**** BEGIN LOGGING AT Tue May 31 02:59:58 2016 May 31 03:59:31 * DougReeder checks in May 31 04:03:14 Herrie, I’ll install http://build.webos-ports.org/luneos-testing/images/mako/luneos-dev-package-mako-20160529160310-testing-0-410.zip unless you tell me there’s something newer to try. May 31 04:09:19 DougReeder: No that one should be fine! May 31 04:09:34 * DougReeder nods May 31 04:10:38 Andolamin: ping May 31 04:10:48 DougReeder: It was strange it ever worked actually May 31 04:11:55 Herrie: pong May 31 04:11:57 Because according to API it shouldn't. Could be that glib just got a bit more picky in a recent release May 31 04:12:23 Andolamin: Thnx for PR. Not sure what you mean that contacts app doesn't work for you? We didn't merge DougReeders PR yet, so not sure what you tested? May 31 04:14:47 I checked his PR and pointed the recipe to his commit. May 31 04:15:30 Ah ok May 31 04:15:35 Any specific errors? May 31 04:16:37 A number of dependency not found errors. May 31 04:18:03 Where can I find the list of errors? May 31 04:22:09 https://gist.github.com/Andolamin/3708b23b677cdc6f8b4d8d9f8a4c1663 May 31 04:23:00 Looks like they're because the source references enyo-webos, but the directory structure has enyo-luneos May 31 04:23:16 It should have both. May 31 04:24:27 enyo-luneos is more stuff that LG didn’t care about. May 31 04:25:00 Yeah, in that case it looks like it's just missing from https://github.com/DougReeder/org.webosports.app.contacts/blob/master/.enyoconfig May 31 04:27:14 Yeah, it seems a clean build doesn’t catch that. May 31 04:27:25 So, I’ll update my PR. May 31 04:28:56 Andolamin: But you tested it with another app then that worked fine? May 31 04:28:57 Hrm, need to stash some feature changes to Contacts May 31 04:29:19 DougReeder: I can merge PR and just add them? May 31 04:29:47 Let me update my PR first. May 31 04:29:50 We just need https://github.com/DougReeder/org.webosports.app.contacts/blob/master/.enyoconfig#l16 and https://github.com/DougReeder/org.webosports.app.contacts/blob/master/.enyoconfig#l22 to be added no? May 31 04:30:38 It should be a quick fix. May 31 04:32:02 There may be a few other lacunae that slipped through. May 31 04:52:22 DougReeder: OK May 31 04:52:32 I merged Andolamin's stuff. Let me try locally to build first May 31 04:53:08 After my new image is done May 31 05:00:26 DougReeder: How does this look? https://github.com/Herrie82/org.webosports.app.contacts/commit/8e158b870b8632f5732367258b8a519d80432181 May 31 05:01:30 That’s correct, but I have a few more corrections as well May 31 05:07:10 Remind me never to update XCode on a day I’m trying to actually program May 31 05:09:30 DougReeder: :S Updated shouldn't be painful :S May 31 05:09:54 s/updated/updates May 31 05:10:14 It wouldn’t let me run any git commands until I accepted the terms for all of XCode! May 31 05:11:01 :S May 31 05:11:23 Well with my change it at least seems to build OK :P May 31 05:11:47 Pushed fixes. May 31 05:13:24 Pull the latest, and when you test it, be sure to edit a contact. May 31 05:14:05 Oh, run `enyo init` after pulling the latest, of course. May 31 05:15:33 DougReeder: Build system should do that for me :P May 31 05:15:53 Good. May 31 05:17:10 DougReeder: How do I see it behaves? May 31 05:19:00 Oh, just that you can successfully edit a contact. May 31 05:19:46 * DougReeder You need to run luna-send -n 1 palm://com.palm.service.contacts.linker/forceAutolink ‘{}’ May 31 05:19:54 before the changes will show up May 31 05:21:07 DOugReeder: I'm just testing on browser now May 31 05:21:12 Seems it loads OK May 31 05:21:21 * DougReeder nods May 31 05:21:22 So seems the recipes Andolamin created are OK May 31 05:22:17 Let me merge your PR and we can take it from there May 31 05:22:44 I think the issue we have with the auto linker not working MIGHT be due to LunaAppManager segfaulting May 31 05:22:48 Looking into the cause for that May 31 05:23:04 * DougReeder nods May 31 05:37:30 That build ( http://build.webos-ports.org/luneos-testing/images/mako/luneos-dev-package-mako-20160529160310-testing-0-410.zip) indeed fixes the albums not beign created bug, but now it seems that tab lebles in the launcher are missing May 31 05:38:30 DougReeder: Well we didn't change anything for that May 31 05:38:36 I noticed that as well on my build May 31 05:38:47 I suspect that's LunaAppManager crashing May 31 05:38:53 * DougReeder makes a moue May 31 05:41:09 DougReeder: It might start behaving on a few reboots May 31 05:41:29 * DougReeder grimaces May 31 05:43:15 FWIW, the contacts autolinker doesn’t appear to be run automatically. May 31 05:43:41 I’m off to bed. May 31 05:44:10 DougReeder: We'll sort the segfaults first and see if things get auto fixed that way :P If not we'll dig deeper :) May 31 05:44:16 DougReeder: Good night! May 31 07:21:21 morning May 31 09:33:24 Tofe: ping May 31 09:48:06 JaMa: morning May 31 09:48:11 Thnx for clarification. May 31 09:48:35 That's exactly what I did for debugfs but it doesn't seem to boot for me? May 31 09:49:04 Or this boot it supposed to take extremely long? May 31 09:50:55 This was on qemux86 May 31 09:51:41 I'm not too happy with my 128GB SSD since it doesn't allow me to build multiple targets, so a 480 GB one is on it's way so this will be history shortly :D :P May 31 10:07:16 Herrie|Pre3: maybe you misunderstood what this debugfs is :) May 31 10:07:37 it's just debug symbols for remote debugging with gdb May 31 10:08:08 JaMa: Yeah I know May 31 10:08:16 so you don't boot it anywhere, you just configure gdb to use it when searching for debug symbols when attached to already booted gdb session running on target May 31 10:08:24 But my VirtualBox should still boot? May 31 10:08:35 boot what? May 31 10:08:43 To LuneOS and then I attach gdb to a process May 31 10:08:59 what image are you booting? May 31 10:09:21 JaMa: I created a vbox image like normal, just with the debugfs now May 31 10:09:31 That shouldn't boot to gui? May 31 10:09:41 Just first time I'm toying with this. May 31 10:09:48 sorry maybe I didn't have enough coffee today May 31 10:10:17 but normal vbox image shouldn't be influenced at all by IMAGE_GEN_DEBUGFS = "1" May 31 10:10:20 IMAGE_FSTYPES_DEBUGFS = "tar.gz" May 31 10:10:31 It could be I misunderstand, my build environment is causing the issue or the image itself somehow. May 31 10:10:41 JaMa: Yeah I added those to local.conf May 31 10:10:50 it just creates "companion" tar.gz but shouldn't change anything in vmdk image May 31 10:11:06 so if it doesn't boot, then it's probably not caused by this addition to local.con May 31 10:11:09 f May 31 10:11:20 I see that in tmp-glibc/deploy/images I have an additional .tar.gz May 31 10:11:43 I just take the .zip as always, unpack it and import the appliance as usual. May 31 10:12:06 And it wouldn't boot for me. Just seems to be busy forever on booting kernel May 31 10:12:25 this should be unrelated to IMAGE_GEN_DEBUGFS May 31 10:12:39 JaMa: Hmmmz May 31 10:12:46 you can check buildhistory reports before and after local.conf modification files-in-image.txt should be the same May 31 10:13:20 Well when I only remove this change in local.conf it boots again :s May 31 10:13:34 So it's either my build environment or something else... May 31 10:13:42 Not sure you could test locally? May 31 10:14:19 running the build now lets see May 31 10:14:24 OK :) May 31 10:14:52 but it was booting for me week or 2 ago when I was fixing that llvm issue May 31 10:15:17 My SSD is getting full rapidly, so ordered a 480 GB one y'day with 10 yr warranty y'day so should be arriving later this week. May 31 10:15:33 So I can at least exclude that as a cause May 31 10:15:52 Just building 2/3 different targets on 120GB isn't feasible :P May 31 10:16:41 are you using tmpfs for tmp-glibc/work? May 31 10:17:07 JaMa: Just have a simple Ubuntu 16.04 with nothing special setup May 31 10:17:17 Just followed the setup guide from wiki May 31 10:17:25 I have 3TB for OE builds, but none of that on SSDs, just work dir in tmpfs for speed May 31 10:17:59 I would be concerned about ssd lifetime when doing a lot of OE builds on it (especially the work directory) May 31 10:18:26 Well price of SSD is OK nowadays and with 10 warranty I'm willing to take the chance := May 31 10:18:28 :P May 31 10:29:22 ok :) May 31 10:30:04 And I don't build that much :P May 31 12:36:17 JaMa: my work machine is a laptop with ssd, and has been fine for the last couple years. overall, that might shorten it's lifetime... but.. oh well. ssds are getting cheap now i guess? May 31 12:36:40 i'm guessing that having 16gb ram is probably not enough to reasonably build a system and put work in tmpfs May 31 12:37:45 EricBlade: My build system is limited to 8GB of RAM physically ;) May 31 12:38:16 Got some Sandisk Extreme Pro SSD. They come with 10yrs ;) And I prefer Sandisk over Samsung :P May 31 12:38:25 Just have better experience with them :P May 31 12:38:28 16GB is enough to do the build, but not very convenient (you need to be careful about disk space) May 31 12:38:30 i've been real happy with my samsungs and intels May 31 12:38:35 with 32GB it's much easier to manage May 31 12:39:01 depends on how many builds you do May 31 12:39:16 I think I write over 100GB to work directories daily May 31 12:39:29 i've got 2 intels that were from the "when 180gb drives were about $200 each" era, and 2 samsungs from much more recent times, and both have been really good for me May 31 12:39:36 8GB works, just should keep the parallel compiles down ;) May 31 12:40:19 Got internal compiler errors over the weekend that seemed to be due to too many parallel compiles :P May 31 12:41:24 And didn't have a swap setup. I guess I need some of that :P May 31 12:44:03 EricBlade: My 480GB is approx $200 now too :P May 31 12:45:16 JaMa: Any suggestions for SWAP size with my 8GB memory? May 31 12:45:48 32GB ram is around the same price and even a bit faster for work directory and will keep your ssds healthy for longer May 31 12:46:35 JaMa: Motherboard has 8GB limitation :( May 31 12:49:10 that's pity, something to check during next mb/cpu upgrade :) May 31 12:49:15 It's MiniITX with 2* SO-Dimm so not that much choice :P May 31 12:49:20 JaMa: Yeah :P May 31 12:49:36 Well it's a small powerful build PC :P May 31 12:49:45 I would use 8 or 16GB of swap depending on how much disk space you have May 31 12:50:11 JaMa: I can use my 128GB SSD for SWAP once I have my 480GB SSD :P May 31 12:50:16 It should fly then I guess :P May 31 12:50:25 my big powerful build PC still feels slow to me, so I won't comment on that :) May 31 12:50:52 and you build in virtual box, right? May 31 12:51:10 If QtWebEngine & QtBase would build more efficiently that would help :P May 31 12:51:27 JaMa: No switched to native 16.04 now on my dedicated build PC :P May 31 12:51:39 On my main PC was doing in VirtualBox May 31 12:52:47 ah ok May 31 12:52:49 good May 31 12:53:44 The more swap the better right? I mean I could assign 32GB probably on this 128GB SSD. Or even more. It's otherwise anyway idling :P May 31 12:55:54 Would there be an advantage of 32/64 over 16GB speed wise? May 31 13:03:49 EricBlade: You know a thing or 2 about Messaging in webOS don't you? May 31 13:05:49 Herrie|Pre3: probably no speed difference just more wearing of the ssd May 31 13:06:02 it will use more ram for disk cache and then write more in swap May 31 13:06:51 and when something goes crazy, it will take a lot longer for OOMK to kill it May 31 13:07:13 Herrie|Pre3: well.. i did. what's up? May 31 13:09:25 JaMa: OK I guess I'll just test a bit with 16 and 32 to see how the speed is. May 31 13:10:37 EricBlade: Just missing some stuff of how things are supposed to work together... We rewrote messaging app from scratch in Enyo 2 since it wasn't open sourced. We have pidgin in the images, however I'm not sure how to hook all this up. Filmor would look a May 31 13:10:37 t it, but seems he went AWOL May 31 13:28:02 hook up what to where May 31 13:33:35 EricBlade: So we can use pidgin backend in Enyo 2 Messaging app. I guess we'd need some kind of "glue" for that? May 31 13:33:40 I'm just clueless :P May 31 13:33:48 So I thought to ask you :P May 31 13:48:21 OK seems I start to understand it a bit more.... We have imaccountvalidator and imlibpurpleservice in our build it seems. May 31 13:48:47 So it might just be a matter of setting up the appropriate templates? May 31 13:49:12 For each of the account types (i.e. borrow some from filmors legacy work)? May 31 14:19:28 Morning May 31 14:19:49 Tofe: Late morning for you today :P May 31 14:19:53 Herrie|Pre3: I might be a bit long to answer these two weeks, as I'm on vacation May 31 14:19:57 Moved to the US overnight :P May 31 14:20:00 ? May 31 14:20:06 Tofe: Nice :D May 31 14:20:07 nope, Finland :) May 31 14:20:14 Tofe: Hehe :) May 31 14:20:53 I brought a TP and a N4 with me with latest testing, but well, I don't have a build machine with me, so it's just light testing or hacking May 31 14:21:16 Tofe: That's dedication :P May 31 14:21:30 :) May 31 14:21:34 I hope my N4 will resurface soon... They said 4-6 weeks and it's 6 weeks today.... May 31 14:21:43 So already sent them a chaser :P May 31 14:21:49 Maguro seems messed up :s May 31 14:22:00 Oh, you can't bring it back to life? May 31 14:22:03 So I'm bound to qemu and TP for now :P May 31 14:22:14 Tofe: Well it randomly refuses to flash... May 31 14:22:32 Could be lxc 2 but I think that should affect flashing? May 31 14:22:33 ah... May 31 14:23:02 no, I don't see how that could have messed it up that much May 31 14:23:22 must be a coincidence, and some hardware failure, maybe the usb port May 31 14:25:05 Tofe this morning it was complaining about the fact it couldn't copy the update script while it's clearly present in the .zip and .zip is not corrupt :s May 31 14:25:12 Something fishy for sure :P May 31 14:25:40 Decided to ditch my 120GB SSD and ordered a 480 GB one for build system :P May 31 14:25:53 Running builds for 2 targets on 120 is no fun :P May 31 14:26:08 Sandisk Extreme Pro has 10yr warranty :P May 31 14:26:12 So well :D May 31 14:27:46 Herrie|Pre3: I have also a 480GB SSD, but from Crucial; let's see which one falls first ! :P May 31 14:28:53 ok, going out for a while, see you later :) May 31 14:29:37 Tofe: OK enjoy! May 31 17:06:17 JaMa: Any luck on the debug symbols qemu image? May 31 17:06:55 Herrie|Pre3: I've just booted the image built for qemux86 on krogoth (+ only one change in jansa/krogoth for that pulseaudio config you reverted) May 31 17:07:19 and 1st boot shown the ui OK, but without any icons in launcher May 31 17:07:34 2nd boot shown 1st time wizard, which got stuck when finished all steps May 31 17:07:44 3rd boot is the charm and everything seems OK May 31 17:09:52 JaMa: Ok, hmmz strange mine doesn't boot May 31 17:10:04 The icon launcher is probably LunaAppManager segfaulting May 31 17:10:16 That's why I want the debug symbols to debug with gdb May 31 17:12:20 Herrie: i would be willing to bet that there was something in legacy that connected libpurple to db8. i don't know what that might've been May 31 17:18:45 EricBlade: Ah ok May 31 17:20:24 If I understand libpurple correctly, I don't know that it would need to talk directly to db8 May 31 17:21:30 The messaging app would store persistent data in db8 May 31 17:22:20 I think https://github.com/webOS-ports/imlibpurpleservice should probably take care of that? May 31 17:22:32 morphis: ^ Do you remember how all these messaging bits fit together? May 31 17:23:40 well, if you're following the old architecture, something took messages coming in from libpurple, and put them into db8, and messaging had a watch on the db8 kinds, and would then display things. Presumably that same thing would also watch for { status: "pending", folder: "outbox" } as part of the message kinds, and then send those messages out via libpurple .. once sent ,it would set the status to "successful" May 31 17:24:30 Yes, that sounds right. May 31 17:24:35 EricBlade, i think we pretty much do the same right now for SMS May 31 17:25:10 see https://github.com/ericblade/webos-messaging-connector/blob/master/service/serviceEndPoints.js from about line 389 .. although that's a very simple example, and probably doesn't demonstrate every possible thing .. and probably not even correctly. May 31 17:27:43 i'd have to go look through synerGV to find all the other fields that i was aware of.. looks like that simple example only manipulates things that are folder: outbox and status: pending .. May 31 17:30:29 IIRC, only the messaging service should write directly to DB8. The messaging app only reads and calls the messaging service. May 31 17:32:27 We probably want one service which talks to both the SMS backend and libpurple. Is there currently a service just for SMS? May 31 17:33:02 DougReederPhone: Hello. Have you ported any Enyo 2.7 apps that display an AppMenu successfully? May 31 17:33:29 Not that i recall May 31 17:33:29 That is, ported any apps to Enyo 2.7 that... May 31 17:33:35 actually, the messaging app in legacy only watched database, and manipulated database. May 31 17:33:57 it was up to the individual synergy services to also watch db for changes May 31 17:34:06 Hrm May 31 17:35:06 The key scenario is associating an incoming message with a conversation. May 31 17:35:27 so when a user wrote a new message in messaging, it would figure out what db kind the account was associated to, and then put a new entry with that kind with { .... , folder: outbox, status: pending } .. then it was up to the service to watch for those May 31 17:35:31 and handle them however May 31 17:36:07 if we have the old conversation linker, i suspect that it could be shot and put out of our misery May 31 17:37:14 after i discovered with the old one that it was possible to send messages to the wrong people, i never trusted it again May 31 17:39:26 For incoming messages, was the synergy service just writing them to an "inbox", which. The conversation linker moved to a conversation? May 31 17:39:31 as far as associating to a conversation goes, i'd probably use some kind of as unique as you can get away with string, to identify the conversation somehow .. as an identifier .. May 31 17:39:38 that sounds correct May 31 17:41:17 How was performance? Was there a noticeable delay between message arriving on the wire, and when the messaging app was notified? May 31 17:41:37 once the database got pretty big it could take quite a while, or if db8 was really busy May 31 17:41:47 Hrm May 31 17:41:54 i suspect that /batch was not used well, if at all May 31 17:41:59 in legacy May 31 17:43:15 https://github.com/ericblade/synergv2/blob/master/service/serviceEndPoints.js from line 746 is where the google voice connector handled putting incoming messages to db May 31 17:44:09 Currently, does our SMS service talk directly to the messaging service, or just to DB8? May 31 17:44:30 so, conversationthreader would take anything in the "inbox" folder and then attach it to a conversation based on whatever criteria it used May 31 17:44:54 good luck understanding conversation threader service if you want to try :-) May 31 17:46:20 the IRC Synergy connector that I played with for quite a while broke down pretty hard with conversation threader May 31 17:47:39 it broke down pretty hard in lots of ways that i didn't understand at the time, and i'm not sure i would now either May 31 17:49:10 Is SMS currently working? May 31 17:50:48 i don't see SynIRC anyway.. ugh. May 31 17:51:21 *anywhere May 31 17:51:39 i must have it somewhere .. but i've got gigs of junk that i was trying to sort through, before the incident last week :| May 31 17:53:27 Elvispre, all I've converted is Contacts, which doesn't currently have one. May 31 17:59:45 EricBlade: Garfonso re-did the chatthreader from scratch if I recall correctly May 31 17:59:55 So he might have killed legacy bugs and introduced new ones :P May 31 18:02:03 So, Herrie, you're probably right that imlibpurpleservice is what needs to be fixed up. May 31 18:04:35 Here's the chatthreader code: https://github.com/webOS-ports/org.webosports.messaging/blob/master/service/javascript/utils/MessageAssigner.js May 31 18:04:38 Might need some work May 31 18:10:16 hmm. only thing i see right off hand that i hate, is that it uses Futures, but I suspect that we didn't have more modern things available at the time it was written ? May 31 18:11:08 i'm also not sure what triggers it and how May 31 18:13:09 Garfonso: ^ May 31 18:13:46 Futures are a clunky form of Promise. Over time we can convert. In fact, we should probably modify the Futures code so a Future is a "thenable". May 31 18:15:04 i'd have to spend more time than i have right now to look at that to understand what it's doing... currently in a meeting :-) May 31 18:21:17 DougReederPhone: That's what I thought. I did study your GitHub PR for Contacts and couldn't see anything. (I can pack and install an app with an AppMenu, it's just that nothing happens. I don't know if it's my code or the infrastructure that's the problem.) May 31 18:32:31 Elvispre, do Enyo 2 May 31 18:33:07 ... Enyo 2.5 AppMenus work in the same builds? May 31 18:34:43 DougReederPhone: Yes. Calculator and Settings are Enyo 2.5, for example. The little app of my own that I am experimenting with also works fine with 2.5. May 31 18:35:35 Sounds like it's the implementation of AppMenu May 31 18:37:47 Yeah, I'm just not certain I have included everything I need to. May 31 18:38:19 No error messages are sent to the log? May 31 18:40:40 Obviously, you call Luna services with Contacts. Is enyo-luneos all you needed to include? Or did you need enyo-webos as well? (I studied your PR, but I forget now.) May 31 18:41:42 If you use enyo-luneos, you almost certainly need enyo-webos as a prerequisite. May 31 18:42:23 Yes, I do have both... May 31 18:46:19 DougReederPhone: Top tip to look in the logs. I just did that and there is a WebApplication:relaunch wth "palm-command:open-app-menu. So it's only after that that nothing happens. Maybe it's not my code. May 31 18:48:50 * DougReederPhone nods May 31 18:49:27 You might ask JayCanuck for a worked example. May 31 18:52:30 * elvispre might just do that. May 31 18:52:37 elvispre: It might also just be our rendering May 31 18:52:54 If Enyo guys changed somethign that stuff got broken in our handling of it May 31 18:53:05 That's something Tofe could look into probably May 31 18:53:36 Herrie: That's why I hesitate to poke JayCanuck in the chest and tell him his library doesn't work ;-) May 31 18:55:38 EricBlade: Basically it seems we do a watch here: https://github.com/webOS-ports/org.webosports.messaging/blob/master/service/javascript/utils/ActivityHandler.js May 31 18:58:16 You may need require('enyo-luneos') in the index.js May 31 18:58:40 ... to hook up the event listeners May 31 19:00:23 DougReederPhone: I saw JayCanuck say that and I have tried it without success. (I may now try taking it out again to see if that affects what is in the logs.) May 31 19:01:02 * DougReederPhone nods May 31 19:01:37 A worked example is a reasonable request May 31 19:02:21 i always forget about _rev May 31 19:02:59 still it seems to me that that query should probably only be on things in "pending" status, and possibly folder === outbox || inbox May 31 19:03:46 i could be wrong though May 31 19:05:34 Like I said it was rewritten from scratch so can probably use some refinement. Only used for SMS so far May 31 19:34:53 * DougReeder pokes himself May 31 19:44:24 DougReeder: New nightly should have the Enyo 2.7 Contacts App :D May 31 19:46:34 Yay! May 31 19:49:00 I didn't test it yet, but building a TP build locally with debug symbols so I can investigate the segfault :) Jun 01 02:53:01 so.. ooc.. does luneos have a working camera/mic on anything? **** ENDING LOGGING AT Wed Jun 01 02:59:59 2016