Thursday, 2014-10-23

*** zalan has joined #nemomobile00:17
*** gabriel9 has quit IRC00:21
*** jonwil has quit IRC00:23
*** vakkov has quit IRC01:34
*** jonwil has joined #nemomobile01:36
*** Dynamit has quit IRC01:56
*** vakkov has joined #nemomobile02:03
*** DrCode has quit IRC02:06
*** DrCode has joined #nemomobile02:29
*** Morpog_PC has joined #nemomobile02:44
*** Behold has quit IRC03:11
*** zhxt has quit IRC03:21
*** zhxt has joined #nemomobile03:34
*** martyone__ has joined #nemomobile04:03
*** cloanta has joined #nemomobile04:12
*** cloanta_ has quit IRC04:15
*** cloanta_ has joined #nemomobile04:15
*** cloanta has quit IRC04:17
*** Oksana has quit IRC04:24
*** cloanta has joined #nemomobile04:30
*** cloanta_ has quit IRC04:32
*** piggz has quit IRC04:37
*** piggz has joined #nemomobile04:38
*** cloanta_ has joined #nemomobile04:42
*** cloanta_ has quit IRC04:43
*** cloanta has quit IRC04:45
*** Morpog_PC has quit IRC04:58
*** hurrian has quit IRC05:01
*** Guest4374 is now known as Termana05:27
*** WWDrakey has joined #nemomobile05:28
*** sletta has joined #nemomobile05:50
*** VDVsx has quit IRC05:51
*** VDVsx has joined #nemomobile06:00
*** Pat_o has joined #nemomobile06:04
*** Sinbad has quit IRC06:05
*** zhxt_ has joined #nemomobile06:10
*** Termana has quit IRC06:10
*** tigeli has quit IRC06:10
*** Jonni has quit IRC06:10
*** w00tc0d3 has quit IRC06:10
*** kido has quit IRC06:10
*** tigeli has joined #nemomobile06:10
*** trollixx has quit IRC06:10
*** zhxt has quit IRC06:11
*** trollixx has joined #nemomobile06:11
*** amccarthy_ has joined #nemomobile06:13
*** Jonni has joined #nemomobile06:13
*** kido has joined #nemomobile06:14
*** amccarthy has quit IRC06:14
*** Termana has joined #nemomobile06:15
*** Termana is now known as Guest9592506:16
*** Guest95925 has quit IRC06:20
*** Morpog_Jolla has joined #nemomobile06:22
*** VDVsx_ has joined #nemomobile06:25
*** VDVsx has quit IRC06:25
*** Sfiet_Konstantin has joined #nemomobile06:36
*** sp3000 has quit IRC06:40
*** Morpog_Jolla_ has joined #nemomobile06:44
*** Morpog_Jolla has quit IRC06:45
*** Dynamit has joined #nemomobile06:49
*** Sfiet_Konstantin has quit IRC06:57
*** ryukafal1 has joined #nemomobile06:59
*** chriadam_ has joined #nemomobile07:00
*** alterego_ has joined #nemomobile07:01
*** Morpog_Jolla has joined #nemomobile07:03
*** Morpog_Jolla_ has quit IRC07:03
*** morphis has quit IRC07:05
*** Svetlana has joined #nemomobile07:05
*** ryukafalz has quit IRC07:05
*** alterego has quit IRC07:05
*** tigeli has quit IRC07:05
*** WWDrakey has quit IRC07:05
*** chriadam has quit IRC07:05
*** teve has quit IRC07:05
*** teve has joined #nemomobile07:08
*** hedayat has joined #nemomobile07:15
*** morphis has joined #nemomobile07:16
*** Morpog_Jolla has quit IRC07:16
*** Morpog_Jolla has joined #nemomobile07:16
*** Svetlana has quit IRC07:17
*** Morpog_Jolla_ has joined #nemomobile07:20
*** Morpog_Jolla_ has quit IRC07:21
*** Morpog_Jolla has quit IRC07:21
*** mjones has quit IRC07:21
*** veskuh has joined #nemomobile07:28
*** cxl000 has joined #nemomobile07:37
*** piggz has quit IRC07:56
*** mjones has joined #nemomobile08:03
*** notmart has joined #nemomobile08:03
*** tigeli has joined #nemomobile08:19
*** faenil has joined #nemomobile08:21
*** ChanServ sets mode: +o faenil08:21
locusfmorning faenil08:21
faenilmorning gusy08:22
hedayatfaenil: I gave up! :P08:26
faenilhedayat: oh noes! :D08:26
hedayatfaenil: 'systemd --user' is not designed to be run manually08:27
faenilhedayat: what did you discover08:27
faenilAard said he could try help by having a look at their git history...let's see if he's awake :)08:27
hedayatfaenil: the problem is that, people were "abusing" 'systemd --user' as a session manager, but newer systemd versions have broken that setup08:28
hedayatfaenil: I've found a working solution, but that's more or less the same as Jolla guys, or the xlogin method I shown before08:29
hedayatfaenil: You should not try to start 'systemd --user', instead, you should login, and it'll be started automatically by logind08:30
faenilhedayat: good info ;)08:30
*** tigeli has quit IRC08:30
faenilso we're back to square 108:30
faenilwe have to try agetty to autologin :)08:30
hedayatfaenil: No, why that? A simple .service with PAMName=login would do the work, and that process should run until user decides to power off the phone08:31
faenilhedayat: which is exactly what Jolla is doing :P08:31
faenilbut we wanted to look for something less hacky, if possible08:31
faenilI don't believe systemd's ideal way is to have a dumb app running to keep your machine alive08:32
hedayatfaenil: Yes. e.g. It was almost enough to put a while in start-user-session, and include PAMName=login/User=nemo in its service file.08:32
hedayatfaenil: The problem is that, the whole solution is a hack,08:32
hedayatfaenil: 'systemd --user' is NOT a session manager, but Nemo is using it in this way08:33
hedayatfaenil: The 'while true; do sleep(LOGN TIME); done' is what is supposed to be The session manager08:34
faenilI see..08:34
hedayatfaenil: 'systemd --user' is designed to be a 'user manager', or 'user service manager'. It is designed to start user services the first time he log in, and terminate them when he closes all his sessions.08:35
faenilis that so different from a session manager? :/08:35
*** tigeli has joined #nemomobile08:35
hedayatfaenil: yes, completely.08:35
faenileducate me :)08:36
hedayatfaenil: actually, having /usr/lib/systemd/user/ to start UX means this:08:36
hedayatfaenil: when I ssh into Nemo using root user, it'll try to bring up yet another UX for root!08:37
faenilthe main difference is the user should be able to  have multiple sessions08:38
faenilbut in our case it's 1:1 ...08:38
faenilhedayat: while pam_systemd only starts a new UX if there is no user@ *at all* ?08:39
hedayatyes. And 'systemd --user' is NOT supposed to be running in user's sessions. It is not attached to user's session by default, and it creates an independent session for itself.08:39
hedayatfaenil: yes08:39
hedayatfaenil: If the phone allowed multiple logins, a user should be able to login twice and get two separete UXes,08:40
hedayatfaenil: but in the current setup, the second login would bring up nothing08:40
faenilwhich "current" setup08:40
hedayatbringing up UX using systemd user services08:41
faenilI read about systemd --user having its own session, but that was not clear to me08:41
hedayatbecause the second login would not run a new 'systemd --user'08:41
faeniland that's where we're headed, a system where second login doesn't start up anything08:41
hedayatfaenil: yes08:42
faenilbut the fact that systemd --user is detached from the user session means you'll never be able to have multiple UXes, doesn't it08:43
faenilor, alternative question, why should it be on a separate session?08:43
hedayatfaenil: but if you want this, you should actually not permit the user to login twice in the first place, (or have a session manager who'll decide how it should start a new session), rather than using systemd --user as the session manager08:44
hedayatyes it does.08:44
faenilah, isn't that quite a limitation? :/08:44
hedayatas I said, because it is designed to manager user "services"08:44
hedayate.g. user's http server08:44
faenilyes, that's ok08:45
faeniland my current user session is not a user service?08:45
faenilit's quite strictly related to my user08:45
hedayatfaenil: limitation? compared to what?! AFAIK, 'systemd --user' is something others don't provide at all. And a generic 'session manager' is something that apparently, nobody provides. But I think systemd will provide it in future (or maybe in its recent versions).08:46
*** tigeli has quit IRC08:46
hedayatfaenil: nope, it's not a user service, it's a user session service!08:46
faenilI just miss background, sorry08:47
hedayatfaenil: If you have 3 sessions, you'd need 'session services' to be running 3 times. but 'user services' should be run just once.08:47
faenilyes,that's clear08:48
faenilah ok, I see what you mean now08:48
hedayatfaenil: e.g. If you have 3 ssh logins to a system, it'll run 3 instances of shell for you, but the 'user service manager' would just run a single http server instance when you create your first session, and will terminate it when you close your last session.08:49
hedayat'systemd --user' does the latter job, but not the former.08:49
faenilyes, yes, it's clear ;)08:49
hedayatsorry, I had already typed most of it, so I didn't like to throw it out :D :P08:49
faenilnp ;)08:50
faenilrepetita iuvant ;)08:50
hedayatUnless systemd provides a session manager in later versions, continuing with this setup is likely to be troublesome from time to time. This is an unsupported setup. So, maybe it's better to have your own session manager, as currently every Desktop DE does.08:52
hedayatWell, that's my opinion. :P08:53
faenilhedayat: ok, so, in the meanwhile, the hack is the best solution? XD08:54
*** Svetlana has joined #nemomobile08:54
hedayatfaenil: I think so.08:54
faenildamn Stskeeps :D08:54
hedayatbut I guess there will be some work to be done until all user services run successfully. there could be some permission issues.08:56
faenilI guess that's been taken care by jolla already, if we decide to hack the same way08:57
*** Svetlana has quit IRC08:58
hedayatfaenil: Yes, it should be. Since I get lots of errors in my setup :P09:01
faenilhedayat: so, you started the discussion saying you were stuck, but it seems your ideas are quite clear, so where's the problem :D09:01
faenilah :D09:01
faenilcan you write all the fixes you have in place? just for the record09:04
hedayatfaenil: I gave up in finding a 'proper solution' for the current setup. :)09:04
faenilhedayat: what do you mean by "current setup"? if you want to change setup, you can09:05
hedayat"current setup" means running user session using systemd service files! a simple 'proper setup' is what you'd do in non-systemd days09:05
hedayate.g. having a script, which is run after user logs in (e.g. a service file with PAMName=login which is wanted by graphical target), and starts user services in the desired order!09:06
faenilsystemd already starts services in the desired order, you just have to tell it which order you wish to have09:07
hedayatI've not checked how it is done in Harmattan.09:07
faenilso you haven't tried the jolla way yet (i.e. dummy sleeping binary + service file with PAMName and User)09:08
hedayatyes, but as I said, it is not designed to start 'session services', at least currently. and it might cause problems here and there, and needs hacks like what Jolla has done.09:08
locusfI saw yesterday on hadk p3110 implementation that the session didn't stop, unlike on n909:09
faenilhedayat: but if we agreed jolla's hack is the way to go, what are we waiting for .)09:09
faenil(or, at least, it seems we couldn't manage to find any less hacky solution)09:10
hedayatfaenil: currently, yes I'm using something like what Jolla has done. But I just get UI work. So, I would suggest Jolla hack. :) I can describe my changes, but it doesn't look less hacky.09:10
locusfit has to be something else that kills the session in which Jolla way might have no difference at all09:11
faenilhedayat: so you *are* using jolla's hack but still have permission issues?09:11
faenillocusf: mm09:11
hedayatlocusf: What do you see in N9's logs? Don't you see something like this: "user@100000.service timed out"?09:12
locusfhedayat: yes I saw it09:12
hedayatfaenil: I'm using my own hackes, but the concept is similar :P09:12
locusfhedayat: but my p3110 is identical with the difference of the hardware adaptation09:12
locusfall other packages exactly the same09:12
locusfeven with old dbus etc09:12
hedayatlocusf: hmm, this service times out when one of its dependencies (in /usr/lib/systemd/user/) doesn't 'start' (the way systemd expects).09:13
locusfalthough the init system for hadk devices is different, but still systemd is called09:13
locusfhedayat: indeed09:13
hedayatlocusf: I'm not sure, but probably even if one of the user services fail for a reason, user@ might time out09:13
hedayatlocusf: or one of the services doesn't 'start'09:14
locusfhedayat: yes09:14
*** vakkov has quit IRC09:14
faenilremember the ks doens't have mce, iirc09:14
locusfso should we instead try not to implement Jolla hacks but figure out which one it is that doesn't start09:14
locusfmy p3110 didn't have mce either09:14
locusfuntil I installed it yesterday to make the display to sleep on power button pressing09:15
hedayatlocusf: hmmm, I'm not sure if you can get rid of 'while' hack,09:15
locusfbut the session still survived for a long time09:15
faenillocusf: you have user-session@service?09:15
locusffaenil: on p3110, I can check?09:15
hedayatlocusf: as far as I've understood, without that while, user loging session will be closed, and user@.service will be stopped by logind09:15
faenillocusf: yes, if you're using a mix of old and new systemd user setups, then I don't know what to expect09:16
faenilhedayat: I agree09:16
locusffaenil: no mixing needed, its still the old way09:16
hedayatlocusf: I've the same "user@.service time out". There is a hack to prevent it from timing out: Add "TimeoutSec=0" in user@.service. :D :P09:17
faenilah, but then no's obsoleted, I think we shouldn't use it09:17
faenilhedayat: lol :D09:17
locusffaenil: I've got user-session@.service09:18
hedayatlocusf: I've failed to use 'the old method' successfully. Each time I tried to ran "systemd --user" without logging in (which will cause it to be started automatically), I failed with some permission errors due to how cgroups/etc are managed09:18
faenillocusf: which is the old way, said to be "obsolete"09:18
locusffaenil: okay09:18
locusfso its still a mystery why my session doesn't time out like it is on n909:19
faenillocusf: I think it's risky that systemd moves on, jolla moves on, and we stay in the past09:19
locusffaenil: yeah we should move on though, but lets do so in a way that doesn't break other stuff09:19
hedayatfaenil: locusf: if the old way actually works fine, it's great! But I've read that it won't. At least, it seems that people failed to use it on their desktop unless they downgrade (see the Arch forum I linked above).09:19
locusfif we just fix n9 then other stuff probably doesn't work09:20
faenilhedayat: it's not great, we go an untested path09:20
locusfI'll continue reading to my exam now, bbl09:20
locusfjust wanted to say this :)09:20
faenilit could seem to be working now, at least to boot ui, but what will happen in the future with the rest of the services ?09:21
faenillocusf: have fund :D09:21
locusffaenil: thanks09:21
hedayatlocusf: I've used the hack I said, and the user session doesn't time out. However, my "systemctl restart user@100000.service" never ends. I think I'm able to try to restart user services one by one to see which one doesn't start.09:21
faenilhedayat: or just look at the list of services and see which one is failing ? :D09:21
locusfhedayat: ok you do that, its still important to get this one fixed for good :)09:21
locusfn9 is still the prime piece of hardware which a real adaptation exists09:22
hedayatfaenil: I agree with your arguments about using 'untested' paths, and this is why I think we should not keep using 'systemd --user' for managing user session! :D :P09:22
faenilhedayat: do we have a choice?09:22
hedayatfaenil: Even if it works, it's a kind of 'abuse'.09:23
hedayatfaenil: well, not in short term.09:23
faenilhedayat: so, should we just wait for spring?09:24
*** sp3000 has joined #nemomobile09:25
hedayatfaenil: :) IMHO, we should check systemd roadmap to see if they are going to provide a proper session manager, and plan accordingly.09:26
faenilwhen I asked in #systemd in the last few days09:26
faenilwhat I got was "use gdm"09:26
faenil"I don't have gdm"..."wat??"09:26
SK_workfaenil: wait what ?09:28
faenilwhat? :D09:28
SK_workuse gdm ? :D09:28
hedayatIf that's an 'official answer', this means that we probably need a session manager sooner or later.09:28
*** msava has quit IRC09:28
faenilhedayat: I don't know how official it is...but it's the only decent answer I could get09:28
faeniltry asking again in #systemd maybe you have better luck ;)09:29
faenil(srsly, try ;) )09:29
hedayatI might try, but certainly not using IRC. :P09:30
faenilbtw, "Systemd: System and Session Manager" mmmmmmmmmm09:30
hedayatfaenil: I'd most certainly receive the same answer :)09:30
faenilhedayat: nah, different people, I don't know how official was the guy who replied09:31
hedayatfaenil: Yes, but see: It might become in future, but currently it really is "System and Service Manager" :)09:32
*** msava has joined #nemomobile09:32
faenilinteresting, different pages different descriptions09:32
faenilwith quite different semantics as well09:33
faenilmaybe in 216 it's a session manager already :)09:34
hedayatfaenil: I've also seen it described as a "session manager", and that lead me to spend time trying to get it work. But then, I found that it is not. And now, everywhere I look, I don't see it described as session manager! Not in its man pages, not in its rpm description. :P09:34
faenil[11:35] <faenil> good morning people [11:35] <faenil> is systemd a session manager? or is it in the plans to provide one?09:36
hedayat215 doesn't seem to have, so only 216 remains.09:38
faenil[11:39] <faenil> and if it has a session manager, when was it introduced? [11:39] <eto> faenil: systemctl --user and systemd --user09:39
hedayatIMHO, documents, and behavior, doesn't agree with eto.09:42
faeniljoin #systemd :P09:45
*** hurrian has joined #nemomobile09:45
*** hurrian has quit IRC09:45
*** hurrian has joined #nemomobile09:45
*** VDVsx_ is now known as VDVsx09:46
*** Sinbad has joined #nemomobile09:46
hedayatlocusf: faenil: What is OOMScoreAdjust=? Does it really work?!! It causes '/usr/lib/systemd/user/booster-qt5.service/ to fail. And since it is in autostart mode, it remain in 'activating' state, which cause the user@.service to remain in 'activating' state, and so times out. Commenting out the option fixed the problem!09:47
hedayatOr maybe it doesn't work as this service is not running as root. ?!09:47
locusfhedayat: its something which doesn't work on n9, I have removed it from eg. dbus service09:47
faenilhedayat: yes, it was commented out in many other files already09:47
faenilhedayat: and yes, it is supposed to work, but it doesn't on Nemo, we don't know why yet09:48
locusfit doesn't work on n909:48
hedayatlocusf: Is this hardware dependent?! I wouldn't think so. Or related to kernel?09:48
locusfon hadk devices it works09:48
locusfhedayat: I wouldn't know really :/09:48
faenillocusf: ah, it works on other devices with nemo?09:49
locusffaenil: yes09:49
locusfno OOM errors there09:49
locusfbut the platform variables etc still make the booster-session-qt5 to fail because it only launches in xcb mode09:50
locusfso booster-* services are still to be fixed09:50
hedayatIt seems that you've applied Jolla hacks to your own Nemo image. Do you have problems with seeing battery charging level in UI?09:50
locusfhedayat: well I don't see it at all since I don't have the appropriate statefs packages available09:52
locusfhedayat: I really haven't applied anything :/09:52
hedayathmm. It works for me "systemctl status statefs" show the level. But UI doesn't, which should be some permission issues...09:53
hedayatlocusf: :)09:53
hedayatwhat is "The Maliit server". Currently the only user service which fails to run.09:56
faenilhedayat: about systemd being a session manager09:57
faenil[11:55] <zdzichu> it is, but something has to start the session\09:57
hedayatfaenil: Yes, i'm there. :)09:57
*** arcean has joined #nemomobile09:57
*** faenil has quit IRC09:58
*** r0kk3rz has joined #nemomobile10:01
locusfhedayat: its the virtual keyboard manager10:01
hedayatlocusf: thanks, so it should run! :P10:03
locusfmaliit fails on the same reason as booster does, missing platform to run on10:07
locusfthis is probably due to the env vars failing10:07
hedayathmmm After remoing OOM settings, apparently all booster's run!10:07
*** metallisto has quit IRC10:08
locusfmaliit is a tricky one10:08
locusfsince its quite old to be honest10:11
locusfand the keyboard is only fitted for the n910:11
*** sp3000 has quit IRC10:21
*** sp3000 has joined #nemomobile10:34
*** Guhl has joined #nemomobile10:35
hedayatlocusf: sorry for interrupting, but: OOMScoreAdjust works for system services (e.g. system wide dbus.service), but fails for user services. Could be due to permission issues, but I wonder how it works on other hw.10:42
locusfhedayat: ok :)10:42
hedayathmmmm some of the userland apps are still not ported to wayland, and need xcb plugin.10:46
locusfalso maliit is complaining about xcb missing10:48
locusfbut maliit did work previously without xcb10:48
hedayataha, now I understand what you meant by "missing platform" :P10:48
locusfthe apps should be either started with -platform wayland or something is missing from the environment10:49
hedayatand booster*s, as you said.10:49
*** sp3000 has quit IRC10:49
*** olesalscheider has quit IRC10:51
hedayatlocusf: Isn't nemo user expected to be part of 'users' group?10:54
hedayat(apparently needed for accessing battery information)10:55
locusfhedayat: I don't know, but I'd assume that it should be10:56
hedayatrebooted my phone to see if it enables battery status in UI10:57
locusfI hate it that I've got some graphics glitches on my n9, I could help out if I could interact with my device :p10:58
*** olesalscheider has joined #nemomobile10:58
hedayat:) HW problem?10:58
locusfapparently so10:58
locusfif it renders well on your devices10:59
hedayatwell, now I've a working(=non-freezing) Nemo with battery status.10:59
hedayatlocusf: adding nemo to users group fixed the problem10:59
locusfgreat :)10:59
locusfnow just upstream the changes and see about building a new image from those11:00
hedayatlocusf: you mean, to mer?11:01
locusfoh yeah, upstreaming wouldn't make a difference since we don't have webhooks enabled to the new repositories11:01
locusfhedayat: yeah and nemo packages11:01
hedayatOK, but it seems to be too early. Settings doesn't run :P :D.11:02
locusfit runs just fine on my p311011:03
hedayatIt did run before my changes. I don't know what I have broken.11:03
jonwilDoes anyone know if any copies of the old meego-on-N900 repositories (the ones back when Meego was using binary blobs for pulseaudio on the N900) still exist anywhere? I need the pulseaudio blobs and bits for some reverse engineering work connected to the Neo900 :)11:05
locusfnow some lunch11:05
*** Sveta has joined #nemomobile11:07
*** hedayat has left #nemomobile11:16
*** Sveta is now known as Svetlana11:18
*** Svetlana has left #nemomobile11:24
*** r0kk3rz has quit IRC11:27
*** r0kk3rz has joined #nemomobile11:43
*** martyone__ has quit IRC11:50
*** vakkov has joined #nemomobile11:57
*** Guhl has quit IRC12:05
*** DrCode has quit IRC12:05
*** DrCode has joined #nemomobile12:08
*** faenil has joined #nemomobile12:20
*** ChanServ sets mode: +o faenil12:20
*** vakkov has quit IRC12:21
faeniljonwil: maybe this can be helpful?
jonwilnope, those dont help, I need meego not nemo12:26
jonwilmeego has the pulseaudio blobs, nemo does not12:26
jonwilat least as far as I know12:27
*** hurrian has quit IRC12:36
*** ryukafal1 is now known as ryukafalz12:38
*** hurrian has joined #nemomobile12:38
*** hurrian has quit IRC12:38
*** hurrian has joined #nemomobile12:38
*** sletta has quit IRC12:42
faeniljonwil: what about
Stskeepspa blobs were redistributable for nokia devices atm12:43
*** phaeron has quit IRC12:43
jonwilooh good, I think I should be able to mount that meego image file via my Linux Maemo dev vm12:44
Stskeepsthey may be unsuitable for new pa12:44
jonwilat least I hope I can12:44
jonwilwe dont plan on updating PA anyway12:45
Stskeepsremember: hardfp vs softfp12:45
faenilrozhkov_: ping12:57
jonwilnow if I can just figure out how to mount this disk image...13:02
jonwilwell its not a disk image, its an image file for a ROM chip :)13:02
jonwilbut I still dont know how to mount it or even what filesysem it is13:03
bencohare you talking about the mmcblk0p file ?13:03
bencohuse bunzip2 on it, then losetup to attach it to a loop device, then kpartx to create partition devnodes, then mount on the partition you're interested in13:04
*** sp3000 has joined #nemomobile13:05
jonwilwhat do I pass to losetup?13:05
rozhkov_faenil: pong13:07
faenilrozhkov_: can you try loading on your jolla?13:07
bencohjonwil: losetup /dev/loop0 /path/to/uncompressed/image13:07
rozhkov_faenil: yes, loaded13:08
faenilrozhkov_: now scroll13:08
faeniluntil you see all white13:08
faeniland see how long it takes before it shows something13:09
jonwilok, what do I pass to kpartx?13:09
jonwilkpartx -a /dev/looop0?13:09
jonwilok, got that13:10
jonwilso now I got the /dev/loop0p0/p1/p213:10
jonwilmake that p1/p2/p313:10
rozhkov_faenil: yes, it takes time. That's known problem13:11
faenilok, good :) do you already know the cause?13:11
jonwilhmmm, kpartx -a isn't doing anything13:11
jonwilkpartx -l says it would create loop0p0/p1/p2 but kpartx -a doesn't make them13:11
rozhkov_faenil: not really, simply didn't have time to look at it :(13:12
bencohjonwil: see /dev/mapper/13:12
*** niqt has joined #nemomobile13:12
bencohyeah, tricky :)13:12
jonwilnow which one of those 3 points is which13:12
jonwiland how do I mount em?13:12
jonwili.e. what fs type13:13
faenilrozhkov_: ok ;) what about the occasional 1sec delay in scrolling? is that known?13:13
faenili.e sometimes you scroll, and the page starts scrolling after half a sec, or a full sec13:13
faenilit doesn't always do it13:13
bencohjonwil: I dont remember, so either someone here does, or just try them all :)13:14
bencoh(I hope it's not jffs/ubifs)13:14
jonwilmount /dev/mapper/loop0p0 /blah mounted something13:14
jonwilso I think we are good13:14
rozhkov_faenil: yes, if IPC channels between threads are congested with messages then the delay becomes noticible. Usually happens on JS heavy sites13:15
faenilrozhkov_: ah I found a way to reproduce it...just start scrolling from an area which also has side scrolling (on homepage)13:15
jonwilok, so p1 is the root filesystem, p2 is swap and p3 is the kernel13:15
faenilrozhkov_: not sure it's that, it only does it when starting from a sidescrollable area13:15
faeniland it does it 100% of the times in that case13:16
jonwilnow to learn how an RPM based system stores the information about which packages are installed and what files are part of those packages13:18
faenil"rpm -qa" are those installed, "rpm -ql package" are the files of a package13:19
faenil"rpm -qf file" return the package which has that file13:19
jonwilgreat if you can run that command on the system but you cant do that on a mounted disk image13:19
faenilI think those are the most useful ones :)13:19
jonwilyou need a live system13:19
faenilah ok, sorry wasn't putting the info into context ;) there's rpm db, but not sure how you can read that13:20
*** flash1 has joined #nemomobile13:20
faeniland not sure it's already filled in when img is created (I think it is rebuilt during image building phase, so it should be there already, but not sure)13:21
rozhkov_faenil: handling scrollable areas could be better for sure. The support for it was added in gecko31 (b2g code path) and u9 is the first update featuring scrollable areas13:21
faenilrozhkov_: no need to justify ;) I just want to report a bug, and ideally have you file it :D13:21
faenil(if you can reproduce it)13:22
*** flash1 has quit IRC13:22
faenilthere are a few sidescrollable areas on homepage so it should be easy, just tap on one, and do *vertical* scrolling13:22
faenilthe system will wait like 1sec13:22
faeniland then start scrolling both the horizontal scrollable area, and the page (vertically) :)13:23
rozhkov_faenil: ic, it first scrolls horizontally (until scroll velocity == 0) and then starts to scroll vertically.13:25
faenilnot here13:25
faenilit does nothing for 1sec, and then scrolls both at the same time13:25
faenilrozhkov_: ah no actually you're right ;)13:27
faenilit is exactly that13:27
faenil(though I'm sure they were scrolling both at the same time before :/ )13:27
rozhkov_faenil: technically speaking there scrolling, but "panning" and "flicking"13:28
rozhkov_*there is no13:28
faenilok, flicking13:29
faenilbtw, images don't flick or pan, they scroll13:30
faenilit's the user who flicks or pans13:30
faenil(afaik :p )13:30
rozhkov_faenil: alright :) i got too used to the PanZoom controller's terminology13:31
faenil:D whatever, we know what we're talking about :D13:32
jonwilnote to self, asking an FTP client to make a copy of /dev is not a good idea, it will pull data until you run out of disk space13:32
faenilflicking and panning are gestures, so I don't think it can be said about moving objects but only the user who performs them, that's my reasoning :)13:32
faenil(I could be wrong, ofc ;)13:33
rozhkov_faenil: when I do pan there is no delay, but when the PanZoom controller goes to the "flicking" state after "panning" it does scroll in an interesting way indeed13:34
rozhkov_there is no bug for that I'm aware of13:34
rozhkov_but it might exist in Mozilla's bugzilla13:35
faenilrozhkov_: actually, panning here has a fixed large delay13:36
faenilwell, not that large but...easy to notice13:40
faenilrozhkov_: btw congrats, browser went from the laggiest to the smoothest sailfish app :)13:40
faenil(with update9)13:40
faenilit's much more usable now :)13:42
*** zalan has quit IRC13:43
*** zalan has joined #nemomobile13:43
rozhkov_faenil: thanks :) I've just check the browser with gecko35 - it doesn't do flicking for scrollareas at all and start flicking without delay13:45
faenilrozhkov_: oh that's cool :) but we're only moving to the next LTS right?13:46
*** jpetrell has quit IRC13:46
rozhkov_faenil: I don't really know. At the moment I'm fighting with a serious bug in 31 which seems to be fixed in the latest engine.13:47
faenilrozhkov_: what about the gcc compilation issues, did anyone take a decision?13:47
rozhkov_faenil: i'm not aware of any13:48
faenilmm :/13:48
*** zalan has quit IRC13:48
faenilStskeeps: any decision on your side?13:48
* Stskeeps mumbles about faenil and hilighting just as he's about to go out the door.. :P13:49
faenilhahah, 6th sense :P13:49
Stskeepsfaenil: it should be feasible to make this stuff build somehow under gcc 4.8, can you show me the errors again?13:49
faenilit's not feasability, it's that it's probably many many commits which have to be cherrypicked from Mozilla13:49
faenil+ those I committed13:49
faenilas Mozilla fixed most of them between gecko 31 and 3313:50
faenilso, either somebody wastes a week worth of compilation looking for all the fixes, or you move to the next engine, and it builds with my fixes13:51
faenilbut, as you forgot about the situation details, I think the answer to my question is "no" :D13:51
*** phaeron has joined #nemomobile13:52
faenilsome patches are here
MerbotMozilla bug 1027251 in General "Remove public destructors of NS_*_INLINE_* refcounted classes, outside of a finite explicit whitelist" [Normal,Resolved: fixed]13:53
Stskeepsit's not so much forgotten but tightly packed away at tapes in a warehouse :P13:54
*** VDVsx has quit IRC13:55
faenilok, understood, it will take months13:56
*** sletta has joined #nemomobile14:00
*** phaeron1 has joined #nemomobile14:07
*** phaeron has quit IRC14:07
*** zalan has joined #nemomobile14:08
*** sletta has quit IRC14:10
*** VDVsx has joined #nemomobile14:10
*** sletta has joined #nemomobile14:20
*** sletta has quit IRC14:20
*** vakkov has joined #nemomobile14:22
*** jonwil has quit IRC14:22
faenilfor the record, if I understood correctly and if all goes well14:30
faenilwe're moving to gecko 38 in MAY201514:30
*** SK_work has quit IRC14:35
*** krnlyng has quit IRC14:39
*** krnlyng has joined #nemomobile14:39
*** m4g0g has joined #nemomobile14:42
locusfso quite a while until we got a working browser then14:43
Stskeepsdid anybody package qtwebengine yet?14:43
faenilStskeeps: thought about that, but first qtwebengine release is with qt5.414:44
faeniland how long will it take for mer/jolla to adopt 5.4? :D14:44
Stskeepsubuntu had some binding too14:44
Stskeepsoxide or smt14:44
w00tyou might be able to get it running with older qt, with a little patching .. i don't think it uses tooooo many internals14:45
w00tas for 'when'.. I think 5.3 (and 5.4) will be much easier jumps to make14:45
Stskeepsfamous last words14:46
w00tqtdeclarative is already half at 5.314:46
faenilw00t: fingers crossed :)14:46
w00tStskeeps: i'm sure it won't be problem free14:46
w00tbut the churn is quite a bit smaller14:46
faenilw00t: I think I'll try in a couple of months or so ;) (to get qtwebengine working)14:50
faenillocusf: we can always fork gecko-dev from nemomobile and move it to 33 xD14:51
faenil(while we wait)14:51
locusffaenil: right14:51
*** msava has quit IRC14:56
faenilI wonder why BBM still doesn't work on used to work until update 6...14:59
faenilmy suspect is Jolla changed the android identification name in that update, and BB is filtering devices based on their name14:59
*** Termana has joined #nemomobile15:13
locusfdoes anyone remember which -platform did maliit take as input in order to work?15:21
*** SK_work has joined #nemomobile15:21
locusfotherwise we might not have a onscreen keyboard15:21
*** jjardon has quit IRC15:28
w00tlocusf: wayland QPA plugin. but you also need to make sure that inside maliit-framework, that it's using the lipstick codepath, not the wayland one (which is not really fully baked yet)15:29
locusfw00t: how do I check?15:30
w00tdon't remember :)15:31
locusfah :)15:31
*** Morpog_PC has joined #nemomobile15:31
locusfglacier keyboard does work but its scaled to n9 size :)15:32
locusfoh yeah it did have fixed widths and stuff15:35
locusfdon't really remember, its been quite a while since I hacked with it15:35
Morpog_PCnah, it had u :D15:35
locusflol u15:35
Morpog_PCremember? ;)15:35
locusfyeah unfortunately15:36
locusfcouldn't test on n7000 due to root password being locked15:38
Morpog_PCrootme? :D15:39
locusfnope :D15:39
locusfalso n7000 is still sadly eglfs implementation15:41
*** m4g0g has quit IRC15:43
locusfthis leads to glacier gallery not working15:45
*** SK_work has quit IRC15:48
*** suy has quit IRC15:50
*** m4g0g has joined #nemomobile15:50
*** suy has joined #nemomobile15:53
*** flash1 has joined #nemomobile15:56
faenilmmm no, BBM segfaults when launched :( awwww, damn aliendalvik15:57
*** filippz has joined #nemomobile15:59
*** veskuh has quit IRC15:59
*** flash1 has quit IRC16:02
*** niqt has quit IRC16:07
*** gabriel9 has joined #nemomobile16:13
*** sp3000 has quit IRC16:31
*** phaeron has joined #nemomobile16:38
*** vakkov has quit IRC16:39
*** phaeron1 has quit IRC16:40
*** Behold has joined #nemomobile16:49
*** phaeron has quit IRC16:53
*** piggz has joined #nemomobile16:53
*** piggz has quit IRC17:05
*** piggz has joined #nemomobile17:06
*** simbrown has quit IRC17:07
*** arturo182 has joined #nemomobile17:07
*** piggz has quit IRC17:10
*** piggz has joined #nemomobile17:12
*** filippz has quit IRC17:15
*** faenil has quit IRC17:16
*** vakkov has joined #nemomobile17:18
*** piggz has quit IRC17:20
*** piggz has joined #nemomobile17:24
*** flotron has joined #nemomobile17:25
*** piggz has quit IRC17:27
*** piggz has joined #nemomobile17:27
*** flash1 has joined #nemomobile17:27
*** piggz has quit IRC17:29
*** flash1 has quit IRC17:29
*** piggz has joined #nemomobile17:29
*** piggz has quit IRC17:36
*** piggz has joined #nemomobile17:37
*** piggz has quit IRC17:40
*** piggz has joined #nemomobile17:40
*** hedayat has joined #nemomobile17:44
*** plfiorini has joined #nemomobile17:46
ZogGsup dogs17:52
ZogGlocusf, \o17:52
*** notmart has quit IRC17:52
locusfZogG: hey :)17:53
*** m4g0g has quit IRC17:58
*** shentey has joined #nemomobile18:06
*** gogeta has joined #nemomobile18:18
*** metallisto has joined #nemomobile18:18
*** piggz has quit IRC18:19
*** gogeta has quit IRC18:26
hedayatlocusf: hi again. I take back some of my comments. Settings didn't start, because I have not set the QT related environment variables for user@.service. After setting the variables for 'systemd --user' service, things started to work:18:28
hedayatSettings run, all booster* services run successfully, and even maliit-server runs.18:29
hedayatFrom user services, only messageserver5.service doesn't run. What's this? :P18:29
hedayatAnd from applications, at least PIN Entry doesn't work now.18:30
hedayatbut I've not yet tried all apps!18:30
hedayatThere is no 'failed' system wide services. So, currently, I only have messageserver5.service which fails18:31
*** gogeta has joined #nemomobile18:33
*** jhonnaiker has joined #nemomobile18:35
*** jhonnaiker has quit IRC18:35
*** faenil has joined #nemomobile18:35
*** ChanServ sets mode: +o faenil18:35
faenilhedayat: it seems your memory is as bad as mine :D18:36
faenilbut glad to hear things started to line up!18:36
hedayatfaenil: :P oh.. why?18:36
faenilhedayat: because you forgot to modify user@ with env vars18:36
hedayatah... yes, actually I've added those vars to start-user-session@.service, because at that time I intended to start 'systemd --user' from there,18:37
faenilI see18:37
hedayatlater, that I realized that it can't be (at least, easily) started manually, I forgot to add them to user@. service.18:38
hedayatanyway, is messageserver5 expected to be running always?18:38
hedayatIs it for SMS?18:38
hedayatIt doesn't start, because qmf-accountscheck complains that there are no accounts, so it won't be started.18:39
hedayatShoud messageserver5 be started even if there are no accounts? (And what kind of accounts it refers to?)18:39
hedayathmmm looks like that it is not related to SMS but other types of account like EMail. So, maybe it's reasonable to not start if there are no accounts.18:41
hedayatOK. Is my SIMCard expected to be recognized and function in Nemo?18:41
faenildon't remember, it used to work pre-wayland18:42
faenilbut since we moved to wayland last year, I've never tried18:43
hedayatPrevious Nemo recognized my sim, and accepted its PIN and I was able to call with it. Sailfish recognizes my PIN, and even detects incorrect PIN, but after that, it always says that there are no sim cards available!!!18:43
faenilhedayat: but I remember you had to run the PIN app manually to get it to work18:43
faenilah :(18:43
faenilah but wait, maybe it's because of ofono18:44
faenildo you see "cannot connect to RILD" error messages in log?18:44
hedayatfaenil: Is kernel(3.5) support  complete?18:44
faenilfilippz is the guy to ask18:44
faenilbut one of the last things he worked on was to get HSI to work iirc18:45
hedayatYes, it repeats indefinitely! What is rild?18:45
faenilok, I think ofono is using android driver by default, which is rilmodel18:45
hedayatfaenil: In the "N9 kernel update" page, it seems that it should work. :P18:45
faenilI couldn't understand *what* changed the default modem, I tried asking in #ofono and nobody replied18:46
faenilbut Jolla is using a conf file18:46
faenilfor ofono.service18:46
faenilwhich just adds "--no-plugins=EVERYTHING_BUT_RILMODEM"18:46
faenil xD18:46
faenilit's like 4 lines of plugins18:46
hedayathmmm :)18:46
faenilso I think that's the first thing we have to try ;)18:47
faeniljust swap our modem with rilmodem in that line18:47
faeniland hope that ofono will cooperate18:47
hedayatfaenil: ok.18:47
*** Guhl has joined #nemomobile18:48
faeniland that could also explain why it doesn't work in sailfish either :)18:48
*** M4rtinK has joined #nemomobile18:49
hedayatfaenil: hey, I have a ... basic! question :P What are the icons in the bottom bar in glacier? I know the 'battery icon', clock, something which is probably GPS, bluetooth, and antenna status. There are 3 icons, for which I have no idea!18:49
faenilhedayat: I have never tried the new glacier homescreen yet! (shame on me) locusf is the developer ;)18:50
hedayatfaenil: :) OK!18:50
faenilhedayat: create something like this18:51
faenilI don't know which one is n9's modem (maybe n900? dunno)18:51
faenilbut just delete the one you think is n9 modem, and replace it with "rilmodem"18:52
faeniland make sure that ofono.service is actually loading files from that dir :)18:52
faenilenv files*18:52
faenilmaybe n9's driver is "phonesim"18:56
faenilor not18:56
hedayatfaenil: the --no-plugins option is a command line option, right?18:58
faenilyes, it's loaded by the .service files18:58
faenilit replaces "OFONO_ARGS"18:58
faenilcheck if you have it in your service file18:59
*** LjL has quit IRC18:59
*** LjL has joined #nemomobile19:00
hedayatyes, it's there. Only the env file is missing.19:02
flotronHello pe19:04
faenilflotron: o/19:04
faenilI'd try n900, then isimodem19:05
*** sp3000 has joined #nemomobile19:06
*** phaeron has joined #nemomobile19:09
Aardfaenil: isimodem19:10
faenilAard: thanks19:10
faenilthis is not encouraging
hedayatfaenil: seems to be n90019:15
faenilhedayat: Aard said it should be n919:16
faenilisimodem, sorry19:16
Aardboth n900 and n9 have isimodem19:16
hedayatfaenil: Aard: OK, apparently ofono calls it n900!19:17
Aardiirc long time ago it was done in two separate drivers, but eventually got merged into one. the n9 isimodem driver has some known bugs19:17
faenilah ok..but we have to choose between "isimodem" and "n900" ofono plugins19:17
faenilmaybe n900 then :p19:17
hedayatfaenil: Aard: ofonod[973]: plugins/udev.c:add_modem() /devices/platform/omap_ssi.0/ssi0/port0/ssi_protocol/net/phonet0 (n900)19:17
faenilyay :)19:17
faenilat least it did something19:17
hedayatfaenil: And the ofonod option is "--noplugin=" rather than "--no-plugins="19:18
hedayatYes it does recognize the device, and points to correct sysfs/devfs entries19:18
faenilhedayat: but...I gave you a trusted my irc msg? XD19:19
hedayatOK, apparently backend is working fine. But I'm unable to start neither PIN Entry, nor dialer apps. Dialer just opens a white window. PIN Entry doesn't open anything.19:20
Aardtry with a sim without pin, and then work your way up :)19:22
faenilhedayat: log says nothing about pin entry?19:23
hedayat:) OK.19:23
hedayatfaenil: let me see19:23
hedayatfaenil: localhost [740]: [D] OfonoSimIf::pinRequired:86 - "-->OfonoSimIf::pinRequired"19:29
faenilmore, more19:29
*** flotron has quit IRC19:29
hedayatfaenil: apparenlty, --noplugins option is just useful to reduce the clutter. oFono works fine without it, as it tries all plugins and one of them succeeds.19:29
hedayatnothing more, yet. :P let me run it in debug mode again.19:30
faenilhedayat: in the logs I saw it never stopped trying to connect to RILD19:30
faenilwhich means it had no other choice19:30
hedayatfaenil:  Aard: Isn't there an easy way to enter PIN? Like echoing it into some sysfs entry? :D :P19:31
faenilcheck pinentry source ? :P19:31
hedayatfaenil: It seems that it always try to connect to all modems! I can find the "pinRequired" log even before adding the option.19:32
hedayatfaenil: :D19:32
Aardhedayat: iirc there's a script in the ofono-test package19:32
faenilthat can be, but then why does it keep connecting to rild? there must be something we're missing19:32
hedayatAard: thanks. I'll check it19:32
hedayatfaenil: maybe. anyway, the option is useful anyway. I don't like seeing an error about RILD to appear again and again in the logs.19:33
faenilhehe :)19:33
faenilI'm still wondering why they removed n950 patches19:35
faenilI can't even find them because they don't have n950 in the commits!19:35
hedayatfaenil: does n950 have different modem?19:36
faenilprobably something different, don't know19:36
hedayatfaenil: do you know what are valid "pin_type"s?19:38
hedayatfaenil: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.UnknownMethod: Method "EnterPin" with signature "ss" on interface "org.ofono.SimManager" doesn't exist19:44
hedayatfaenil: I guess the interface is changed19:44
hedayator maybe not! The same error happens when trying /usr/lib/ofono/test/enter-pin19:46
hedayatfaenil: The error says that there is no "SimManager"!19:48
faenilwhich one? the one you pasted says unknown method19:49
hedayatSimManager can be found in ofonod binary. I wonder what the problem is19:49
*** zhxt_ has quit IRC19:49
hedayatfaenil: oops, you are correct. sorry19:50
*** gabriel9 has quit IRC19:54
*** ejjoman has joined #nemomobile20:07
*** zama has quit IRC20:07
*** zama has joined #nemomobile20:08
faenilhedayat: upstream changes please :D20:10
*** zhxt_ has joined #nemomobile20:11
*** ejjoman has quit IRC20:15
hedayatfaenil: I guess filippz has applied a number of them, right? applying Jolla method?20:16
faenilI think all his fixes are in the ks20:17
faenilso you can check from there20:17
*** phaeron has quit IRC20:19
faenilwhat matters is that someone sends the changes upstream so that we can focus on the next issues and more people can join :)20:24
*** shentey has quit IRC20:25
hedayathmmm, I don't see how autologin is handled. Well, ok. What is the upstream for Nemo packages? Mer?20:26
faenilif you want to talk to locusf and filippz before doing it, that's ok as well ;)20:28
hedayatfaenil: I should prepare my nemo-mobile-session changes. It needs some cleanups.20:29
hedayatfaenil: There was other changes like ofono one, and adding user nemo to users group20:29
hedayatfaenil: I'm not sure where should I propose them. specially the nemo group. Where such things are handled? :P20:31
faenilnot sure about the nemo group, but the rest are probably for the link I pasted20:32
faenilhedayat: oh look, user group changes
faenilsame package20:33
faenilalso have a look at
faenilplease pretty please also add a comment about why you had to add it to that group20:37
hedayatfaenil: OK, apparently everything I need is in this package.20:38
hedayatfaenil: OK!20:38
faenilyes, because that's the stuff jolla didn't update as they don't care about it :D20:38
*** plfiorini has quit IRC20:52
*** LjL has quit IRC20:54
*** LjL has joined #nemomobile20:59
*** gogeta has quit IRC21:12
*** gabriel9 has joined #nemomobile21:21
*** gogeta has joined #nemomobile21:23
hedayatfaenil: OK, I'm going for sleep. I hope to be able to create a pull request tomorrow or the day after it; as I'll disappear for about a week from Nemo world!21:29
hedayatgood night21:30
faenilhedayat: thanks! I think I'll disappear soon as well21:30
faenilgood night!21:30
*** hedayat has left #nemomobile21:31
*** vakkov has quit IRC21:31
*** faenil has quit IRC21:33
*** phaeron has joined #nemomobile21:37
*** gabriel9 has quit IRC21:41
*** gabriel9 has joined #nemomobile21:42
*** zalan has quit IRC22:00
*** Morpog_PC has quit IRC22:02
*** Pat_o has quit IRC22:03
*** zalan_ has joined #nemomobile22:04
*** zalan has joined #nemomobile22:05
*** zalan has quit IRC22:09
*** vakkov has joined #nemomobile22:13
*** gabriel9 has quit IRC22:15
*** gabriel9 has joined #nemomobile22:16
*** jonwil has joined #nemomobile22:33
*** zalan has joined #nemomobile22:34
*** phaeron has quit IRC22:39
*** r0kk3rz has quit IRC22:54
*** gabriel9 has quit IRC22:55
*** phaeron has joined #nemomobile22:55
*** gogeta has quit IRC23:08
*** Oksana has joined #nemomobile23:21
*** olesalscheider has quit IRC23:35
*** DrCode has quit IRC23:39
*** olesalscheider has joined #nemomobile23:39
*** DrCode has joined #nemomobile23:40
OksanaIn long-term, why is systemd used? Why not upstart (like in Maemo 5) or runit or something else, smaller and easier to understand?23:48
*** lemketron has joined #nemomobile23:53
lemketronHow does one get an account on ?23:53
OksanaYes, and have the same account23:56
lemketronThanks!  I searched all over but could find no mention of that, and saw no "Register" link on any of the bugzilla pages.  Sorry if I somehow overlooked that link.23:56
OksanaDuckDuckGoing :
lemketronAh, there is "New Account" on Merproject, but not on Nemomobile23:57

Generated by 2.11.0 by Marius Gedminas - find it at!