Broken out from bug 1016371 comment 12. In B2G jobs, the logcat is dumped to emulator-5554.log - which is then cat'd to the end of the test logs: http://mxr.mozilla.org/build-central/source/mozharness/scripts/marionette.py#486 http://mxr.mozilla.org/build-central/source/mozharness/scripts/b2g_emulator_unittest.py#398 ...and looks like: http://mozilla-releng-blobs.s3.amazonaws.com/blobs/mozilla-inbound/sha512/9b03fded118afb88b46b278daaa50e9af513883c959aabe34bdbdd3f52f1ccce3067bf9ab9733cf4363d8e76cc52e7c39cd5284c4a4e5e571687ae952abc1beb 06-19 08:22:50.461 35 35 I qemu-props: connected to 'boot-properties' qemud service. 06-19 08:22:50.461 35 35 I qemu-props: receiving.. 06-19 08:22:50.491 37 37 I DEBUG : debuggerd: Jun 18 2014 22:50:15 ...which doesn't match that used elsewhere, including that dumped to the similar logcat-emulator-5554.log by Android 2.3 emulator test runs: http://mozilla-releng-blobs.s3.amazonaws.com/blobs/mozilla-inbound/sha512/4656d70117528e7d67b23a9bbbcaef17d858a34602807be4aa80f19ffc1f47a87fdcde26cee89982c2780c793a50f7a289cbbe600ba30f4f6dfcb93abb50ebd3 06-19 02:25:19.187 I/DEBUG ( 34): debuggerd: Jan 23 2014 05:01:52 06-19 02:25:19.258 D/qemud ( 41): entering main loop 06-19 02:25:19.298 I/Netd ( 33): Netd 1.0 starting This difference in format is making bug 1016371 and bug 1016251 harder than they need be, since we're trying to filter out multiple logcat formats (whilst still trying to avoid false positives).
What is confusing me however, is that the only references I can find to 'emulator-5554' or '5554' are as 'device_id' and 'emulator_port' in: http://mxr.mozilla.org/build-central/source/mozharness/configs/android/androidarm.py#43 http://mxr.mozilla.org/build-central/source/mozharness/configs/android/androidx86.py#41 A search for those: http://mxr.mozilla.org/build-central/search?string=(emulator_port|device_id)®exp=on&tree=build-central ...only finds a hit in android_emulator_unittest.py: http://mxr.mozilla.org/build-central/source/mozharness/scripts/android_emulator_unittest.py#567 However android_emulator_unittest.py is only for Android tests on Android 2.3 emulator? Looking for other places where we invoke logcat, gives: http://mxr.mozilla.org/mozilla-central/source/testing/mozbase/mozdevice/mozdevice/devicemanager.py#127 http://mxr.mozilla.org/build-central/source/mozharness/mozharness/mozilla/testing/device.py#582 http://mxr.mozilla.org/build-central/source/mozharness/scripts/android_panda.py#328 ...but none of those seem relevant.
Jonathan, I'm trying to work out where the logcat call is located, that populates the emulator-5554.log file on B2G test runs, since the output there is in the wrong format compared to Android. I don't suppose you know? (My searches for 'emulator-5554' and '5554' which lead to 'device_id' and 'emulator_port' didn't get me anywhere) Thanks! :-)
It's currently here: http://mxr.mozilla.org/mozilla-central/source/testing/marionette/client/marionette/emulator.py#496 Bug ahal's patch in bug 997244 will be moving that.
Thank you :-) http://mxr.mozilla.org/mozilla-central/source/testing/marionette/client/marionette/emulator.py#502 502 cmd = [self.adb, '-s', 'emulator-%d' % self.port, 'logcat', '-v', 'threadtime'] We'll need to switch threadtime to time, but I'll wait until bug 997244 has landed.
Depends on: 997244
Created attachment 8450268 [details] [diff] [review] Mozrunner should use logcat format 'time' not 'threadtime' This switches the logcat call in mozrunner to use 'time' format, to be consistent with what we do elsewhere in automation.
Attachment #8450268 - Flags: review?(ahalberstadt)
Attachment #8450268 - Flags: review?(ahalberstadt) → review+
Thank you :-) remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/e5c7ac63a017
Status: NEW → RESOLVED
Last Resolved: 4 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla33
You need to log in before you can comment on or make changes to this bug.