IRC log of #harmattan for Saturday, 2012-02-25

itsnotabigtruckbindi: sorry, was out for lunch00:03
itsnotabigtruckor ejabberd, but prosody is newer and lighter00:03
bindiitsnotabigtruck: i'm already using openfire, and only to gateway the messages00:03
itsnotabigtruckdon't use openfire, it's no longer maintained00:03
itsnotabigtruckand it has some severe bugs that probably won't ever be fixed00:03
itsnotabigtruckalso it has a lot of bulk (like the web admin interface) that you don't really need00:04
itsnotabigtrucki have to admit, openfire's web interface is really nifty, but it's also dead weight00:04
*** pinheiro has joined #harmattan00:04
rZrdjszapi: back00:07
*** rZr is now known as RzR00:07
RzRdjszapi: so you want a app section into home:rzr:harmattan:apps ?00:08
*** xnt14 has joined #harmattan00:10
*** aquarius has joined #harmattan00:11
itsnotabigtruckis there anything like ^ that accepts something other than an SVG as input?00:23
itsnotabigtrucki need to do an icon for which no svg exists00:23
ieatlinti recall there being photoshop icon templates released by nokia to help you do the harmattan styling00:25
aquariusNot sure if this is the right place to ask this. I've recently acquired an N9, which I rather like, but there are a couple of browser issues: primarily that the browser's a bit slow sometimes, and that it doesn't support HTML5 audio. Are these things likely to be resolved by harmattan?00:26
itsnotabigtruckieatlint: i'm hardly a graphics pro though and i don't have PS installed right now00:26
ieatlinti used gimp00:26
ieatlintit worked well enough with the ps templates00:26
itsnotabigtruckaquarius: for html5 audio...i think that might be in pr1.2 (??)00:27
ieatlintwas mostly just icon sizing, and then creating the rounded rect with the proper shading gradient00:27
itsnotabigtruckfor the slowness i doubt that's going away00:27
*** Venemo_N950 has joined #harmattan00:27
itsnotabigtrucki think i'm going to skip harmattanizing the icon for now00:27
*** Venemo_N950 has quit IRC00:27
*** Venemo_N950 has joined #harmattan00:27
bindiitsnotabigtruck: so how do calls on xmpp work?00:27
itsnotabigtrucki just need to do the necessary stuff to bind the desktop icon to the app uninstaller00:27
itsnotabigtruckbindi: jingle00:28
itsnotabigtruckwhich is an a/v layer for jabber00:28
bindiitsnotabigtruck: how does it work on the N9? how does it work on android?00:28
itsnotabigtruckno idea about android00:28
itsnotabigtruckwith n9, you can make a call from your contacts list00:28
aquariusitsnotabigtruck, that's helpful, thanks! I could test that, perhaps...?00:28
itsnotabigtruckhowever, i'm not sure you can actually perform a call yet from N9 with a server that isn't gtalk00:28
itsnotabigtrucki've been meaning to write the gui layer to do that, but haven't yet00:28
bindiaahhh how can calls and messages in the same application be so hard00:29
itsnotabigtruck(and nokia should have done it in the first place *grumble grumble grumble*)00:29
itsnotabigtruckaquarius: supposedly pr1.2 is coming out next week00:29
itsnotabigtruckof course everyone associated with nokia refuses to say anything00:29
itsnotabigtruckdon't quote me on the html5 audio, i'm not sure00:30
ieatlintrumours say n9 pr1.2 is coming out 2/2900:30
ieatlintbut totally unconfirmed00:30
aquariusis there OTA upgrade?00:30
aquariusI'm really starting to like this N9, so I'd like to do some development for it. :00:30
ieatlintaquarius: ota, yes... but based on past experiences, backup everything off the device before doing it00:31
aquariusooh. there's the voice of someone who was burned last time :) I'll bear that in mind!00:31
bindiitsnotabigtruck: it seems that newer versions of asterisk support simple - i'm gonna try it00:31
aquariusat the moment I'm vacillating between my nexus S and the N9, because I'm too chicken to cut down my SIM to a microsim :)00:32
ieatlintaquarius: in the 1.0 to 1.1 upgrade, a LOT of people were burned00:32
Jarei hope they have fixed some annoying bugs in the contacts app00:32
itsnotabigtruckaquarius: i'd recommend reflashing the phone instead of using the ota upgrade00:34
itsnotabigtrucklatter is too risky00:34
itsnotabigtruckthough the former will nuke all your apps00:34
aquariusso ota might do something worse than removing apps? like, say, bricking the phone?00:34
itsnotabigtruckbricking the phone, no...even if that happened it would be recoverable00:35
itsnotabigtruckbut typical issues have been only parts of the system being upgraded, or arbitrary stuff breaking00:35
itsnotabigtruckbindi: prosody is sooooo much better than asterisk :p but ok00:37
itsnotabigtruckmake sure to use SSL on your server00:37
itsnotabigtruckin fact, enforce the use of ssl on your server00:37
itsnotabigtruckget a free ssl certificate at
aquariusitsnotabigtruck, ah, right. Still a bit worrying. I appreciate the warning :)00:37
itsnotabigtruckcan someone point me to how to link the .desktop file to the app so it can be uninstalled from the home screen?00:38
itsnotabigtruckthe low level harmattan documentation is really seems like there's a general implicit assumption that you won't be poking around in the bits of stuff because qt creator automates everything00:38
itsnotabigtruckah, it's
Jarenow why would someone compare asterisk to an xmpp server00:40
Jarebindi: i would recommend trying Freeswitch. It's a much fresher platform than Asterisk00:41
bindidoes it support SIMPLE?00:42
bindi(already building asterisk)00:42
bindiif this goes poop, i'll try it00:42
itsnotabigtruckjare: freeswitch is bloody gigantic and breaks every rule in the book when it comes to proper packageable software00:43
itsnotabigtruckthe only way to run it is out of a folder in /opt full of hacked-up library dependencies00:44
itsnotabigtruckthat said, asterisk is terrible too00:44
itsnotabigtruckalso bindi: can't you load your distro's package?00:44
bindiitsnotabigtruck: way old version there00:45
Jarewell if you're in a need of an xmpp server then freeswitch or asterisk isn't definitely your choice00:45
bindino, i don't need that00:46
Jareif you need a solid voip server, then i would recommend freeswitch00:46
itsnotabigtruckJare: s/he is comparing them because both do both text and voice00:47
itsnotabigtruckexcept that voice is a second class citizen on xmpp and text is a second class citizen on sip00:47
itsnotabigtruckbut if you need both and want just one system, then you have to choose00:47
*** jabis has quit IRC00:49
*** stroughtonsmith has quit IRC00:53
Jaremmkay, I'm using Openfire for xmpp and Freeswitch for sip. Voice features in xmpp servers don't meet my requirements and chat features in sip servers are lacking features too00:54
Jareand I hate the fact that telepathy sofiasip doesn't support presence on the N900:55
Jareeven though sofia sip stack does support it00:55
itsnotabigtruckjare: hmm...might it be possible to swap out telepathy-sofiasip for one that does?00:57
bindiJare: freeswitch configuration, how hard is it?00:58
bindii'm stuck at asterisk not working with my old configuration files, and they're giving me a huge book to read which is a suicide00:58
itsnotabigtruckbindi: prosody, doooo eeeetttttt00:58
bindiitsnotabigtruck: but prosody haz no SIMPLE00:59
bindii've got an application for android here that does calling and messaging with sip and simple00:59
itsnotabigtruckerm, of course not, it's an xmpp server...00:59
itsnotabigtruckthen change apps00:59
bindidon't have any voice apps for android00:59
itsnotabigtruckthere is absolutely no reason to use simple, ever00:59
itsnotabigtruckit's garbage00:59
itsnotabigtruckit's been obsoletefor close to a decade00:59
bindiyeah well all i care about is getting this to work :p01:00
bindiJare: do you use simple with freeswitch?01:01
*** rm_work has quit IRC01:03
Jarebindi: freeswitch uses xml in the configuration files and if you don't like configuring them manually you can use a web gui like bluebox. Btw. bluebox works with Asterisk too01:04
bindii don't like reading heavy documentation01:05
bindiold config nolonger works01:05
bindiUnable to create channel of type 'SIP' (cause 20 - Unknown)01:05
* bindi kills self01:05
bindior maybe it does01:06
*** jabis has joined #harmattan01:06
*** fiferboy has quit IRC01:07
Jarewell both systems are designed for big implementations, so the complexness kinda comes with it01:07
itsnotabigtruckdoesn't bluebox involve giving some website root access to your server to instal it?01:08
itsnotabigtruckno thanks01:08
itsnotabigtrucki'm sticking with my recommendation of using xmpp and prosody because it's orders of magnitude lighter than this voip software you're trying to load01:08
itsnotabigtruckand the voip industry hasn't had its heads on straight for a long time01:08
Sazpaimonso last week i bought a server from 1and1 to replace my older 1and1 server thats now being sold for less01:08
Jareumm no? It's a bunch of php(?) scripts that you install on your webserver01:08
Sazpaimoni call to ask what the problem is01:09
Sazpaimonand they tell me that last friday they got their last eqipment shipment and already exhausted it01:09
Sazpaimonand they dont have an eta on their next shipment01:09
itsnotabigtruckjare: ah, i was thinking of whistle (by the same outfit)01:10
itsnotabigtruckyou're supposed to enter your credentials into a "deploy" page01:10
Jareright, I wouldn't do that :)01:10
*** jabis has quit IRC01:11
bindiahha it works now01:12
bindiwas my error01:12
*** admiral0 has joined #harmattan01:23
*** gabriel9 has joined #harmattan01:23
itsnotabigtruckhow do you mark additional files in the debian directory for inclusion in the debian tarball01:25
ieatlintyou can specify that in the qt project file and it will package them01:26
itsnotabigtruckieatlint: not using qt creator...01:28
itsnotabigtrucki wasn't under the impression qt creator even *can* generate source packages01:29
ieatlintah... yeah, i've always loathed packaging debs01:30
ieatlintactually i've basically loathed packaging in any format01:31
itsnotabigtruckieatlint: the source package thing is a real problem01:32
itsnotabigtrucksource packages are essential for obs and being all opensource friendly in general01:32
itsnotabigtruckbut everyone is being coddled on qt creator, which requires significant finangling to produce them iirc01:32
ieatlinti would call it shielded vs coddled :P01:32
itsnotabigtruckbtw the files were actually automatically included, i was looking at the wrong tar.gz01:33
ieatlintbut yeah, i haven't touched obs for about 2 years01:33
itsnotabigtruckanyway, at least qt creator is a step forward from dpkg -b manually created debs01:33
itsnotabigtrucklike everyone did with maemo501:33
ieatlintthere were helper scripts as i recall01:34
itsnotabigtrucki believe i have milkytracker almost 100% harmified01:34
*** niqt has quit IRC01:34
itsnotabigtruckargh, turns out it still isn't integrating with meegotouchhome01:37
itsnotabigtruckieatlint: any idea how meegotouchhome knows that an icon matches up with an uninstallable package?01:37
itsnotabigtruckXB-MeeGo-Desktop-File-Name: milkytracker didn't do it01:37
ieatlintnope, not looked into it01:38
*** Free-MG has quit IRC01:38
*** gabriel9 has quit IRC01:45
*** NIN101 has quit IRC01:45
*** jabis has joined #harmattan01:54
itsnotabigtruckugh, i'm getting really tired of pkgmgr leaving around stale dpkg lock files01:57
* ieatlint agrees01:59
ieatlintbeen bitten by that too much01:59
itsnotabigtruckalso, it looks like if you install a package via dpkg, the icon is never bound to the package02:06
itsnotabigtruckyou have to use apt or pkgmgr for that02:06
*** Natunen has quit IRC02:12
RzRitsnotabigtruck: make sure you're killing it before using apt02:16
itsnotabigtruckrzr: is pkgmgr actually doing anything in that case?02:20
itsnotabigtruckwhat happens is that it creates the lock file after doing an operation through pkgmgr02:20
itsnotabigtruckand then never removes it, ever02:21
djszapiRzR: yes, harmattan:applications02:21
djszapimy vote is for harmatttan:applications, and not harmattan:apps02:22
RzRi'll upload my 1st app then02:22
itsnotabigtruckoh yeah, what did you guys end up deciding on02:22
itsnotabigtruckw/r/t splitting up the repo02:22
djszapibeford: have you figured out the issue ?02:23
djszapiitsnotabigtruck: nothing special, I proposed having an applications repository as the first step02:23
djszapiRzR: we need to warn the people02:23
djszapiI will extend the summary to the middleware repo02:23
itsnotabigtruckimo calling the utilities repo "middleware" is rather lame02:24
itsnotabigtruckthat word is tainted02:24
djszapiRzR: please write to the forum once you created the applications repository.02:24
djszapiitsnotabigtruck: I do not think so since that what it is02:24
djszapiplus play nice and consistent with the rest of community (nemo, plasma active, mer and so forth)02:25
itsnotabigtruckexcept it's not...if i understand correctly, it's for non-graphical programs and libs02:25
itsnotabigtruckthat isn't middleware02:25
itsnotabigtruckmiddleware is a label tacked onto software that is a) boring, and b) usually costs money02:25
djszapino, you misunderstood02:25
itsnotabigtruckfor example, smart cards often don't have drivers, they have "middleware"02:25
djszapimeh, I do not care about your ranting.02:25
itsnotabigtruckso you have to pay for the drivers separately02:25
itsnotabigtruckor stock markets buy "message bus middleware"02:26
djszapiI know qt, middleware, etc is bad.02:26
itsnotabigtrucki never said qt is "bad", not everything is binary you know02:26
itsnotabigtrucki'm just saying here that middleware is a poisoned word02:27
djszapiit is your opinion, which is against about everybody else out there.02:27
djszapiwe, nemo, mer and plasma active.02:27
*** Venemo_N950 has quit IRC02:28
ieatlintwait, you're criticizing the toolkit as being middleware?02:28
itsnotabigtruckresult #3: "Oracle Fusion Middleware"02:28
ieatlintis this one of those "i only write in assembly and draw stuff by hand to the framebuffer" things?02:28
itsnotabigtrucki'm not criticizing middleware, i'm criticizing use of the word middleware02:28
RzRdjszapi: fell to edit :
itsnotabigtrucki don't understand what you don't understand about words having unwanted connotation02:29
djszapiRzR: thanks.02:29
RzRi am uploading redactilo02:29
ieatlinti'd call qt the toolkit.. not middleware02:29
djszapiRzR: very cool.02:29
RzRbut please keep existing apps02:29
RzRin the current shared repo02:29
RzRfor now at least02:30
djszapiieatlint: except that qt is not in the repo.02:30
ieatlintmeh, apt is just middleware, i install everything directly ;)02:31
djszapinot following.02:32
ieatlintmeh.. i'm not funny anyway02:33
*** tbf has quit IRC02:33
RzRdjszapi: let's wait and see where it'll lead now02:34
djszapiRzR: fair enough. Write to the forum we will move the apps soon.02:34
djszapiin that repository only02:34
djszapiso it is sort of "depricating" the apps in the community repository.02:34
djszapiand in few weeks, we can clean up02:35
RzRwell i dont think this is a so good idea02:35
RzRmost pple will get confused02:35
djszapiabout ?02:36
RzRtoo many sources02:36
RzRand unstable undefined policy02:36
RzRso now we have pple using02:36
RzRno one switch to02:37
RzRand i feel we should put middleware into02:37
djszapiRzR: no because 1) it is more than libs 2) we should not break that too for nowe02:37
djszapijust one small step this time02:38
RzRso i would keep the crap into home:rzr:harmattan02:38
itsnotabigtruckfor names: how about: system, utils, support, tools, core02:38
djszapithere is no "core.02:38
itsnotabigtruckany one of those instead of middleware02:38
djszapithat is provided by Nokia02:38
itsnotabigtruckok, then pick something else from the list02:38
djszapiand no, I would not personally like that many repositories02:38
itsnotabigtruckthose are alternatives02:38
itsnotabigtrucki.e. there would be 1 repo out fo those02:39
* RzR feels skizo , maemo means meego, harmattan means maemo6 and shared repo is not apps4meego and apps4meego is not .....02:39
djszapiI do not think we need that for now02:39
djszapimaemo means meego ?02:39
RzRsorry i said a mistake02:40
RzRi wanted to say meego is maemo02:40
RzRis it clearer now ?02:40
itsnotabigtruck@everyone: give it a try and let me know if it works02:40
RzRitsnotabigtruck: cool is it finger friendly ?02:40
itsnotabigtruckrzr: nope :p02:40
djszapiRzR: I honestly dislike breaking the current community repository term02:40
djszapithe smallest break the better, which is harmattan:applications for now.02:41
RzRa4m or shared one ?02:41
RzRitsnotabigtruck: i did try to use it on a android tablet , it was less usable than on n8x002:41
itsnotabigtruckrzr: i'm planning on switching to the low-res 320x240 ui and resizing it to fit the N902:42
itsnotabigtruckwell, if i end up devoting any more time to this02:42
RzRdjszapi: wanna test redaktilo ? and clone the sources ?02:42
itsnotabigtruckbut right now it has these teensy 3mm tall buttons :p02:42
djszapiRzR: yes, but I do not have time, sorry.02:42
RzRdjszapi: this is just a basic text editor ... can be useful someday02:43
djszapiRzR: I would propose four weeks for the interim period about the applications in the community repository.02:43
itsnotabigtruckalso btw what do you guys recommend for version numbering on modified versions of upstream debian packages02:43
itsnotabigtruckright now i stuck a +harm1 suffix but i'm not sure if that's really kosher02:43
RzRI dont think i will take more time to polish it since it is usable enough for me now02:43
djszapiRzR: people have four weeks to add one line, if they really use apps in there for some reasons.02:43
RzRyou're going to kill that repo :)02:44
djszapiwhat do you mean ?02:44
RzRi feel the users should tell us what to do02:44
RzRif they dont care or understand of those extra repo .. then noone will use it02:45
itsnotabigtruckwhy couldn't/can't we have something like extras :|02:46
djszapido not understand an apps repo concept, and adding one line ? :)02:46
djszapithen mer, nemo and plasma active users would die out :D02:46
RzRmer users and n9 users may differ too02:46
djszapiI am sorry, but I cannot accept it.02:46
RzRstart a survey and let them decide02:47
djszapiif we have users not "understanding" one liner addition, that is very sad, whereas I do not think that is the situation.02:47
RzRi am not against your proposal02:47
djszapiokay, I will write the proposal soon.02:47
djszapito the forum.02:47
RzRthis is a good start yes02:48
djszapithough, not not tmo because I dislike that trollish site.02:48
RzRf.m.o then ?02:48
djszapiI would prefer that, yes.02:48
RzRno problem i'll send the link on tmo02:49
RzRthis project is kinda weird sometimes ...02:49
djszapiwhy are you so negative ?02:49
RzRi am not02:49
RzRthere is so much fragmantation02:50
RzR2 forums02:50
RzR2 supported repo02:50
RzRand ours02:50
djszapi2 supported repo ?02:50
djszapiwhat do you mean ?02:50
RzRovi and a4m02:50
djszapia4m is not supported by Nokia02:50
RzRyes i know02:51
djszapialso, I understand people why they moved out of the tmo site.02:51
djszapimany people dislike that site (including me).02:51
itsnotabigtruckheh, i was 4 shells deep just now02:51
RzRbut i fell there is some redundancy ...02:51
itsnotabigtruckdjszapi: well, it's authorized by a nokia package - not supported by nokia, but approved by nokia02:52
RzRactually i prefer tmo to fmo , because i fell that tmo may be more future proof02:52
djszapiRzR: I disagree02:52
djszapiespecially because the wiki page is already on meego.02:52
itsnotabigtruckand FMO is more or less kaputt, and could disappear when lifo/nokia/intel get sufficiently bored with meego02:52
itsnotabigtruckTMO is still quite active and probably isn't going anywhere02:52
djszapirzr: so if any, I feel tmo the reason of the "fragmentation".02:53
RzRcould be02:53
RzRitsnotabigtruck: i agrea w/ u . so i am installing that tracker on n95002:53
djszapiRzR: and I second Stskeep's opinion about it, really.02:53
RzRi trust him too02:54
RzRitsnotabigtruck: try again02:54
RzRitsnotabigtruck: this icon is nice but does nothing on click :)02:55
djszapiitsnotabigtruck: that is just silly saying02:55
itsnotabigtruckrzr: hmmm02:55
djszapimeego is for the linux foundation, not Nokia/Intel, and lifo (what?)02:55
RzRi needed to reinstall sdl :)02:55
itsnotabigtruckrzr: run /opt/milkytracker/milkytracker from a shell02:55
itsnotabigtruckoh ok02:55
itsnotabigtruckuse pkgmgr install-file -f, that takes care of dependencies and the icon-package association02:56
itsnotabigtruckdjszapi: linux foundation = lifo02:56
itsnotabigtruckthe point is, meego is even more dead than harmattan is02:56
djszapiso the thing is that it is already a community thing, not company.02:56
itsnotabigtruckneither intel or nokia have any intention to use straight meego02:56
djszapiitsnotabigtruck: the term is not.02:56
djszapiactually apps4meego has quite prospereous future.02:57
RzRitsnotabigtruck: using n950 hw kb is ok02:57
itsnotabigtruckand intel and nokia control the website, plus lifo sorta (but they're mostly just acting as a proxy)02:57
itsnotabigtruckrzr: cool :)02:57
djszapiitsnotabigtruck: ... and takes a little effort to clone the sites, and actually some people probably already did ?02:57
RzRitsnotabigtruck: were you at the tizen fosdem prez of DF ?02:57
itsnotabigtruckrzr: i wasn't at fosdem, so no02:58
djszapiitsnotabigtruck: also, it is highly unlike a big company like Nokia or Intel would close it (especially) without warning02:58
RzRdjszapi: sorry02:58
djszapiRzR: no02:59
RzRIf i understood well, it looked that those corps dislike throwing their money away into dead communities...02:59
djszapiI did not even take the free tizen t-shirt.02:59
itsnotabigtruckthere's lots of examples of big companies shutting down community websites with no notice at all02:59
itsnotabigtruckor with a few days warning02:59
djszapithat would be plain silly.02:59
djszapiand bad reputation for Nokia or Intel.03:00
itsnotabigtruckyou wake up one day and the link redirects to the company website, or a "the forum has been shut down, go buy all our playsets and toys" or whatever03:00
djszapiso I do not think they "can" do it.03:00
djszapiand as I said, some people already cloned it anyway03:00
*** M4rtinK has quit IRC03:00
RzRanyway i wish we'll be able to use obs for side projects too03:00
itsnotabigtruckbtw what's the most proper way to insert the aegis manifest from debian/rules03:01
djszapiI do not know how this closing is any relevanty...03:01
djszapitmo can be closed even earlier...03:01
itsnotabigtrucksince it doesn't work automatically on OBS03:01
djszapisince it is not company driven03:01
djszapiso it is not their purpose to not get bad reputation03:01
djszapiso imho, meego is even safer in this sense.03:01
djszapiespecially because maemo was dead earlier than meego03:02
RzRitsnotabigtruck: can u package an sample .mod into milky.deb ?03:02
djszapiand as I said: "tmo" is very trollish.03:03
itsnotabigtruckrzr: just unzip something from
RzRthis is the price of freedom03:03
djszapimany people feel like avoiding that tone happening in there (including me).03:03
RzRitsnotabigtruck: i am unziping mine :)03:03
itsnotabigtrucki did find the audio to be a bit stuttery...if you have any ideas on how to resolve that i'm all ears03:03
itsnotabigtruckthe problem seems to be really high cpu usage03:03
djszapiso arguing about meego being that is a very weak argument since maemo was even earlier dead.03:04
itsnotabigtruckpulseaudio, the kernel, and milkytracker each take a big chunk of cpu time03:04
itsnotabigtruckdjszapi: but maemo being dead doesn't have anything to do with being dead03:04
itsnotabigtruckboth of them have x-factors03:04
djszapiyes, it has.03:05
djszapithe author thinks one day it is cool to close, it is gone03:05
itsnotabigtruckbut i'm basing my opinion on that i think it's less likely that the owners of will get bored and shut it down03:05
djszapiand since it is not driven by big company, bad reputation is not a problem03:05
djszapiso it is even less safe03:05
itsnotabigtruckthan it is that intel/nokia/lifo will take the final steps to terminate meego03:05
djszapiso you just remind me, meego is the way of going03:05
RzRitsnotabigtruck: it's loading but silent03:06
itsnotabigtruckbesides, harmattan is much more maemo than meego, despite the name03:06
RzRno its ok03:06
itsnotabigtruckfor all intents and purposes it's maemo 603:06
itsnotabigtruckharmattan stuff isn't really very welcome in meegoland03:06
djszapithis is very weak arguing03:06
djszapiespecially because it is called meego03:06
djszapidoes not matter what is in the background03:07
itsnotabigtruckit's called meego, but it is maemo03:07
itsnotabigtruckwhat something is called has very little to do with anything03:07
djszapiplus apps4meego already uses it, so does meego community obs and so on.03:07
itsnotabigtruckmarketing departments aren't governed by logic :p03:07
djszapiyou do not get the point03:07
djszapiit is not about marketing.03:07
djszapiit is about clear consistency.03:07
djszapiand it seems apps4meego, meego community obs, meego forum etc happened to be created by the community.03:08
djszapiand honestly: tmo is something I would not follow03:08
itsnotabigtruckharmattan = direct successor to maemo5, used to be called maemo6, produced by the same company as maemo503:08
itsnotabigtruckmeego = has almost no commonality with either harmattan or maemo503:08
djszapiI think I know the details much better than you :)03:08
djszapibut you are just hijacking03:08
djszapinot concentrating to the point, so I leave you with this feeling. :D03:09
RzRitsnotabigtruck: i was able to load that .xm track
itsnotabigtrucki'm sure you do, which is why i find your positions so strange03:09
RzRbut djszapi is somehow in truth03:09
djszapitwo is a trollish site03:09
djszapimany people like03:09
RzRbecause if noone use they will shut it down03:09
djszapieven "known" community members03:10
djszapiit is driven by who btw ...03:10
djszapiI do not even know03:10
RzRso it makes sense to keep both project/communities alive03:10
RzRbut i fear it will cause a division effect03:10
djszapiwait a second: it is not "djszapi"'s truth.03:10
djszapiI was not starting to pick up the meego term everywhere.03:11
djszapiI just adopted myself what the community wanted.03:11
djszapiand as I said, I find tmo trollish enough to take part in there.03:12
RzRactually i dont care03:12
djszapiat least Intel and Nokia put a lot of effort to create a healthy community, and that has signs still.03:12
RzRhealthy but a bit in coma state :)03:13
djszapinot really.03:14
djszapimany people went to mer, nemo and pa for continuing something cool03:14
RzRthe year you died twice dont expect your next reincaranation (tizen) to be nirvana :-)03:14
djszapiand not back to "tmo", and that community.03:14
djszapiRzR: seriously, just read what Stskeeps said about the maemo bits03:15
djszapiI second his every word.03:15
RzRhey guys I know the solution !!!03:15
RzRlets open talk.harmattan.org03:15
RzRand they kill it for talk.failedcommunities.org03:16
RzRand then ...03:16
RzRwell enough talk , let's get back to reality03:17
RzRitsnotabigtruck: do you want me to try to build that tracker since it's working using n950 kb03:17
RzRitsnotabigtruck: it used to depend on jack03:18
*** jaywink has quit IRC03:18
djszapiRzR: also, I will already put the apps I maintain into the apps repository.03:21
RzRwhich ones ?03:21
djszapias I said, the ones I maintain.03:22
RzRthe kde based one I suppose03:22
djszapinot just, I packaged around 100 packages last summer, if not more.03:22
djszapiand I recall there were some applications in there.03:22
RzRi need to explain something03:22
RzRdid u notice the propose of the 2 builders ?03:23
djszapithe KDE packages will be eliminated either way03:23
djszapinot just the applications, but also the libraries and so forth03:23
RzRone will depend on the core packages03:23
RzRand the other will depend on others repo03:23
djszapiI will collaborate most likely with the PA guy, but we need to discuss it in 10 days at the PA sprint.03:24
djszapiso I will pull out the whole KDE stack most likely from the community repository.03:24
RzRthat's how i solved so unresolvable deps03:24
RzRok i trust you about all kde subprojects since i am not using it on harmattan03:25
RzRbut i did try  aktive on a lenovo s103t03:25
djszapinot you wrote more times "aktive", just felt to correct it. :)03:27
RzRi think you should correct the project name03:27
RzRyou kde guys educated us to put k anywhere03:28
RzRnow you have to deal with users :)03:28
djszapinot really03:29
djszapiactually that is a very ancient story, already03:29
djszapialmost nobody puts "k" into such terms03:29
djszapiyou are 5-6 years outdated :P03:29
RzRmy editor has a K in it :-)03:29
djszapi... which was born like hell long ago ? :)03:30
RzRmy customers are KKK party03:30
itsnotabigtruckrzr: i'll upload the source package03:36
itsnotabigtruckit would be nice to have a forum dedicated to harmattan03:37
itsnotabigtrucki mean, one that's actually half decent03:37
itsnotabigtruckproblem is that it's damn near impossible to get a new forum going03:38
*** aquarius has quit IRC03:44
RzRitsnotabigtruck: just reading this
RzRitsnotabigtruck: do you know about serious native wp7 hacks ?03:49
itsnotabigtruckrzr: nope, never was really that interested in WP703:49
itsnotabigtruckbut i know such hacks exist03:49
itsnotabigtruckwell, sort of, iirc there's been setbacks03:50
itsnotabigtruckfrom my experience with openzdk i want to have nothing to do with winCE again03:50
RzRis there some hope we could use qt someday ?03:50
RzRyes I can believe you03:51
itsnotabigtruckwell, there's a port of qt to wince already, but it's based entirely on native winCE apis (the ones used with winmo6 etc.)03:51
itsnotabigtruckboth zune hd and wp7 have their own heavily customized frontends that don't interoperate with the native apis at all03:51
RzRI did have to support directshow on winmo5 IICR03:51
djszapiRzR: yes03:51
djszapipeople have been working on Qt at Nokia for a quite while.03:51
RzRbut WP7 is an other beast03:51
RzRI met some nokia+ms pple at some "nokia day"03:52
RzRthere were no hope for me03:52
itsnotabigtruckdo you guys know the correct way to add the aegis manifest at the end of the build03:53
itsnotabigtruckthe wiki suggests hardcoding the package name into debian/rules, that will not do :p03:53
RzRthat nok day everything that was related to nokia was presented as "our current partner" by the ms guy03:54
RzRitsnotabigtruck: what privilege do you need for that app ?03:54
itsnotabigtruckrzr: GRP::pulse-access, but that's besides the point03:54
itsnotabigtruckevery app that requires asserting privileges requires manually adding the manifest to build on OBS, and i can't figure out how03:55
RzRdo you have a debian/*.aegis file ?03:56
djszapipeople can help more if you tell what rules you use ...03:56
djszapidh or cdbs ...03:56
djszapiotherwise really hard to help with.03:56
itsnotabigtruckfound it, > aegis-deb-add -control debian/foo/DEBIAN/control .. debian/foo.aegis=_aegis03:58
itsnotabigtruckyou have to put that immediately after dh_builddeb03:58
itsnotabigtrucki was wondering why aegis-deb-add had that funky control argument, that's why03:58
itsnotabigtruckand rzr: that only works on scratchbox builds03:58
itsnotabigtruckit's ignored on OBS builds because OBS doesn't contain the harmattan version of debhelper/etc.03:58
itsnotabigtruckor something along those lines03:59
djszapithough, that is not the most correct line.03:59
djszapibut it is all documented, and there are examples (many) out there.04:00
itsnotabigtruckdjszapi: that's straight out of
djszapithey write "[ ! -f debian/$(notdir $@).aegis ] || aegis-deb-add -control debian/$(notdir $@)/DEBIAN/control .. debian/$(notdir $@).aegis=_aegis " line04:01
itsnotabigtrucklook a little bit further down04:01
djszapibut yes, it is clearly documented :)04:01
RzRdjszapi: lets cross our finger but does it worth it ?04:01
djszapiitsnotabigtruck: why would I ?04:02
djszapiyou asked for rules file entry, and that is what it should04:02
RzRi will try to build that qml on android now04:02
djszapiRzR: worth what ?04:02
RzRanything related to ms04:03
itsnotabigtruckdjszapi: because that's not the rules file entry mentioned in the nokia page i linked04:03
RzRnokia said i should get a wp7 device04:03
djszapiRzR: yeah, why not, if it supports qt ?04:03
itsnotabigtruckyou're looking at the " command" part04:03
RzRbut i never recieved it04:03
djszapiitsnotabigtruck: no04:03
RzRand i did not feel the need to ask for it :)04:03
djszapiitsnotabigtruck: please try to read, it is not that hard ;)04:03
djszapi"Add the following lines into the rules file of your application package:"04:03
RzRdjszapi: i'll reconsider the question the say i'll a decent qt demo :)04:04
itsnotabigtruckread the next paragraph04:04
itsnotabigtruckdon't be pedantic when you aren't sure04:04
djszapiitsnotabigtruck: seriously, just read04:04
djszapithey say add a call in general04:05
djszapiand they describe the exact syntax above04:05
djszapiI cannot believe it is that hard ...04:05
djszapiRzR: I saw something, but I am afraid I cannot talk about it yet.04:05
*** RzR is now known as rZr04:05
itsnotabigtruckdjszapi: you still haven't read the next paragraph04:05
itsnotabigtruckso stubborn...04:05
rZrno problem i will not torture you to get those infos04:06
itsnotabigtruckin any case, rzr: what's the proper version number suffix to use04:06
djszapiitsnotabigtruck: I stop helping you.04:06
djszapithat is what everybody did into the rules file04:06
djszapiwhat I said, in the platform04:06
djszapiand you even say the opposite again04:06
itsnotabigtruckdjszapi: what you pasted is for CBS and autotools, only04:06
djszapiI am quite fed up with this attitude04:06
itsnotabigtruck> See the corresponding instructions below based on whether you want to use the command, or a manually created debian/rules file for building the package.04:07
itsnotabigtruckit has two sections - one for CDBS, one for not-CDBS04:07
rZritsnotabigtruck: major.minor.revision-DebianVersionIfNonePut0maemo$packageversion (i also add ~userN on quick and dirty packages)04:07
djszapiI do not really understand what you are talking about, but I do not care anymore either04:08
djszapi[ ! -f debian/$(notdir $@).aegis ] -> this is obviously not cdbs related usage.04:08
itsnotabigtruckare you sure about the notdir part?04:08
djszapialso it is really hard for helping to anyone if he does not share basic information04:09
djszapilike what he is trying to ask for help with ...04:09
djszapi(as I asked in the beginning)04:09
itsnotabigtruckrzr: 0.90.85+dfsg-2 -> 0.90.85+dfsg-2maemo1 ?04:10
djszapiand if it is not cdbs, it is not enough for you anyway04:10
*** admiral0 has quit IRC04:11
rZrlooks good to me04:11
itsnotabigtruckfor what it's worth, it looks like notdir *is* a make feature, not cdbs, so that does indeed work in debian/rules04:12
itsnotabigtruckso it looks like you're right in that the other version is valid (and it is more general, as it checks for the existence of the aegis file)04:12
djszapiwhen you ask for packaging help, *please* mention what packaging rules you use04:12
djszapiotherwise you waste everybody's time.04:12
itsnotabigtruckbut at the same time, the other version isn't wrong (it just assumes the aegis file is present)04:12
djszapiit is not the most correct, that is all04:13
djszapiyou can stick with less correct, for sure, but that is not my stance.04:13
djszapinor platform stance apparently since everybody uses it as I said.04:14
itsnotabigtruckmy point was, you were saying "this is inferior, look at the document", and the document says "if you're using debian/rules use <fragment i just pasted>"04:14
itsnotabigtrucknow, it turns out either works04:14
itsnotabigtruckbut every little thing turns into this adversarial mess, argh04:14
djszapiplease do not waste others' time04:15
djszapijust let others know what you try to ask for help with.04:15
djszapifor future reference04:15
*** admiral0 has joined #harmattan04:16
* djszapi does not understand why one would run a command on a file if the file does not exist04:17
itsnotabigtruckdjszapi: if i know the file does exist04:18
djszapiexcept that if you remove in the future etc04:18
djszapiplus quite inconsistent with the rest04:18
djszapisince that is what everybody uses.04:19
djszapi(even the documentation writes that exact syntax with the rules example)04:19
*** admiral0 has quit IRC04:22
rZri am gone now04:32
djszapirZr: I would personally like being notified with nominations to a maintainership.04:33
djszapiotherwise everybody can just get everybody into, and would go for mess.04:34
djszapiso I propose a nomination process.04:35
rZrthought it would be good to not restrict to good will04:35
rZrbtw packages can be pushed and approved04:35
djszapiwell, it does not work like that anywhere for severalreasons.04:35
rZrso no need to give ownership04:36
djszapiand I would personally like to see packages for reviews.04:36
djszapias a first step to the nomination.04:36
rZrwell i am tired i am taking note of that04:36
rZrthis sounds good04:36
djszapibecause then we can assess whether someone is skilled in the area and trustworthy.04:36
rZri know yes that's the risk04:37
rZrbut at the same time , i can not handle all the review tasks can you ?04:37
djszapiit does not mean we need to lose the quality04:38
djszapiwe should do what we can do, but with quality, not more than that.04:38
rZrthe quality is for a4m04:38
djszapiwhen more people join, there is going to be more review.04:38
djszapiI heavily disagree with that04:38
djszapithe community repository should also be qualified as much as possible.04:39
djszapiI spent huge amount of deal about cleaning up the crap in there already...04:39
rZrthen the goal changed04:39
djszapinot really.04:39
djszapiit was always a goal to have as much quality as possible.04:40
itsnotabigtruckrzr: so much for being gone :p anyway, later!04:40
rZri thought that was just a q&d solution just to get things done before proposing to a4m04:40
itsnotabigtruckfinally starting to have OBS figured out...with any luck it'll build on the first try04:40
djszapirZr: when I proposed the idea of the community repository back then last summer04:41
rZrwell its late04:41
djszapiI have never had in mind it is an unqualified repository.04:41
djszapiit was actually quite the opposite.04:41
rZrcan we resume later ?04:41
djszapitry to make the best quality we can do, but do not miss the "get something done" as the Surround lost it.04:41
rZrthen think about a :libs subproject04:41
djszapiI do not think ":libs" makes any sense04:42
djszapias I said earlier.04:42
rZrwith higher quality than what it is04:42
rZri am gone04:42
djszapiitsnotabigtruck: please do not put anything against the documentation into the community repository04:43
itsnotabigtruckfirst, i'm putting something into my own repository04:44
itsnotabigtruckand second, nothing is against the documentation04:44
djszapiit is04:44
djszapiand if you do that, I will not be happy in the community repository.04:45
itsnotabigtruckthird, argh, the build completely bombed out04:45
itsnotabigtruckthen you should reread the documentation instead of picking arguments with everyone on this channel04:45
djszapithe documentation gives the exact syntax that /everybody/ uses.04:47
djszapiand afair, you were picky about a quite non-programmer issue, like how an icon look like ...04:47
djszapinot so far ago...04:47
*** admiral0 has joined #harmattan04:47
djszapiand the main point is that, one of the community repository maintainer, I dislike it, so please do not spread it with inconsistent things.04:50
djszapibecause I will need yet another hard time once, the inconsistency arises, like many times in the past. Cleaning up is a huge job, and nobody apparently does not that apart from me in there. :/04:52
*** djszapi has left #harmattan04:57
*** djszapi has joined #harmattan05:00
Sazpaimonwhy are people complaining that N9 is only getting flash 10.1 and not 1105:03
Sazpaimoni see no benefit flash 11 as a run of the mill NPAPI plugin will bring to the N905:03
Sazpaimon10.1 already has hardware based video decoding05:03
psycho_oreosI suppose they want the latest and greatest05:03
Sazpaimonall 11 really does is slightly improve it, and add support for 3D rendering over gles05:04
SazpaimonI dont know about you, but Ive never played any 3D games made in flash05:04
Sazpaimonnor would I want to05:04
Sazpaimonas long as 10.1 is complied against gles, I'd be happy05:05
*** adlan has quit IRC05:05
itsnotabigtruckyeah, i tried out flash 3D on my pc a while back and it wasn't exactly blindingly fast05:05
itsnotabigtrucki can't imagine it's particularly amazing with the N9's hardware05:06
Sazpaimonon flash 11?05:06
*** heymaster has quit IRC05:06
Sazpaimonflash 11 has better hardware support for 3D05:06
Sazpaimonespecially on mobile, where it uses opengles 2 directly05:06
Sazpaimoni dont see why a developer would make 3D games in flash though05:06
Sazpaimonespecially for mpbile05:07
SazpaimonI think after flash 10.3, adobe went completely balls in with AIR instead05:08
Sazpaimonballs deep i should say05:08
Sazpaimonballs in would be moreso though05:08
Sazpaimonbut nobody ever goes balls in05:08
Sazpaimonthats just asking for trouble05:08
itsnotabigtruckthe build almost completed...then it failed because zlib1g-dev wasn't installed05:15
itsnotabigtruckwell, at least OBS using an absolutely minimal install helps grind out these build dependency mistakes from upstream05:15
itsnotabigtrucki wonder how debian's build system works05:16
itsnotabigtruckSazpaimon: once i get OBS to build this, load up milkytracker!05:17
itsnotabigtrucki should really stop watching the build log, a watched pot never boils05:18
*** oberling_ has joined #harmattan05:20
*** oberling has quit IRC05:24
itsnotabigtruckmake: aegis-deb-add: Command not found05:36
*** rcg has quit IRC05:36
itsnotabigtrucknot another one >_>05:36
befordnew sdk thing looks prettier05:39
itsnotabigtruckbeford: hmm, should i update05:40
befordnot sure05:40
befordi have not really used it yet, just started it05:41
itsnotabigtruckhmm, just noticed in navifirm: 059M182: RM-819_NAM_US_CV_VAR191237_BLACK_V1 (that's for the lumia 800)05:43
itsnotabigtruckwonder if that means there's a us version of the 80005:43
itsnotabigtruckthe sdk connectivity tool package doesn't contain a changelog :(05:51
*** rcg has joined #harmattan05:52
itsnotabigtruckit built, yay06:00
itsnotabigtruckonly took a few millenia for it to finish :p06:00
*** Natunen has joined #harmattan06:01
itsnotabigtruckbeford: install this06:03
befordyou didn't want to install mine :C06:06
befordok downloading06:06
itsnotabigtruckbeford: lol06:06
itsnotabigtrucki didn't? which one is it06:06
befordi'm debugging it atm, i'll let you know :)06:07
itsnotabigtruckoh, the ovi crash thing earlier06:07
itsnotabigtruckwell, if you put it somewhere i'll try it out, though enigmatic crash bugs are never any fun06:07
itsnotabigtruckoh yeah, i should load up rzr's mod file and play it06:09
itsnotabigtruckoh no...the resource policy isn't working now06:10
befordlooks like an hexeditor06:11
itsnotabigtruckhaha yeah06:12
*** Natunen has quit IRC06:12
*** Natunen has joined #harmattan06:13
beforddoes it sound or something? :P06:14
*** virtuald has quit IRC06:16
itsnotabigtruckbeford: the volume can't be adjusted, that was the problem06:17
*** virtuald has joined #harmattan06:17
itsnotabigtruckoh, as for making it play something...grab a  MOD or XM file and load it into your mydocs06:17
itsnotabigtruckthen load it into milkytracker and hit play sng06:17
*** pinheiro has quit IRC06:23
beforditsnotabigtruck, .it files are supported?06:25
*** admiral0 has quit IRC06:26
itsnotabigtruckbeford: according to the package description...except i tried to load one and it was a no go06:33
itsnotabigtruckbeford: alright, package is now updated to 2maemo206:36
itsnotabigtruckit *should* have a legit resource policy file now06:36
itsnotabigtruckwhich means the volume buttons will work as one might expect06:36
befordthat means06:36
itsnotabigtrucki had it working but forgot to fix the path after optifying06:40
itsnotabigtrucki uploaded calendarrr into OBS too, but need to import an updated libtimed-dev also :(06:41
befordnot sure if its me or it sounds the same06:41
itsnotabigtruckbeford: it does sound exactly the same06:43
itsnotabigtruckyou can just change the volume now06:43
befordi mean, it's not changing06:43
itsnotabigtruckhm, i wonder what 'Create as maintenance project' does06:43
itsnotabigtruckoh, that's strange06:43
befordsounds the same in 'beep' mode and in max voulume06:43
befordi believe06:43
itsnotabigtruckwait, if you press the volume buttons with milkytracker focused06:44
itsnotabigtruckdo you get the blue volume bar or the pink one06:44
itsnotabigtruckwell, magenta one06:44
itsnotabigtruckand you updated to 2maemo2?06:44
itsnotabigtruckcat /usr/share/policy/etc/syspart.conf.d/milkytracker.conf06:44
itsnotabigtruckmake sure it says /opt/milkytracker/milkytracker06:44
befordi think so06:44
befordlet me see06:44
itsnotabigtruckalso if you left it running during the update, exit it and start again06:44
befordi could have downloaded it and not installed06:44
itsnotabigtruckapt-cache show milkytracker should tell you what ver you have loaded06:45
befordok reinstalled let me see06:45
*** npm has quit IRC06:45
*** npm has joined #harmattan06:46
befordyea same itsnotabigtruck06:46
befordVersion: 0.90.85+dfsg-2maemo206:46
beford$ cat /usr/share/policy/etc/syspart.conf.d/milkytracker.conf06:47
beford[classify gaming]06:47
itsnotabigtruckbeford: hmm06:54
itsnotabigtruckand you're sure you closed milkytracker then started it again?06:55
itsnotabigtruckremember that since this app isn't exactly natrive it'll do things like linger around after you close it with a popup box06:55
befordunless its running when I close it06:55
befordkillall -9 milkytracker ?06:55
itsnotabigtruckif you had a modified file open06:55
itsnotabigtrucklol, sure06:55
befordno process06:56
befordsame, bluebar06:56
itsnotabigtruckok, this is weird06:56
itsnotabigtruckoh hmm, is this on pr1.1 or pr1.206:56
itsnotabigtruckpr1.0 or pr1.106:57
itsnotabigtruckok, so it's not that either06:57
itsnotabigtruckand if you start a game (e.g. angry birds), you get the magenta bar?06:57
itsnotabigtruckvery strange...07:02
befordwow, i pushed something free like a week ago on ovi, and the download from china is like 350 and next country 907:04
itsnotabigtruckhaha, wow07:04
itsnotabigtruck^ what on earth happened there.../usr/include/qt4/QtCore/qvector.h:331: error: assuming signed overflow does not occur when assuming that (X + c) < X is always false07:05
itsnotabigtruckthis isn't even my code, that's a nokia open source package i imported07:05
befordmaybe nokia uses a differente gcc07:06
itsnotabigtruckit works in scratchbox also07:07
itsnotabigtruckbut yeah, maybe obs is loading something different07:07
*** hardaker has quit IRC07:12
*** MikaT has quit IRC07:14
*** MikaT has joined #harmattan07:14
itsnotabigtruckthe "awesome" thing is that it looks like the warning (which is being treated as an error) is coming from qt07:14
*** Corsac has quit IRC07:15
*** Corsac has joined #harmattan07:16
*** adlan has joined #harmattan07:16
*** benares_98 has quit IRC07:26
*** heymaster has joined #harmattan08:00
*** beford has quit IRC08:09
djszapiitsnotabigtruck: if you take the time to make a little bit of google, as I did earlier, you figure out thiago mentioned it is a bug either in qt or either in gcc. It is actually just a warning, so it seems warning turning into errors switch is on.08:17
djszapiI do not know why you would turn all the warnings into error, but can be solved with pragma statements.08:19
itsnotabigtruckdjszapi: right, it's not a bug per se in gcc, but it's undesirable behavior that's a result of the gcc devs being stubborn08:20
itsnotabigtruckthe real problem is that -Wall -Werror is a bad combination08:20
djszapiohh...also gcc devs are stubborn08:20
djszapiseems many known and big projects are stubborn or bad for you :)08:20
itsnotabigtruckwell, they are...if you look into this, it's based on an extremely strict interpretation of the c++ spec08:20
djszapias I and thiago also said08:20
djszapiit is either a gcc or qt bug08:21
itsnotabigtruckin gcc-land, ((unsigned int)-1 + 1) > 0xFFFFFFFFU08:21
itsnotabigtruckit's not a bug, it's incorrect behavior that the spec lets them get away with08:21
djszapiwe tried to figure out back then, but we wre not able to find anyhing bad in the qt code after even rearranging the code, but it does not mean it is not a qt bug since.08:21
itsnotabigtruckbecause it allows optimizing out a comparison08:21
djszapiI do not care about your continous "bad" and "stubbornness" tone.08:22
*** JackaLX has quit IRC08:22
djszapiI told you the situation.08:22
djszapialso the workaround, not much more to comment on.08:22
itsnotabigtruckthe problem is that all the normal solutions involve modifying the source package08:23
djszapialso, "c-obs" does not use different gcc than the Nokia SDK so c-obs imports that altogether.08:23
itsnotabigtruckwhich is what i ended up doing, which makes the fix easy08:23
djszapinot at all08:23
djszapiwhy would a gcc option be source package change ?08:23
* djszapi fails to understand08:23
itsnotabigtruckbecause timed uses qmake, and qmake forces all the gcc options08:23
itsnotabigtruckthere's ways to get around that but they're more than a bit complicated08:24
djszapihuh ? not.08:24
itsnotabigtruckQMAKE_CXXFLAGS='-Wno-strict-overflow' dpkg-buildpackage -rfakeroot = no change in the actual gcc command line08:24
djszapiseriously, what are you talking about ?08:24
itsnotabigtrucktried a bunch of other variables too (QMAKE_CXXFLAGS_RELEASE, CXXFLAGS, QMAKE_CFLAGS_RELEASE, etc.)08:24
djszapiyou should never really modify the source package according to a build system08:25
djszapineveruit is a debian patch normally.08:25
djszapinot source08:25
itsnotabigtruckthe problem isn't the build system per se, it's a problem with the source package being fragile08:25
itsnotabigtruckdue to ill-advised use of -Werror08:25
djszapithe author of the source package does what he wants.08:26
itsnotabigtruckwell that's all well and fine until somebody actually has to use it for something08:27
itsnotabigtruckand code exists to be used08:27
itsnotabigtruckand a) it's a native package, so no debian.tar.gz patches08:27
djszapiheh, even the author is wrong now :)08:27
djszapiwell, I give up.08:27
itsnotabigtruckb) qmake ignores any and all environment variables, so that route won't work08:27
*** heymaster has quit IRC08:27
djszapiI am not interested in hearing everybody is wrong. :)08:27
itsnotabigtrucklike i said before, not everything is binary08:28
djszapiit is a valid stuff to turn everything into error, many projects do.08:28
itsnotabigtruckby your logic, if i say harmattan has bugs, i'm saying it's horrible software that should not be used08:28
djszapino you swear about everything.08:28
itsnotabigtruckthe fact of the matter is a) timed's package has a bug, b) i need to build it, c) so i need to implement a workaround08:28
djszapieven it is quite invalid08:29
djszapiyou cannot even be convinved with a text from an official page.08:29
djszapido whatever you want really. Pardon, I tried to help.08:29
djszapiwith this few minutes google issue.08:29
itsnotabigtruckno, you launched into a tirade about this out of nowhere08:29
itsnotabigtruck*after* i already got around the issue08:29
itsnotabigtruckit's certainly not an ideal solution...that's why i'd like to know a more proper way to override the source package's build flags08:30
itsnotabigtruckbut you haven't suggested one08:30
itsnotabigtruckjust chided me for daring question the author of the original package08:30
djszapiI already told you what a debian developer does.08:30
djszapiyou said it is wrong08:31
djszapiokay, do whatever you want ...08:31
* djszapi afk08:31
itsnotabigtruckto add a debian/patches patch?08:31
itsnotabigtruckbut it's a native package08:31
itsnotabigtruckthere is no debian/patches08:31
djszapi1) No, I have not said that08:31
itsnotabigtruckwhat a debian developer would do is fix the source package and NMU it, but this isn't debian and there's no such thing as NMUing a package into nokia08:32
djszapi2) Why on earth would you make native package for harmattan?08:32
djszapi3) Good luck with native packages on c-obs08:32
itsnotabigtruckit's not mine, it's *timed* from pr1.108:32
djszapiwhy would you change timed at all ?08:33
* djszapi fails to see this mess08:33
itsnotabigtruckbecause c-obs comes with 2.50, i need 2.5108:33
itsnotabigtruckso i'm building 2.53 so i can access the new shiny headers08:33
*** RST38h has joined #harmattan08:33
itsnotabigtruckon second thought copying the header files into my project is probably no worse than this mess08:34
djszapigood practice for c-obs: never ever build native package08:34
itsnotabigtruckwhy do you say that?08:34
djszapisecondly, I still think it is possible to change the rules file, but if not you need to patch it anyway08:34
djszapiso you could still make the debian patch08:34
djszapifor instance in the rules file...08:34
djszapito make a replacement.08:34
djszapiit is not that hard really...08:35
djszapisimple sed command really...08:35
djszapiand no, the author is not wrong08:36
djszapihe should not really develop the platform due to c-obs limitations, if any.08:36
djszapior whatever build system08:36
djszapiactually I am happy about him doing it this way.08:38
djszapisince it means he cares about warnings, too (a.k.a. quality)08:39
djszapiI wish more authors did it ...08:39
djszapiyou basically say that, do not write tests either, because most of them fail in Harmattan.08:40
itsnotabigtruckwell, what i'm saying is that fail on warning is a dangerous option08:42
itsnotabigtruckit can make sense to enable it on e.g. debug builds08:42
itsnotabigtruckbut the problem is that you make your build extremely fragile08:42
djszapiit is a very valid option.08:42
itsnotabigtruckabsolutely anything can cause it to fall apart08:42
itsnotabigtrucklike a minor increment to a gcc version08:42
djszapithe more author use the better08:42
itsnotabigtruckwarnings are noisy08:42
itsnotabigtruckhigh level warnings are very noisy08:42
djszapibut softwares are full of (serious) warnings08:42
djszapiif a gcc version gives a warning, fix the software08:43
djszapiand since this is not the common case, why lower the quality ?08:43
itsnotabigtruckthis particular warning (strict-overflow) is a perfect example of what i mean by noisy and dangerous08:43
* djszapi fail to understand08:43
djszapiit is one liner to fix up in packaging08:43
itsnotabigtruckchanging *anything* can cause it to kick in, because it's detected after inlining and optimization08:43
djszapiwhat is so dangerous in one line ?08:44
djszapiin order to lose the quality ?08:44
itsnotabigtruckso whatever affects gcc's inlining heuristic can fail the build at random08:44
itsnotabigtruckand parts of qt are not exactly perfectly warning clean08:44
itsnotabigtruckthis warning isn't even coming out of timed, it's coming out of qt08:44
itsnotabigtrucktotally out of control of either myself or the timed developers08:45
djszapithe error message is quite clean08:45
itsnotabigtruckby "warning clean" i mean as in, library header files don't cause warnings ever08:45
itsnotabigtruckonly user code08:45
itsnotabigtruckso if qt's header causes a warning in your otherwise valid code, there's nothing clean about that08:46
*** JackaLX has joined #harmattan08:46
djszapiI am really not sure what you are talking about ...08:46
djszapithis is one line to fix.08:46
itsnotabigtruckbut for qt to never cause a warning ever is an impractical standard, because gcc's warnings change all the time08:46
djszapiand you did not still do it, just arguing...08:46
djszapiand saying gcc, author etc wrong08:46
itsnotabigtrucki already did it, i'm waiting for the build08:46
djszapiwhy not just put the damn one line ?08:46
itsnotabigtruckbut you also said my solution is incorrect08:46
djszapithen why argue ?08:46
djszapiif it is fixed ?08:46
djszapiyou like arguing even if it is fixed ?08:47
itsnotabigtruckerm, now you're flipping this around, i told you it was fixed 15 mins ago (actually i ran into another issue, but i think that isn't an upstream problem)08:47
djszapithe proper fix (at least what almost all the debian developer does) is rules file change, or debian patch08:47
itsnotabigtruckyou were saying this solution is improper, somehow08:47
djszapiyes, modifying the source and not packaging is quite wrong08:48
itsnotabigtruckalso, this is very weird, dpkg-genchanges is deciding it's on i386 all of a sudden08:48
djszapiso your source is not in sync anymore with the official one ...08:49
itsnotabigtruckdjszapi: right, now it's timed_2.53+0m6.0itsnotabigtruck108:50
* djszapi has probably never seen a debian developer so far, or even a newbie packager to modify the source tar ball directly.08:50
itsnotabigtruckhow many times do i have to say that it's a native package: there's only one way to modify it, which is to modify the source tarball08:51
itsnotabigtruckand increment the version, which i did08:51
djszapias I said, *not*08:51
djszapisimple grep does the damn trick in the damn rules file.08:51
itsnotabigtrucksince the rules file is inside the source tarball, modifying the rules file is modifying that08:52
djszapino no08:52
djszapiyou were saying "source"08:52
djszapiI was saying "debian packaging"08:52
djszapiand you said debian packaging modification is not possible08:53
djszapithis isclearly a debian packaging modification08:53
djszapinothing related to the source/upstream.08:53
itsnotabigtruckwell, i was saying "source package", which in this case is a DSC and a tarball, only08:53
itsnotabigtruckwhere the tarball has both the source code -and- the debian packaging together08:54
djszapiwhy the heck did you tell me debian packaging is not good what I said ?08:54
djszapiif you actually followed my advice...08:54
petterihappy saturday morning #harmattan08:54
itsnotabigtruckhey petteri08:55
djszapiif you say in the beginning, you are right "debian packaging" is enough, this time waste does not happen08:55
djszapibecause then I get the situation you agreed with me already at that time instead of arguing08:55
djszapisource is confusing especially because you just later said it is native package.08:56
djszapiso as for me, it seems you were not clear which caused the misunderstanding.08:56
itsnotabigtrucki probably could have been more precise, i suppose /me shrugs08:56
djszapithat is the problem, you shrug if you waste someone's time who tries to help..08:57
djszapiyou was not clear with cdbs/dh previously either08:57
itsnotabigtruckas well as mine...if you don't want to spend time on irc, then don't08:57
djszapibe precise, really hard to help you.08:57
djszapiwhen you make vague statements.08:58
itsnotabigtruckpetteri: check out milkytracker on harmattan09:00
itsnotabigtruckit's probably a lot more useful if you have an n950; on my n9 it's close to impossible to use the microscopic buttons09:00
itsnotabigtruckbut it's a cool tech demo :D09:00
*** heymaster has joined #harmattan09:14
itsnotabigtruckawesome, timed = built09:14
*** risca has quit IRC09:18
*** risca has joined #harmattan09:18
*** risca has quit IRC09:18
itsnotabigtruckit's a shame OBS builds a new VM installation from scratch for *every build*09:24
itsnotabigtruckanyway, djszapi: apologies if i was getting on your nerves all today :(09:55
itsnotabigtruckon the bright side, everything is finally building now09:55
*** heymaster has quit IRC10:03
*** gabriel9 has joined #harmattan10:27
*** diggy has joined #harmattan10:32
*** heymaster has joined #harmattan10:39
dm8tbritsnotabigtruck: that's part of the concept, clean build environment for every package.11:17
dm8tbritsnotabigtruck: also it shouldn't be that much overhead for a well scaled system. most things will be in RAM or cached. even the environment itself.11:19
*** gabriel9 has quit IRC11:40
*** DocScrutinizer has quit IRC11:40
*** DocScrutinizer has joined #harmattan11:40
*** Termana has quit IRC11:44
bindiitsnotabigtruck: i've come to a conclusion, i think i'll just use xmpp along with sip11:46
bindialthough that means more battery drain11:56
* bindi kills self again11:56
*** Hq` has quit IRC12:04
*** M4rtinK has joined #harmattan12:19
*** NIN101 has joined #harmattan12:23
*** Guest45488 has joined #harmattan12:28
*** heymaster has quit IRC12:36
*** Hq` has joined #harmattan12:43
*** djszapiN9 has joined #harmattan12:47
*** ieatlint has quit IRC12:51
*** ieatlint has joined #harmattan12:51
*** artemma has joined #harmattan12:54
artemmaHi All12:56
*** Hq` has quit IRC13:04
djszapiX-Fade ping13:08
*** zz_gri is now known as gri13:08
*** auenf has quit IRC13:13
*** auenf has joined #harmattan13:14
*** aparaatti has joined #harmattan13:19
aparaattihow can I add a user in terminal?13:20
djszapiuseradd ?13:21
aparaattiuseradd not found...13:21
aparaattii have n913:21
djszapiwhy do you need that, if I may ask ?13:21
aparaattiyup there it is, thank you13:22
*** trx has joined #harmattan13:26
aparaattican I set up something like sudo in the terminal?13:26
*** djszapiN9 has quit IRC13:27
aparaattiok... this is so cool :D13:27
aparaattiI got ssh connection working for the first time...13:30
*** niqt has joined #harmattan13:32
leiniraparaatti: you should only need to log in as developer...13:34
leinirturn on developer mode, open up the sdk connection tool, select wlan or usb according to what you're using, then log in with ssh on the ip address there, username developer, password as listed13:35
aparaattileinir: the root ssh access was blocked at some point13:35
leinirwell yes, because root isn't really useful on a device controlled by aegis ;)13:36
aparaattipassword as listed?13:36
djszapiaparaatti: sdk connection tool will give you a password to developer.13:36
aparaattiI made a own user now, so is there some reason I shouldn't use that?13:37
leiniraparaatti: sorry, yeah, should've been a bit more precise there - the same page that gives you the ip address also lists a password :)13:37
aparaattioh indeed it does13:38
djszapiaparaatti: yep. :P It is not a multi-user system in general. Can you please answer my question above about why you need a separate user ?13:38
aparaattiI was first trying to connect through wifi... and could do it by making a new user13:38
aparaattithen I realized the sdk mode could be used for faster connection. I just want to edit the theme with a prober keyboard13:39
djszapiI think you can do that without a new user.13:39
aparaattiyes, the developer method works. So maybe I'll stick with that13:41
leinir*nods* It's what people will be able to help you with in the future too, should you need it :)13:41
*** artemma has quit IRC13:55
djszapiX-Fade lbt
*** Arkenoi has quit IRC13:58
*** Guest45488 is now known as Termana13:58
*** Arkenoi has joined #harmattan13:59
lbtdjszapi: I think we need a maemo community team to start taking ownership of that side of the OBS13:59
djszapilbt the Harmattan target ?14:00
lbtpersonally it's not something I have any knowledge about - I don't even have a harmattan device :(14:00
djszapilbt is it a difficult role ?14:00
lbthard question...14:01
aparaattileinir: to be able to do anything usefull might be a long way to go for me, but I'm learning the sdk and playing with the phone on freetime anyway... thank's for information.14:01
leinirHey, it's what we do, yeah? :)14:01
leinirAll gotta start somewhere :)14:01
lbtit's a responsible role - and requires the community to trust the occupier14:01
lbtI'd hoped to be more involved in it as we migrated towards meego ... but you know the history14:02
lbtI have to focus on Mer now as the future14:02
lbtand I simply don't have the bandwidth to handle Harmattan too14:03
lbtI think the maemo council needs to address this14:03
lbtobviously there will be lower level OBS tasks that need doing - and I can help with implementation14:04
*** blueslee has joined #harmattan14:04
lbtbut I don't have the knowledge to make the decision you ask for14:04
lbtI need someone to 'own' the OBS for Harmattan - and right now that is just X-Fade14:05
djszapiI do not know anybody with too much knowledge about OBS side. At least, I did not see anybody here on IRC. Perhaps on forums, or somewhere else.14:06
djszapiOBS side bits*14:06
djszapiand I do not have time for that either. :/14:06
djszapiand this is probably the issue for many.14:06
djszapiWecan post something to the forum later.14:06
djszapior perhaps obs/meego levlists, if any.14:07
lbtforget meego dev lists14:07
lbttotally dead14:07
lbtthis is a harmattan only issue14:08
lbtI know that the OBS was discussed as the 'official' community build system a long time ago - not sure how it evolved from then on14:08
bluesleeanyone tried firefox from ovi with flash?14:09
djszapilbt: who helps for the PA guys regarding the issues, if any ? You and X-Fade there, too ?14:10
lbtdjszapi: is there a context for that reply?14:10
djszapilbt: attachment.14:10
djszapibut I sent a mail to them this morning, so they will probably write an answer there on Monday.14:10
lbtwrt PA - that's primarily me14:11
djszapilbt: the context is the reproducing steps.14:11
djszapiand my bugreport14:11
djszapimmmh, and the other bugreport to the Nokia SDK team in the mail.14:11
djszapiso the solution would be for us if either they sign the Nokia SDK repository, or we de-integrate their repositories.14:11
djszapiThe former would be the least effort for everybody.14:12
djszapibut they ar eobjecting to that heavily, and I have no idea why.14:12
lbthonestly I don't understand the issue so can't comment14:12
djszapiso if they cannot help, we need to fix the issue up somehow on our own.14:12
lbtwhat I meant was - do you have a link to the conversation where that reply was made14:12
djszapilbt fair enough, it needs some understanding, yes.14:12
djszapinot atm because I cannot somehow get my emails.14:13
lbtah, didn't know it was a private email14:14
djszapiyeah, it is internal company discussion14:15
djszapibut this is in a nutshell what I wrote:14:16
djszapi"The problem is that the Nokia repository is not signed. It should be fixed in my opinion (Marcell "CC"-ed from the SDK team), but it is still not a guarantee for external repositories out there."14:16
djszapiMarcell can state his opinion as the Nokia SDK TPO on Monday, hopefully. I would not personally like to copy everything out.14:16
lbta quick read seems to suggest that the SDK binaries shouldn't go on devices as they're potentially sb-hacked14:16
lbtso rebuilding such binaries from clean source in an sdk-alternative project sounds sane14:17
djszapi(They have never written it anywhere previously, and I think that is why X-Fade began integrating them).14:17
djszapiso I think it is not the fault of the community.14:17
lbt*nod* it looked like an easy fix but maybe it wasn't quite the best one14:18
djszapithere is going to be some transient procedure14:18
djszapibut they do not help us, the earlier we do the better.14:18
djszapiwe do not need to put everything from there to our repository, because there is some overlap between the apps/tools repository and the Nokia SDK repository.14:19
djszapiI am not sure how much, but there are some.14:19
lbtis all the SDK source available ?14:19
djszapithere are no free packages, too.14:19
lbts/all/enough of/14:19
infobotlbt meant: is enough of the SDK source available ?14:19
djszapiI think it is managable.14:20
djszapiwe need to copy/paste14:20
djszapiinto the current community repository or somewhere else.14:20
lbtyes - that's scriptable too14:20
djszapijust very unhandy knowing the fact the signing would be less effort on their side, much less. :)14:20
djszapiand even according to the debian recommendation.14:21
djszapilbt: another question, is it only you and X-Fade behind apps4meego or are there more people there ?14:21
lbtanyhow....I have to go finish the Mer SDK ... trust me, it's a lot more open :)14:21
lbtapps4meego != OBS14:21
lbtthere are more people in apps4meego14:22
lbtI'm not really involved there very much (other than BOSS side)14:22
djszapiyes, Mer is more open.14:22
djszapiI am looking forward to seeing a well-working Ui stack on top of it.14:22
lbton OBS it's me, X-Fade and phaeron14:22
lbtNemo is really close14:22
djszapiPA is a bit buggy yet :P14:23
djszapianyway, thanks. I do not bother you more. :)14:23
*** diggy has quit IRC14:26
*** e-yes has quit IRC14:27
Jarebindi: depends on your needs, but you could connect for example freeswitch to an xmpp server and use only xmpp or sip on a mobile phone14:42
*** pinheiro has joined #harmattan14:50
aparaattiis vim available some where?14:51
aparaattifor n9 that is...14:51
ZogG_laptopaparaatti: yes i think so14:51
djszapiaparaatti: nope14:52
psycho_oreosthere's vi I think in harmattan-dev repository14:52
djszapiyou need to use the community reppository for that14:52
djszapiI packaged it last summer14:52
*** blueslee has quit IRC14:52
djszapipsycho_oreos: nope14:53
djszapiit is available by default14:53
aparaattiok... how can I add the community repository?14:53
djszapiaparaatti: hold on14:53
ZogG_laptopdjszapi: if you don't see it - it doesn't mean there is no vim14:53
psycho_oreosdjszapi, interesting14:53
djszapiecho 'deb ./' >>/etc/apt/sources.list.d/nicks.list14:53
aparaattiyes.. there is vi... but I'm not used to that14:53
djszapiaparaatti: yes, vim is unfortunately missing, but on the other hand, it is understandable why.14:53
ZogG_laptopaparaatti: ^ here is the link14:54
djszapiaparaatti: after that: apt-get update14:54
djszapiand apt-get install vim14:54
aparaattioh... so is ovi somekinf od frontend for apt?14:55
djszapiaparaatti: it is an app store supplied by Nokia.14:55
*** denism1 has joined #harmattan14:56
djszapiaparaatti: apt frontend is the packager manager.14:57
aparaattiok, now I have a working vim :)14:59
djszapicool =)14:59
Jareimo ovi store downloads debs from nokia and installs them using the package manager, so it's like the software center on ubuntu15:00
djszapinope :P15:01
djszapithat is the Ovi client ;-)15:02
aparaattisurprised that there is apt-get by default...15:03
*** arcean has joined #harmattan15:05
djszapiaparaatti: the developer mode would be a bit useless without apt ;-)15:06
aparaattiso I need to update the packages from commandline.. and is it save to do upgrades etc. ? if I have installed stuff from the community repository15:06
Jareor is it nokia store client nowadays :B15:06
djszapiI hope I wrote it out to the community repository, we do not provide any guarantee altogether for that.15:08
aparaattiyes.. so if something goes wrong I can recover the phone by flashing?15:08 Testing - Drop Cache application is available (mostly a UI wrapper for echo 3 > /proc/sys/vm/drop_caches and sync). Any feedback / review is really appreciated. (as a wow number - for me 250 mb more free memory become available)15:10
aparaattibut propably there is some people using the same repo with n950?15:10
djszapiaparaatti: there are no known issues15:11
*** gri is now known as zz_gri15:11
aparaattidjszapi: i can settle with that :) so is there a web page or other info somewhere on the community repos etc.?15:13
aparaattithank you15:14
djszapidenism1: do you need to do it after each reboot, or have you put it into an upstart job after the reboot ?15:16
djszapiAlso, can we install from testing using the apps4meego client ?15:17
denism1djszapi: you can do it at any time, it is pretty useless and really for the end users might be used just as a WOW application - together with some UI resource monitor (like RscMonitor).15:18
denism1djszapi: for developers and QA testers - it might be interesting just to see how fast Maps, for example, starts without a cache at all.15:18
denism1djszapi: yes, apps4meego - install "Apps Testing enabler" before and TESTING applications will become visible.15:19
djszapidenism1: is the apps4meego client available in ovi ? If yes, what is its name ?15:21
denism1djszapi: no it is available on website, it depends to aegis-community-source-policy package also (so apps repository has to be enabled on your device)15:22
djszapiI will try when it is available from Ovi. ;-)15:22
*** longthen1 has quit IRC15:24
*** longthen has joined #harmattan15:26
denism1yes, and if really about this apps4meego - Drop Cache is using GRP::root and UID::root credentials in aegis manifest and works fine :) - really nice, but any app should be tested very well.15:27
denism1I have tried apps4meego just today - but I really like it. Very powerful and nice.15:29
djszapiyes, QA is, of course, needed.15:30
*** djszapi has left #harmattan15:36
*** Hq` has joined #harmattan16:01
*** aparaatti has left #harmattan16:14
*** tarantism has joined #harmattan16:15
*** niqt has quit IRC16:21
*** Free-MG has joined #harmattan17:01
itsnotabigtruckdm8tbr: the problem is that most of the build time is spent building the vm on all but the largest packages17:02
itsnotabigtruckand it could be using clones or snapshots or somethng17:02
itsnotabigtruckon the positive side doing it this way makes it really strict about missing buildeps so those get caught17:03
dm8tbritsnotabigtruck: that would destroy the flexibility of the system and at the same time bloat your storage beyond belief17:03
*** zz_gri is now known as gri17:03
dm8tbryou can't keep enough variants of rootfs snapshots on fast enough storage17:03
itsnotabigtruckdm8tbr: well, since i only personally care about harmattam, that's 1 image, but if you include a bunch of meego images that's still not that many images17:04
itsnotabigtruckbut it's all kind of a boot point since obs is hardcoded to work this way i assume17:04
dm8tbritsnotabigtruck: yes, what you just described is a case for e.g. SB217:05
itsnotabigtruckthe idea is anything beyond core system/build-essential stuff gets added after the snapshot17:05
itsnotabigtruckthen obliterated after the build17:05
itsnotabigtruckalso s/boot/moot, heh17:06
itsnotabigtruckjust got up, /me is teh tired17:06
*** lbt has quit IRC17:11
*** lbt has joined #harmattan17:12
*** lbt has joined #harmattan17:12
*** hardaker has joined #harmattan17:23
*** gri is now known as zz_gri17:48
bindiis the harmattan OS opensource btw?17:51
ZogG_laptopnot really18:01
*** heymaster has joined #harmattan18:06
itsnotabigtruck@everyone: download that and get tracking! :D18:11
itsnotabigtruckbindi: not really to a meaningful extent, though it contains a lot of open source components18:12
itsnotabigtrucka decent chunk of the nokia system software is open source, but the user interface is what matters most of the time and that's all closed18:13
*** pinheiro has quit IRC18:15
*** pinheiro has joined #harmattan18:17
itsnotabigtruck now there's an n9-apps page too18:23
phakois it a harmattan problem that if I use a QSortFilterProxyModel in a QML listview I get the rowcount of the underlying model with the filtered lines being empty?18:32
phakoor a general QML issue?18:32
merlin1991anybody know what the updates to developer mode and connectivity tool bring?18:48
itsnotabigtruckmerlin1991: apparently the sdk connectivity tool looks different now18:57
itsnotabigtruckdownload milkytracker!18:57
bindidoes skype drain much battery?18:58
phakoseems to work on desktop :-/18:59
SpeedEvilbindi: yes - lots and lots19:00
DocScrutinizerLOL yeah19:01
SpeedEvilIt both keeps the processor, and teh network active.19:01
bindiwhat's my best bet if i want voip and IM in the same program, and not drain the battery? >.<19:02
* SpeedEvil needs to see if his cisco thingy will do SIP19:03
SpeedEvilAnd talk to his ISP19:03
DocScrutinizerand actually that's why "skype just works", it's doing a lot to work around NAT, firewall, whatnot else, and that means to keep traffic flowing to not let the NAT sessions expire19:03
SpeedEvilYeah. Skype sucks.19:03
SpeedEvilThere isn't however a better global solution.19:03
bindibeen playing with SIP19:04
DocScrutinizerSIP works great for me19:04
bindiwhat server software do you use?19:04
DocScrutinizerserver software??19:04
bindii thought you'd run your own server :p19:04
SpeedEvilDocScrutinizer: I mean my ISP offers me a SIP phone number. I want to make that my 'landline'19:04
abDocScrutinizer, more to that. It also decrypts the executable code in the memory and locks it down from swapping out to avoid crackers to get to their secrets. This doesn't make kernel swappiness more happy19:05
DocScrutinizerSpeedEvil: yo, understood19:05
DocScrutinizerab: indeed :-P19:05
SpeedEvilHmm. Actually - I wonder if it would make more sense to put that on the Pi.19:05
SpeedEvilIf I get.19:05
itsnotabigtruckSpeedEvil: you! install milkytracker!19:06
itsnotabigtrucki'm trying superliminal messaging about my port :p19:06
bindiDocScrutinizer: does sipgate let you do instant messages (SIMPLE)? also, do they use TCP? I read that TCP is better for battery life19:06
* DocScrutinizer waves and heads out for RL19:06
DocScrutinizerbindi: both yes, afaik19:07
DocScrutinizerSIMPLE seems to work, TCP should be supported. SIMPLE is not connected to a SMS gateway as it's done on some other SIP providers19:07
DocScrutinizerI cloudy recall checking that when Michel implemented IM nto twinklephone19:08
DocScrutinizeranyway I use automatic transport for sipgate19:09
DocScrutinizerprobably uses UDP19:09
DocScrutinizerusually use keep-alive REGISTER, and set register timeout to 120s19:10
DocScrutinizeralso avoids beoing offline for hours after IP addr change19:11
DocScrutinizerwhich is an evil effect usually ignored by SIP users and providers19:11
DocScrutinizerSIP registrar will try to send INVITE to outdated IP whenever you roam to a new DHCP lease19:12
DocScrutinizershort registration timeout will cure that19:13
bindii thought you should keep the timeout biggie to consverve battery19:15
DocScrutinizerwhen a SIP call comes in, some 5 devices start to ring here XP19:15
DocScrutinizerone short data burst every 2 minutes doesn't hurt much19:15
DocScrutinizeryou need some sort of keepalive anyway19:16
bindido you have a code for me? :p19:16
DocScrutinizerso why not using REGISTER for that?19:16
bindii'm at and it wants a code19:16
bindican't select basic19:17
DocScrutinizerooh, a area code19:18
bindiah :p19:18
DocScrutinizertry 08919:18
DocScrutinizeror 091119:18
bindi089 works19:18
bindi- Only for residential customers residing in Germany19:19
DocScrutinizerthey will check19:19
bindiright, what can i do then? ever configured asterisk server? :D19:19
*** hardaker has quit IRC19:20
itsnotabigtruckthere's other sip providers19:21
itsnotabigtrucki'd recommend against asterisk, for one it has problems with secure calls19:23
itsnotabigtruckthough the N9 doesn't even support secure calls because the telepathy devs cba to have anything to do with security19:23
itsnotabigtruck(neither can almost any other generic telephony library)19:24
itsnotabigtruckbindi: here's one
itsnotabigtruckthe problem is that unlike other sip servers, asterisk forces all calls through the asterisk server19:25
itsnotabigtruckand it wants to be able to decrypt everything19:25
bindiseems legit :-D19:25
itsnotabigtruckyeah, that site is kinda lame...but it's from the devs of the blink sip client19:25
itsnotabigtruckno phone number integration, that gets you *just* a sip address19:26
bindijust what i want19:26
bindibut i like to be in control :p19:26
bindimight try ot19:26
DocScrutinizerasterisk is no SIP server but a SIP B2B-UA19:26
DocScrutinizereven s/SIP//19:28
DocScrutinizeras * also can interface POTS to SIP, SIP to <random other transport>19:28
DocScrutinizereven act as TAM19:28
bindiright well, only problem i have with it atm is failing to configure it to work with instant messages19:30
DocScrutinizerso basically you can think of * as a SIP proxy rather than a registrar/server19:30
*** piggz has joined #harmattan19:31
phakomeh, used wrong offsets in beginInsertRow.19:31
DocScrutinizerand, unlike native SIP UA on maemo, it supports direct IP2IP mode19:31
DocScrutinizerI.E direct SIP INVITE to which in turn points to my UA19:32
DocScrutinizerwhich I consider the true native way of SIP19:32
itsnotabigtruckDocScrutinizer: of course it's a sip server, it's just also a b2bua19:33
itsnotabigtruckthat isn't mutually exclusive19:33
DocScrutinizerwell, kinda19:33
DocScrutinizeranyway, the common notion you'd need an * somewhere to use SIP is a misconception19:35
itsnotabigtruckbut you do need a sip server of some sort somewhere, and they all have big problems19:35
itsnotabigtruckthere's asterisk, freeswitch, and all the fragmented versions of openser19:35
DocScrutinizerbasically two SIP UA should be able to talk to each other without any registrar or server in between, and esp without need for a b2b ua19:35
bindiwhat is a b2b ua? :<19:36
itsnotabigtruckand if you don't want b2bua that rules out the first one19:36
itsnotabigtruckand freeswitch is a bloated mess19:36
DocScrutinizerback to back user agent - basically two sip-phones ass to ass19:36
itsnotabigtruckand the multiple forks of openser are poorly maintained19:36
DocScrutinizersipgate uses openser19:37
itsnotabigtruckbindi: it's a SIP server that takes the place of the callee's phone19:37
itsnotabigtruckand redirects the actual conversation (media path) through itself19:37
itsnotabigtruckthat's why it's so picky about what it'll let you send through it19:37
itsnotabigtrucka normal SIP server acts more like a router19:38
DocScrutinizerlike a DNS19:38
itsnotabigtruckasterisk doing that allows it to do various pbx-y things like intruding in on the conversation with prompts and stuff19:38
itsnotabigtruckbut it also means no secure conversations and (i guess) no simple19:38
itsnotabigtrucknormally with SIP the two ends negotiate on a codec (so both parties have to share a codec), asterisk can do conversion unlike a real sip server, also19:39
itsnotabigtrucklots of stuff uses openser, it's close to the only's just not very good19:39
itsnotabigtruckwell, it's ok, but it's not getting any better19:39
itsnotabigtruckif you've looked at the openSIPS and kamailio projects, you'll see they're complete disasters19:39
bindii just want something that works :)19:39
bindiyes, kamailio, ugh19:39
DocScrutinizeruse sipgate19:40
DocScrutinizeror freeworlddialup (fwd)19:40
DocScrutinizeror sipphone.com19:40
DocScrutinizeror *any* other of the zillion SIP providers out there19:40
bindii'm paranoid, they all probably sniff my dataz :(19:41
DocScrutinizerbindi: that'S exactly ther point: they CAN'T19:41
itsnotabigtruckbindi: if you use your N9 to talk your dataz can be sniffed no matter what, no zrtp19:41
DocScrutinizeras you data is going straight to peer19:41
DocScrutinizerunlike with * in the middle19:42
itsnotabigtruckDocScrutinizer: well, theoretically, sipgate/etc. could kick in b2bua just for you19:42
itsnotabigtruckand log your media path19:42
itsnotabigtruckit would be detectable, but realistically no one would check19:42
DocScrutinizeryou bet I check19:42
itsnotabigtrucki don't think you even *can* check on the N9, it's not like there's a gui for that19:42
DocScrutinizerbut yes, they could do it on a per-case basis19:42
DocScrutinizerwhen special interest19:42
itsnotabigtruckso i doubt you check, not every time19:42
bindiDocScrutinizer: registering for sipgate a bunch of times with fake addresses seems like hard work :(19:44
binditrying to get free communication between friends going on here :p19:44
DocScrutinizerwhen my UA asks server "please connect me to 001800123456" then I got no clue what's $IP in the answer "send your INVITE to $IP:$port"19:44
*** beford has joined #harmattan19:45
bindibtw, what client do you use on your pc?19:45
DocScrutinizerbindi: can't you find a provider in your own geographical area19:45
bindidoubt there's any finnish one, already tried googling19:45
DocScrutinizeron my PC? twinkle of course19:45
DocScrutinizerit's partially my baby, you don't think I'd ever use anything else ;-D19:47
bindiwell that clearly won't run on windows, ekiga is failing with TCP19:47
DocScrutinizerthe awesome DE l10n is entirely by me19:47
itsnotabigtruckbindi: use blink, it's probably the best windows client19:48
itsnotabigtruckit doesn't have zrtp, but neither does the n919:48
DocScrutinizeralso the script hook concept, mostly19:48
itsnotabigtruckit has mac and linux versions also19:48
DocScrutinizerand partially audio19:48
bindiso DocScrutinizer you seem to know much about asterisk19:48
itsnotabigtruckas for twinkle, the time for unportable software passed a long time ago19:48
DocScrutinizernot really19:48
bindiever configured asterisk 10 to support instant messages?19:48
itsnotabigtruckalso, it looks like twinkle isn't actually maintained anymore, and the website isn't very fashionable19:49
DocScrutinizeritsnotabigtruck: unportable????19:49
bindiummm, blink is giving me "bonjour SDK"19:49
DocScrutinizertwinkle been not only portable but in fact ported since rev0.119:49
itsnotabigtruckDocScrutinizer:'s linux, only, and the authors seem to have all intent to make it stay that way19:49
bindido i want something made by apple19:49
DocScrutinizeritsnotabigtruck: nope, it's definitely NOT linux only19:49
DocScrutinizerit's general unix19:49
RST38hEHLO gentlemen19:50
RST38hHow is suffering tonight?19:50
DocScrutinizerand the main problem with twinkle is the GUI based on Qt3 and using a mutex concept that conflicts with Qt419:51
RST38hitsnotabigtruck: yeah! but how do you press the tiny UI elements???19:51
DocScrutinizerand neither Michel de Boer nor myself have any "intent to make it stay that way"19:52
DocScrutinizerit's just we both don't have the time and motivation to do a major rewrite, or add pulseaudio just for the whiners19:52
DocScrutinizertwinkle "just works" the way it is, and Michel moved on to other hobbies19:53
itsnotabigtruckRST38h: lol, that's kinda the problem19:53
bindiwell ever since i started using tcp with *, phone and pc both connect fine, but calling times out on both19:53
DocScrutinizeritsnotabigtruck: if the website isn't fashionable enough for you, go .....19:54
itsnotabigtruckDocScrutinizer: heh. ok, i was probably being a little harsh19:55
itsnotabigtrucklast time i looked into twinkle i was a bit irritated because i needed a ZRTP client for windows, and twinkle is one of the few ZRTP clients that exist, and it's for linux only19:56
*** risca has joined #harmattan19:56
DocScrutinizertwinkle still used in a remarkable number of call centers etc, and I don't see any need for 'maintenance' as long as it just works19:56
DocScrutinizermodulo the Qt3 problem19:57
DocScrutinizerwhich noone of "the authors" is going to tackle, as we boldly failed in trying19:57
DocScrutinizerwould need a complete rewrite of the GUI part19:58
*** faenil has joined #harmattan19:59
DocScrutinizerand Michel always hated the GUI development duties, he's been interested in the SIP core only, which still is reference class quality and used by *all* SIP providers to check their stuff20:00
itsnotabigtruckalright, fair enough20:01
DocScrutinizeritsnotabigtruck: you're free to port it to windows20:01
itsnotabigtrucki was being rather unfair to your program :/20:01
itsnotabigtruckin any case, there's no sip client for windows, nor sip server for any platform, that i really consider recommendable, and that's incredibly frustrating20:02
itsnotabigtruckand fixing these things usually = a gigantic project20:03
itsnotabigtrucke.g. for the sip server, that pretty much means writing one from scratch, and a lot of the sip clients out there aren't exactly easy to work on20:03
itsnotabigtrucke.g. i thought about putting zrtp support into blink, but the rtp code is a bunch of messy python20:03
itsnotabigtruckbinary protocols + python = usually not such an awesome idea20:04
*** heymaste_ has joined #harmattan20:04
*** heymaster has quit IRC20:07
*** risca has quit IRC20:07
DocScrutinizertwinkle using some lib for RTP, I am not the biggest fan of that lib either20:14
*** risca has joined #harmattan20:14
*** decibyte has quit IRC20:14
DocScrutinizerbut there's libzrtp for it20:15
DocScrutinizer"recently" AIUI they came up with even some echo cancellation in that RTP lib - I already gave up on following the development at that point in time20:16
*** decibyte has joined #harmattan20:16
DocScrutinizerMichel did a last little fancy to add the config to twinkle settings, I heard users don't like it too much20:17
DocScrutinizeritsnotabigtruck: (portable) Michel rejected all my pushing to implement IPC via dbus. Rationale: dbus is not available on all arbitrary platforms. So he used named pipes20:21
DocScrutinizerIIRC we even made sure endianness is not an issue for twinkle20:23
itsnotabigtruck if any of you don't have DX:HR yet, it's $10 right now20:24
DocScrutinizerso if it wasn't for the friggin Qt3 GUI, twinkle was easy to port even to maemo (though I seem to recall some of the libs were a major PITA and in the end all those who tried porting to Mac or N900 or Freerunner gave up due to the libs)20:25
DocScrutinizeranyway one last word about this antiquity SIP client: you can build a plain comandline version that works quite nice and shouldn't need any KDE/Qt/whatever stuff as dependency20:32
itsnotabigtruckDocScrutinizer: right, but it probably has a lot of posixisms20:32
itsnotabigtruckwhich are unfortunately hard to avoid without doing a lot of work to break out platform specific stuff20:32
itsnotabigtruckand if a program was written for *nix it's almost impossible to root all of that out20:32
DocScrutinizeryeah, it at least never was meant for non-*nix platforms20:33
itsnotabigtrucki'm surprised that porting the non-gui libs to maemo would be an issue though...i'd have imagined that's a matter of rebuild and you're set20:33
DocScrutinizeryeah, we all thought that20:34
itsnotabigtruckand the thing is that windows support is a big deal on a sip client, given that you have to get other people to use it, and most of those other people use windows20:34
*** faenil has quit IRC20:35
DocScrutinizerwell, seems the the libccrtp devels are not the brighter half of devel folks20:36
DocScrutinizeror was it libcommoncpp2?or are they even the same bunch?20:38
itsnotabigtrucki think libcommoncpp = a support library, libccrtp = rtp implementation on top of commoncpp20:40
itsnotabigtruckDocScrutinizer ^20:40
DocScrutinizerone of both gave all the guys that tried to port twinkle quite some headache20:41
itsnotabigtrucklooks like rzr accepted milkytracker, so it'll be in the repo in a few mins20:42
DocScrutinizerport to Mac or non-*nix platforms that is20:42
itsnotabigtruckcalendar-live-icon is however unresolvable due to the whole libtimed-dev problem20:42
*** NIN102 has joined #harmattan20:46
*** NIN101 has quit IRC20:46
*** risca has quit IRC20:56
*** risca has joined #harmattan20:56
*** risca has quit IRC20:56
Aranelitsnotabigtruck: <credential name="UID::root" /> <for path="/opt/app/bin/" />21:00
Aranelitsnotabigtruck: "sh" returns "cp: can't create '/etc/init/apps/meesaver.conf': Permission denied"21:00
itsnotabigtruckAranel: hm...who owns /etc/init/apps/21:01
itsnotabigtruckremember that just being root doesn't allow you to write anywhere, you need CAP::dac_override for that21:01
itsnotabigtruckbut i'd expect that /etc/init/apps/ is owned by root, so you should be able to write there21:01
pawhy is smartcam free for a lot of platforms but not meego?21:02
itsnotabigtruckpa: do the other ones have ads? if so that's why probably21:02
pai dont know, i just know they are available on sourceforge21:02
Aranelitsnotabigtruck: Root has the rw privileges for that folder, but even after aegis privileges that sh script can't write to there.21:03
itsnotabigtruckAranel: how about doing devel-su (don't do develsh afterward) and echo test > /etc/init/apps/meesaver.conf21:10
Aranelitsnotabigtruck: in the .sh file itself?21:10
itsnotabigtruckalso remember that giving UID::root to stuff through aegis is dangerous, you're literally making something setuid root21:10
itsnotabigtruckAranel: nah, just do it from the hsell21:10
Aranelitsnotabigtruck: well I was looking for a way to enable user add/remove my app daemon on UI itself21:11
Aranelitsnotabigtruck: when user clicks Add-to startup it copies it's daemon .conf file to /etc/init/apps, if user wants to remove it runs a rm on that .conf file.21:12
jonniAranel: well you can add root and user credentials to your binary and do setuid(0) when you create conf file and then setuid back to user21:12
jonnino need to do shell scripts21:13
Aranelitsnotabigtruck: It works well for devel-su root, but doesn't via .sh script.21:13
jonnimost likely your .sh script failed because you didnt set root group, as making root.user conf most likely fails, as its not root.root21:14
itsnotabigtruckjonni: does that work without CAP::setuid?21:14
itsnotabigtruckAranel: the point was to test...the fact that it works from the shell but not the script is very weird21:14
itsnotabigtruckput accli -I in your script and make sure it prints out UID::root21:15
jonniitsnotabigtruck: lets, me check, but if I remember right it doesnt need that cap21:15
Aranelitsnotabigtruck: nah, it didn't gave UID:root. It's still UID::user21:16
itsnotabigtruckAranel: there we go21:16
itsnotabigtruckare you sure your manifest looks like21:17
itsnotabigtruck<aegis><request policy="set"><credential name="UID::root" /><for path="/opt/app/bin/" /></request></aegis>21:17
itsnotabigtruckmaybe throw in a <credential name="GID::root" /> just to be sure21:17
jonniand if you run sh it will always fail21:18
itsnotabigtruckok, add the GID::root also, and combine them into one request21:18
jonniyou need to sh with full patch21:18
itsnotabigtruckoh, and yeah21:18
jonnish /opt/appt/bin/addtoboot.sh21:18
jonniaegis doesnt give credentials if you dont use full path21:18
itsnotabigtruckjonni: hm? i thought either one won't work, you have to execute it through bitfmt_whatever21:18
itsnotabigtruckso just /opt/app/bin/addtoboot.sh21:18
pabut does N9 have FM radio tuner or not?21:18
jonnipa yes FM radio works nicely on N921:18
itsnotabigtruckpa: yup, and there's apps to tune it coming out now21:18
pacool.. i read some page saying it didnt have21:19
jonniitsnotabigtruck: I just installed the FM radio deb today, and radio works nicely21:19
pais ti free?21:19
jonniyes its free21:19
AranelQML Radio with javispedro's packages are great ^^21:19
pacool :)21:19
ZogG_laptopis there any transmission options as well?21:22
jonniZogG_laptop: not yet atleast, since javis only has N950 and that hw doesnt have transmission21:22
itsnotabigtruckAranel: less /var/lib/aegis/restok/restok.conf, and make sure everything is listed there for your package21:22
jonninot sure if N9 has transmission hw21:22
itsnotabigtruckjonni: don't they both have wl1273 inside?21:22
itsnotabigtruckand that chip includes fm xmit21:23
itsnotabigtruckthe issue was whether there's an antenna attached to it21:23
jonniI dont remember, havent looked the hw specs for a while21:23
Aranelitsnotabigtruck: my manifest.aegis now looks like , but for some reason accli -I still doesn't show UID and GID root.21:25
itsnotabigtruckAranel: you can't have more than one request for a particular binary21:25
itsnotabigtruckbut you can have more than one binary in a request21:25
itsnotabigtruckso combine those all into 1 request block21:25
itsnotabigtruckwith 2 credentials and 2 fors21:26
itsnotabigtruckthen install it and check restok.conf21:26
pabut does the N9 even have an FM radio transmitter on board???21:28
itsnotabigtruckpa: it does, but it may or may not actually be connected to anything21:28
pagsmarena says..21:28
pabut how to use the transmitter?21:28
itsnotabigtruckalso i'm not sure if the drivers to interface with it actually exist either21:28
pai mean for what?21:28
*** NIN102 has quit IRC21:28
itsnotabigtruckthere's been efforts to use the wl127x fm transmitter on some android phones, but after a bit of searching i couldn't find any positive results21:29
itsnotabigtruckif the android people make some headway it should be possible to adapt that for N921:29
pabut what to use it for?21:29
itsnotabigtruckassuming there's an antenna connection21:29
pai mean what use could it have?21:29
itsnotabigtrucktransmitting fm audio :p21:29
paah, you need an antenna21:29
pabut there+s no place to plug it..21:30
itsnotabigtruckallowing you to listen to your phone from your car21:30
Aranelitsnotabigtruck: Package: meesaver Request: UID::root, accli -I still doesn't show that cred :|21:30
*** NIN101 has joined #harmattan21:30
itsnotabigtruckproblem is that fm transmitters are uniformly super weak21:30
pareceiving of course has a point21:30
itsnotabigtruckso unless there's some way to put this thing into overdrive21:30
pabut could it be used to transmit in short range?21:30
itsnotabigtruckit isn't terribly great anyway21:30
itsnotabigtruckyeah, that's what it's for21:30
palike using it for walkie talking or something like that?21:30
itsnotabigtruckno, it's for tuning your car's radio into music/audio from your phone21:31
itsnotabigtruckway too weak for two way communications21:31
paah now i see21:31
payou deliver audio via fm21:31
itsnotabigtruckfm transmitters are usually too weak even to have a clear signal in your car, a couple feet away21:31
itsnotabigtruck(from the antenna)21:31
itsnotabigtruckmy understanding is that the fcc has pressured fm transmitter vendors to lower the transmit power well below the legal maximum, which is already very low21:32
itsnotabigtruckso they're worthless unless you can defy the fcc and amp them up21:32
pabut if it works in the car, wel, that might be ok?21:33
itsnotabigtruckno, that's the problem, they don't even work in the car21:34
itsnotabigtruckin almost all areas, there's interference over the entire fm spectrum21:34
itsnotabigtruckeven if it isn't a strong enough signal to produce any discernible audio21:34
itsnotabigtrucktypical fm transmitters are too weak to overpower interference from an extremely distant high power station21:34
trxi use the fm transmitter all the time in my car21:35
itsnotabigtruckso the result is that even if you can find a completely unoccupied frequency and tune into it, you'll get interference in a couple miles21:35
Chaz6I'm glad for fm still, dab is rubbish21:35
trxand it works brilliantly21:35
itsnotabigtruckit depends on the car too, some cars have way more sensitive radios21:35
itsnotabigtruckand it depends on the transmitter, some are stronger than others21:35
itsnotabigtruckeven if you don't mod them21:35
itsnotabigtruckand also the antenna placement in the car21:35
itsnotabigtruck(front vs back)21:35
trxbut i have detached my car antenna and attached a piece of copper wire instead21:36
itsnotabigtruckah, see21:36
itsnotabigtruckthat's a good idea...21:36
itsnotabigtruckChaz6: what's so bad about it?21:36
itsnotabigtruckit's kinda ridiculous how they deployed hd radio in the us instead, which is totally incompatible with everything else21:37
itsnotabigtruckthat is one nice thing about fm, it's universal21:37
Chaz6itsnotabigtruck: it's so scratchy21:37
Chaz6I only have to walk through the room for it to go bad21:37
Chaz6when it goes bad it's far worse than fm21:37
itsnotabigtrucki've had ok results with hd radio, but there's very little that supports it21:37
Chaz6It's like scratching nails on blackboards21:37
itsnotabigtruckthe zune hd is pretty much the only portable with hd radio21:37
Chaz6The best signal we can get here apart from fm is satellite21:37
Chaz6dab is the worst21:37
Chaz6satellite radio in the uk is 192kbs21:38
Chaz6dab is only 128kbs21:38
itsnotabigtruckback at home there's a station that broadcasts local indie music, on AM21:38
itsnotabigtruckso dab being scratchy is probably nothing compared to that, lol21:39
itsnotabigtruckAranel: did you sort your stuff out?21:39
itsnotabigtruckhm, i better get going21:39
paby the way, is the fm radio receiver on sourceforge or ovi developer?21:40
Aranelitsnotabigtruck: combined them to 2 <request>'s,
* RST38h ritually murders a hamster21:41
itsnotabigtruckAranel: that's still not valid...21:41
itsnotabigtruck1 request, with 2 credential elements, and 2 for elements21:41
itsnotabigtruckyou can only have 1 for element for a particular path in the entire file21:41
itsnotabigtrucki.e. don't repeat them like you did21:41
itsnotabigtruckit's very simple, you're requesting UID::root + GID::root, for /opt/ and /opt/blah2.sh21:42
itsnotabigtruckalso are you sure about the /bin/bin?21:42
itsnotabigtruckthat doesn't really look right...21:42
itsnotabigtruckthis document is pretty cryptic but if you can make your way through it, it does a pretty good job of describing the aegis manifest format
Aranelitsnotabigtruck: like this one? , yup I'm sure about that directory, it's "bin" twice.21:43
itsnotabigtruckAranel: closer, get rid of the 2nd two credential elements21:43
itsnotabigtruckconceptually you're making a single request for privileges, requesting UID::root and GID::root, applying them to the two .sh files21:44
Aranelitsnotabigtruck: oops
itsnotabigtruckAranel: looks legit21:45
itsnotabigtruckso install that, then check restok.conf21:45
Aranelitsnotabigtruck: sh and accli -I  failed, restok.conf shows as Request: UID::root21:46
Aranelitsnotabigtruck: sh /opt/app/bin/bin/ gave permission denied, accli -I didn't show required creds.21:47
itsnotabigtruckAranel: drop the sh21:48
itsnotabigtruckaegis doesn't do anything if you run the interpreter instead of the script21:48
itsnotabigtruckif you were doing that, that's the whole problem21:48
itsnotabigtrucki think jonni was a bit off earlier, if you run sh, it doesn't matter what you use as the path21:48
itsnotabigtruckbecause aegis considers you to be executing sh, not /opt/app/bin/bin/addtoboot.sh21:49
Aranelitsnotabigtruck: /opt/MeeSaver/bin/bin $ ./ gave "permission denied" too.21:49
itsnotabigtrucktry literally /opt/app/bin/bin/addtoboot.sh21:49
Aranelitsnotabigtruck: /opt/MeeSaver/bin/bin $ /opt/MeeSaver/bin/bin/ , "permission denied" again21:50
itsnotabigtruckwhat about the accli output?21:50
Aranelitsnotabigtruck: accli -I /opt/MeeSaver/bin/bin/ shows as UID::user and GID::users21:51
itsnotabigtruckerm, wait21:51
itsnotabigtruckyou have to run accli -I from the script21:51
itsnotabigtruckmodify and reroll the package21:51
itsnotabigtruckyou can't modify it on the phone either, that will break the hash and cause running it to produce Operation not permitted21:51
itsnotabigtrucki don't think specifying an argument to accli -I even does anything21:52
*** jaywink has joined #harmattan21:52
itsnotabigtruckAranel ^21:52
itsnotabigtrucki gotta roll, bbiaf!21:52
Aranelitsnotabigtruck: sure, thanks for helping me out =)21:53
Aranelitsnotabigtruck: re-rolled the package, /opt/MeeSaver/bin/bin/ gave UID::user and GID::users again.21:54
DocScrutinizerFMTX on N9 won't work, no matter what tools you use to drive it: it has no antenna attached21:59
*** dymaxion has quit IRC22:00
*** dymaxion has joined #harmattan22:01
jonniAranel: aranel you should check restok.conf file untill you see the root in there, and then you might need to reboot the machine, since sometimes if you update your package from different credential to another reboot is needed22:04
Araneljonni: it shows UID:root for my package on restok.conf, but doesn't for accli -I on those sh scripts.22:06
jonniAranel: and you have rebooted your N9?22:06
jonnias there is that aegis cache thing that it remembers the older credentials if path is same22:06
Araneljonni: by the way, If I grant UID::root and run "python /opt/app/" from it, would have the same creds or not?22:09
Araneljonni: I'm thinking of using os.setuid() as you suggested.22:09
jonniif I remember right python inherits the credentials, so I would say yes it would have the same creds22:10
Araneljonni: os.setuid(0) > OSError: [Errno 1] Operation not permitted22:11
Araneljonni: d'oh :|22:11
jonni(python is not installed on device by default, so you need also add  depends: python on your debian control file....22:11
jonniAranel: really hard to debug without seeing your sourcecode22:12
Araneljonni: it's a slot named "addboot", os.setuid(0), os.setgid(0), os.system("sh /opt/path/to/"), os.setuid(29999) and os.setgid(29999)22:14
jonnic++, perl, python, sh or what22:14
jonniwithout seeing the sources there is no idea22:14
Araneljonni: well it's more than one file, but wait a moment22:14
jonniand you rebooted N9 after giving root credentials to aegis manifest?22:14
*** xmlich02_ has joined #harmattan22:16
jonniand manifest file for everything? or git repo?22:17
DocScrutinizerjonni: (cache) I thought the kernel aegis cache works via inodes22:17
DocScrutinizernot paths22:17
*** arcean has quit IRC22:17
itsnotabigtruckyou shouldn't need to do any setuid stuff22:18
itsnotabigtruckAranel: could you post the restok extract for your package22:18
Araneljonni: for manifest.aegis, "python /opt/path/to/" is the single line of main.sh22:19
itsnotabigtruckAranel: it should normally be called packagename.aegis22:19
DocScrutinizeranyway AFAIK kernel aegis cache gets setup during boot, so any later modifications in restok won't take any effect until reboot?22:19
itsnotabigtruckDocScrutinizer: no, because dpkg reloads the aegis state22:19
DocScrutinizeraaah, ok22:19
itsnotabigtruckAranel: wait, does your script have a shebang line in it?22:20
itsnotabigtrucki.e. #!/bin/sh22:20
Aranelitsnotabigtruck: "The Aegis manifest file is named manifest.aegis inside Qt Creator projects. The file located is in Qt Creator project files directory qtc_packaging/debian_harmattan/. At the start of a build, Qt Creator copies this file to debian/<package-name>.aegis. "22:20
Aranelitsnotabigtruck: harmattan-docs said that was okay :|22:20
itsnotabigtruckoh ok22:21
jonnior do you have the .deb file, might be easier just to debug that one22:21
itsnotabigtrucki haven't even touched qt creator, so yeah22:21
itsnotabigtruckwhat about the shebang22:21
Aranelitsnotabigtruck: nope It doesn't have that line22:22
Aranelitsnotabigtruck: restok.conf
itsnotabigtruckAranel: there's the answer22:23
itsnotabigtruckput in the #!/bin/sh22:23
jonnimost likely if you prefix with "sh " or "python " you just drop your privileges, you should remove those22:23
itsnotabigtruckalways always always include a shebang line in all scripts with +x permission22:23
itsnotabigtruckjonni: the problem was that executing a script without a shebang line invokes magic behavior in the shell22:24
itsnotabigtruckso it was doing the equivalent of sh script.sh22:24
itsnotabigtruckinstead of passing it to the kernel and having aegis do its thing22:24
Araneljonni: I'm running it as /opt/MeeSaver/qml/meesaver/main.sh22:24
itsnotabigtruckalso Aranel: where's the GID::root22:24
itsnotabigtruckthough that isn't actually needed22:25
itsnotabigtruckbut i thought you put that in, strange that it isn't there22:25
itsnotabigtruckanyway, try the shebang22:25
jonniah yes, if .sh is missing #!/bin/sh then that could explain it22:25
Aranelfor python code, it didn't change anything, OSError: [Errno 1] Operation not permitted again.22:26
Aranelnow I'll try it on external scripts ( to see if it works on them.22:26
itsnotabigtruckAranel: did you include the shebang line in the python script...22:27
itsnotabigtruckthis isn't really that complicated22:27
itsnotabigtrucka) an executable script is a script with +x permission22:27
itsnotabigtruckb) executable scripts MUST have a shebang line with the interpreter to use22:27
itsnotabigtruckfor python that's /usr/bin/python, for shell it's /bin/sh22:28
itsnotabigtruckc) aegis privileges set on a script only apply if you execute the script22:28
itsnotabigtruckpaths don't matter, but executing it vs not executing it does22:28
Aranelitsnotabigtruck: yup it works for external scripts now ^^22:30
*** rantom has quit IRC22:30
Aranelitsnotabigtruck: and accli -I shows UID::root too.22:30
merlin1991d) aegis privileges on a script only work if dpkg builds a hash for the script, and that is only done if it has a shebang22:30
*** beford has left #harmattan22:33
*** beford has joined #harmattan22:33
*** Free-MG has quit IRC22:35
DocScrutinizerI doubt that, but it's not applied when it isn't 'executable' (read: missing the shebang)22:39
DocScrutinizeraegis privileges AIUI get applied to executables only22:39
DocScrutinizeractually: on execution22:39
DocScrutinizerwhich implies it's an executable22:40
itsnotabigtruckAranel: also, for your python, you can leave the shebang out and use /usr/bin/python x22:41
itsnotabigtruckonce you've already acquired the permission through the script22:41
itsnotabigtruckbut if the script does nothing but run the python, that's unnecessary22:41
merlin1991DocScrutinizer: you can doubt it, but I'm right :D22:42
DocScrutinizeras itsnotabigtruck just explained, a script without shebang will make ld-linux barf up when shell tries to feed it, and then shell takes over and starts its magic which means it's actually invoked like "sh <yourscript>"22:42
merlin1991upon installing your .deb the manifest get's read and each "path" in the manifest that has no sha1 has in the digsigsums file gets ignored22:43
*** piggz has quit IRC22:43
DocScrutinizerand I guess aegis is implemented/hooked-in to ld-linux22:43
merlin1991the digsigsums file in term get's created by some code that checks each file in the build for elf header or a shebang22:43
merlin1991and only hashes those22:44
*** Mek has quit IRC22:44
*** eean has quit IRC22:44
*** spenap has quit IRC22:44
*** lfranchi has quit IRC22:44
*** torarne has quit IRC22:44
*** MohammadAG has quit IRC22:44
*** jkk_ has quit IRC22:44
*** jpwhiting has quit IRC22:44
*** acidjunkie has quit IRC22:44
*** jkt has quit IRC22:44
*** aikakana has quit IRC22:44
merlin1991so bascially even if there would be no problem with ld,... you'd never get your agegis manifest past dpkg -i because it would ignore your settings22:44
*** Mek has joined #harmattan22:45
*** eean has joined #harmattan22:45
*** spenap has joined #harmattan22:45
*** lfranchi has joined #harmattan22:45
*** MohammadAG has joined #harmattan22:45
*** torarne has joined #harmattan22:45
*** jkk_ has joined #harmattan22:45
*** jpwhiting has joined #harmattan22:45
*** acidjunkie has joined #harmattan22:45
*** jkt has joined #harmattan22:45
*** aikakana has joined #harmattan22:45
merlin1991hm that should have been: upon installing your .deb the manifest (_aegis in the .deb) gets read and only paths with a valid checksum (sha1 saved in the digsigsums file) are actually added to aegis22:46
*** heymaster has joined #harmattan22:47
Aranelyay it works both for adding and removing daemon now =) Thanks guys :) Now I have to figure out why app shortcut doesn't work properly and it's off to Ovi Store ^^22:48
*** heymaste_ has quit IRC22:48
merlin1991btw the package in question that does the hashing inside scratchbox is refhashutils (see for its crypitc invocation by dpkg-buildpackge)22:50
*** merlin1991 has quit IRC22:51
*** merlin1991 has joined #harmattan22:52
merlin1991hm what would be the easiest operation (done from python) that requires an aegis token?22:57
DocScrutinizersu ?23:03
*** Iktwo has joined #harmattan23:08
DocScrutinizeros.system("su") actually23:13
IktwoHi, I've made a very simple wifi scanner using wireless tools, I want to share it with you but I don't know how to request root access in my app, I can use it normally when launche with QtCreator, but in the phone I need to open it as root.. can you help me? then I'll post it on TMO Thanks23:14
*** risca has joined #harmattan23:18
infobot , or "The purpose of this framework is: ... to make sure that the platform meets the requirements set by third party software that requires a safe execution environment.", or, or
DocScrutinizerIktwo: you have still a long way to go ;-)23:19
DocScrutinizeractually it seems to me the last topic on this chan, ending like 2 lines above your post, was about exactly same problem23:20
beford~!seen admiral023:21
IktwoDocScrutinizer: I understand how aegis works, I'm not planning to release my app in the store.. I just tought it may be useful for some else and want to share it23:21
beford~seen admiral023:21
infobotadmiral0 <> was last seen on IRC in channel #harmattan, 2d 2h 59m 52s ago, saying: '?'.23:21
IktwoI saw an app (didn't test it) that let's you change the background by a editing a protected file, it requires developer mode if I remember correctly23:22
Iktwoso I tought there's was a way of lauching my app as root or as a developer, I'm not talking about any hacking, I just need to know if it's possible by editing the .desktop launcher or smething..23:23
AranelIktwo: *now I'm talking by blind guess and may say something stupid but..* aegis-exec, maybe?23:24
Sazpaimonthinking of buying a white N9 case23:25
Sazpaimonas in the housing23:25
Sazpaimonare they incredibly difficult to swap out?23:25
IktwoAranel: thanks I'll see if that works23:27
*** xmlich02_ has quit IRC23:27
*** longthen1 has joined #harmattan23:32
*** longthen has quit IRC23:34
itsnotabigtruckwhoa, i was scrolled way up23:37
*** aergus has joined #harmattan23:37
itsnotabigtruckAranel: so what's the app?23:37
*** rigow has joined #harmattan23:37
itsnotabigtruckIktwo: remember, alignment23:38
Iktwoa WiFi Scanner using wireless-tools23:38
itsnotabigtruckmake sure that the padlock is aligned with wifik, the channel is aligned with iktwo, the bssid is aligned with the ssid and the bar23:38
Iktwoit's very simple, I just parse the output of iwlist wlan0 scan23:38
rigowhi all, is there any FAQ on the latest dev-update? My openssh-server on my N9 doesn't start anymore. Google doesn't offer help23:39
itsnotabigtruckiktwo: as for the aegis stuff, what's the problem?23:39
itsnotabigtruckfor root access, you can just add that to your aegis assertions23:39
itsnotabigtruckjust remember to be careful, you're making something setuid root that way23:39
itsnotabigtruckactually, hmm23:39
itsnotabigtruckyou probably don't need to be root23:39
itsnotabigtruckyou need CAP::net_admin23:39
itsnotabigtruckor one of the other caps23:40
itsnotabigtruckcheck if the scanner works under develsh, without devel-su23:40
Iktwoif I run the app from QtCreator it works I can succesfully use iwlist wlan0 scan, if I just launch my app from it's icon it doesn't work23:40
itsnotabigtruckok, put in an aegis manifest asserting that capability23:40
itsnotabigtrucksee if it starts working23:40
Iktwoit work's if I do develsh first23:40
itsnotabigtruckok, then that's it23:41
Aranelitsnotabigtruck: It's a small app to select a random lockscreen background (slideshow) every X seconds from given folder23:41
Aranelitsnotabigtruck: nothing big/useful in fact =) just to educate me about Harmattan development.23:41
itsnotabigtruckyou need something like <aegis><request><capability name="CAP::net_admin" /><for path="/opt/path" /></request></aegis>23:41
itsnotabigtruckiktwo ^23:41
itsnotabigtruckoh, also23:42
itsnotabigtruckcheck what's in the current aegis manifest23:42
*** leinir has quit IRC23:42
itsnotabigtruckyou're probably using qt creator, and qt creator normally attempts to guess what capabilities you need23:42
itsnotabigtruckthat means if you add your own manifest you might lose capabilities from what you had23:42
*** leinir has joined #harmattan23:42
itsnotabigtruckon the other hand i don't think a simple gui app with no audio needs special capabilities23:42
*** leinir has quit IRC23:42
*** leinir has joined #harmattan23:42
Iktwooh.. let me check23:42
itsnotabigtruckIktwo: remember that many things that involve interacting with the kernel require linux capabilities23:43
itsnotabigtruckon a normal linux system, becoming root gives you all the capabilities23:43
itsnotabigtruckwith aegis, they're two separate things23:43
DocScrutinizernot on aegis :-P23:43
itsnotabigtruckdevel-su gives you root, develsh gives you some (but not all) capabilities23:43
itsnotabigtruckand having root doesn't mean you have root access, you're just the root user23:44
DocScrutinizeratually on aegis/HARM root has less capabilities than user, it seems23:44
itsnotabigtruckso manipulating the wireless state is a network admin action23:44
Iktwothanks again :)23:44
itsnotabigtruckso it requires CAP::net_admin23:44
DocScrutinizerthough honestly that's a poor design when iwlist needs CAP:net-admin to *list*23:45
itsnotabigtruckDocScrutinizer: well, it probably lists by telling the wifi chip to do a scan23:45
itsnotabigtruckso conceptually you're changing the wifi state, sort of23:46
itsnotabigtrucki guess23:46
DocScrutinizerut what's so extremely nasty in asking wireless chip to scan?23:46
itsnotabigtruckwell, you could do it a zillion times in a loop :p23:46
DocScrutinizerno, you can't, when driver design is OK23:46
Iktwommm it's weird, I added CAP::net_admin and now sometimes networks are shown sometimes not, but running under develsh works perfectly.. I'm gonna upload it and the code maybe someone want to test it23:47
itsnotabigtrucki mean, in this case that isn't aegis' fault, it's the linux kernel's fault23:47
DocScrutinizerbut well, you need CAP::blind-user to flash the indicator LED ;-P23:47
itsnotabigtruckiktwo: hrm...23:47
Aranelitsnotabigtruck: do you know how QtCreator creates app shortcuts? It generates a desktop file everytime, does it -kind of- wrong and even If I change that file, generates it again on next  build.23:47
*** heymaster has quit IRC23:47
itsnotabigtruckIktwo: try asserting all the network capabilities23:47
itsnotabigtruckthen remove them one by one23:47
IktwoAranel: I use QtCreator and it doesn't generate my files every build..23:48
*** heymaster has joined #harmattan23:48
itsnotabigtruckalso Iktwo: make sure you're doing error checking23:48
itsnotabigtruckif iwlist bombs out, it'll return a non-zero value23:48
*** zk8 has joined #harmattan23:48
itsnotabigtruckcheck for that23:48
AranelIktwo: don't you have yourappname.desktop on your project folder?23:48
itsnotabigtruckit might be a good idea to log the iwlist output somewhere too23:49
itsnotabigtruck(not sure what you're using to spawn iwlist)23:49
IktwoAranel: yes, I think that if you use the default project it creates something like app.desktop and app_harmattan.desktop23:49
itsnotabigtruckiwlist probably reports error messages on stderr23:50
IktwoAranel: I edit the _harmattan one and it work's23:50
Iktwoitsnotabigtruck: I'm not getting output at all.. when no networks are shown, also if I open a terminal and as a user I do /usr/sbin/iwlist wlan0 scan it works23:51
itsnotabigtruckIktwo: are you monitoring stderr?23:51
itsnotabigtruckerrors aren't reported on stdout23:51
itsnotabigtruckhowever, you should only use stderr for debugging anyway, use the return value for the actual success check23:52
Iktwooh.. then no, no error checking..23:52
itsnotabigtruckhow are you spawning the process? i've had luck with glib's process spawning api23:52
itsnotabigtruckbut qt probably has something too23:52
rigowlooks like there is no FAQ23:52
itsnotabigtruckrigow: don't think there is one, just ask! :D23:53
itsnotabigtruckIktwo: anyway, you need to know how it's failing to know how to make it not fail23:53
itsnotabigtruckif it's not failing with "operation not permitted" it's probably not aegis23:53
itsnotabigtruckoh, also, idea:23:54
itsnotabigtruckfrom develsh: aegis-exec -c -a CAP::net_admin -a CAP::net_raw ... /usr/sbin/iwlist wlan023:54
itsnotabigtruckyou can use aegis-exec to constrain the caps used when launching iwlist23:54
itsnotabigtruckmuch faster than permuting your manifest and reinstalling over and over23:55
rigowitsnotabigtruck, my sshd stopped working23:55
rigowfigured that there is no /etc/init.d/ssh anymore23:55
rigowstarting sdk on WLAN I get connection refused and ps aux shows no running sshd23:56
* rigow finds that strange23:56
rigowI think I will try to re-install the openssh server23:56
rigowbut on meego-terminal, this is really not pleasant ;)23:57
itsnotabigtruckrigow: harmattan uses upstart, not sysvinit23:57
itsnotabigtruckso everything's in /etc/init23:57
itsnotabigtruckdo ls /etc/init/ssh.conf23:57
itsnotabigtruckfrom meego terminal23:57
rigowkernel 3.023:57
itsnotabigtruckwait, what?23:57
itsnotabigtruckalso, you can restart ssh with: /sbin/initctl restart ssh23:58
itsnotabigtruckthere isn't a kernel 3.0 for the N9 i'm aware of23:58
itsnotabigtruckrigow ^23:58
rigowah, just reacting on upstart, now also in opensuse 12.123:59
rigowand has kernel 3.223:59
itsnotabigtruckfwiw harmattan ships with 2.6.32.something iirc23:59

Generated by 2.15.1 by Marius Gedminas - find it at!