**** BEGIN LOGGING AT Mon Jul 13 02:59:59 2015 Jul 13 04:34:34 Morning Jul 13 05:03:35 morning Jul 13 10:19:27 morphis: ping Jul 13 10:19:42 nizovn: pong Jul 13 10:30:29 morphis: on tenderloin seems bcattach and hciattach are needed to run on every bluetooth switch. so i tried to handle it using udev Jul 13 10:30:29 at this moment rule is not good: https://github.com/nizovn/utilities/blob/wip/tenderloin/bcattach/files/udev/tenderloin-bluetooth.rules Jul 13 10:30:29 i tried better way firstly, using SYSTEMD_WANTS, SYSTEMD_READY (to start service), but don't know how to stop service Jul 13 10:30:29 and in the end seems i broke something and bluetooth automatically turned itself off after hciattach Jul 13 10:30:31 but dirty rules work. Do i need to make them better? need some help with systemd Jul 13 10:31:33 nizovn: with bluetooth switch you mean power cycle? Jul 13 10:31:44 yes Jul 13 10:31:55 power on Jul 13 10:32:33 hm let me sync a minute about this Jul 13 12:14:01 ka6sox: ping Jul 13 13:25:11 Herrie|Veer: ping Jul 13 13:26:24 Tofe: pong Jul 13 13:28:50 Herrie|Veer: I can't seems to validate the addition of a bookmark in the browser Jul 13 13:29:11 only the Cancel button works, it seems Jul 13 13:29:17 (on N4) Jul 13 13:29:30 Tofe: morphis broke some stuff when he implemented multi-window... Jul 13 13:29:54 It was working before. I didn't have a chance to really look into it to fix it Jul 13 13:30:04 oh, ok, weird that some parts work and some don't Jul 13 13:30:16 I do reproduce the maps vkb issue Jul 13 13:30:22 Also browser history is broken Jul 13 13:30:31 You can luna-send the URL :P Jul 13 13:30:36 hehe :) Jul 13 13:30:52 As workaround, beats typing it :P Jul 13 13:31:07 According to nizovn it's OK with hw kb on emu Jul 13 13:31:17 but I see the text "Enter a query" disappear in the text field, isn't that a first success ? Jul 13 13:31:32 So it must be something about how vkb is being dealt with Jul 13 13:32:02 is the text placeholder working differently from the suggestions ? Jul 13 13:32:15 I would have thought both are connected to the same signal Jul 13 13:32:58 Well the placeholder is there initially and does the query. Just any changes or input don't trigger it Jul 13 13:33:18 I tried to use the webinspector, I don't see anything odd in there Jul 13 13:33:42 If you open the url on desktop you can see the desired behavior Jul 13 13:33:54 yes; what JS function should be called? Jul 13 13:33:58 It works everywhere except for LuneOS :P Jul 13 13:34:18 Well that's all Google obfuscated stuff :p Jul 13 13:34:32 oh, ok, even the query field Jul 13 13:34:51 no wait, the query field must be enyo Jul 13 13:37:17 That's Enyo. Just the whole pac-container stuff is google injectie Jul 13 13:37:23 mmmh Jul 13 13:37:35 trying onKeyUp instead of onKeyDown might be sufficient Jul 13 13:38:35 but how to give it a try? Jul 13 13:38:52 I can update the GH pages branch for you? Jul 13 13:40:24 Or actually you should be able to do it as well :) Jul 13 13:40:42 We can always delete the branch and recreate if needed :P Jul 13 13:42:07 I'll create my temporary branch, maybe even on a fork Jul 13 13:43:36 OK you know how the gh-pages stuff works? Jul 13 13:44:03 I'll figure it out soon enough Jul 13 13:44:49 Well if you fork it to your own repo you can just change it directly in the gh-pages branch Jul 13 13:45:08 Then in the url just replace webos-ports with tofee Jul 13 13:50:38 Even when that fixes it, the webappmgr/qml-launcher should be adjusted to correctly deal with the behavior? Jul 13 13:50:52 Because LuneOS is different from all others :P Jul 13 13:53:22 https://bugs.launchpad.net/oxide/+bug/1300382 <-- seems like the ubuntu guys had the exact same issue Jul 13 13:53:56 their analyze is that we should send a key event from our qml-webapp wrapper Jul 13 13:56:07 Yeah I guess so. Could that also somehow be related to the double taps that seem to be sent in Enyo 1 apps? Enyo 2 apps seem to be fine... Jul 13 13:57:00 Seems like they fixed it in ubuntu-keyboard? Jul 13 13:58:23 we already have their fix for backspace in our vkb Jul 13 13:58:48 but we don't have the part of the fix they did in their browser (oxide), I'm trying to retrieve the commit Jul 13 13:59:31 I wonder if the browser needs a fix too since the hw kb works. Could be just the vkb? Jul 13 14:02:02 it's qt input method Jul 13 14:02:22 https://code.launchpad.net/~osomon/oxide/im-fake-key-events/+merge/213816/+preview-diff/499643/+files/preview.diff Jul 13 14:04:41 I don't think we handle anything like that in Browser currently Jul 13 14:09:08 no, and I don't like where they did the fix Jul 13 14:10:15 but maybe that's the best we can without patching webkit Jul 13 14:11:02 I suggest to patch webkit. Jul 13 14:11:26 Anyway not likely there will be many updates to qtwebkit anymore. Jul 13 14:12:02 Isn't patching webos-keyboard sufficient? Jul 13 14:14:57 I'm not sure, it looks like Qt has no way for the input method to send the hw key events like up and down Jul 13 14:15:06 but I'll check Jul 13 14:37:05 Herrie|Veer, pong Jul 13 14:44:21 Seems we had someone edit the main page of the wiki, i blocked the user, made the page protected and reverted the change. Not sure if it was out lack of setting the rights properly or it was unauthorized access somehow. Any way for you to check? Jul 13 15:08:55 Herrie: I think I have a simple fix proposal Jul 13 15:13:16 I propose this: https://github.com/webOS-ports/webos-keyboard/pull/76 Jul 13 15:13:51 I didn't find any regression, and it made m.imdb.com proposals work; I didn't yet test maps, but it should work as well Jul 13 15:22:18 mmh it doesn't fix the maps, that's weird Jul 13 15:24:32 Can you send me ipk to test tonight? Jul 13 15:24:43 sure Jul 13 15:24:48 Could someone tell what's the difference in signal handling between m.imdb.com and our maps app? or is that hard to tell? Jul 13 15:25:40 It should be similar I'd say Jul 13 15:27:03 following the analysis from ubuntu guys, I introduced some more keypress events, even for the usual characters Jul 13 15:27:13 that fixes imdb suggestion popup Jul 13 15:27:29 but enyo must have a slightly different way of doing it, I guess Jul 13 15:28:08 NB: I only sent the keypress, not the keyrelease; but it seemed like maps uses the keydown signal, so that should be fine... or maybe not :) Jul 13 15:30:30 I think it uses keydown :P Jul 13 15:30:38 But difficult to say Jul 13 15:32:03 isn't it what we see line 3, in views/main.js? Jul 13 15:33:50 I think that might be used in a different context Jul 13 15:34:13 At least I don't see it directly connected to the input field Jul 13 15:35:05 ok Jul 13 15:35:44 Seems more related to when you select something from the search suggestions but can be misreading it Jul 13 15:37:24 btw, I find it very hard, or maybe impossible, to change the URL once it has been validated Jul 13 15:38:12 ? In browser or? Jul 13 15:38:17 yes Jul 13 15:38:29 That's on N4? Yeah I lacked the left right navigation there Jul 13 15:38:39 I can select it, copy it, cut it, but not simply edit it Jul 13 15:38:46 Trackball on TP is usefull there Jul 13 15:38:55 Well you can but not scroll left or right Jul 13 15:39:12 We need something for that :P Jul 13 15:39:24 no no, I mean, I can't even get a simple cursor Jul 13 15:39:31 can you? Jul 13 15:39:35 I can I think need to check Jul 13 15:39:44 PS: it's a 2 weeks old build Jul 13 15:39:46 I think I managed last weke Jul 13 15:40:05 ok Jul 13 15:40:35 maybe it's just me, at least I hope :) Jul 13 15:44:47 Well anyway that should be a minor fix in case it doesn't work properly. Jul 13 15:50:26 I guess so Jul 13 15:51:00 I tried to replicate legacy behavior Jul 13 15:52:18 But browser needs some work/cleanup again in general Jul 13 15:52:24 Was my first QML thing Jul 13 15:52:35 So learned a thing or 2 since then :) Jul 13 15:53:50 Yes, it was the same thing for me :) Jul 13 15:56:02 Like how to use BorderImage :P Jul 13 17:00:00 Tofe: For cursor positioning: 1 tap = select all, 2 taps shows copy/paste, 3rd tap allows cursor setting Jul 13 17:01:22 Copied that from TP ;) Jul 13 17:27:57 Tofe: luna-send -n 1 luna://com.palm.applicationManager/open '{ "id": "org.webosports.app.browser", "params":{"target": "http://webos-ports.github.com/org.webosports.app.maps/debug.html"}}' Jul 13 17:46:33 Tofe: I can confirm that with your patch, for me it also doesn't work on N4 Jul 13 19:21:31 Herrie: ok, that was quite expected Jul 13 19:21:54 Still, as it now works with imdb, I think it's an improvement Jul 13 19:31:18 Yeah true Jul 13 19:32:04 Maybe should do key-release as well or both? Jul 13 19:33:17 not sure yet Jul 13 19:38:01 WEll I'm happy to guinea pig test when you send me IPK ;) Jul 13 19:38:28 I'll have to guess what's wrong first :) Jul 13 19:46:11 WEll you can just make an IPK with key-release and with key-press and release ;) Jul 13 19:46:39 I'll try, I'll try Jul 13 19:47:00 but it may be also that Enyo looks if the event contains an actual keycode Jul 13 19:48:03 Well I can ping the Enyo mailing list ;) Jul 13 19:48:14 If you got a proper question for me ;) Jul 13 19:48:22 not yet Jul 13 19:48:41 I'm subscribed to their mailing list Jul 13 19:50:27 Tofe: https://github.com/enyojs/enyo/blob/master/source/dom/keymap.js#L23 Jul 13 19:50:32 Might be of help? Jul 13 20:09:47 maybe Jul 13 20:41:36 Herrie: sending also the keyRelease seems to help, let's see if there is a win-win setup Jul 13 20:48:28 looks like it works! I've update my PR, and the ipk is on its way **** ENDING LOGGING AT Tue Jul 14 02:59:58 2015