[00:16] I've built ubuntu-touch for my Galaxy Nexus. Upon booting, logcat says "failed to connect socket for 'dev/socket/upstart-text-bridge'". Why wasn't that socket created while booting? [00:17] is adbd running? are the android container running? [00:18] adbd is running. i can adb shell and adb logcat. How do I access the container to know it is running? [00:19] lxc-info -nandorid [00:21] lxc-info -nandroid -> android doesn't exist [00:23] Humm, then it looks like it didn't setup the lxc correctly... do you have a last_kmsg that i can read? (and did you run that command in adb shell btw) [00:24] yeah. I initially did run lxc-info in adb shell. oops. within adb, it can't find lxc-info [00:26] that fine, i was just confirming that you wasn't running it on you host computer. [00:27] your* [00:31] randy1 do you have a last_kmsg or syslog i can read? paste it to pastebin.com or something [00:32] http://paste.ubuntu.com/10607342/ [00:34] randy1 it looks like you rebooted it before everything was loaded. [00:37] remember last_kmsg is for the reboot that was before you pulled, if you want the current kmsg do adb shell dmesg > log.log && adb pull log.log [00:37] or do a reboot and then pull last_kmsg [00:39] http://paste.ubuntu.com/10607377/ [00:39] that is my current log [00:41] init: untracked pid 2259 exited [00:41] that is the only thing i can find... humm [00:43] yeah, but that is a long time after the kernel booted. what should be happening in the kernel? After around 19sec, nothing seems to happen. [00:43] at some point the kernel should mount the ubuntu container, right? [00:44] yeah [00:44] but it should mount system.img too [00:44] initrd: mounting system.img (user mode) [00:45] that's form my dmesg [00:45] from* [00:45] that's what I thought [00:45] can you check is system.img is even existing? in /data [00:46] yep. /data/system.img [00:47] is it mounted? what do you get when do ls in shell? [00:49] where is it supposed to get mounted? [00:50] it doesn't appear to get mounted. there is no /var dir, for example. [00:51] btw there is two system.img one induces ubuntu root system (/data/system.img), and the other one is android (placed inside data/system.img in var/lib/lxc/android) [00:52] try to mount it manually, mkdir cache/system && mount -o loop data/system.img cache/system [00:53] cd cache/system [00:53] ls [00:54] alright, had to add the fs type, but it is now mounted at cache/system [00:55] it's all there. [00:55] is it possible to kick start it now that it is mounted? [00:56] no, it's not mounted in root [00:56] or maybe, but i don't think so [00:57] okey, so the kernel is not mounting system.img, now we need to find out why :P [00:58] can you try with a kernel already build and see if that works? [00:58] yeah, I probably messed up the kernel somehow. [00:58] i don't think there's a prebuilt kernel for the galaxy nexus [00:59] i attempted to boot with one of the other kernels (maybe mako) but it wasn't successful [00:59] yeah, i don't think that will work [01:00] are you compiling with a mac? I don't know, I don't think that will make any different, but you never know [01:01] no, it's ubuntu, just running on my mac [01:01] ah, okey [01:02] did you edit the kernel config? [01:02] while compiling my kernel, I applied the update for the kernel config that is mentioned in the porting guide [01:02] ah, this is port [01:03] then, did you use this tool? https://github.com/mariogrip/phablet-porting-scripts/blob/master/kernel/check-config [01:03] well, that's the only place that mentions editing the kernel config [01:03] yeah [01:03] i defintely read that without those kernel configs, the ubuntu container wouldn't start. [01:05] try to add debug to kernel command line, the you get more outputs [01:05] after runnning that script (check-config), I just ran make, which I'm pretty sure should have used the updated .config file, but this is pretty new to me. [01:05] ok [01:06] btw, did you ran check-config with -w config? [01:06] yeah [01:06] ok [01:09] darn. so I just ran check-config again on my .config file, and it didn't 'pass'. It must have written to the config file elsewhere, but I guess I need to copy that file onto .config [01:09] let me rebuild my kernel. :) [01:10] :) [01:11] mariogrip thanks for the help. I'll be back on after I build the kernel again. [01:13] randy1, you'r welcome, btw i will be going to bed soon, so if i'm not online i will tomorrow [01:13] you're* [05:57] So now that my device has detected what time zone I'm in, how do I get it to accept alarm times as this time zone? Whenever I create an alarm in the "clock" application, it defaults the time to UTC instead of my time zone. === marcusto_ is now known as marcustomlinson_ === marcustomlinson_ is now known as marcustomlinson [06:31] How would I go about pasting my wifi password in Ubuntu Touch? [06:32] I can copy it from my router web page, or from a text document I have placed on the device, but when I press and hold in the wifi password box there's no option to "paste". [07:20] Get Wifi Password Into Ubuntu Touch | http://askubuntu.com/q/597333 === chriadam is now known as chriadam|away [07:58] good morning [08:28] oSoMoN, o/ [08:30] rpadovani, \o [08:32] oSoMoN, seen your last reply to the new tab view. I don't get how you plan to achieve the same behavior with column and flickable.. How could you fix the position of the section header and push it out when bump the next section? [08:33] rpadovani, I don’t think we could, indeed… But the design spec doesn’t strictly specify this behaviour as required either [08:35] oSoMoN, oh, well, if isn't specified by design I can totally simplify that page [08:36] oSoMoN, thanks, I'll try during the week [08:36] rpadovani, given the numerous issues you’ve bumped into with the current implementation, I think that’s the easiest solution [08:37] indeed === chihchun is now known as chihchun_afk [09:50] hello! have anyone ported Ubuntu to Motorola xt894 (Droid 4/maserati) with the new guide? [09:51] I've just received a phone call from my bank's fraud dept. alerting me to some prepay phone transactions from Europe on my account... As the only thing I've purchased from Europe is the bq phone, I'd like to let everyone in here know to check their bank statements for small transactions they didn't make... I'm not being accusational, just being safe. [09:59] Good morning all; happy Monday, and happy Freedom of Information Day! :-D [10:24] I have problems with setting up a porting env for s4 jfltexx...can somebody help me [10:26] Isotop7: maybe check the XDA forums for porting your device [10:27] thanks :) [10:34] well...on xda there is no info for s4 or porting in general...just people pretending to have working ports but none of them tells how...i did the basic structure but i have errors while make -j4 saying that values where overriden...i did some research on these errors but i found nothing which leads me to a conclusion...the porting guide is nice, but without further informations on which files out of aosp or cm tree are exactly needed [10:34] , noobs like me are pretty much screwed... [10:49] does anyone know where i can get more informations regarding porting to another device and problems with it? === chihchun_afk is now known as chihchun [11:20] 1 === chihchun is now known as chihchun_afk [11:34] i need help regarding porting ubuntu touch to my s4 jflte(xx)...are there any further sources of information besides the porting guide? === chihchun_afk is now known as chihchun === MacSlow is now known as MacSlow|lunch === chihchun is now known as chihchun_afk === Guest36153 is now known as vik === alan_g is now known as alan_g|lunch === dandrader is now known as dandrader|afk [13:14] Elleo, can i get a review when you have time? https://code.launchpad.net/~ken-vandine/content-hub/lp1429899/+merge/252448 [13:19] kenvandine: sure, is there any apps using this that I can test against? iirc it was being used in messaging app? is that actually landed or is it still in a branch somewhere? [13:19] Elleo, not really :) [13:19] heh [13:19] the apps don't use it anyway, it was really for the sdk [13:20] the sdk is going to filter that out [13:20] so apps don't even see it [13:20] but it'll be used to get the uri signal [13:20] kenvandine: presumably without the sdk part it still should fix the statesaver issue though? since its now going through the uri dispatcher path [13:21] kenvandine: so if I write a quick test app that uses statesaver I should be able to reproduce issues without this branch and not see them with this branch? [13:21] not really [13:21] oh :/ [13:21] Elleo, turns out that was working out of luck [13:21] heh [13:21] in that case I guess I'll just run it through the test plan and make sure it doesn't break any existing behaviour [13:23] yeah [13:23] just make sure it doesn't cause regressions [13:23] which is all i did [13:23] it'll be up to the sdk guys to make sure their change fixes the problem [13:24] okay [13:24] kenvandine, so ignoring my review comments and approving settings changes anyway? ;-) [13:25] seb128, i know... but i'm not going to block on that [13:25] it's one line [13:26] kenvandine, yeah, jgdx could also easily revert that 1 line in the mr :p [13:26] maybe he'll fix it before i actually land it :) [13:26] jgdx, ^^ [13:26] * jgdx runs around franticly [13:27] seb128, but i'd rather it be on the approved MP list so i can more easily keep track of what i want to land... but it'll need the FFe [13:27] and it's a UI change [13:27] kenvandine, right [13:28] I'm not sure I know what mr you're talking about. [13:28] the other vibrate branch [13:29] you had a spurious change in it, just white space that seb128 had wanted fixed [13:29] seb128, you want it back to WRONG? :p [13:29] jgdx, kenvandine, no big deal, but random spacing changes in unrelated mrs screw bzr blame and logs [13:30] jgdx, I would rather have that line associated with a "fix spacing" commit than with a "add other vibration" [13:30] seb128, I can do that. kenvandine? [13:30] jgdx, because associating code with unrelated changes leads to confusing history logs [13:30] yeah, that'd be better [13:30] same MP is fine, just separate commit [13:30] +1 [13:30] jgdx, you probably have a plugin for that don't you? [13:31] I usually uncommit; revert unwanted change; commit; change, commit; push --overwrite [13:31] if you overwrite make sure to do it on your branch though :p [13:31] not on trunk [13:32] oh sigh... facebook is dropping xmpp too? [13:32] what is the world coming to! [13:32] seb128, lol [13:32] seb128, Haaaha [13:32] ;-) [13:35] seb128, I'll do another branch. Is that OK? So we get the blame right. [13:35] jgdx, wfm [13:35] jgdx, feel free to run your intender on all the sources maybe while you are at it [13:36] seb128, indentations are untouched I believe, it's only trailing spaces. But sure, I remove trailing spaces for all files. [13:37] jgdx, oh, right, thinko ... thanks ;-) === MacSlow|lunch is now known as MacSlow [13:41] seb128, err, it transforms tabs into spaces as well. Was that what you meant? [13:42] jgdx, no, I was just wondering if you had more "trailing spaces" changes to commit because we can as well batch them in one mr [13:42] yap, coming [13:42] but I guess it's your editor doing it for you want you edit a file? [13:55] seb128, yes === dandrader|afk is now known as dandrader [14:01] kenvandine: approved :) [14:01] Elleo, thx === chihchun_afk is now known as chihchun === alan_g|lunch is now known as alan_g [14:30] zsombi, nik90, I'd like to land something for https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1426519 and https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1431384 this week [14:30] Ubuntu bug 1426519 in indicator-datetime (Ubuntu) "Need a mechanism for specifying what happens when an ical menuitem is clicked" [Medium,Confirmed] [14:30] Ubuntu bug 1431384 in indicator-datetime (Ubuntu) "Need a mechanism for specifying an ical event's icon" [Medium,Confirmed] [14:30] zsombi, nik90 mentioned that you might want to give some feedback on the former, relating to https://bugs.launchpad.net/ubuntu/+source/ubuntu-ui-toolkit/+bug/1426361 [14:30] Ubuntu bug 1426361 in ubuntu-ui-toolkit (Ubuntu) "Alarms API should expose a unique permanent alarm ID that clock app can use to open an alarm" [High,In progress] [14:30] do you two have time to talk about it today? [14:32] charles: Yeah I have time to discuss this today [14:37] zsombi: If you can let us the time when you would be free, we can organize a hangout to discuss this quickly [14:38] nik90, zsombi, I'll be free from meetings & errands in ~90 minutes and will be open for this any time after that === salem_ is now known as _salem === dandrader is now known as dandrader|lunch === salem_ is now known as _salem === dandrader is now known as dandrader|lunch [17:05] zsombi, ping === _salem is now known as salem_ [17:36] charles: nik90: sorry guys, I had some business today afternoon, I'm here for a short time [17:36] zsombi, nik90, I'd like to land something for https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1426519 and https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1431384 this week [17:36] Ubuntu bug 1426519 in indicator-datetime (Ubuntu) "Need a mechanism for specifying what happens when an ical menuitem is clicked" [Medium,Confirmed] [17:36] Ubuntu bug 1431384 in indicator-datetime (Ubuntu) "Need a mechanism for specifying an ical event's icon" [Medium,Confirmed] [17:36] zsombi, nik90 mentioned that we should talk to you about the former, relating to https://bugs.launchpad.net/ubuntu/+source/ubuntu-ui-toolkit/+bug/1426361 [17:36] Ubuntu bug 1426361 in ubuntu-ui-toolkit (Ubuntu) "Alarms API should expose a unique permanent alarm ID that clock app can use to open an alarm" [High,In progress] [17:37] charles: yes, I started some code for that, but then I saw your post, so I'll assign an ID to some tag to the alarm which you can consider as ID [17:37] charles: so far I thought I could use teh QtOrganizerItem ID, but that one you don't see [17:38] charles: and that si not future proof enough - thinking of the new backend we wil have [17:38] s/si/is [17:38] charles: but Alarm service will not launch the app [17:39] zsombi, ? [17:39] charles: nik90: we have the UriHandler, Clock app could use that to intercept the call from i-dt [17:39] charles: so you can launch the app with upstart, right? [17:40] zsombi, yeah it'll be launched through calling url_dispatch [17:40] zsombi, that's why I'd prefer that datetime be fed a url rather than an ID, so that the same code will work for clock, and for calendar, and for reminders [17:41] * nik90 reads the backlog [17:42] eg, the app-id plus the alarm id as a parameter to the url [17:42] I'm sure tedg has some ideas how a url to launch clock-app via url_dispatch should look [17:43] charles: uhm... I'm not sure we woudl need to specify any special URL... [17:43] it should be calendar:// (the same thing evolution already uses) [17:44] dobey, not for alarms it wouldn't? [17:44] charles: well you could make it alarm:// or whatever instead [17:45] you'd need some mechanism to differentiate from alarms to calendar to reminder [17:45] charles: what if there would be a separate tag specifying the app as well? you then build this URI and amen [17:45] :) [17:45] zsombi: I think the url should be something like alarms://

and then that launches the necessary app as required [17:46] charles: an event is an event. if you tap on the event it will go to the event in the calendar i presume, regardless of whether it has a reminder or not, no? [17:46] although I am not sure how urldispatcher would know if it is clock app, calendar app or any other 3rd party app to open [17:46] well evolution is kind of lame in that regard it seems [17:46] No, we don't have a way to parse the URL [17:46] dobey, language is confusing here because the words are so overloaded :-) [17:46] only goes to the calendar view, and not the event itself [17:47] It needs to be in the protocol really. [17:47] dobey, if we're talking about a reminder that was created in the reminders app, tapping on it should launch that instead of the calendar app [17:47] charles: i think if you want to treat (actual) alarms specially in this regard, then it should be alarm://foo, and events/dates/whatever should be calendar://foo to get to that event/date/whatever in the calendar app [17:48] charles: i thought the reminders app was notes (aka evernote)? [17:48] tedg: then how does one handle multiple apps handing alarms created by them? we dont want i-dt launching the clock app regardless of which app created that alarm [17:48] Exactly, each target should probably have it's own protocol. [17:48] dobey, if a URL is missing, we'd probably want to assume the default is calendar-app, e.g. events imported from google [17:48] nik90, One would be "clock:///" and one would be "myapp:///" [17:48] charles: evernote (reminders-app) should probably be the https://evernote.com/whatever URL [17:49] dobey, Depends on the app author, some of basically taken the "whatever" and put their own protocol on it to simplify the routing. [17:49] charles: that's how i'd do it for meetup/trello/etc… if i could provide backends for the calendar, or events to the indicator somehow [17:49] charles, zsombi: Then we assign one of the alarm tag url values as "app-id:///alarm-id"...this way i-dt can just pass that to the urldispatcher which would then open the correct app [17:50] nik90, Then it would open the app, but not the item in the app. [17:50] tedg: sure, but in general i think if it's an app that implements a native interface for a web service, it should handle the http(s) urls and parse them [17:50] dobey, App author's choice :-) [17:50] dobey, that would be mzanetti's call [17:50] so that when i click a link in irc, it opens the app instead of the browser, or if i have event and not the app, it opens the browser [17:50] tedg: well we are also passing the alarm-id, which the app can then parse and take it to the correct page [17:50] tedg: yes, some make bad choices [17:50] nik90, You can't pass the alarm-id in an appid:/// URL [17:50] here, I just want to agree on the general structure so that it's something url_dispatch() can handle and that clock app et al can consume [17:51] tedg: we should document this as the recommended way i think :) [17:51] hmm [17:53] tedg, so alarm://${alarm-id} ? [17:53] +1 [17:53] Wait, you need three slashes [17:53] no you don't [17:54] Depends on the format of the alarm ID. [17:54] charles: but that won't allow 3rd party apps to handle their alarms (which was the issue we started with) [17:54] You WANT three slashes. [17:54] nik90, They'd use a different URL. "myapp:///$myid" [17:54] you don't NEED any slashes === chihchun is now known as chihchun_afk [17:54] nik90, I think you're right [17:54] dobey, Correct, but you really want 3 or zero here. [17:55] tedg, what would be a url-dispatch friendly way of letting 3rd party apps work here? [17:55] * tedg is confused. [17:55] tedg: i'm not sure why you're insisting on three. [17:55] You want the app setting the alarm to provide it's own URL. [17:55] For clock app, it can use alarm:///$(alarm id) [17:56] For calendar, it can use "calendar:///$(event id)" [17:56] ted, how does the protocol get resolved to $app in the case of 3rd parties? [17:56] For foo-app, it can use "foo:///$(foo id)" [17:56] charles, Each app will have to register its protocol in its package. === chihchun_afk is now known as chihchun [17:56] charles, That's done with a url-dispatcher hook. [17:56] ok [17:56] it should be calendar:/// [17:56] i think [17:57] dobey, I find it's easier to talk about that way. People don't think things are URLs without a few slashes for better or worse. [17:57] dobey, App author's choice :-) [17:58] tedg: the only reason file:// needs the third slash is because on sane file systems, / is the root of the path. we're not talking about filesystems here, so the third slash is not relevant/useful [17:58] tedg, partially. I think in the specific case of calendar://, dobey is talking about making it play nice with evolution as well [17:58] dobey, The data after the second slash needs to conform with domain name naming requirements. [17:59] charles: exactly, it should use the same thing we've been using for N years for this already :) [17:59] dobey, For instance, QtURL won't let you use capital letters there. [17:59] tedg: domain names can't start with a slash [17:59] dobey, +1 [17:59] evolution uses things like "calendar:///?startdate=%s&enddate=%s" [18:00] Sure, but from datetime's perspective it's all opaque. [18:00] String from iCal entry passed to URL Dispatcher. [18:00] i was trying to find exactly what evolution was already using, but my search-fu is failing me [18:00] cursory ag'ing looks like startdate, enddate, source-uid, comp-uid, and comp-rid are what evolution can handle as url parameters [18:01] tedg, ideally this will be opaque to datetime [18:01] ah so evolution just does everything as a parameter to /? [18:02] tedg, but in the case of a missing url we'll want to fall back to calendar-app for reasons mentioned above, so datetime can't escape knowing calendar:// [18:02] i guess that's what we get with 17 year old software [18:02] charles, Sure, makes sense, but I think we shouldn't admit it publicly :-) [18:02] dobey, that's how datetime launches evolution now, with "evolution \"calendar:///?startdate=%Y%m%dT%H%M%SZ\"" [18:02] tedg, :-) [18:03] charles: yeah. that is a bit nasty/annoying though [18:03] dobey, what's the better way? [18:04] it would be great to shed the "evolution \"" substring but afaik url dispatch isn't working on desktop yet [18:04] charles: well, i mean, when i click on an event in the indicator, it just opens the evolution calendar to that start time. it doesn't show me the event (which is what i wanted to see) [18:04] well, xdg should be able to handle calendar:// fine i think [18:04] oh, i see your point. Yes it would be better to use the id [18:04] tedg: how can I register an app to be able to handle uris dispatched? - just a side Q [18:05] ah, but it doesn't [18:05] but that's just the fault of whomever failed to register calendar:// as a uri scheme for xdg-open === dandrader|lunch is now known as dandrader [18:05] evolution, presumably [18:06] zsombi, https://developer.ubuntu.com/en/apps/platform/guides/url-dispatcher-guide/ [18:06] tedg: dobey: I need in one test to launch a test app with a special URI, which disables the StateSaver [18:06] tedg: ah, brilliant!! [18:07] zsombi: an autopilot test? [18:07] dobey: an AP test, yes, which should launch a simple test app, created otf [18:08] zsombi, Don't you want to restore the state and then sent the signal for a URI? [18:08] zsombi, The URI response could be state specific [18:08] zsombi: if it only has to work on >= vivid, maybe look at using ubuntu-app-test to launch such an app instead [18:09] tedg: well, the issue is with teh ContentHub - we agreed with kenvandine that he will launch the app with content-hub:///true, so when taht comes, that should disable the state restoration [18:10] kenvandine, ? Did we drop the having a specific exec line? [18:10] Seems like it shouldn't be a URL really. [18:11] tedg: dobey: here's the bug, and the MRs linked https://bugs.launchpad.net/ubuntu/+source/ubuntu-ui-toolkit/+bug/1429899 [18:11] Ubuntu bug 1429899 in ubuntu-ui-toolkit (Ubuntu) "StateSaver should not restore when app launched from content-hub" [High,In progress] [18:12] tedg, not decided yet [18:12] tedg: there was no exec line, but kenvandine added this URI, we want to keep the UITK cleen from dependencies [18:12] but the uri is just so the sdk can see it [18:12] tedg, did you see i got the trust session to work? === om26er_ is now known as om26er|food [18:13] kenvandine, It seems like it'd be better to pass as a parameter than trying to parse a URL. [18:13] kenvandine, I did not, where can I see it? [18:14] tedg: by parameter you mean argument? [18:14] +t [18:14] Yes, foo-app --no-restore-state === alan_g is now known as alan_g|EOD [18:15] tedg, nothing you can see yet :0 [18:15] i had mentioned it to you on irc the other day [18:15] kenvandine: ^^ what do you think of that? [18:15] You'll need to be able to implement the same thing on other toolkits. And passing I URL is silly. [18:15] that's fine with me [18:15] a [18:15] kenvandine, Spring break :-) [18:15] kenvandine, Super exciting though. [18:16] kenvandine: ok, let's do it so. We have components for apps to parse the Arguments, so disabling can also be done in the app [18:17] kenvandine: however I can also handle that in the StateSaver backend, so it'll be global disabled if that si specified [18:17] zsombi, Seems like something we could do for app developers, no? Seems like a sane default. [18:17] Ah, okay. [18:17] * tedg was hasty [18:17] :) [18:19] zsombi, so just change it to --no-restore-state ? [18:19] kenvandine: yep, just commenting on the bug === om26er|food is now known as om26er === dandrader is now known as dandrader|afk [18:33] zsombi, so i'm still just going to be passing that arg as an uri to ubuntu-app-launch, but later when/if we change to have a separate exec line we'll do it there [18:33] i think we'll only have separate exec lines for imports [18:33] so this way we can still pass it [18:34] and you want to get the arg change when we focus the app if it was already running [18:34] kenvandine, What about exporters that don't have applications? [18:34] i.e. my package of scope, importer, exporter [18:34] Remember, applications are dead :-) [18:35] they don't need a separate exec line [18:35] Where do I put any exec line? [18:35] they'll just get the signal and arg passed [18:35] i.e. I don't have a desktop file. [18:35] true [18:36] we'll need to figure that out :) [18:36] but hopefully we'll still be using UAL for that, so that'll be your problem [18:36] :) [18:36] heh [18:36] i only need to figure out the exec line for imports [18:37] tedg, so scopes will already be running anyway, so we just need to make it focus [18:37] and since they are running they'll get the signal [18:38] kenvandine, but in both cases you could start from a running app. For example in messaging attaching a picture, or gallery exporting a picture. You need to overlay in both directions, so you can't use UAL. [18:38] kenvandine, Scopes are almost never running. [18:38] they will be when they get focused :) [18:38] kenvandine, Default idle timeout is 10 seconds. [18:39] No, they get refreshed and killed. [18:39] so we can't use UAL to start the scope? [18:39] so you're saying the scope will provide a real app too right? [18:39] that doesn't die after refreshing [18:39] No, you can pass a URL, scope://scopeid [18:39] kenvandine, No, why would it? [18:39] to handle the content request [18:40] I'm specifically thinking of pinboard.in. I'd have a scope that shows my book marks. And then I need a way to pull and push them via content hub. [18:40] I don't need an app that lists them. [18:40] The scope does that. [18:40] so saving one and getting one [18:40] but you would need to pick one right? [18:40] it would need to keep running until the user picks one? [18:41] Sure, so those would be QML "apps" but they'd not have an app icon. Only brought up when someone asks for a URL. [18:41] Yes, by content hub in the trusted session. [18:41] yeah, so they'll have a desktop file :) [18:41] just hidden [18:41] We don't hide desktop files [18:41] why not? [18:41] and yes we do... [18:41] Because it's broken by design. [18:41] signon-ui does [18:41] It's not using UAL. [18:41] or at least it use to [18:42] true [18:42] but it won't show up as an app in the dash [18:42] apps scope [18:42] Something needs to be an app or not. Apps being "hidden" means they end up in a grey area of bad design. [18:42] well a way to figure out the exec line [18:42] we need some way for apps to define APP_EXEC [18:43] why not use the same format we use elsewhere [18:43] I don't care what the format is, I care that it doesn't have a desktop hook. [18:43] I'm happy for the file format to be a desktop file. [18:43] sure [18:44] as long as UAL can handle the smarts for me :) [18:44] we don't want everyone to have to write code for figuring that out [18:44] No, but it'll probably be your untrusted helper exec helper, not UAL specifically. [18:45] It's content-hub specific [18:45] we are only going to be using the trusted session for imports [18:45] but i guess i could have my own helper for the others too [18:45] the "open with" and "share with" implies you probably want to stay in the other app when complete [18:46] open a document, then read it instead of switching back [18:46] so not tied together [18:46] and those should be the full versions of the apps, not a custom UI for handling it === dandrader|afk is now known as dandrader [18:47] It seems to me that the exporter should decide if it's the full app or not. Not you :-) [18:47] sure [18:47] They could be optimized for the task. [18:47] they could already do that though [18:47] create a separate app in the click package [18:47] that does it [18:47] it'll be the one that gets registered [18:47] So in that case they use the same Exec line. But having the choice is kinda critical. [18:52] i guess we don't support actions in the .desktop files yet? [18:54] that's what i had originally wanted [18:55] but not sure that really helps with the concern tedg has [18:55] well it would be much better than having multiple "apps" in a package [18:55] indeed [18:55] but if one of them is a scope and one is a qml app [18:55] I do want to support actions, but I think those'll end up on the launcher. So you don't want "Import" on the launcher :-) [18:55] the scope doesn't have a desktop file [18:56] yeah [18:56] scopes do have .desktop files [18:56]