| === Jucato_ is now known as Jucato | ||
| === Isla_de_Muerte is now known as NwS | ||
| === pleia2_ is now known as pleia2 | ||
| === Elleo_ is now known as Elleo | ||
| === hggdh is now known as vittnet | ||
| === vittnet is now known as ayd | ||
| === salem_ is now known as _salem | ||
| === enthusi_ is now known as enthusi | ||
| === chihchun_afk is now known as chihchun | ||
| === autra_ is now known as autra | ||
| === _salem is now known as salem_ | ||
| javier4 | abeato: hi. I'm porting UT to an mt6795 based device. I'm issuing some problems with ril. Can you help me? | 12:45 |
|---|---|---|
| abeato | javier4, sure, first of all make sure you are selecting the mtk ofono driver | 12:46 |
| abeato | javier4, for that, change /etc/init/ofono.override | 12:47 |
| javier4 | abeato: i tried both mtk both mtk2. The first one give a generic error. This is the log in debug mode of ofono with the second one. | 12:48 |
| javier4 | https://www.irccloud.com/pastebin/siM2juou/ | 12:49 |
| javier4 | Sorry, wrong link | 12:50 |
| javier4 | Give me a minute | 12:50 |
| abeato | :) | 12:50 |
| abeato | javier4, anyway, is that list-modems output for the device? | 12:51 |
| javier4 | Yes. :) | 12:51 |
| javier4 | abeato: and this is the debug | 12:52 |
| javier4 | https://www.irccloud.com/pastebin/VmO4u6KB | 12:52 |
| abeato | javier4, ok, I see that you are not able to register apparently, although things do not look too bad in fact | 12:54 |
| abeato | javier4, try to add these two variables to ofono.override: | 12:54 |
| abeato | OFONO_RIL_HEX_TRACE="" | 12:54 |
| abeato | OFONO_RIL_TRACE="" | 12:54 |
| abeato | then grab again the trace | 12:54 |
| javier4 | abeato: sorry for the delay. Syslog decided to be less cooperative. I hope I'll be able to post the log soon. | 13:26 |
| mterry | tedg: bug 1669546 -- are you familiar with those issues mentioned in last few comments? Do you know of an easy fix for disabling systemd units in certain xdg desktop sessions? | 13:30 |
| ubot5 | bug 1669546 in mediascanner2 (Ubuntu) "[MIR] mediascanner2" [Undecided,Fix committed] https://launchpad.net/bugs/1669546 | 13:30 |
| tedg | mterry: Well if it gets bus activated there's not much we can do, but otherwise it should only start if required by something else, like unity8. | 13:32 |
| tedg | mterry: It shouldn't be "disabling" as much as "not enabling" | 13:32 |
| mterry | tedg: sounds like it should always be dbus-activated instead of just started by upstart | 13:33 |
| tedg | I'm not familiar with the specific issues they're talking about though. | 13:33 |
| javier4 | abeato: here we are, finally | 13:33 |
| tedg | mterry: Then if someone starts it, it'll be there. I mean, that's the joy of dbus activation no control and random services. | 13:34 |
| javier4 | abeato: http://pastebin.com/SwhPJ5jv | 13:34 |
| mterry | jamesh_: ^ you around? Do you have some of the context for mediascanner2? | 13:35 |
| tedg | mterry: I mean you could always add a preexec to check for the environment, but it seems to me if someone asked for it they want it running. | 13:35 |
| javier4 | abeato: Probably you need to know that this tree doesn't use the standard rild daemon, but a mtk-customized one. | 13:35 |
| mterry | tedg: yeah I think most things in unity7 won't start it by default -- though if the user downloads a touch-written snap it may become started and they'd have the problems again | 13:36 |
| mterry | seb128: so mediascanner2 -- sorry for surprising you on that, I didn't know it had a bad history with unity7. Would a patch that cleanly exits if it detects non-unity8 environment work do you think? | 13:41 |
| seb128 | mterry, hey, no worry, yes it would be better to have it exit if it's not tested/useful on those sessions | 13:42 |
| abeato | javier4,hmm, this is suspicious: MTK2_RIL_UNSOL_RESPONSE_REGISTRATION_SUSPENDED | 13:43 |
| tedg | I'd say just make it not dbus activated... it doesn't make sense to not have it running on u8 (you want it to scan things in the background) and it doesn't make sense to have it running anywhere else. Doesn't seem like a candidate for dbus activation. | 13:43 |
| dobey | i think the main issue would be that it scans everything on start, and causes lots of disk i/o, which is not very nice to do | 13:43 |
| abeato | javier4, probably you need to send a request to get the registration to re-start | 13:43 |
| abeato | javier4, iirc this was done by mtk2 driver at some point, maybe there is some change in this device. It looks like there is no hook for that event | 13:44 |
| mterry | tedg: most of it isn't dbus-activated. There's a main daemon that is upstart/systemd run. Then a couple things that are dbus activated (extractor or whatever) | 13:44 |
| javier4 | abeato: user NeKit point me to this patch to mtk.c driver (not mtk2) | 13:45 |
| javier4 | https://github.com/rilmodem/ofono/commit/2c83eb87980b08435b072ab4209030c192ad354d | 13:45 |
| javier4 | but it's not easy to replicate it for mtk2 | 13:45 |
| abeato | javier4, how old is the device SoC? any idea? depending on that you should use mtk or mtk2 | 13:46 |
| abeato | javier4, or, if none fits completely, try to port that patch to mtk2 | 13:46 |
| javier4 | ~end 2015. It's Helio X10 | 13:46 |
| abeato | javier4, hm, I'd say mtk2 would be better, but not 100% sure | 13:47 |
| abeato | javier4, try to do something similar in mtk2 to what was done in mtk | 13:47 |
| javier4 | porting that patch is really colpmex. mtk2.c is just a copule of functions, while mtk.c redefine a lot of ril.c ones. | 13:47 |
| abeato | javier4, try register to the event with g_ril_register() in the networking atom, then send RESUME_REGISTRATION in the hook | 13:48 |
| javier4 | i.e. to port that patch I had to add mtk_data struct to mtk2.c, but that struct uses a lot of mtk-customized function to initialize its fileds. | 13:49 |
| abeato | javier4, no need to patch mtk2.c :) | 13:49 |
| === salem_ is now known as _salem | ||
| === _salem is now known as salem_ | ||
| javier4 | abeato: drivers/rilmodem/network-registration.c? | 14:04 |
| abeato | javier4, yes, would be a good place: you could override delayed_register() in new file drivers/mtk2modem/network-registration.c | 14:08 |
| javier4 | abeato: that's what I was thinking about. Could you explain the overriding chain of ofono? I.E. when a function is not present in one of the customized driver, it fallbacks to ril one? Same for plugins? | 14:10 |
| abeato | javier4, drivers/{mtk2modem,rilmodem}/voicecall.c would be a good example | 14:53 |
| abeato | javier4, note that how it works is that you create the mtk2modem driver in mtk2.c, and that some of the methods of that driver are taken from rilmodem/voicecall.c | 14:54 |
| abeato | javier4, ups, actually not true, you do not need to modify mtk2.c :) | 14:55 |
| javier4 | ok. Then it includes rilmodem/voicecall.h, and probably links the object file from rilmodem/voicecall.c, i think. | 15:05 |
| javier4 | But now, in my g_ril_register, what should I pass as function? Should I write something new like "ril_reg_suspend_notify()"? | 15:07 |
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| sayeed99 | is ubuport 5.1 really buildable atm?? | 15:41 |
| sayeed99 | I am facing a lot issues. And I am not sure if it because I am trying to force it to build with jre 1.8 | 15:42 |
| sayeed99 | anyone?? | 15:45 |
| === chihchun is now known as chihchun_afk | ||
| === genii_ is now known as genii | ||
| === JanC_ is now known as JanC | ||
| mterry | jamesh_: when you get a chance, I'd appreciate a review of https://code.launchpad.net/~mterry/mediascanner2/only-in-u8/+merge/320505 -- it's blocking mediascanner2 being in main / on image | 17:57 |
| dobey | mterry: eh? i thought there was a way to do that with the systemd service file now? | 18:02 |
| mterry | dobey: is there? I looked at systemd unit documentation, didn't see an easy way to filter by env variable or desktop session type | 18:03 |
| dobey | mterry: oh, that's what i thought from the discussion on this earlier | 18:04 |
| mterry | Not to my knowledge, but would be happy to be proven wrong. Though the env variable parsing is a little tricky, I'm not sure built-in systemd support would handle regexes | 18:05 |
| mterry | Though if it natively understood desktop sessions, enough to offer something like OnlyShowIn=, that would be swell | 18:06 |
| dobey | 09:35 < tedg> mterry: I mean you could always add a preexec to check for the environment, but it seems to me if someone asked for it they want it running. | 18:06 |
| mterry | dobey: yeah a preexec key doesn't exist in systemd, only upstart | 18:06 |
| dobey | oh | 18:06 |
| tedg | mterry: ExecStartPre | 18:07 |
| joe_w | hi, had a problem with dekko about a year ago, loading then not doing much else (not connwcting to server) having the problem again....was told to delete some files it worked last time but cant find the files (cant remember) can any one remind me...bq aquirous thanks | 18:07 |
| tedg | mterry: man systemd.service | 18:08 |
| mterry | tedg: !? I will search for that, couldn't find it in docs or google, cool. Though may not be any easier than what I ended up doing | 18:08 |
| dobey | mterry: ExecStartPre= | 18:09 |
| mterry | dobey, tedg: very cool, thanks. But looking at the docs for that key, I think the in-code method is easier (for this particularly complicated env check) and consolidates the logic between upstart/systemd (which mediascanner2 still installs file for both) | 18:12 |
| dobey | mterry: i guess ExecStartPre=echo "${XDG_CURRENT_DESKTOP} ${XDG_SESSION_DESKTOP}"|grep -i unity8 | 18:13 |
| mterry | That fails on Unity82 :P | 18:13 |
| dobey | mterry: that should do what you need i guess, or make a make a shell script | 18:13 |
| dobey | we'll have mediascanner 22 by then :) | 18:14 |
| mterry | Is there a problem with the in-code method that you so favor a systemd-line? | 18:14 |
| dobey | special env var, and extra code that requires running mediascanner itself to decide if it should be run. which seems weird | 18:16 |
| dobey | i think using the systemd method might still allow it to be started by dbus activation too, when an app does need it | 18:17 |
| mterry | Scanning daemon doesn't dbus-activate itself I don't believe, there is a separate dbus-activated daemon that spins up if needed | 18:17 |
| mterry | And a dbus-activated extractor. Both of which should be fine to spin up if something happens to request them in u7 | 18:18 |
| mterry | Added comment to MP explaining options; I still prefer in-code though | 18:23 |
| TartanSpartan | Hey, does anyone have a Meizu Pro 5? | 18:57 |
| TartanSpartan | Anyone? | 19:20 |
| === ayd is now known as hggdh | ||
| === salem_ is now known as _salem | ||
| carryononon | hi i'm trying to install ubuntu touch on my kyocera. Are there any premade ubuntu-touch images for that phone? | 23:32 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!