**** BEGIN LOGGING AT Tue Jan 26 02:59:58 2016 Jan 26 07:30:59 morning Jan 26 07:31:32 Seems KyleMaas_Laptop is been adding quite some things to issue tracker Jan 26 07:32:01 Tofe: Did you do a build overnight with the new components? Jan 26 08:10:52 Morning! Jan 26 08:12:10 Herrie|Veer: I have compile connman, ofono, but my voicecall rebase failed because of a not-found ngf-qt5 package, so I'm trying to see if it's needed or not Jan 26 08:16:37 Tofe: Ah ok. Jan 26 08:17:38 I don't think we need ngf Jan 26 08:17:47 Seems something specific for Mer? Jan 26 08:18:30 Cannot find too much about it though. Did run into some guys doing Open Source smartwatch OS with OE based on Mer/Sailfish called AsteroidOS :P Jan 26 08:26:15 Tofe: Legacy had a popup in Preware to ask if we want to make Preware default handler for IPK. However since we have Preware shipped I think we can make it the default handler anyway? Same for our PDF and other apps that we supply? No need to explicitly a Jan 26 08:26:15 sk users permission? Jan 26 08:30:08 Herrie|Veer: I disabled ngf, it's fine; now I need to disable commhistory as well, otherwise we have a dependency on qtcontacts, and I don't want that now Jan 26 08:30:33 Yeah Jan 26 08:30:41 commhistory we don't need either Jan 26 08:31:26 That's what I saw in VoiceCall: They had quite some commits relating to that. I guess we might need some patches in our repo to apply to disable the things we don't want from upstream Jan 26 08:46:22 Herrie|Veer: for the defaults, I think it's sensible to put ours as default as a starting point Jan 26 08:47:07 Tofe: OK. Let me PR that. I'm not sure what to do with the existing ones that we don't have yet (mainly video & audio ones). Remove them completely? Jan 26 08:47:34 See luna-init/files/conf/command-resource-handlers.json Jan 26 08:50:45 if we don't have a default, then we ask upon installation of an adapted app, if possible Jan 26 08:54:28 Tofe: PR is in but seems I broke GH with this small text files change :P "Sorry, this diff is temporary unavailable due to heavy server load" Jan 26 08:54:43 :D Jan 26 09:04:23 voicecall done Jan 26 09:04:45 which means, all done. Jan 26 09:09:16 Tofe also the libofonoqt and connmanqt? Jan 26 09:10:41 Yes Jan 26 09:10:49 :D Jan 26 09:10:51 No idea if it works, though Jan 26 09:10:58 Question is if it'll work of course :P Jan 26 09:11:07 I guess that's for tonight to try? Jan 26 09:13:06 Got my build VM ready too now :) Jan 26 09:13:23 So can test some stuff locally going forward :) Jan 26 09:18:17 Herrie|Veer: you'll need to use this: https://github.com/Tofee/voicecall/tree/tofe/rebase2 (latest commit) Jan 26 09:18:42 and this: https://github.com/Tofee/meta-webos-ports/tree/fido Jan 26 09:19:14 And it should be the same as I Jan 26 09:20:16 wait, I've made things easier Jan 26 09:20:26 use this: https://github.com/Tofee/meta-webos-ports/tree/fido-voicecall Jan 26 09:22:50 Tofe: I'll let you test first :P Jan 26 09:23:54 Well, maybe all wil be fine! Jan 26 09:24:37 For VoiceCall you think we should maintain a fork or just do patches on upstream? Jan 26 09:24:57 Changes seem quite a few, but nothing really that warrants an own fork imho? Jan 26 09:27:27 I'm not a big fan of the patches structure, but when we don't really have own code besides disabling upstream stuff I guess it makes sense to use it? Jan 26 09:33:20 We can switch to a patch, eventually Jan 26 09:33:32 Tofe: Yeah ;) Jan 26 09:36:58 You saw my PR for luna-init? Jan 26 09:41:44 yes; I merge it now? Jan 26 09:42:29 Yeah won't break anything I'm sure... It even passes jsonlint now :P Jan 26 09:42:41 Which it didn't before due to , after last time :P Jan 26 09:42:44 (oh nice, I have to rebuild webengine...) Jan 26 09:42:55 merged Jan 26 10:16:12 Tofe: LOL :P Jan 26 10:16:22 How come? Jan 26 10:27:23 Well, I guess it's QtMobility Jan 26 13:34:25 Tofe: ping Jan 26 16:05:58 Herrie: I finally got where I could download a new VM image yesterday, so I figured I'd do some testing. Jan 26 16:07:22 KyleMaas_Laptop: Yeah a fresh view is always welcome Jan 26 16:07:39 I already fixed some today :P Jan 26 16:08:26 Yeah, I saw that. Glad to see it. Sometimes the UI stuff doesn't seem too important, but it's still the main face that users see, so it makes a huge difference. Jan 26 16:08:34 Emulator is pretty nice in general. That you don't get packages in Preware might be because of x86 and targets are arm Jan 26 16:09:12 I need to test that myself. Preware app in general needs some TLC. It has some legacy things we don't need and does log-spam Jan 26 16:09:20 And not very happy with it layout wise Jan 26 16:10:12 I'd think that would be bypassed by the "Ignore device compat." setting. Jan 26 16:10:57 KyleMaas_Laptop: For example that dates back to when App Tuckerbox & Profile servers were still running. Jan 26 16:11:50 Where you could force specific device flags to be ignored Jan 26 16:12:03 architecture != device ;) Jan 26 16:12:15 Ah. I figured that disabled arch checking. Jan 26 16:12:41 Considering how many webOS apps were completely arch-independent, I figured that setting would be for both. Jan 26 16:12:46 KyleMaas_Laptop: No it would allow you to see Pre x apps on TP and vice versa Jan 26 16:12:53 Gotcha Jan 26 16:13:06 So, I have a couple new general questions: Jan 26 16:13:15 KylesMaas_Laptop: I think we should have x86 apps in Preware though :P Jan 26 16:13:30 Herrie|Veer: you probably need "else" here https://github.com/webOS-ports/org.webosports.app.filemanager/pull/9/files#diff-263bd4e90179a0121b87178bf6129983R333 , as "i" in next if block will be -infinity for empty file Jan 26 16:13:52 1. I seem to remember an effort a while back in the Open webOS project to port to an installable Android apk. Is there anywhere with information about that, and is that something LuneOS could work with? Jan 26 16:13:58 KyleMaas_Laptop: SCP and SSH are your friends with virtualbox ;) Jan 26 16:14:17 So you can access, copy files etc Jan 26 16:14:31 nizovn: Hmmz could be... I'll test a bit more. Jan 26 16:14:42 Herrie|Veer: and i think we should use name units as KiB, MiB, GiB Jan 26 16:14:53 2. I assume the prebuilt images for the existing device (non-qemu) platforms include proprietary blobs. How are we able to do that, and are there other platforms we could port to? Jan 26 16:15:22 nizovn: I disagree with that :P Jan 26 16:15:36 then you should use 1000, not 1024 Jan 26 16:15:37 Herrie|Veer: Is there no mechanism to mark an app as noarch in the Preware feeds? Jan 26 16:15:49 Because some people don't know what a proper definition of a KB is we need to change :P Jan 26 16:16:09 KyleMaas_Laptop: I think we simply didn't look into it yet :P Jan 26 16:17:47 Afaik 1000 is used for marketing (hd's to "boost" capacity), programs/OS-es always use 1024? Jan 26 16:19:19 KyleMaas_Laptop: Since we don't really have PDK apps all our apps should run on every arch currently. Jan 26 16:19:57 1000 is used everywhere, as least in si. ls seems follow right units Jan 26 16:21:08 nizovn: that it's used everywhere doesn't make it correct LOL :P Jan 26 16:21:34 I'm old school from cassette tapes on my C64 and 5,25" floppies := Jan 26 16:21:36 :P Jan 26 16:21:50 But we can make it an option somewhere LOL :P Jan 26 16:24:29 Herrie|Veer: IIRC Apple now uses -bibytes to measure file size and free space. Xubuntu uses 1000-byte kilobytes. Dunno about Windows. Jan 26 16:24:36 it's just a standard :P please explain why hdd capacity 1GB=1024^3 bytes, but cpu's frequency 1GHz=1000^3Hz :Ь Jan 26 16:26:14 nizovn: Hey I didn't invent the 0 and 1 principle so we end up with 2 ^ :P Jan 26 16:30:49 Herrie|Veer: If you're using 1024, I'd recommend using KiB, MiB, GiB, etc. too. From a user perspective, it takes the guesswork out of which unit you mean. Jan 26 16:32:10 Would be TiB then or?nn TeB? Jan 26 16:33:43 Herrie|Veer: TiB Jan 26 16:34:20 Herrie|Veer: https://en.wikipedia.org/wiki/Tebibyte Jan 26 16:38:21 Herrie|Veer: Also, I think it'd be good to add a bounds check to that. If it ever hits a PiB, you'd go out of bounds on the suffix array. Jan 26 16:38:43 Not likely for a while on a phone or a tablet, but just good as a general practice. Jan 26 16:38:52 KyleMaas_Laptop: It will be a decade before we get there :P Jan 26 16:39:12 But yeah, feel free to improve upon it. Jan 26 16:39:20 So you're saying you don't think LuneOS will be around in a decade? Come on, man! Have some confidence in your work! :) Jan 26 16:39:43 KyleMaas_Laptop: Quite sure it will be in some way or form :D Jan 26 16:50:25 Herrie|Veer: Untested, but: https://github.com/webOS-ports/org.webosports.app.filemanager/pull/10 Jan 26 16:57:41 Herrie: Any comment on this? http://issues.webos-ports.org/issues/1049 Jan 26 17:00:32 KyleMaas_Laptop: I can provide you legacy com.palm.app.musicplayer ipk from 3.0.5 Doctor if you don't have it. You can SCP this to qemux86 and opkg install it, then do a luna-send -n 1 luna://com.palm.applicationManager/rescan ' Jan 26 17:00:44 {}' in Putty Jan 26 17:01:36 You'll then have a working app. Throw some MP3s to /media/internal with SCP and you'll have mediaindexer doing it's work Jan 26 17:01:57 What triggers mediaindexer to reindex? Jan 26 17:02:15 It permanently scans media/internal for changes I think :) Jan 26 17:02:29 Or at a very regular interval :) Jan 26 17:03:20 How do I deploy a newly-compiled version of mediaindexer to the VM? Jan 26 17:05:23 For that matter, how do I compile it? I assume cmake, but while I'm okay at C/C++, I'm not great at build systems. Jan 26 17:05:44 cmake complains about missing webos_modules_init. Jan 26 17:08:17 Ah. https://github.com/openwebos/cmake-modules-webos Jan 26 17:16:18 KyleMaas_Laptop_: it's simpler to use our build environment, it will generate ipk for you Jan 26 17:16:21 http://webos-ports.org/wiki/Build_for_qemux86 Jan 26 17:18:17 nizovn: Yeah but some quick sanity check without a full build environment would help :P Jan 26 17:18:46 you don't need to compile whole luneos-emulator-appliance, only mediaindexer Jan 26 17:19:31 i doubt it's possible to satisfy all build dependencies Jan 26 17:23:04 nizovn: Ah ok. Just I'd like to have some environment where I can code properly besides vi and notepad++ :P Jan 26 17:23:27 So something like QtCreator or similar where I can quickly test things :P Jan 26 17:24:12 But I guess it goes with single recipe too. Need to remember how to do that again, I'm sure it's in the log. My Ubuntu VM is ready with a recent build :) Jan 26 17:25:12 Ouch. Well, I guess mediaindexer work is going to have to wait, then. Jan 26 17:25:45 KyleMaas_Laptop_: You need a decent machine to build Jan 26 17:26:11 4 cores, 80GB hd and 8GB of memory is quick enough :P Jan 26 17:26:49 KyleMaas_Laptop_: I got an environment up here, maybe I'll give it a go, just have close to 0 C++ experience :P Jan 26 17:26:59 It's not the build I'm worried about. This machine would be capable of it. But I usually only have it out for short periods of time, which precludes a long build process like WebKit (based on past experience with my desktop, which runs Gentoo) until I can get set up in one place for a long enough period of time. Jan 26 17:27:28 Herrie: pong Jan 26 17:27:47 KyleMaas_Laptop_: We don't do too much webengine changes so that doesn't get rebuild every time :P Jan 26 17:28:05 Sure. But it has to be built once, at least. :P Jan 26 17:28:32 Tofe: Was just wondering if you have an easy way for me to toy with MediaIndexer on Windows with QtCreator maybe :P ? But nizovn said it won't be easy probably Jan 26 17:28:37 KyleMaas_Laptop_: but what's 4 hours, in a whole life? Jan 26 17:28:51 KyleMaas_Laptop_: Useful for overnight :P Jan 26 17:28:51 I'm currently on battery power. My recollection is that on my desktop, it takes about 6 hours to build WebKit. Now, this machine has a better processor and more RAM than my desktop, but I still don't have that much time of high power consumption. Jan 26 17:29:48 KyleMaas_Laptop_: You're with ka6sox in middle of nowhere in Alaska up a mountain somewhere LOL :P ? Jan 26 17:30:20 Herrie|Veer: I thought mediaindexer was more a backend thing, not a frontend? Jan 26 17:30:23 Herrie|Veer: I am fairly mobile at the moment. :) Jan 26 17:31:45 Not quite middle-of-nowhere, but there's a reason it took me a while to get to a non-metered WiFi network for long enough to download the new VM image. Jan 26 17:31:59 Tofe: Yeah it's some c++ service that runs in background Jan 26 17:32:38 Herrie|Veer: If you get a build going of just that component, throw documentation of how to do it into the README file please. Jan 26 17:32:46 Herrie|Veer: not sure it would be easy to put on QtCreator... as soon as it gets to C++ dependencies, it's pretty complicated Jan 26 17:33:08 Herrie|Veer: So far, I know you would at least need this: https://github.com/openwebos/cmake-modules-webos Jan 26 17:33:30 I suspect you would also need db8 libs. Jan 26 17:34:25 Past that, I don't know. But I can say it would be really, really helpful for new developers like myself if as many LuneOS components as possible could be built separately. Jan 26 17:34:53 Actually, I should probably get going again. Jan 26 17:39:07 Tofe: I'll play a bit Jan 26 17:39:18 Your build with ofono etc ready to test? Jan 26 17:49:02 well, seems like my build wasn't all correct Jan 26 17:49:36 I mean it built, but I think there's a patch that was missed somehow Jan 26 17:54:57 Herrie: ok, there's a little glitch in the libconnman-qt patch for remove the MeeGo prefix Jan 26 17:58:56 Ok, all good now Jan 26 17:59:05 Let me PR that Jan 26 17:59:57 Unfortunately, it doesn't fix the issues I've remarked so far (empty timestamp, empty VoiceCallManager informations) Jan 26 18:02:06 ah, I forgot to put the voicecall adaptation as a path, one moment Jan 26 18:27:19 Herrie: https://github.com/webOS-ports/meta-webos-ports/pull/164 Jan 26 18:36:55 Tofe: Nice work! Jan 26 18:36:59 Ready to merge? Jan 26 18:37:07 I noticed the 0:00 1 jan 19:00 timestamps as well :P Jan 26 18:39:06 JaMa: ping Jan 26 18:39:35 Ready Jan 26 18:41:47 Would be good if JaMa or morphis could have a quick 2nd look but lgtm Jan 26 18:42:09 sure Jan 26 18:43:35 Could be that voicecallmgr doesn't do the dates properly. I think Mer might use the commhistory for this somehow? Jan 26 18:44:13 no, it listens to DBus, and with dbus-monitor I can see it doesn't get a correct date right from the start Jan 26 18:45:12 ah but maybe it's the voicecallhandler backend... Jan 26 18:49:07 From log I'm quite sure oFono gets it properly and from there it should be passed to VoiceCallManager I guess somehow? Jan 26 18:54:31 I'm learning the inside of voicecall :) Jan 26 18:55:29 ... which leads to qofono Jan 26 18:59:46 Which will probably lead to oFono :P Jan 26 19:00:21 Well some logging here and there should be able to help Jan 26 19:02:06 Looks like ofono's voice CallManager never sends the startedAt property Jan 26 19:13:20 That looks like something that might be upstreamable :P Jan 26 19:16:44 ok, actually ofono only sends startedAt when the call becomes active Jan 26 19:17:49 And even so, when looking at voicecall's dbus, it looks like ofono's startedAt message gets lost along the way Jan 26 19:18:33 Tofe: Hmmz we'd need some timestamp for when a call doesn't get answered (i.e. rejected or missed) too? Jan 26 19:18:57 ah no wait, when disconnected it also sends it Jan 26 19:22:06 ah no re-wait, I'm wrong; the start time isn't actually initialized until the voicecall has become active. Jan 26 19:22:26 Ok, all this looks really fishy Jan 26 19:23:01 I propose, when a call ends, to take "current_time - duration" as the time we log Jan 26 19:29:55 i'm sort of dropping in on hte middle of this, but i'd think a database would store the start time and end time, and you can easily derive from that Jan 26 19:30:57 We have start time and duration Jan 26 19:31:06 well, we store that Jan 26 19:54:13 Are we the only ones who use ofono ?... Jan 26 19:54:23 sorry, QOfono Jan 26 19:55:29 Tofe: Not sure... I think it came from upstream somewhere? Jan 26 19:55:31 I was looking for why the "incoming" bool property wasn't correctly set, then I spotted the bug, which is that it tries to read the property too soon Jan 26 19:55:56 So a thing so basic as "isIncoming" is wrong Jan 26 19:56:19 I'm a bit worried... I hope it's just bad luck Jan 26 19:56:44 Pretty easy to fix, hopefully, but well Jan 26 19:59:24 I'll ask the Mer guys Jan 26 20:04:45 Tofe: Yeah all links on GH point to them Jan 26 20:05:14 Plus the hosting Jan 26 20:05:54 Tofe: https://github.com/plasma-mobile/plasma-phone-components Jan 26 20:06:05 Has some interesting read on PhoneSIM and how to use it :) Jan 26 20:06:57 yes, well spotted Jan 26 20:07:28 Tofe: Seems libqfono is not very actively maintained ;) It might explain some of the odd things you have with your Jolla if there are other things like this Jan 26 20:08:41 It's actually in VoiceCall, but in the (Q)Ofono plugin Jan 26 20:09:48 Tofe: You expect a fix soon or I should kick off a nightly already :P ? Jan 26 20:10:36 Answer from #mer: most people use telepathy nowadays Jan 26 20:12:08 Problem for me is that I don't know why we chose ofono; was it the only mature stack at the time? something related to dependencies? hidden issues with telepathy? Jan 26 20:13:21 Ah, I wish morphis was there to ask him :) Jan 26 20:16:43 Tofe: I think oFono is the most mature Jan 26 20:17:25 I need to read a bit I guess; I just discovered Telepathy actually uses ofono Jan 26 20:17:34 but extends it Jan 26 20:17:40 Yeah seems so :S Jan 26 20:17:50 oFono is quite actively developed sitll it seems Jan 26 20:18:19 So maybe we could try switching to a VoiceCall/Telepathy/Ofono stack ? Jan 26 20:19:18 Stskeeps told me Telepathy-ring was the place to be, now Jan 26 20:21:02 Tofe: Lets ask morphis about that when he's around Jan 26 20:21:07 Leave things as they are for now? Jan 26 20:21:18 He should be around tomorrow or day after I guess Jan 26 20:22:28 yup Jan 26 20:23:03 You feel like a quick QML one? http://issues.webos-ports.org/issues/1085 Jan 26 20:26:44 could be it's related to the lock status Jan 26 20:27:54 Tofe: Yeah I don't have much chance to debug with a device now Jan 26 20:28:03 Since I'm on laptop Jan 26 20:28:20 lockMode: "none" Jan 26 20:30:37 I don't know yet; it needs more debug Jan 26 20:39:21 Here is the Telepathy connector that could link us with ofono: https://git.merproject.org/mer-core/telepathy-ring Jan 26 20:39:50 However, we'd have to handle the recipe ourselves, there's nothing in OE currently Jan 26 20:40:11 though there is for telepathy itself Jan 26 20:40:21 But let's wait for morphis Jan 26 20:41:48 Herrie|Laptop: hey! you already had that discussion with morphis! http://logs.nslu2-linux.org/livelogs/webos-ports/webos-ports.20151027.txt Jan 26 20:45:01 Reading that, I understand that the voicecall/ofono plugin choice was made for 2 reasons: 1. using telepathy would mean adding two more API layers, and we would use only the ofono part of telepathy, that's to say 10% of it; 2. the other telepathy plugins don't fit in webos' architecture, where all goes through node.js Jan 26 20:45:17 Tofe: Yeah Jan 26 20:45:20 So he chose the lightest and straightest solution Jan 26 20:45:24 So I guess current way makes sense ;) Jan 26 20:45:44 I would otherwise have been surprised of morphis :) Jan 26 20:46:32 So, let's just fix the little bugs we encounter. It's nothing really scary so far, just a code that could get more love. Jan 26 20:46:59 Tofe: Yeah, so I guess you can do some fixes and PR to upstream too (Even though they probably won't use it), but maybe others will like Plasma, Tizen or whomever ;) Jan 26 20:47:16 haha, tizen Jan 26 20:47:20 sorry. Jan 26 20:47:40 Yeah LOL Jan 26 21:11:28 ok, gonna leave it for today -- gn8! **** ENDING LOGGING AT Wed Jan 27 02:59:59 2016