Age | Commit message (Collapse) | Author |
|
UI tests used to install APKs programmatically during tests,
which is not stable as observed due to security issues.
Besides, that way also deviates from manual test cases.
Instead, we will install APKs before tests and deprecate
the programmatical method to install an apk in test runtime.
Change-Id: I059265ffcd9da7ac949506cf7467779b8503cf46
|
|
|
|
UiDevice adds executeShellCommand method to run shell commands.
We use the API instead of ShellUtil fixture in the framework,
and ShellUtil fixture will be deprecated once we migrate all
usage.
Change-Id: Ib90afb585c482def56038e8d2a88d08b4bf2be80
|
|
|
|
Change-Id: Ic3de4a7c6cca402733b468897453b7ad6129d93b
|
|
|
|
VpnTest takes a bit more than 60 seconds sometimes.
Change-Id: I942bfc4e90c82487ae0f1c5d5449b08b54a3e492
|
|
Change-Id: Iedd49c4f93b13df7ad59de91df38292d76581247
|
|
- sdcard.img is required to run ui tests
Change-Id: Ia10ae0a9a1082331aa6013d31f05f1c47f2138b1
|
|
Change-Id: I5f35d8d68d6440eec620b6b6929f5b3c8027d008
|
|
... based on buildbot results on x86, api23, default system images.
Change-Id: I878903976db9aeb86ab8f8f4a962f456195f0ed0
|
|
Change-Id: I7b4dd9f8e017582bae16bb298cc379d85421eb7b
|
|
Change-Id: I1823d8fbfe0430863d0b9766eb5b79d742ffc0f7
|
|
Change-Id: I434def8ffba823c57960036faddd53279c85cde2
|
|
|
|
Change-Id: I4f28a8b82bccfaa279280e118f3bee93aaca4864
|
|
Change-Id: I08614b2a195e43a88520848b8d584ced25bde243
|
|
Change-Id: I020f55f71b14626f41fe568359af9ef97d720b69
|
|
- The purpose is to save disk space
- Always clean up system image directory and only download before test step
Change-Id: If01f9435a47ccab1d4bdbea7f6a94ed72217e7b4
|
|
|
|
Also refactor some string and remove dead code in test_ui.
Change-Id: I88472223812c1090957a3ad1235f0851847f918d
|
|
- Create ui scheduler that only watches for emu-master-dev.
- Add one linux builder for ui test.
- This CL starts only adds single triggering branch and single linux builder,
we will add more branches and builders when it's stable.
Change-Id: I24db4b2a4c29b05d53863b0830a52e30b5b3ebfc
|
|
Change-Id: I08eb8c5ff4fc637825dff7a22a98f1800c15a5b7
|
|
Change-Id: I47e14d521044da9859022c7e9c1be32d22a04fe7
|
|
and 24-hour format."
|
|
|
|
Change-Id: I8fb6db737436fbf421e5552bc2737cf72dd4f1ca
|
|
24-hour format.
Change-Id: Ic303c092fb004d6dafe27f7c1c7f35ce5301b044
|
|
Those results used only 3 runs of CTS tests, which is not enough
data to conclude 'gotBroken' or 'gotFixed'. This CL changes
those two designations to 'flaky', which is the best conclusion
with such a small sample.
Summary of the information in this file:
751 Bad
347 Flaky
0 gotBroken
------
1098 Total bad
0 gotFixed
119881 Good
------
120979 Tests that were run
6036 Not executed
------
127015 Total tests
Change-Id: I5def124e28d07d5f29913d5cbad551e508b6fa3e
|
|
My previous CL
https://android-review.googlesource.com/#/c/241410/
introduced code to exclude certain CTS tests. While I tested the
translation locally before submitting, there was an error that was only
revealed last night, when the code ran on the BuildBot machine: I got
the directory in which the CTS plan files live wrong ("plan" -->
"plans").
This should fix that.
Change-Id: I09e8c0281fcdc8fc7320eb9691e57f640758b57d
|
|
|
|
Until a few days ago, CTS tests were not running to completion.
These results are based on the few tests that were run after
that issue was resolved:
2016-06-19_10-46-23 build_0-rev_2988904
2016-06-20_17-31-08 build_3-rev_2991947
2016-06-21_19-05-57 build_6-rev_2997187
Summary of the information in this file:
751 Bad
127 Flaky
63 gotBroken
------
941 Total bad
157 gotFixed
119881 Good
------
120979 Tests that were run
6036 Not executed
------
127015 Total tests
I have changed the format of this text file. The changes should make
it easier to compare revisions *in the future*. (But the changes make
it very difficult to compare with versions prior to this one.)
The format changes are:
1) Put the flakinessResult on the same line as the fullName.
This helps keep 'diff' tools sync'd.
2) Alphabetize using sqlite3's sort order. (Rather than picking
some arbitrary sort tool each time.)
Change-Id: Id3507787183c4d640e51f55c295fd92adb823a33
|
|
Some CTS tests have been found to have persistent negative side-effects:
they may either crash the emulator, or leave it in a state of diminished
performance. In order to get useful information from the rest of the
test run, this CL allows us to specify tests to exclude.
The set of tests to exclude is expressed as a data structure in the
test_cts.py file. The mechanism we have available is that a CTS test
plan file specfies a set of test packages in the plan, but also allows
you to specify individual tests to be excluded. The data structure is a
dictionary mappin plan names (e.g., "CTS") to another dictionary, this
one mapping test package names (e.g., "android.calllog") to a sequence
of excluded tests in that package (e.g.,
"android.calllog.cts.CallLogBackupTest#testSingleCallBackup").
When we're asked to execute a test plan, we check whether that plan has
exclusions in this data structure. If so, then we create a new test
plan, derived from the .xml for the original. We assume the original
plan has no exclusions. We copy line by line; if the line specifies a
test package with exclusions, we insert those exclusions in the output.
We direct cts-tradefed to execute the newly-created test plan.
I've tested (via the test code now commented out in the file) that the
translation code produces the desired output: for the current CTS.xml,
and the exclusions given in the CL, it produces the desired output: a
CTS-emu-stable.xml file that differs from the original only by the added
exclusions. If there are suggestions on how to test further, I would
welcome them.
Note: I also eliminated what seemed like a redundant second call to
create_test_case_for_avds(). If I'm missing a reason this might be
necessary, let me know. Also, I'm not completely fluent in python, so
feel free to critique coding style decisions.
Change-Id: I57c0154439760fac4019adf5aeaa202031cb49e4
|
|
prior to test running."
|
|
test running.
Change-Id: Ic91053485f7fdcb190c6b1fc7115baabcee55ba4
|
|
- Sometimes it takes longer than 30 minutes to complete update,
especially when there are updates in system images.
Change-Id: I679a3c5e7f4b78fae26dde906178425ffe8e28f8
|
|
|
|
Change-Id: Ie071cf782bd7b5305c12fd4bf5dee89455f57703
|
|
This reverts commit 81435e862497e36cc76ef17814764d97def66bd5.
|
|
use "android-24", "android-N" has become obsolete.
Change-Id: Idcd682dcad8a002e6b4c5c1beef01fc32abdffc9
|
|
on Windows builders
Change-Id: I3fd1606e0db263fdec1e7723affc04a60448536c
|
|
- Observed some weird issue after deleting tmp files,
Need to revert this change after figure out the reason
Change-Id: Ic4703ea5e824f84c7c51c6315a897e8002771e26
|
|
- Print out from adb changed since adb version 1.0.36
Calculate the speed by script, to remove dependency on adb
Change-Id: Ib08d103a5f0c8672f1045ff3185576bfe2953886
|
|
on Ubuntu 15.04
Change-Id: I802ef17d79e706e589f50191a5b4fe5616f802f5
|
|
- When running local debug, we sometimes need to make a copy of
CTS AVD, and append suffix to the orignal name. this change avoids
running CTS on such AVDs
Change-Id: I887605cd93c710792c070263d0952791f5694a45
|
|
Change-Id: I25afaa7990f0f45b68b98b178fe6290b5dd64ac1
|
|
Change-Id: I26d27a00e463b20d068e72f9540d1b225b4b0a5f
|
|
Organzie Res class into three categories for better management.
Refactor existing resource-id identifiers to force the rule
that all resource-ids must be acquired from Res class.
Change-Id: I4804f92f0f84ba70e3c25f066b743a56606c19d2
|
|
Verifies Set date and Set time fields are editable.
Change-Id: I9481f609bfbf16861a45e67eb2c083b55b8851e5
|
|
|