user/network would be more appropriate?08:59
* djszapi is still waiting for the savior to let him know how to initialize a property from c++ code...08:59
Sazpaimoni wasnt sure, utilities seemed better for me08:59
djszapiwithout entering the changed signal handler.08:59
Sazpaimonshould I make the libfreerdp-dev package user/development and not user/utilities09:00
itsnotabigtruckmaybe the -dev package should be non-user/09:00
djszapisame with dbg09:01
itsnotabigtruckSazpaimon: use devel09:01
Sazpaimonuh, why09:01
djszapiand network for the libraries09:01
itsnotabigtrucki think the -dbg is debug09:01
Sazpaimonthere is no user/devel09:01
itsnotabigtruckSazpaimon: user/ packages should be interesting packages to the end user only09:02
itsnotabigtrucki.e. conceptually they should be "apps"09:02
itsnotabigtrucklibraries are not user/09:02
Sazpaimonso devel/development09:02
itsnotabigtruckneither are header and symbol packages09:02
itsnotabigtruckno, no prefix09:02
itsnotabigtruckthe program itself: user/network09:02
itsnotabigtruckthe devel package: devel09:02
itsnotabigtruckthe debug package: debug09:02
itsnotabigtruckthe lib package: libs09:02
itsnotabigtruckthat will hide the "uninteresting" packages from meecatalog etc.09:03
Sazpaimondjszapi said use network for the libraries09:03
itsnotabigtruckthey can still be installed from apt-get09:03
itsnotabigtrucktell djszapi to look at the backlog :p09:03
itsnotabigtruckSazpaimon: noooooo09:04
itsnotabigtruck> the program itself: user/network09:05
Sazpaimonyeah i changed that09:05
itsnotabigtruckfreerdp-x11 is NOT A LIBRARY, so why would it go in libs?09:05
Sazpaimonoops, i wasnt paying attention09:05
djszapistill not okay09:05
itsnotabigtruckthis isn't terribly complicated, my suggestion is simply based on mixing the meego wiki info and the debian policy manual09:05
itsnotabigtruckjust apply some logic, almost everything you're asking is easily solved based on ubuntu/debian/meego/harmattan/whatever packaging guidelines09:06
Sazpaimonokay, so make freerdp and freerdp-x11 user/network09:06
*** risca has quit IRC09:07
itsnotabigtruckSazpaimon: and everything else stays put, sounds about right09:08
Sazpaimon okay09:08
Sazpaimonmaybe i should get some coffee09:08
Sazpaimonbecause ive had a headache all day09:08
itsnotabigtruck:( that's no good09:08
Sazpaimonand coffee is the only thing that helps09:08
itsnotabigtrucksorry if i was being a little harsh just now09:08
Sazpaimonno its fine09:08
itsnotabigtrucki usually get headaches precisely because i haven't had coffee, but by then it's too late09:08
djszapidbg is still not okay :)09:09
djszapinor others09:09
djszapionly 2 and 28 are ok09:09
itsnotabigtruckSazpaimon: could you explain to djszapi that only freerdp-x11 qualifies as a "user package"09:09
Sazpaimonbasically what he said09:09
itsnotabigtruckand that the user/ prefix is not to be indiscriminately applied to anything and everything09:09
djszapihe = ?09:09
itsnotabigtruckSazpaimon: he's ignoring me right now09:10
*** risca has joined #harmattan09:10
Sazpaimonnow i have to be the middleman and he'll ignore *me*09:10
Sazpaimonbut ok09:10
Sazpaimon <itsnotabigtruck> Sazpaimon: could you explain to djszapi that only freerdp-x11 qualifies as a "user package"09:10
Sazpaimon <itsnotabigtruck> and that the user/ prefix is not to be indiscriminately applied to anything and everything09:10
itsnotabigtrucki was kinda hoping you'd rephrase that x_x09:11
Sazpaimontoo tired atm09:11
djszapiSazpaimon: please do not post what he says. I have a reason why I would not like to listen to him.09:11
djszapiI warned him many times.09:11
* Sazpaimon rolls eyes09:11
Sazpaimonokay fine09:11
Sazpaimonmake freerdp-dbg and freerdp-dev user/development09:12
Sazpaimonand make the libs user/network09:12
djszapidoes it build that way (preferrably without a warning) ?09:12
djszapicould you please try it out ?09:12
*** heymaster has quit IRC09:13
Sazpaimonthe source package builds find09:13
djszapi(scratchbox build is enough for testing it)09:13
Sazpaimonyeah hold on let me put it in scratchbox09:14
itsnotabigtrucki still don't get this, user/ is specifically for packages that are interesting to end users09:14
itsnotabigtruckand make sense to uninstall individually09:14
itsnotabigtrucka package that is only installed as a dependency of others, i.e. a library, should never be user/09:14
itsnotabigtruckand a devel package shouldn't be installed on-device anyway09:14
itsnotabigtruckthink about it this way: if you're a user scrolling through meecatalog, would you have a reason to want to install the package09:15
itsnotabigtruckthat's true for freerdp-x11, and not so true for the others09:15
itsnotabigtruckSazpaimon: just found that wiki page, most of those guidelines for maemo are still in effect on harmattan09:21
Sazpaimonitsnotabigtruck, i really dont feel like arguing with djszapi tonight09:21
Sazpaimonor in general09:21
Sazpaimonalso for some reason quilt isnt working on scratchbox09:21
Sazpaimonit can't run awk09:21
Sazpaimondespite awk being installed and runnable09:21
djszapiwhy would you argue ?09:22
djszapiif we make one package installable for the user on the device with pkgmgr, we do not do any good09:22
djszapiwe should be able to install any packages with package manager.09:22
djszapiis it not sane ?09:22
Sazpaimonokay im not sure whats wrong with scratchbox09:24
djszapiSazpaimon: if you need explanation, I can tell you in private why I ignored him after tons of warns, if you wish, but I would not like to pollute this channel with it anymore.09:25
Sazpaimondjszapi, I really dont want to get involved09:25
djszapithat is why I said, to please never refer to him.09:25
Sazpaimonbash: /usr/share/quilt/compat/awk: No such file or directory09:25
Sazpaimonlrwxrwxrwx 1 maemo maemo 17 Mar 17 03:24 awk -> ../../../bin/gawk09:25
Sazpaimon/usr/bin/gawk: ELF 32-bit LSB executable, ARM, version 1 (SYSV), for GNU/Linux 2.6.16, dynamically linked (uses shared libs), stripped09:26
Sazpaimonthats preventing quilt from working09:27
* Sazpaimon shrugs09:28
djszapiDoes anybody know how to initialize a qml element property from C++ code without entering the chanted signal handler ?09:35
djszapiSazpaimon: also, I would probably not push it against the current "community repository".09:37
djszapisince we have just been about to split and not to increase.09:37
djszapiwe will probably have a utils repository for these things.09:38
djszapiso just push it against the applications repository for now.09:39
ieatlintdjszapi: i don't think you can without doing something questionable09:42
djszapiieatlint: hey, how are things ?09:43
djszapiDo you feel a bit better ?09:43
djszapiieatlint: do you have any "questionable" idea on the tip of your tongue ?09:43
ieatlinti'm fighting with my employer, but today is a day where my ethnic heritage says i need to be drunk and eat cheap meat09:44
djszapiyou at least have employer :p09:44
ieatlintso the questionable ways aren't necessarily applicable09:45
*** lamikr has quit IRC09:45
ieatlinti've seen it done with custom qml elements, but the property value changes won't be detected inside of qml09:46
ieatlintessentially, just change it entirely on the c++ side09:46
ieatlinta quick look suggests may have some help09:47
*** risca has quit IRC09:47
ieatlintbut i've never used it... and it may not actually be different than the signal/slot method09:47
djszapiieatlint: I would like to play a little sound effect, if the selection is changed by the user (really just a little one), but it is generated while launching the application.09:48
ieatlintoh, i see, you don't want the on<value>Changed signal called for initializing the property?09:48
ieatlintthere are other ways to handle that :P09:49
djszapiyes since it is just an initilization after all.09:50
ieatlinti've done things as simple as Text { property bool init: false; text: "test"; onTextChanged: { if( init ) doSomething(); }; Component.onCompleted: init = true }09:50
djszapiI do not intend to change, just initialize.09:50
djszapiyes that was my fallback idea.09:50
ieatlintyeah, not sure there is a better option09:50
djszapiI could not just believe my eyes, I cannot initialize a qml property from c++.09:50
ieatlintwell, the qdeclarativeproperty thing could, but it'd be messier, and less efficient09:51
ieatlintand, actually, i'm not sure it wouldn't still cause the changed signal09:51
djszapimyProperty: myContextPropertyClass.myMethod();09:52
djszapiI would think it is a proper initialization syntax.09:52
*** psycho_oreos has quit IRC09:53
ieatlintshould be fine... the Component.onCompleted signal won't be called until after the properties are initialized09:53
djszapiieatlint: I will just go for the fallback, I believe.09:54
ieatlintit's best, as any other workaround will be messy, potentially result in more function calls, and will be a LOT harder for someone else to read09:55
djszapiwell, the nastyness of the fallback is that 1) I cannot declare js global variable in the same qml file 2) I cannot have read only property.09:55
djszapiwhich might still be better than spending yet another X hours with this.09:56
ieatlinton the plus side, i have beer09:56
djszapialcohol free, right ? ;)09:57
ieatlintno, i said beer09:57
* djszapi still has some alcohol free beers in the fridge09:58
ieatlintalcohol free "beer" falls somewhere between budweiser and decaf coffee09:58
*** jaywink_ has joined #harmattan09:58
*** blueslee has joined #harmattan10:10
*** blueslee has quit IRC10:11
*** adlan has joined #harmattan10:12
djszapiieatlint: there is an interesting thing10:16
djszapiI cannot hide the infobanner if I still "initialize" the property from C++10:16
djszapithe showed up infobanner remains for ever.10:16
djszapiuntil a new show, and then show up with the new text as such10:17
djszapiso it seems the initialization workaround is still not fixed :/10:17
*** cvaldemar has quit IRC10:19
ieatlintnot sure i follow10:21
ieatlintare you using that c++ method, or the js method of using a boolean and then setting it via Component.onCompleted?10:21
djszapiieatlint: the initial value comes fro mC++10:22
ieatlintalso, if it's an infobanner that you display when the value is changed, why is there an initial variable?10:22
djszapiit is like a setting that you store on the disk so when the application is launched, it restores the selected item.10:22
djszapithe infobanner is entirely different thing10:22
djszapibut it takes effect on that apparently.10:23
ieatlintand somehow the Component.onCompleted signal isn't being called in the order you want?10:23
djszapiI think it is better if I show the code10:24
ieatlintline number for the relevant element?10:25
djszapiand 96, for sure10:28
ieatlinttry removing lines 58-60, and at line 104, add else { initialized = true; }10:28
*** cvaldemar has joined #harmattan10:29
djszapiwell, that cannot be a solution, so what do you expect out of it ?10:29
djszapithe problem is not with the sound10:30
djszapisound works as expected after this workaround10:30
ieatlintoh, but randomly the infobar doesn't hide?10:30
djszapithe problem is that if I try to initialize a qml property like that, the infobanner gets dismissable.10:30
djszapinot randomly.10:30
djszapiit never hides10:30
djszapijust for a new show10:30
djszapiso it disappears and immediately appears10:30
djszapibut I guess that is how show is supposed to work10:30
ieatlintso hide isn't being called is your issue10:31
djszapiit is called10:31
djszapiconsole.log enters that line10:31
djszapibut it does not take effect10:31
djszapithe time is cool before showing up10:31
ieatlintit's as if the timerEnabled property of the infobanner isn't true (which is the default value)10:33
ieatlintor the timerShowTime is not set to a sane integer10:34
djszapiboth are ok10:34
* djszapi has the feeling an auxilliary property will be the workaround10:35
ieatlinti see "khangmanEngineHelper.hintHideTime * 1000;" ... you sure hintHideTime is <10?10:35
djszapiofc I could set the selectedIndex to the return value in the onCompleted, sleep for a bit, and then set to true10:35
djszapithat works, but super hackish :)10:35
djszapiieatlint: yes, I printed out10:35
djszapiit prints out the things according to the settings page10:36
djszapi1000, 3000, 5000 etc10:36
djszapi    function hide() {10:36
djszapi        animationHide.running = true;10:36
djszapi    }10:36
djszapisorry, for spamming:     NumberAnimation {10:37
djszapi        id: animationHide10:37
djszapi        target: root; property: "scale"; to: 0; duration: 200; easing.type: Easing.InExpo10:37
djszapi    }10:37
ieatlintcould the show signal somehow be looping?10:37
djszapiI even tried to call show() and then put a hide() statement immediately right after10:37
ieatlintwell, not sure... could be a repainting issue, but i wouldn't say that's likely10:38
djszapiieatlint: I restore certain settings on the settings page:
djszapiI do it that way.10:42
*** heymaster has joined #harmattan10:43
ieatlintyeah, that shouldn't do anything10:43
djszapithough if I use it that way10:44
djszapiI am afraid initialized true runs earlier than the signal handler code10:44
djszapimeaning that I will hear the  sound10:44
ieatlintunless that is a loop... you set the hide time, detect the hide time was changed, set the slider amount, which causes a signal change, which sets the hide time?10:44
djszapiI could put some sort of sleep in there sure ... but that is really nasty :)10:44
djszapino loop involved.10:45
djszapithe point of that code is to avoid the loop in fact :)10:45
djszapiwithout that, there is a loop10:45
ieatlintwell if the Component.onCompleted is being set too soon, then you can eliminate it10:45
djszapibut with binding qml elements there are not.10:45
ieatlinteg, onSignal: { if( initialized ) doSomething; else initialized = true; } -- the initializing changed signal then sets it initialized10:46
ieatlintso it theoretically can't be too soon10:46
djszapiit would be nice if property initialization from C++ was fixed in QML2.10:49
djszapiieatlint: I am not sure I understand the onSignal line10:51
djszapiwhy that is good from InfoBanner pov.10:51
ieatlinti don't see the direct connection between your infobanner and your initialized bool10:55
ieatlinti see where the initialized bool is set and checked10:55
ieatlintbut not how it interacts with the banner10:55
djszapime neither10:55
ieatlinthowever, if you believe that the Component.onCompleted signal is being set too early, that would be a way around it10:56
djszapimight be an infobanner bug as well10:57
ieatlintif i had to guess, the the show() method of infobanner is being called in a loop somehow, or there's a repainting issue... but i wouldn't eliminate other possibilities10:59
djszapiI do not honestly know how to solve this ituation11:03
djszapiwe do really wanna have a simple way in qml2 to initialize a qml property from C++.11:06
djszapiauxilliary variable (property) did not help :/11:10
*** heymaster has quit IRC11:10
djszapiis there a handy sleep function available for qml/js or shall I just put an empty for loop in there for this goal ?11:12
djszapipreferrably qml/qt. js probably has one.11:12
djszapiieatlint: ahh not I gotcha...11:14
djszapiso it is viable to set the thingie inside the onCompleted, but do not set the true there.11:14
ieatlintit's an idea anyway11:15
ieatlintand i think there is a simple sleep() function11:15
djszapiinitialized == true ? : initialized = true11:16
ieatlintalternatively, you can use a timer11:16
ieatlintessentially, yes11:16
*** NIN101 has joined #harmattan11:19
*** psycho_oreos has joined #harmattan11:28
*** Termana has quit IRC11:29
*** psycho_oreos has quit IRC11:30
djszapiieatlint: sorry, it is broken :p11:30
paitsnotabigtruck, not yet.. i had a couple of ideas more to try, but essentially no. do you know where i can find this package i need?11:30
jonniwhat package are you refering to?11:32
djszapiieatlint: pls wait11:33
djszapipa: though, I do not seriously recommend working against the security platform11:33
djszapiway cleaner workflow to work how it is intended.11:34
ieatlinthah, yeah, i just get a black screen indicating a qml error :P11:34
djszapiyeah, stupid assignment vs. binding character11:34
ieatlintsimple game, but it works11:43
padjszapi, well i dont11:43
pai just need aegis-su11:43
pafor testing out scripts on the fly11:43
pathanks , anyway!11:43
jonnipa: aegis-su is under NDA, only for nokia internals.11:44
*** koe has quit IRC11:44
*** koe has joined #harmattan11:44
djszapipa why would you need aegis-su ?11:44
djszapihow do you know what it does in the first place ?11:45
djszapiyou should not.11:45
djszapiif there is such a tool at all.11:46
*** M4rtinK has joined #harmattan11:46
*** Guest92708 has joined #harmattan11:48
pajonni, interesting. it is mentioned in the nokia sdk docs11:52
paas well as aegis-manifest11:52
pathat is also not  available anywhere11:52
djszapipa: they write that which package ships that11:53
djszapiand I would highly hope it is just a documentation mess up11:53
pai can link it11:53
djszapiin any case, you do not need that package.11:54
paand djszapi , again, the reason i want aegis-su, is because im trying to write a script, and i dont want to repackage it every time. i was just dropping it into the device mydocs, and running like python (as it is in the very beginning of development)11:54
djszapiaegis-manifest is available in scratchbo11:54
pabut i need some aegis credentials11:55
djszapirequest it from the manifest file11:55
djszapiand you do not play against the security platform11:55
djszapionce you request, you have it.11:55
paessentially what this guy is saying:
djszapiyes, we read that last summer11:56
djszapirequest the cred you need11:56
djszapifigure out what you need11:56
pathere you go11:56
pai know what i need11:56
djszapiI found that article earlier11:56
djszapiwhich is again, hopefully a bug11:57
pabut how do i request the credentials "on the fly"?11:57
djszapiyou do not need that11:57
djszapiwhy would you need that ?11:57
pawell basically you are telling me that i have to package everything together with an aegis manifest11:57
djszapiyes of course11:58
djszapithat is how we designed the harmattan platform security11:58
pafine, what do i have to do every time i mod that script? repackage and reinstall?11:58
djszapibut it has nothing to do with with aegis-su or whatever you dream of.11:59
padjszapi, what?11:59
djszapiyou have even been told how to do that in different ways11:59
djszapiwhat is wrong about those ?11:59
pai was trying to do just what it+s written in that wordpress article11:59
pais that playing "against the rules"?11:59
*** adlan has quit IRC12:00
*** lbt has quit IRC12:00
djszapiI do not seriously understand what you mean12:00
pain fact i probably what the exact same command line12:01
djszapitons of other devs could manage to run scripts without any super tool.12:01
djszapiI wonder why you cannot do the same.12:01
*** lamikr has joined #harmattan12:03
pawell, i havent read these different ways. if you could repeat me one. For example, can i simply replace the script installed by a package with a new one without reinstalling (from devel-su) while keeping the credentials?12:03
paheh.. no comment12:09
jonniwell you can easily develop scripts, just dont replace anything that has been installed, and run the script with aegis-exec from opensh, that way you get all capas to your script.12:20
pahmm i probably have to understand what you just told me, but thanks : )12:23
djszapijonni: thanks for repeating for zillionth time :)12:25
djszapiI kinda lost my sake for that :/12:25
jonnipa: aegis-exec --help, -a or -p adds all the credentials that you need. people should just read the documentation better.12:26
* djszapi still think that packaging is a cleaner workflow since packaging a script is almost no time anyway12:27
djszapiyou will encounter issues while developing a script in this dirty way when you try to use it nicely.12:28
djszapiwill = might12:28
jonniits maight be cleaner, sometimes its just easier to edit with vi/nano manually on device, at very early stage of script developing.12:28
djszapieasier, as in seen here many times: spares few seconds but causes headaches asking others here for days12:29
pasorry, didnt read, i tried something like this: aegis-exec -a TrackerReadAccess python testcontacts.py12:30
pahowever, i still get the same problem as before: i still get my little list of available qcontactmanagers (tracker, memory, aggregated,etc..)12:31
pabut when i try to create any of them, i get .QtMobility.QContactManager.Error.DoesNotExistError12:31
pai thought it was due to lack of credentials12:32
djszapipa seriously just go and learn about security12:32
djszapias it was advised many times on this channel, take a look at the syslog12:33
djszapiif no complains about missing credentials, nothing is missing12:33
djszapiand it is like few seconds to check12:33
djszapiinstead of poking others for days12:33
djszapijonni: also I pinged you in pm12:35
*** DocScrutinizer has quit IRC12:35
djszapi(not sure how often you read those)12:35
*** DocScrutinizer has joined #harmattan12:35
jonniah thanks, didnt notice it, now responded :)12:37
pajonni, thanks. i tried it, but in my case it doesnt print out anything. and i am not sure what syslog he meant, but in my syslog there's nothing about it12:51
jonnipa: are you calling python with the capa, or the actual script?12:52
pano no i am calling the perl script you pasted12:52
jonniaegis-exec -a TrackerReadAccess python   vs  aegis-exec -a TrackerReadAccess ./testcontacts.py12:52
djszapinot sure what syslog ? :)12:52
djszapithat is a pretty standard where it occurs.12:53
pajonni, at the moment i am trying to check it with aegis-exec -a TrackerReadAccess ./example.pl12:53
pawhich doesnt print out the credential12:53
jonnipa: and you runned the command from opensh? if not then it return empty :)12:54
djszapistill at the same topic ? run the dirty way behind opensh ? ;p12:54
pahah, that might be12:54
pabut opensh, is the one that djszapi pasted before? i read it can cause malfunctions.. and basically i dont have it installed yet.12:56
djszapiyes, it is.12:56
* djszapi has just packaged a script and it took few seconds12:57
*** heeeegua has joined #harmattan12:57
paman, the problem is not packaging it, the problem is editing it12:57
jonniopensh doesnt cause malfunctions, its the stupid users who mess up the system with it who cause malfunctions.12:57
*** diverse_izzue has joined #harmattan13:00
djszapiif the editing the problem, and not the packaging, then why not package it ?13:00
* djszapi does not follow this logic13:00
djszapi(though it is still not about initialization)13:04
djszapiis it just me not seeing Nokia N9 while publishing to Ovi ? I saw that few weeks ago as an option.13:14
djszapiI see only Symbian options.13:15
djszapierr...I chose the qt based app in a rush, heh13:16
djszapiis it possible to change the wrong selection later ?13:16
djszapi(to not re-fill all the metadata again)13:17
*** Sazpaimon has quit IRC13:23
*** Natunen has quit IRC13:27
*** natunen has joined #harmattan13:30
*** dymaxion has quit IRC13:39
*** djszapiN9 has joined #harmattan13:53
*** hhartz has joined #harmattan13:59
*** M4rtinK has quit IRC14:04
*** M4rtinK has joined #harmattan14:06
*** diorahman has joined #harmattan14:08
*** M4rtinK has quit IRC14:11
*** hhartz has quit IRC14:12
*** heymaster has joined #harmattan14:17
paso im trying to package this script, using psa. but it seems that, despite what is written in the doc, the aegis manifest doesnt get added, even when it is not empty14:29
pai assume no one here uses psa, tho? :)14:29
djszapiwhat is psa ?14:34
djszapiwow :)14:37
*** blueslee has joined #harmattan14:41
*** tbf_ has quit IRC15:20
*** blueslee has quit IRC15:24
*** tbf_ has joined #harmattan15:32
pai guess i found the way to trick the aegis with the imp module16:05
*** njsf has quit IRC16:06
*** piggz has joined #harmattan16:07
*** heymaster has quit IRC16:13
*** diorahman has quit IRC16:23
*** faenil has joined #harmattan16:23
*** niqt has joined #harmattan16:59
*** njsf has joined #harmattan17:05
*** hardaker has joined #harmattan17:06
*** heymaster has joined #harmattan17:24
*** javispedro has joined #harmattan17:27
*** javispedro has quit IRC17:27
*** javispedro has joined #harmattan17:27
*** faenil has quit IRC17:30
*** vLassi has joined #harmattan17:41
itsnotabigtruckpa: "imp module"17:50
*** njsf has left #harmattan17:50
*** ZogG_laptop has joined #harmattan17:50
*** djszapiN9 has quit IRC17:55
*** villager has joined #harmattan17:58
*** zk8 has joined #harmattan18:07
*** niqt has quit IRC18:11
*** zk8 has quit IRC18:14
*** arcean has joined #harmattan18:19
*** n9appscom has joined #harmattan18:23
*** M4rtinK has joined #harmattan18:32
payes, with it you can import a script from wherever18:51
pathat you can obviously change without the restriction of the checksum18:51
*** heymaste_ has joined #harmattan18:53
*** lbt has joined #harmattan18:55
*** lbt has joined #harmattan18:55
*** heymaster has quit IRC18:56
*** hardaker2 has joined #harmattan19:02
*** hardaker has quit IRC19:04
Jaffaalterego: belated pong ;-)19:07
itsnotabigtruckwhoa, this looks awesome19:09
ZogG_laptopitsnotabigtruck: game?19:13
*** djszapi has left #harmattan19:15
itsnotabigtruckZogG_laptop: yeah19:21
*** blueslee has joined #harmattan19:27
npmis there a way for an application to automatically add it's own "extra" keys to the virtual keyboard -- for example, cursor-control and copy/paste/select-all19:28
mgedminon the keyboard itself, I don't think so, but on the toolbar above the terminal -- yes19:29
mgedminit's also handled by maliit19:29
mgedminlook at how xterm does it19:29
mgedminthere are .xml files that describe the toolbar buttons; you can have arbitrary keys there19:29
npmbut i want it as general feature on platform. i guess that means it's not implemented yet :-)19:30
mgedminmaliit is a general feature of this platform19:30
mgedminor do you mean you want a plugin of some kind that adds those features to every app?19:30
npmas in there could be an X atom holding some string (e.g. JSON) that can hold the extra bindings, icons, actions)19:30
mgedminthe swype keyboard has arrows and copy/paste iirc19:31
npmso that any app can post a JSON string to X and then the VKB would pick it up whenever the app is being used for typing19:31
itsnotabigtruckmgedmin: i think the idea is to have context dependent layouts19:31
itsnotabigtruckfor example, the @ key is important in an email client, but the ~ key is not19:31
mgedminxterm used to have a custom layout, iirc19:32
itsnotabigtruck-~/\*"';|<> are important in a shell, but @ is not19:32
npmoh yeah i guess if you use swype specifically, there's some way of adding buttons to the bar above19:32
itsnotabigtruckmeego-terminal adds an auxiliary button bar, though it doesn't have all the buttons it ought to19:32
mgedminreplacing , with / iirc19:32
mgedminin the main vkb layout19:33
mgedminbut recent versions turned that off19:33
itsnotabigtruckand it's less convenient than if the english kb layout had a version optimized for shell usage19:33
npmis the way it achieves that documented?19:33
javispedronpm: yes, there is19:33
javispedronpm: e.g. :)19:34
*** hhartz has joined #harmattan19:34
npmjavispedro: looks perfect. but then you have to impose your own keyboard or does it use whatever user has chosen19:35
javispedronpm: it just appears on top of the user keyboard; not sure if it works with swype though (you can check in the terminal app, it is the same feature)19:35
itsnotabigtruckif meego-terminal shows up with a button bar with swipe on, then it works19:36
npmyeah i saw that, guess i need to rtsl19:36
itsnotabigtruckthere's the keyboard, and then a white bar with symbols/arrows above it19:36
javispedrobasically you write a xml file with the button defs and push it to the im-server19:36
itsnotabigtruckerm, swYpe19:36
javispedroMTF has an API for it, dunno about harmattan qt quick components though...19:36
javispedroin any case, if a sample exists, it's on the widgetsgallery19:36
*** hhartz has quit IRC19:37
javispedro that's for MTF19:37
npmgiven that it's my own app, i guess i should just put up some buttons and be done with it :-)19:37
javispedrowhat was the name of the harmattan qt quick components example app?19:37
javispedroI remember they DID some keyboard stuff19:37
javispedrofor ex. I found
npmor is there some magic vulcan-death grip that pastes what's in the cut-buffer when the VKB is up. (wp7 has that little "pastie" round thing w/ two-documents icon when there's something in the cutbuffer)19:40
javispedroit was a "design" decision that they removed the paste button from the vkb19:41
javispedroI think you can probably still make it show up changing the theme, but no idea about themes19:41
npmjavispedro: yeah that example gives different "enter" key prompts and perhaps also lets you set a numeric versus alpha kb19:41
npmjavispedro: that's too bad. the way wp7 handles cut/paste vs vkb is nice. the way it actually handles the cursor is horrid). they need to mate and produce a child...19:42
*** beford has joined #harmattan19:43
npmaka both should have harmattan's cursor-magnifier and saner cut/paste activation. harmattan should have wp7's copy activation/noise/graphics as well as paste.19:44
*** blueslee has quit IRC19:48
* npm wonders if there's a german word for the disappointment you feel when you realize you accidentally plugged in lumia instead of n9 into USB port19:48
npmschlechtenplatformgepluggenheit or something19:50
npm works in firefox on n920:00
*** cvaldemar has quit IRC20:04
*** cvaldemar has joined #harmattan20:06
pai noticed that if i enable a skype account, i get skype contacts among the contacts, while if i disable it these contacts disappear. Do you know whether these contacts are added and deleted every time, or whether is is simply the contact manager that appear and disappear?20:09
paand if so, is it possible to add a second contact manager of type tracker?20:10
*** vLassi has quit IRC20:18
*** piggz has quit IRC20:25
*** blueslee has joined #harmattan20:28
*** hhartz has joined #harmattan20:40
*** arcean_ has joined #harmattan20:41
*** arcean has quit IRC20:43
*** Sazpaimon has joined #harmattan20:43
*** Sazpaimon has quit IRC20:43
*** Sazpaimon has joined #harmattan20:44
SazpaimonrZr, ping20:49
*** arcean_ is now known as arcean20:50
*** natunen has quit IRC20:59
*** Natunen has joined #harmattan21:02
*** ajalkane has joined #harmattan21:06
*** blueslee has quit IRC21:07
*** hhartz has quit IRC21:19
*** luke-jr has quit IRC21:27
*** luke-jr has joined #harmattan21:27
*** pinheiro_ has quit IRC21:38
*** deimos has joined #harmattan21:41
*** Saviq_ has joined #harmattan21:42
*** adlan has joined #harmattan21:42
*** Saviq_ has quit IRC21:42
*** piggz has joined #harmattan21:50
*** lamikr has quit IRC22:05
*** Transformer has joined #harmattan22:07
*** Transformer has quit IRC22:10
javispedrohmrpf, when positioning is active, harm seems to love issuing wlan ap scans..22:28
*** e-yes has joined #harmattan22:38
mgedminmakes sense, I suppose22:44
mgedmin'dear nokia location server, let it be known henceforth that these wifi AP macs indicate this gps position: ..., love, a n9'22:45
javispedroexactly :)22:47
javispedroevery minute22:47
mgedminI suppose this is visible in Battery Usage?22:48
itsnotabigtruckmgedmin: is the data uploaded?22:49
javispedromgedmin: that is the problem I wanted to avoid22:49
javispedromgedmin: I used to leave the n900 for long times recording GPS tracks22:49
javispedrocurrently, with the n950 this is getting terrible battery life :(22:50
mgedminitsnotabigtruck, I know just as much as you do; my guess would be 'yes, if you agreed to use that service'22:50
javispedroitsnotabigtruck: they do, how do you believe it is populated? :)22:50
javispedroevery single phone manufacturer does it these days any way22:51
javispedro(and tablet ofc)22:51
itsnotabigtruckjavispedro: well, any idea where that upload is being done /from/?22:54
itsnotabigtruckalmost every single phone manufacturer is either apple, or using android22:54
javispedroor webos, or wp722:55
itsnotabigtruckso i don't know if that proves much besides, /those oses/ have that22:55
javispedroand symbian22:55
javispedrothe only one I don't know is BB22:55
itsnotabigtrucksaid manufacturers normally don't maintain the database or even write the software that publishes to it22:55
javispedrothere are basically two such databases only22:56
javispedrogoogle's and skyhook's22:56
DocScrutinizerand openBmap afaik22:57
DocScrutinizerwhich is mainly about BTS but also keeps records of WLAN AP22:58
javispedroheya DocScrutinizer22:58
DocScrutinizerjust had a spare 5 min to look here22:59
*** piggz has quit IRC22:59
*** piggz has joined #harmattan23:01
*** javispedro has quit IRC23:14
itsnotabigtruckwell, since it's doubtful nokia is using google's db23:24
itsnotabigtruckthen if the only databases are google and skyhook, then it'd be using skyhook if it's using any wifi positioning at all23:25
itsnotabigtruckbut do you guys *know* it's doing wifi positioning, or is it just "everybody's doing it so the n9 must have it too"23:25
itsnotabigtruckthough this wifi polling is strong evidence for it23:25
itsnotabigtruckbut it's important to know for sure a) is it happening, b) can it be suppressed23:25
mgedminmaybe it's also doing ip geolocation?23:26
mgedmin(b) definitely23:26
mgedminooh, or cell tower geolocation23:27
mgedminit's definitely not relying solely on the gps23:27
mgedminSettings -> Device -> GPS and positioning23:28
mgedmin"Allow debice to locate itself using GPS, WLAN and mobile network"23:28
mgedminalso "Allow use of mobile data connection for higher location accuracy"23:28
itsnotabigtruckmgedmin: hmm, well, if it says that then it definitely has wlan location then23:29
itsnotabigtrucki assumed the latter part has to do with agps23:29
itsnotabigtruckthough that isn't for higher accuracy...just downloding ephemeris data for quicker lock23:29
mgedminI seem to remember a popup when I first launched Maps that asked me about sending data back to Nokia to enhance their location information ...23:29
mgedminI don't trust my memory23:29
*** jaywink_ has quit IRC23:36
*** javispedro has joined #harmattan23:38
javispedrowhen booting maps for the first time there's a prompt about whether you accept "terms and conditions of positioning service"23:40
javispedroand those clearly mention that they gather wifi hotspots, cell towers, etc.23:40
javispedroI even remember they said they also use the information to "predict traffic jams" or something like that23:41
javispedro(read them in spanish and long ago)23:41
javispedroand yeah nokia uses skyhook, ms does use skyhook iirc correctly too23:42
javispedroapple too23:42
javispedrowebos uses google's though23:43
javispedrodespite using bing maps...23:43
itsnotabigtruckjavispedro: so does that "higher accuracy" option toggle use of that then?23:47
javispedroI think that only disables it when not in a wlan23:48
DocScrutinizerthe higher precision is a red herring23:49
DocScrutinizeryou can't get higher precision than with GPS23:49
SpeedEvilThan poor GPS - you can23:50
javispedroe.g. interior of buildings23:50
itsnotabigtruckhm, is there a gconf option that controls use of skyhook23:50
itsnotabigtruckor i guess some hackery would be required then to filter skyhook communication23:50
itsnotabigtrucke.g. put it into leech mode :p23:50
DocScrutinizeryou can get higher speed to first fix, and you can get lower than gps positioning in situations where GPS doesn't work at all23:51
itsnotabigtruckalso any idea if nokia's a-gps scheme is just ephemeris downloading, or is it more complicated than that23:52
javispedroDocScrutinizer: I am really surprised at what the touchpad (using MS's hotpost database, no GPS) can do23:52
itsnotabigtruckthe wiki article talks about two kinds of a-gps23:52
itsnotabigtruckone of which involves sending detailed gps data to a server23:52
javispedroDocScrutinizer: afaik they even do some triangulation based on the RSSIs of the hotspots23:52
DocScrutinizerwell, you know that doesn't worl23:53
javispedroon a urban jungle setting, it does feel more precise than GPS23:53
DocScrutinizerwhile I dunno what exactly is RSSI, assuming it's signal strength23:53
itsnotabigtrucki think it's "receive signal strength indication" or something along those lines23:54
*** tbf_ has quit IRC23:54
DocScrutinizersignal strength is highly variable, depending on local interference aka standing waves, also on shielding effects and directional anteanna behaviour23:55
javispedronote that I'm obviously not saying that it will get you a 5m fix, that's imposible23:55
javispedrobut precisely on areas where GPS might fail because they're full of signal-blocking steel buildings, it tends to work23:56
DocScrutinizertaking all those effects mentioned above you *might* get sub-meter accuracy, or be completely off by a few 100m from actual position with your guestimate23:56
javispedroand the fact that you don't even know if the hotspot's location in the DB is remotely accurate or not23:57
DocScrutinizers/taking/taking into account/23:57
infobotDocScrutinizer meant: taking into account all those effects mentioned above you *might* get sub-meter accuracy, or be completely off by a few 100m from actual position with your guestimate23:57
itsnotabigtruckwell, an obvious naive way to do it is just to interpolate based on the rssi compared to samples of gps position + bssid + rssi23:57
DocScrutinizerjavispedro: on pos of AP: you definitely dunno shit about where it is23:58
DocScrutinizerthe database been constructed from readings like yours23:58
* javispedro nods23:58
DocScrutinizerso they just guess23:58

Generated by 2.15.1 by Marius Gedminas - find it at!