**** BEGIN LOGGING AT Sun Oct 20 02:59:58 2013 Oct 20 10:32:20 morning Oct 20 10:49:33 Garfonso: morning Oct 20 10:49:55 Garfonso: the platformVersion will arrive soon as part of our build images Oct 20 10:50:23 but you need to check /etc/webos-release instead of /etc/platform-version Oct 20 10:51:02 it consists of one line with the format: ${BUILD_DISTRIB_ID} release ${DISTRO_VERSION}-${WEBOS_DISTRO_BUILD_ID} (${WEBOS_DISTRO_RELEASE_CODENAME}) Oct 20 10:51:22 where WEBOS_DISTRO_BUILD ID looks like WEBOS_DISTRO_BUILD_ID = "005-284" Oct 20 10:51:57 DISTRO_VERSION is 1.0 Oct 20 11:08:56 ok. Oct 20 11:10:30 and which is the number to check against the manifest? Oct 20 11:13:27 or will that change, too? Oct 20 11:18:48 Garfonso: WEBOS_DISTRO_BUILD_ID = "005-284 will change for every stage we do Oct 20 11:19:03 so 005 is the platformVersion Oct 20 11:19:11 284 is the number of the build Oct 20 11:21:42 Garfonso: btw. if you want I have some else for you to work on after the update service Oct 20 11:26:11 tell me about it Oct 20 11:27:00 you remember the media db? Oct 20 11:28:15 yes Oct 20 11:28:23 although I never worked with it Oct 20 11:28:38 we're still missing a implementation of it Oct 20 11:28:50 it was done before with a native service implementation Oct 20 11:29:01 but I don't see any disadvantage currently with doing it in js Oct 20 11:30:24 yes, nodejs could be good for this. Oct 20 11:30:36 I worked a bit with filesystem access in node before. Oct 20 11:30:37 Garfonso: afaik http://nodejs.org/docs/latest/api/fs.html#fs_fs_watch_filename_options_listener should help to watch for file system events Oct 20 11:31:01 hm, seems like the webos developer site is not available again Oct 20 11:32:44 Garfonso: I never really looked into but from what I remember the service created db8 objects which were then queried by possible clients Oct 20 11:33:01 sounds reasonable. Oct 20 11:33:22 Garfonso: and it would be very good if we recreate the same format Oct 20 11:33:36 yes Oct 20 11:34:07 it's not urgent to have the service but it's one thing I would like to introduce after the Alpha 3 release Oct 20 11:35:31 ok, I'll keep it on my list and think about it. ;) Oct 20 11:35:40 Garfonso: great Oct 20 11:36:40 Garfonso: any good idea for a name for the service? Oct 20 11:39:49 Garfonso: btw. we need to add a license for the update app Oct 20 11:42:45 why not keep media indexer? Oct 20 11:42:57 ok. how do we do that? ^^ Oct 20 11:46:46 Garfonso: the license thing? Oct 20 11:46:51 btw. media indexer is fine Oct 20 11:46:57 org.webosports.service.mediaindexer? Oct 20 11:47:02 yes, the license thing. Oct 20 11:47:08 ok Oct 20 11:47:28 Garfonso: which license do you prefer? Oct 20 11:47:38 you did the work so it's up to you Oct 20 11:48:44 we just need a COPYING file with the license text in it Oct 20 11:49:41 oh my... I don't know. What licenses do you usually use? Oct 20 11:50:00 for luna-next we're using GPL-3 Oct 20 11:50:43 you can use GPL-3 too or something less strict like BSD or MIT Oct 20 11:52:03 personally I would prefer GPL-3 Oct 20 11:53:32 ok, let's use GPL-3 then. Oct 20 11:57:37 Garfonso: ok Oct 20 11:57:48 Garfonso: I will add the app later to the build Oct 20 12:02:13 actually in the owo emulator it is not working yet. I can't talk to the service Oct 20 12:02:19 don't know why. Oct 20 12:05:23 luna-send tells me that the service is not installed. Oct 20 12:05:38 but I can run it with run-js-service and it is in the right directory Oct 20 12:11:56 hm.. might \r\n line endings be the issue? Oct 20 12:16:17 nope. Oct 20 12:17:03 hm Oct 20 12:17:18 you installed the ls2 configuration files? Oct 20 12:17:24 roles and service configuration? Oct 20 12:17:36 otherwise it can't be started automatically Oct 20 12:22:32 probably that's the issue, then... I just opkg installed the package I also used for the legacy emulator Oct 20 12:48:38 so I need postinst and prerm scripts like these: https://github.com/webos-internals/tweaks/tree/master/control ? Or should I just do it manually and you handle it in the image? Oct 20 12:50:58 Garfonso: just do it manually and I will take care about packaging them correctly Oct 20 12:51:33 Garfonso: just store them in files/sysbus folder Oct 20 12:52:00 Garfonso: https://github.com/openwebos/app-services/tree/master/com.palm.service.contacts/files/sysbus Oct 20 13:04:59 do you have a sample for the new file? Oct 20 13:05:10 the webos-release one Oct 20 13:07:50 Garfonso: ${BUILD_DISTRIB_ID} release 1.0-005-284 (webos-ports) Oct 20 13:08:10 I am unsure whats the value of BUILD_DISTRIB_ID is but the other should match what the result will look like Oct 20 13:10:49 ok. Oct 20 13:16:48 ah.. the new emulator already has it: Rockhopper release 1.0-004-120 (webos-ports) Oct 20 13:29:32 ah great Oct 20 13:34:39 is the rest of the information of any need for the update service? Oct 20 13:34:52 I have a regexp that parses all the parts.. Oct 20 13:35:11 morning Oct 20 13:35:16 hi tofe Oct 20 13:35:40 but that might fail if unexpected characters are in any of the parts I don't need.. Oct 20 13:40:34 ok. work's nicely now on the owo emulator. :) Oct 20 13:42:49 btw: I have to kill luna-next nearly always when I close an app. And often it will die by itself anyway. Is that expected? Oct 20 13:43:27 and also I noticed that you can't scroll in the web-apps (i.e. wikipedia, g-mail). Oct 20 13:43:32 Should I file bugs for those? Oct 20 14:05:29 Garfonso: not really Oct 20 14:05:47 Garfonso: yes, please Oct 20 14:06:00 Garfonso: I already have them on my list but it's better to have a bug for them Oct 20 14:06:54 Tofe: morning Oct 20 15:29:03 Tofe: ping Oct 20 15:32:45 pong Oct 20 15:33:21 my bb is taking some time -- fetching yocto is "a bit slow" (30kB/s...) Oct 20 15:35:05 morphis: I am sure to be completely clear about the current situation on maliit, overlay, dora & dashboard things... To make it short: in current dora, maliit wants to create an "overlay" window, is that it ? Oct 20 15:35:09 +not Oct 20 15:35:14 (not sure) Oct 20 15:38:22 Tofe: yes, but I've found a problem with it today Oct 20 15:38:38 actually I added a patch for maliit to set the window type but forgot to activate it Oct 20 15:39:30 second thing I've found: we can't use the maliit keyboard as it makes the maliit-service segfault after eglMakeCurrent has failed (seems like there is some EGL code in which isn't suitable in our case) Oct 20 15:41:49 mmmh, so the pb is narrowed down to eglMakeCurrent failing, is that it ? Oct 20 15:42:33 yes Oct 20 15:42:50 but it's not a problem anymore as I will switch to use the nemo keyboard which is pure QML Oct 20 15:45:10 well, it's simply a switch of frontend for the maliit server, after all Oct 20 15:47:03 yes Oct 20 15:47:24 Tofe: ah and one thing: we should not close the overlay window from our side once it's not shown anymore Oct 20 15:48:33 Tofe: currently you have some logic to stack the overlays and remove the last one if a new one gets pushed Oct 20 15:48:40 that seems to cause maliit-server to exit Oct 20 15:50:21 to be more exact, I remove it when I get a tap or swipe up gesture; not when the next one is stacked. But yes, that's a mistake for maliit Oct 20 15:51:21 One thing I'm missing on QML side is a signal from the overlay window which would ask to be visible again, or to be hidden Oct 20 15:51:28 btw. I don't see the keyboard yet Oct 20 15:51:39 I saw it once for less then a second Oct 20 15:52:28 quite a low MTBF... Oct 20 15:52:58 yes Oct 20 15:53:08 I will get my recent changes out to you so you can test this too Oct 20 15:53:39 the thing is, I can hide the overlay without destroying the window, but I won't know when to show it again as I won't get any signal from the compositor Oct 20 15:54:40 (if you want to disable the hide-on-tap, simply comment out the "windowManagerInstance.addTapAction" call in OverlaysManager.qml) Oct 20 15:55:31 hm Oct 20 15:56:04 well, I don't need a signal from the compositor itself, just from the window, ofc Oct 20 15:56:06 afaik when the window has lost the focus the input method should be hidden automatically Oct 20 15:58:41 isn't it more precisely when the input entry has lost the focus ? we can still have the window in maximized and discard the keyboard, for example Oct 20 15:59:23 not sure if the window would send a focus change in that case Oct 20 16:00:39 hm Oct 20 16:02:04 let's take the pb from the other side: it is when a Qt input method is asked that we connect to the maliit server, and that triggers some show/create event, isn't it ? Oct 20 16:02:16 Tofe: pushed needed changes to the dora branch of meta-wop Oct 20 16:02:26 yes Oct 20 16:02:30 ok, great, thanks Oct 20 16:03:16 so, when the input method is discarded, that should symmetrically trigger a hide/delete event Oct 20 16:03:39 (I don't know is there is such thing as input method discarding...) Oct 20 16:03:42 if* Oct 20 16:05:56 afaik maliit will set the window to visible = false Oct 20 16:06:16 ah, but that would be sufficient for me Oct 20 16:07:32 currently we don't forward the visibility event when it comes from CompositorWindow's side Oct 20 16:08:20 right Oct 20 16:09:21 I just have to be careful not to create an obvious dependancy loop :) Oct 20 16:14:46 yeah :) Oct 20 16:15:29 Tofe: btw. one thing I saw in the battery indicator code: the icon currently isn't updated when the charging state changes Oct 20 16:15:33 only when the level changes Oct 20 16:15:47 as the level is the only input parameter for the bound method Oct 20 16:16:33 fixed, thanks Oct 20 16:17:03 Tofe: ah and the battery-charged.png image should only be displayed when charging = true Oct 20 16:17:16 as the icon currently indicates that a charger is connected which Oct 20 16:17:32 so if you have charging = false and level > 11 you get a icon which indicates the device is charging Oct 20 16:17:58 Tofe: btw. when http://jenkins.nas-admin.org/job/webos-ports_setup_webos-ports-dev-image_maguro/285 is finished you can also download the image instead of waiting for your local build to finish Oct 20 16:18:21 morphis: yes, I'll do that Oct 20 16:20:40 thanks Oct 20 16:49:24 Tofe: ok, it needs one more fix for meta-qt5 Oct 20 16:49:33 and we have to wait for JaMa to apply it Oct 20 16:50:11 ok Oct 20 16:51:03 Tofe: http://lists.openembedded.org/pipermail/openembedded-devel/2013-October/092817.html Oct 20 16:52:06 I'm happy that you know how to deal with this sort of problems :) Oct 20 16:52:15 Tofe: :) Oct 20 16:52:48 Tofe: I am really happy that I don't have to deal with UI things :D Oct 20 16:53:02 hehe, good organization we have here ! Oct 20 16:53:37 for maliit, would there be a way to make so that maliit doesn't actually hide the window, but simply asks for it ? Oct 20 16:53:46 it would look nicer Oct 20 16:54:28 hm Oct 20 16:54:30 maybe Oct 20 16:54:38 have to look into it Oct 20 16:55:07 however I don't know how to do that in a generic way, as the only thing that you have down there is a QWindow Oct 20 16:55:32 maybe you can send a signal Oct 20 16:56:08 yes, that may work Oct 20 17:38:24 Tofe: btw. one thing we have to think about is how we handle the first use application Oct 20 17:45:46 morphis: yes, that's true. Is it a web app used in fullscreen ? Oct 20 17:51:06 I just remarked one flaw in our window management: when the window destroys itself, we don't manage it well Oct 20 17:51:27 (I'm testing self-destruct on dashboard items, which is quite common I would say) Oct 20 18:39:56 morphis: oops, your build failed Oct 20 18:43:02 morphis: how useful for me would you say is the latest build of qemux86 ? for testing dashboard and/or overlays on something more concrete than Qt desktop Oct 20 18:44:28 morphis: also, I pushed some new stuff to my tofe/dashboard-work branch. They should work, but I'd rather wait for on-device tests before doing the PR Oct 20 18:45:38 From your point of view, the 'stuff' includes not hiding the overlay on tap, and hiding/showing WRT visibility of wrapped window Oct 20 19:35:26 Tofe: hm, I restarted the build but it will not have the input context plugin installed Oct 20 19:39:09 Garfonso: https://github.com/webOS-ports/org.webosports.service.mediaindexer Oct 20 19:44:59 Garfonso: https://github.com/webOS-ports/org.webosports.update/pull/3 Oct 20 19:49:02 Tofe: hm, seems like the build needs manual help Oct 20 19:49:05 will fix that tomorrow Oct 20 19:49:59 ok Oct 20 19:50:08 morphis: I renamed that little list :) Oct 20 19:50:16 Tofe: thanks Oct 20 19:50:20 ok, to late for me Oct 20 19:50:21 gn8 Oct 20 20:09:22 morphis: yes, makes sense. :) **** ENDING LOGGING AT Mon Oct 21 02:59:58 2013