mozilla :: #mobile

14 Mar 2017
07:13xandraI am so glad focus is around now
07:13xandraNames iOS usable
07:13xandraOne day Apple will let everyone write their own browser if they so choose but this is great for now
09:17ahuntsebastian: Ive managed to make CDATA-less po files work:
09:18ahuntIm not sure I really like that solution, since it breaks having html formatting for normal strings (but we dont actually use that in focus) - but it means that we can use CDATA in the english string, the po files have normal html, and the translated strings.xml get CDATA again
09:18ahunt(I should expand my html detection to look for things other than <ul>, since the about page might not have that, but will have other html tags)
09:21sebastianahunt: This sounds good for now! With the time constraints I think we should just go with that and then fix things if they come up in the future. We probably won&#39;t have HTML formatting in normal strings in v1.
09:22sebastianAfter the release there might be more time to look at better tools etc.
09:23sebastianAt every company I have been translating android apps was a pain. It&#39;s astonishing.
09:23sebastianI actually wonder if it would be easier to teach pontoon and co. how to understand Android XML files ;)
09:41walkingiceIf I changed dtd file (android_string.dtd), should I NI anyone in bugzilla to inform the modification?
09:46ahuntwalkingice: usually no - but you might need to change the string id (eg foo->foo1), to make sure translators know that it changed.
09:46ahunt(I&#39;m not 100% sure that&#39;s needed nowadays, can anyone confirm?)
09:47walkingiceahunt: thanks! I inserted new string, so I think no need to change string id
09:50ahuntOh yeah - additions get handled &quot;automatically&quot;!
10:49sebastianahunt: I wonder whether we shouldn&#39;t add FLAG_SECURE for debug/dev builds. Just wanted create screenshots/videos for antlam to review and it&#39;s now a pita. ;)
10:59ahuntsebastian: Maybe we should just make it user configurable? (Could be helpful if people want to review/blog about/etc focus too)
11:00sebastianahunt: I think we decided against a user visible setting in v1
11:00sebastianahunt: But yeah, I&#39;m also concerned about blogs/news. We should make sure to have a good set of screenshots for them ;)
12:11sebastianahunt: <-- is this one ready?
12:27flodsebastian: I wonder, are you still the right person to review this?
12:27firebotBug 1300201 NEW, Move list of Firefox for Android search plugins to JSON
12:27flod(lost track of who does what in the mobile team these days)
12:28sebastianYeah, I can do that. In fact I am currently working on getting the search engines into Focus. ;)
12:28flodinteresting. Asking since that&#39;s the first of a series of bugs to fix Fennec, but then it needs to move to iOS for example
12:29Pikefirst, we take Manhattan
12:30ahuntsebastian: yeah I think it&#39;s ready now!
12:31ahunt(I might refactor the html processing a bit, but that can be part of the about page bug)
12:32sebastianahunt: and is this really the correct string? <string name=&quot;error_malformedURI_message&quot;>.... web addresses....</string>
12:32ahuntSigh, I&#39;m sure I&#39;ve replaced that 50 times but forgotten to actually commit it...
12:34flodI&#39;ve stared at the string for a while yesterday
12:35flodwhy 4 dots? why a space after the first, but not the second dot invasion?
12:40ahuntsebastian: Ive updated it!
12:40ahunt(Plus I added the commit which moves the html loader into a util class since Ill need it for the about page, seeing as I had to repush)
12:53ahuntflod: sorry I missed your comments, thats in:
12:54flodnp, it happens (didn&#39;t realize it was ready for merge)
12:54ahuntThe &quot;Secure Connection Failed came from our Firefox for Android examples in netError.dtd (whereas everything else is lowercase), I wonder if we want to fix that?
12:55flodsee also line 160/164 (all code on one line)
12:55flodif Fennec has clear UX guidelines, probably yes
12:56ahunt(FWIW, android2po seems to do its own thing when we convert to po, so the multiline stuff will disappear again there - but maybe we can fix that somehow.)
12:57flodI&#39;m trying to remember where the copy guidelines for Firefox live, and failing
13:00flodahunt: found it. For Android it says to follow the Android guidelines, which say that title case is evil. So yes, I think you need to fix that in Fennec.
13:02flod(I would also point to the nice &quot;Don&#39;t&quot; under the caps-locked one, but that&#39;s for another day :-))
13:04ahuntBug 1347136
13:04firebot NEW, Don&#39;t title case error page titles
13:39snorpblassey: can you invite &#39;standups&#39; to this channel
13:51snorprnewman: thanks
13:51snorppretty cool system if folks didn&#39;t know about it
14:18snorpgbrown: I think maybe I can run tests now btw
14:45njparkahunt, sebastian: do those error messages need to be localized as well?
14:46sebastianI&#39;ll do another export today
14:48njparksebastian: i see, is there an easy way to induce those errors by any chance? so far I can only induce connection error with invalid url
14:52sebastiannjpark: No, there&#39;s no shortcut to trigger a specific error. Can you file a bug? Maybe we could add a specific error url or something like that
14:52njparksebastian: will do
14:55jryansdoes fennec still use the UI telemetry system? the dashboard seems to be gone (it was under mfinkle&#39;s account)
15:04sebastianjryans: yes, it does. The data can be queries with re:dash now:
15:05jryansok, thanks!
16:12justdaveFennec Nightly sure has been crashy lately. Won&#39;t even have it running and suddenly there&#39;ll be a firefox crash report popping in front
16:14justdavethe crash report looks like it&#39;d be almost useless for debugging it, to me.
16:14justdaveNullPointerException in org.mozilla.gecko.EventDispatcher
16:18justdavelooks like something created a dead callback and then it crashed when the callback happened. Hard to know what caused it without knowing what set it to begin with. I don&#39;t know the code though, that&#39;s just guessing based on that stack trace
16:20rbarkerjustdave: Bug 1346019 ?
16:20firebot FIXED, Crash in java.lang.NullPointerException: Attempt to invoke virtual method &#39;&#39;org.mozilla.gecko.EventD
16:22justdavequite possibly. In which case the nightly I just installed moments ago may fix it :)
17:32sebastiandelphine: flod: A new pile of strings :)
17:35delphineaha, looking :)
17:38flodwow, it already has merge conflicts
17:39flodsome of the multiline strings look pretty confusing unfortunately (when the \n is in the middle of the sentence)
17:39snorpjchen: nsresult JNI mode should catch exceptions right?
17:39snorpjchen: and not abort?
17:41delphineflod: ha didn&#39;t realize that
17:42snorpI&#39;m gonna make it not throw, but still, not sure why we get a bad reference
17:42snorpwhich presumably causes the VM abort
17:42sebastianflod: Those strings should be 1:1 imports from
17:42sebastianWith the same line breaks. Otherwise something went wrong. :)
17:56sebastiandelphine: flod: I pushed a new version. Does this look better?
17:57flodsebastian: the \n is added by the conversion
17:58sebastianflod: Yeah, that&#39;s where we have line breaks in our Android strings. Is this an issue?
17:58flodI&#39;m pretty sure they will show up as literal \n in Pontoon
17:59sebastianflod: But we already have a string with \n
17:59sebastianflod: key = teaser
17:59floduhm, let me check
17:59RyanVMdroeh: ping
18:00flodsebastian: interesting, it doesn&#39;t show up as \n, but as a new line
18:00flodI wasn&#39;t expecting that
18:02flodsebastian: good to merge for me at this point
18:02sebastianflod: phew! :)
18:04sebastiandelphine: flod: After this drop there&#39;s not much left. Only the &quot;About&quot; screen should add some more strings:
18:06ahunt1 string - it&#39;s essentially done (but I need to figure out the html styling for that page)
18:20droehRyanVM: pong
18:22delphinesebastian: that&#39;s good to know :D seems like we&#39;re on time for string freeze date then!
18:22RyanVMdroeh: hey, RE: bug 1305815 - I guess we can land the follow-up patch as-is on 53+, but we&#39;ll probably want a folded patch for m-r
18:22firebot REOPENED, Mechanism to exclude non-NEON versions
18:23droehRyanVM: ok, I&#39;ll put one up shortly
18:23RyanVMcool, thanks
18:23RyanVMfwiw, I had to rebase your patch for the original uplift to m-r, so you may want to use that to start
18:24droehok, will do
19:44snorpsebastian: you have any automated builds going for focus?
19:46firebotBug 1346924 NEW, Inconsistent synced tabs on Android
20:23sebastiansnorp: Yep, on buddybuild.
20:23snorpsebastian: ah, cool
20:24snorpsebastian: can we build GV versions
20:24snorpsebastian: do you run tests against it?
20:26sebastiansnorp: We are already building GV versions too. :)
20:26snorpsebastian: you&#39;re the best
20:26snorpsebastian: url me up
20:27sebastiansnorp: njpark has some UI tests but we are not running them against the gecko version.
20:27nalexandersebastian: link to Focus Android build configurations? I want to know how you&#39;re consuming GV.
20:29sebastiannalexander: For now just as a static AAR build from central:
20:29sebastiansnorp: (Not sure how much you can see?)
20:31nalexandersebastian: oh, just checked into the repo. If you want, there should be a fairly easy to produce URL for fetching that from automation. Next time you update, we can figure it out.
20:32sebastiannalexander: Oh, that&#39;s interesting! I guess we still want to somehow version it to not randomly break our Focus builds.
20:33nalexandersebastian: yes, and that should be possible; it&#39;d be encoded int the URL. At one time I produced a mini-Ivy repository in automation to make this easier, but I think it&#39;s just AAR files now.
20:33sebastianAh, perfect!
20:34nalexandersebastian: might be tricky with the Task Cluster index redirection. I&#39;d have to think about it.
20:34snorpsebastian: hmm, no way to publish buildbuddy artifacts unless you have are the repo owner?
20:34nalexandersebastian: hmm, I see we&#39;re not versioning the GV AAR yet, although we do upload .pom. There&#39;d be work to support this.
20:35nalexandersebastian: file when you get tired of checking in an AAR.
20:37nalexandergrisha: got a moment?
20:38sebastiansebastian: Afaik you can only deploy to testers (email with download link) or Google play
20:38sebastiannalexander: Okay! Right now the gecko version only exists because I wanted to play with it.
21:41mkaplyI&#39;m seeing the &quot;Detected problems with app native libraries: unauthorized access to on Android N built on nightly.
21:41mkaplyI thought we&#39;d resovled this?
21:43nalexandermkaply: snorp and jchen are your contacts, I expect.
21:51ahuntmkaply: are you on a normal N release, or the beta program? IIRC google said they only show the toast if youre on the beta program, but if thats not the case...
21:52ahunt(Or am I confusing this with something else?)
21:52mkaplyahunt: Beta. Thank you. I thought I was going crazy. Are we still going to need to fix that at some point?
21:52mkaplylibrary &quot;; (&quot;/system/lib/;) needed or dlopened by &quot;/data/app/org.mozilla.fennec_michaelkaply-2/lib/arm/; is not accessible for the namespace &quot;classloader-namespace&quot; - the access is temporarily granted as a workaround for http://b/26394120, note that the access will be removed in future releases of Android.
21:54ahuntmkaply: not too sure, snorp would know that : ) (IIRC its something to do with support for older android releases)
21:59esawinmkaply: bug 1316934
21:59firebot NEW, Illegal access to system library
22:00mkaplyesawin: Tx
23:03snorpyeah, that.
23:40Caspy7someone here is reporting several issues on Android
23:40Caspy7dunno if anyone wants to have a look, give some insight
15 Mar 2017
No messages
Last message: 9 days and 7 hours ago