[14:45:21] mdholloway: hello. I have updated the android-emulator plugin on Jenkins :) [14:47:05] hashar: \o/ [14:47:09] thank you! [14:47:27] mdholloway: I should have forked first instead of trying to debug the whole thing :-} [14:47:33] looks like the test is running now [14:47:37] eventually I later found out you already did a bunch of tests on https://integration.wikimedia.org/ci/job/apps-android-wikipedia-periodic-test/ [14:47:45] I have updated that one with android-25 [14:47:50] and keeping ABI: x86 [14:47:57] 00:00:50.725 Installing Google APIs Intel x86 Atom System Image, Android API 25, revision 6 [14:48:04] we will see what happens. [14:48:22] that's a good sign! [14:48:29] tools/android create avd -f -a -s WVGA800 -n hudson_en-US_240_WVGA_android-25_x86 -t android-25 --abi x86 [14:48:35] yes, we definitely want to keep x86 [14:48:36] [android] Failed to run AVD creation command [14:48:39] Error: Invalid --tag default for the selected target. [14:48:40] :( [14:48:53] oh, i think i know how to fix that... [14:49:20] that would be because of the -t android-25 maybe? [14:49:59] yeah, i think it should be google_apis/android_25, but i'd have to double check [14:50:10] it is worth trying :-} [14:50:11] * mdholloway grabs other laptop [14:50:28] I am having some bathroom break / fresh air /coffee and will be back in roughly 10 mins [14:52:08] hashar: ok! i'll be in android team meetings for about a half hour coming up here at 11 (UTC-4) [14:57:40] back [14:57:43] mdholloway: back around :) [14:58:37] hashar: ok! here's the emulator creation command i have working for an API 25 emulator on another machine: [14:59:08] or could it be the patch https://gerrit.wikimedia.org/r/#/c/366253/1/src/main/java/hudson/plugins/android_emulator/SdkInstaller.java is not applied ? [14:59:09] android create avd -f -n emulator -t android-25 --abi google_apis/x86 [14:59:36] ah the abi [14:59:44] so I guess your patch needs to adjust the abi name as well [14:59:48] yep [14:59:59] or we can try changing it in the job config :D [15:00:01] i'll update after meetings [15:00:07] ok [15:00:18] it might work! [15:00:23] worth a try :) [15:00:26] mdholloway: I will be in a meeting by the time you come back but should be back around in ~4 hours ( 19:00 UTC) [15:00:37] * hashar tries [15:00:56] hashar: ok! meeting time for me now. talk to you in a bit! [15:01:02] happy meeting! [15:03:51] apparently changing the abi in the job config did the trick [15:04:01] though it is waiting for the emulator. I pasted a summary on the Gerrit java change [15:04:14] it is waiting for the device now :((( [16:31:07] mdholloway: I have reverted the job config to android-24 / x86 [16:31:20] hopefully it will spawn and find the emulator just fine [16:31:24] hashar: thanks! i was just about to try that myself ;) [16:32:52] looks good [16:33:34] mdholloway: looks like the job is running fine now [16:33:50] mdholloway: I am not there this evening after all. So I wanted to leave the job in a clean state [16:34:12] still have to figure out why android-25 can't spawn the emulator though. Maybe that can be debugged in a copy of the job [16:34:21] at least your java patch seems to work :-} [16:34:49] always a good sign :) [16:35:03] yeah, looks like more debugging is in order. thanks again for all of your help! [16:46:51] dbrant: interesting the zim search is nearly instantaneous on your device. it takes nearly 10s on my 6P [16:47:01] similar on my older galaxy s3 [16:48:01] mdholloway: you might just have some deep directory structures... [16:48:27] we might want to tighten the depth of the search.... it's currently 10, but might easily be made smaller. [16:57:53] mdholloway: we could also restrict the search to where the user is most "likely" to save a compilation, e.g. in the root directory of the storage, or the "Downloads" directory, etc. [16:58:57] yeah, it seems like a device-wide "Downloads" directory is the default (at least when invoking from Chrome; not sure if that's the case when the Download Manager is invoked from another app...) [17:09:41] mdholloway: yep, Rita is looking at it [17:41:16] I'm looking for a place to discuss a problem with the Android app. Will this do? [17:46:36] WillPittenger: you've come to the right place. what's up? [17:46:49] Thanks. [17:46:59] Can we get the desktop version of the pages? [17:47:45] I'm using a 10" tablet. The phone version limits me to one photo and none of the info boxes that commonly appear at the bottom of pages. [17:52:42] WillPittenger: there is a task in Phabricator (our work tracking software) for a more tablet-aware UI, though I can't seem to find it just this minute (dbrant?) [17:53:21] i am not sure that we'd seek to emulate the desktop view, necessarily, though that would ultimately be up to our product designers [17:53:50] It'd be the quickest option. Just change the UA. [17:54:51] I'm guessing you're using the Chrome engine behind the scenes. [17:55:22] WillPittenger: unfortunately we haven't yet devoted much time to designing a proper tablet UI for the app. It's not quite as simple as changing the UA. The way the app fetches content is pretty radically from the website. [17:55:39] *radically different [19:19:02] mdholloway: FYI https://phabricator.wikimedia.org/T152540 [20:33:01] dr0ptp4kt: we still on? [20:33:12] here i am [20:53:01] coreyfloyd: o/ hey, we were in a meeting, what's going on with T152540? [20:53:02] T152540: Migrate to HTML5 section ids - https://phabricator.wikimedia.org/T152540 [20:54:05] i see a reference to it on T170690... [20:54:06] T170690: Extract a References JSON API - https://phabricator.wikimedia.org/T170690 [21:18:44] mdholloway: has something changed in Lint within the last hour? https://gerrit.wikimedia.org/r/366304 [21:19:01] dbrant: not that i know of, and it's passing locally. [21:20:57] dbrant: this is apparently what's failing: https://integration.wikimedia.org/ci/job/apps-android-wikipedia-test/7847/artifact/app/build/reports/lint-results-alphaRelease.html#NewApi [21:21:07] (not even related to my change(?)) [21:22:08] yes, i see... but it's been passing for all previous patches [21:23:00] hmm, the only other place where isDestroyed() is used is in PageActivity.loadPage(), but that method is annotated with @TargetApi(17) [21:23:08] perhaps for that reason? [21:26:30] mdholloway: I'm thinking we can just rename isDestroyed() to something else, and just not override it. It's only used in one other place. [21:36:07] * mdholloway notices the RFC meeting [22:13:42] dbrant: a TargetApi(17) annotation seems appropriate. Just trying to figure out why it's not getting flagged locally... [22:15:53] or rather, wasn't getting flagged anywhere until CI suddenly started caring... [23:11:47] maybe an sdk issue?