mozilla :: #mobile

20 Mar 2017
09:10mr_booCan anyone help me with interacting with javascript layer from android java layer
09:10mr_booCan someone help me with interacting with javascript layer from android java layer
09:32freesamaelI image you need to send a message through this
09:33freesamaelah it looks deprecated. I only have impression I read GeckoThread or so
09:47mib_2ao7x6I believe that singleton class BroswerApp residing in mobile/android/chrome/browser.js hold all the tab information.
09:47mib_2ao7x6I need to get the active tab url from the back end into android java layer
09:47mib_2ao7x6can some one please help me
09:59rnewmanmib_2ao7x6: have you read
10:00rnewmanthe Android side already knows the current tab
10:00rnewmanand there are already messages that keep it apprised of the current URL when navigation occurs
10:10mib_2ao7x6yes @rnewman. already knew holds it. But i have been updating the url in for some case
10:11mib_2ao7x6And i need to cross check that url with is one in back end layer
10:17rnewmanthen follow the existing message sending pattern
10:18rnewman(btw, if you change your screen name in Mibbit, it'll make it easier for people to know who you are)
10:21mib_2ao7x6sure thank you. i just need to knew that message sending patter is synchronous. i.e consider getCurrentUrl from java layer should get the url from back end layer
10:23rnewmanI don't believe the message send pattern is synchronous by default. jchen can confirm.
13:52gobican some one help to get the current URI from backend layer using JNI
13:52snorpgobi: why do you need that?
13:53snorpgobi: we don't use jni for that
13:54gobii have updated mUrl with for some case
13:54gobii need to verify that mUrl with identical with backend active tab nsURI
13:56gobicorrect me if am wrong mobile/android/chrome/content/browser.js is the one that holds the active tabs and BrowserApp object.
13:58snorpgobi: right, but you don't use JNI (directly) to get that
13:58snorpgobi: you send a message
13:58gobisnorp: messaging is synchronous ?
13:58snorpgobi: no
13:59gobii need it to be synchronous. is there any workaround for it ?
13:59snorpgobi: it feels like you're doing something wrong
13:59snorpgobi: if you are changing the url in then you should just keep the old one around for comparison
13:59snorpif that's what you are trying to do
14:00daleharveyhrm this is confusing (android layout stuff), trying to set a custom actionbar, pretty similiar to custom tabs
14:00gobisnorp: thank you. thats how currently i have implemented.
14:01daleharveyI give it a TextView, but whatever text I set, whether the initial value in the layour or via setText, it only ever says "Fennec daleharvey"
14:01daleharvey is the code
14:01snorpgobi: whenever the location is changed from gecko, it gets sent to, so you should always have it available there
14:01gobisnorp: thank you very much :)
14:02snorpdaleharvey: that string is the action bar title, not your textview I would guess?
14:03daleharveysnorp: ah yeh looks like I forgot about setDisplayShowTitleEnabled
15:30njparksebastian: looks like the UI tests are consistently passing on buddybuild, when you have time, could you review my PR?
15:30njparki think it adds about 5 min max to the total build time
15:30sebastiannjpark: Nice! I'll have a look.
15:31njparksebastian: thanks! on buddybuild, you can also see the video of the test playback. they are quick easy sanity tests
15:33sebastianTrashcanTest =)
15:34njparksebastian: i am terrible at naming stuffs :)
15:34sebastiannjpark: I approved the PR and added 2 comments.
15:35njparksebastian: aha, this line, because the espresso was using old annotations and causing conflicts, i think it can be removed, will do that and add proper headers
16:01sarentzcsuciu: !!! you submitted a patch! :-) :-)
16:02csuciusarentz: yeah :D
16:02sarentzcsuciu: more please :-)
16:04csuciusarentz: hope so :-)
16:08csuciusarentz: new beta today?
16:08slerouxcsuciu: new beta today!
16:09csuciusleroux: great!
16:10slerouxI submitted an internal testflight (not beta) just now - feel free to ignore that for now
16:10AaronMTtested out the linkpreview in our webview, we need to work against the context menu for the long press gesture
16:10slerouxIm submitting an actual beta build shortly
17:18ahuntDoes anyone know how fennecs webfont blocking decides to load a webfont regardless of setting (some sites use webfonts for icons, and they seem to get loaded regardless of webfont setting).
17:20ahuntNvm, it turns out thats all in - it seems fonts are loaded when characters not contained in the default fonts are used (and icons use special characters)
17:20firebotBug 789788 FIXED, Not allowing pages to choose their own fonts breaks with icon fonts
17:26njparksebastian, ahunt : I enabled UITest on bb for focus- I'll keep a watch on it to see whether anything bad happens. (I also checked 'treat tests as warnings' so it won't fail your builds for now)
17:27njparkit only runs on focuswebitdebug, on 5.1, 6.0-6.1, and 7.1, takes about 5 min extra
17:27AaronMTnjpark: oh nice
17:28njparkAaronMT: yeah hoping to add couple more tests this week, but about 6 basic sanity tests are added
17:29AaronMTwhich framework?
17:29AaronMToh espresso
17:29njparkAaronMT: yup it's a fusion of espresso/uiautomator
17:30njparkAaronMT: thanks! hoping these would catch the super obvious issues at least
17:32AaronMTtrashcantest lol
17:32njparkI mentioned to sebastian that I am terrible with naming things
17:34njparksome other tests i'm planning are the ones that visit about us/help/your rights page, and one that sends bad url like htps:// (the latter needs fix currently)
17:35AaronMTsounds good
17:49sebastiannjpark: <-- can we delete the branch (there&#39;s a button at the bottom of the PR) - or do you want to keep it for something?
17:49njparksebastian: oh, about to delete right after i see it running on master- I just got rid of it
17:50sebastiannjpark: thx :)
18:07mkaplyIs activitystream_contextmenu.xml used for anything yet?
18:08ahuntmkaply: yup, but only if you enable activity stream (its only used in the page that would replace topsites)
18:08ahunt(And thats not enabled anywhere except opt-in users on nightly)
18:08mkaplyahunt: So if I&#39;m adding something to the context menu (bug 1348412) should I be adding to activity stream context as well?
18:08firebot NEW, Add a menu for setting the homepage
18:17ahuntmkaply: we might want to run it past UX if its going to be enabled for everyone (and well also need an icon for any addition to AS context menu)
18:17mkaplyahunt: I thought Android didn&#39;t do icon menus anymore?
18:19ahuntmkaply: - at least Google have started doing that with their apps (which is why were doing it in AS)
18:19ahunt(Thats a somewhat outdated screenshot, thats all been rearranged since then)
18:20mkaplyThat&#39;s frustrating. There used to be icons and FF looked so much better, but Google said to remove them in material design
18:21mkaplyI&#39;m thinking I just leave this off activity stream for now (because the icon will be the same - Home)
18:21mkaplyBut you&#39;re saying I should get a UX review for any Fennec menu changes?
18:24sebastianmkaply: Yes, please, just flag antlam so that he is aware of changes to the UX. For AS that&#39;s also bryan bell.
18:26sebastianahunt: I wish github had a button &quot;update branch and merge automatically if all checks pass - and redo that until you can merge&quot; :)
18:27ahuntSounds a bit like reviewboard and autoland?
18:28ahuntI guess we could make checks optional (or allow landing if at least one previous build passed checks) - but then you risk landing broken code again
18:28sebastianI guess I made it worse by only allowing rebased commits. But I just really like a linear history ;)
18:28JacobWIn Settings on Firefox-iOS should &quot;Open With&quot; be something like &quot;Open Email Links With&quot;? Or is it going to be opened up to include more than just email links in the future?
18:30ahuntI think the current system works well though (clean commit history, and prevents broken stuff from landing at all) - maybe we just need to coordinate landing sessions to avoid requiring repeated updates?
18:33slerouxJacobW: I think the intention of &#39;Open With&#39; was to make it flexible for when we want to add additional open things
18:52sebastiannjpark: I&#39;m currently waiting 27 minutes for a build and it&#39;s still running:
18:53njparksebastian: hmmm, this could be the issue with google testcloud.. lemme check
18:54njparksebastian: yeah looks like the there was some kind of connection issue, I see them occasionally
18:54njparkmaybe I&#39;ll make it run on only one API like, 7.1
18:55njparksebastian: also, you can modify the test config for your branch only as well, but I&#39;ll change it so it runs one test set less
18:55sebastiannjpark: I wonder whether we should just run them for &quot;master&quot;. This slows down merging pull requests. Although on the other hand it&#39;s good to find issues before merging of course.
18:55njparksebastian: i&#39;m good with either way, I can make it that way too
18:57njparksebastian: yeah video shows the test only took 1:50, somehow it was frozen while transferring the results back
19:04njparksebastian: oh wait, now it&#39;s showing up, the trashcantest failed, wonder why...
19:05njparkhmm, it couldn&#39;t find the erase button, maybe the id is changed?
19:05sebastiannjpark: Does the test depend on a fixed view hierachy? This patch wrapped the layout in another FrameLayout
19:05sebastianbut no id changes
19:07njparksebastian: aha, i should&#39;ve simplified my locator, I was checking for the id=erase which has a parent = main_content which has parent = container
19:08njparkwas paranoid that there may be some other id=erase somewhere
19:08njparkno prob, I can file a quick fix after you merge this to master, test doesn&#39;t affect build status
19:11njparksebastian: to speed things up, maybe I&#39;ll enable UITests on master only, i guess it&#39;s pretty easy to revert if there&#39;s something wrong with the code
19:11sebastiannjpark: okay :)
19:12sebastianIf this would be two separate tasks in GitHub then we could just make only one of them mandatory for merging.
19:15njparksebastian: i turned it off by default, and enabled only on master, you can choose to add UITests to your personal branch if you wish, theoretically I&#39;ll check the test results and let you know, although you&#39;re encouraged to see when there is a test failure :)
19:17mcomellaHow can I get the source for a particular build ID?
19:17mcomellakbrosnan, njpark: ^ Do you happen to know?
19:18kbrosnanmcomella: about:buildconfig assuming it is a mozilla build
19:18njparkmcomella: do you mean for firefox for android?
19:18mcomellanjpark: Yes
19:19sebastianmcomella: Welcome back! Do you need a starter bug? :D
19:19kbrosnanif you built it yourself the ids will only map to your local build
19:19kbrosnaner your local source tree
19:19mcomellakbrosnan: So if I have the build ID (and not the build), should I download the build? If so, where should I DL it? I&#39;m trying to find the source from a crash stats thing
19:19mcomellasebastian: ;) Don&#39;t worry, I&#39;m just a reviewer
19:20kbrosnanmcomella: oh, there might be a tool to map a build id to a checkin. #build might know. however in most cases you should be able to just search for the code using dxr or searchfox
19:21njparksebastian: just sent you a review request
19:22kbrosnan ?
19:22firebotBug 1317191 NEW, Crash in java.lang.IllegalStateException: Expected session to be started before session end is calle
19:22mcomellakbrosnan: Yeah, that&#39;s the one
19:22mcomellakbrosnan: Build ID 20170303025843 had a ton of crashes and I wanted to see what had changed between that and the others
19:23kbrosnanmcomella: it might just be the id of release
19:24mcomellakbrosnan: Yeah, can I map that to the source without downloading the build?
19:24kbrosnan ?
19:25mcomellaOh, it might be the ID of release, not &quot;the release&quot;
19:25mcomellaI see, the crashes increased because it was a larger population
19:25mcomellaOkay, ty :)
19:28njparksebastian: oh btw, I agreed with delphine that I&#39;ll produce screenshots every mon/thurs night, was there any changes regarding l10n since last friday night? (I don&#39;t see it, so just checking)
19:28delphinenjpark: you read my mind. I was *just* about to ping you about this :)
19:28njparkdelphine: i do remember certain things :)
19:29sebastiannjpark: Nope, we want to fix something in the importer (if you look at the about screen, we do not apply the html formatting) - i&#39;ll do another import tomorrow.
19:30njparksebastian: aha, thanks! okay, thursday night it is then
19:30delphinecool :)
19:31njparkdelphine: or does it work as &#39;no monday, so you should do it on tuesday&#39; thing? either way is fine
19:31delphinenjpark: good question :)
19:31njparki guess that means yes, suuure. :)
19:31delphinenjpark: hahahahaha
19:31delphineno so I&#39;m more thinking
19:32delphinethat it doesn&#39;t mean that there is no string change, that localizers aren&#39;t changing screens on their side
19:32delphineso it does make sense for them to see screenshots today
19:32delphinebecause I&#39;m thinking some will have worked on their localization over the weekend
19:32delphinenjpark: so I&#39;d just do it anyways
19:32njparkdelphine: but unless we import it, they won&#39;t see those changes though, no?
19:33delphinenjpark: ah yeah....
19:34delphineok well let&#39;s just do next import then
19:34njparkokay sounds good, tomorrow it is
20:11ahuntsebastian: (In case youre still around : ) ) Ive managed to fix the about string import issue (that was a dumb one - the previous code could only handle cases where there was only one root html tag in the string, but about has multiple <p>s in the root) - Ill prepare a new string import as part of that PR
20:14grishaantlam: will you have some time today to chat about Bug 1346751 ?
20:14firebot NEW, Incoming tab isn&#39;t opened when notification is dismissed
20:19sebastianahunt: Nice! Thanks!
20:43sebastianahunt: Approved
20:43sebastianahunt: This string looks broken though:
20:43sebastian(ar, line 33)
20:49ahuntHmm, I dont even see ar listed under
21:48jwattanyone with a fennec build handy care to check if this bug still occurs?
21:48firebotBug 884787 NEW, crashes Fennec
21:50JanHjwatt: the frame rate is not super spectacular, but so far no crash
21:50JanHjwatt: on a 2GB phone
21:55nalexandergrisha: holy bugmail, batman!
21:55grishanalexander: more to come, i&#39;m afraid
21:55grisha3 years of bugs to clear up :)
22:33JacobWsleroux: would it make sense to change it to &quot;open email with&quot; for now and change it to something else when more things are in there?
22:37ahuntbnicholson: what device did you get the autocomplete issue on? ( ) - and does it work once you have 2 letters?
22:38bnicholsonyeah, the bug is for 1 letter ;)
22:38bnicholsonI guess I should have been clearer
22:38ahunt(I cant reproduce on either an N5/Android 6, or N6P/Android 7.1 - so I wonder whats causing that for you)
22:39bnicholsonoh I misread your question -- yeah, it does work after I enter the second letter
22:39bnicholsonor if I hit backspace and type the same letter again
22:39bnicholsonI&#39;m on a Moto X Pure Edition
22:40bnicholsonusing Gboard, that might be it
22:40ahuntbnicholson: do you get the same issue in fennec (it looks like were reusing most of the autocomplete code from there)
22:41bnicholsonyep, same bug
22:42ahuntI also have Gboard (that seems to be the default on 7.1), so that probably isnt the culprit either
22:42bnicholsonhm, strange
23:48grisharnewman: do you have any concerns about landing Bug 1346438 now that it&#39;s been r+&#39;d by n.alexander?
23:48firebot ASSIGNED, Use &#39;X-If-Unmodified-Since&#39; header when uploading meta/global and crypto/keys to avoid race conditio
23:50rnewmangrisha: he&#39;s more thorough than me, so no
23:52grishaheh, cool
23:53grisharnewman: btw very good article on complexity, thank you for sharing
23:54grishai like that he elaborated on their &quot;do the hard thing&quot; strategy
23:57grishaand the clarity around importance of reducing &quot;accidental&quot; complexity is insightful
21 Mar 2017
No messages
Last message: 5 days and 21 hours ago