These manifests are special in that they don't package their test files
into the test package. Each test listed in an instrumentation manifest
serves as an identifier rather than a file.
--HG--
extra : rebase_source : 0321528a2dc380e57b824746efbcf61d295204a9
========
3cd96669bc
Author: Nick Alexander <nalexander@mozilla.com>
Date: Mon Aug 18 17:43:24 2014 -0700
Bug 1026005 - Review comments.
========
5aefe1f716
Author: Nick Alexander <nalexander@mozilla.com>
Date: Mon Jun 16 11:16:16 2014 -0700
Bug 1026005 - Part 2: Add AccountLoader.
========
61bc36bd7d
Author: Nick Alexander <nalexander@mozilla.com>
Date: Mon Jun 16 11:12:37 2014 -0700
Bug 1026005 - Part 1: Add and broadcast ACCOUNT_STATE_CHANGED_ACTION.
========
9d80df0aa7
Author: Nick Alexander <nalexander@mozilla.com>
Date: Mon Aug 18 17:42:11 2014 -0700
Bug 1026005 - Pre: Allow testing with |mvn integration-test|.
This avoids errors like:
W dalvikvm(24082) Class resolved by unexpected DEX: Lorg/mozilla/gecko/fxa/AccountLoader;(0x42c13358):0x75440000 ref [Landroid/support/v4/content/AsyncTaskLoader;] Landroid/support/v4/content/AsyncTaskLoader;(0x42c13358):0x74a71000
W dalvikvm(24082) (Lorg/mozilla/gecko/fxa/AccountLoader; had used a different Landroid/support/v4/content/AsyncTaskLoader; during pre-verification)
W dalvikvm(24082) Unable to resolve superclass of Lorg/mozilla/gecko/fxa/AccountLoader; (247)
W dalvikvm(24082) Link of class 'Lorg/mozilla/gecko/fxa/AccountLoader;' failed
E dalvikvm(24082) Could not find class 'org.mozilla.gecko.fxa.AccountLoader', referenced from method org.mozilla.gecko.background.fxa.TestAccountLoader.testInitialLoad
W dalvikvm(24082) VFY: unable to resolve new-instance 1299 (Lorg/mozilla/gecko/fxa/AccountLoader;) in Lorg/mozilla/gecko/background/fxa/TestAccountLoader;
D dalvikvm(24082) VFY: replacing opcode 0x22 at 0x0005
W dalvikvm(24082) Class resolved by unexpected DEX: Lorg/mozilla/gecko/fxa/AccountLoader;(0x42c13358):0x75440000 ref [Landroid/support/v4/content/AsyncTaskLoader;] Landroid/support/v4/content/AsyncTaskLoader;(0x42c13358):0x74a71000
W dalvikvm(24082) (Lorg/mozilla/gecko/fxa/AccountLoader; had used a different Landroid/support/v4/content/AsyncTaskLoader; during pre-verification)
W dalvikvm(24082) Unable to resolve superclass of Lorg/mozilla/gecko/fxa/AccountLoader; (247)
W dalvikvm(24082) Link of class 'Lorg/mozilla/gecko/fxa/AccountLoader;' failed
D dalvikvm(24082) DexOpt: unable to opt direct call 0x1df8 at 0x07 in Lorg/mozilla/gecko/background/fxa/TestAccountLoader;.testInitialLoad
I TestRunner(24082) started: testInitialLoad(org.mozilla.gecko.background.fxa.TestAccountLoader)
I TestRunner(24082) failed: testInitialLoad(org.mozilla.gecko.background.fxa.TestAccountLoader)
I TestRunner(24082) ----- begin exception -----
I TestRunner(24082)
I TestRunner(24082) java.lang.NoClassDefFoundError: org.mozilla.gecko.fxa.AccountLoader
I TestRunner(24082) at org.mozilla.gecko.background.fxa.TestAccountLoader.testInitialLoad(TestAccountLoader.java:145)
I TestRunner(24082) at java.lang.reflect.Method.invokeNative(Native Method)
I TestRunner(24082) at java.lang.reflect.Method.invoke(Method.java:525)
I TestRunner(24082) at android.test.InstrumentationTestCase.runMethod(InstrumentationTestCase.java:214)
I TestRunner(24082) at android.test.InstrumentationTestCase.runTest(InstrumentationTestCase.java:199)
I TestRunner(24082) at android.test.ActivityInstrumentationTestCase2.runTest(ActivityInstrumentationTestCase2.java:192)
I TestRunner(24082) at junit.framework.TestCase.runBare(TestCase.java:134)
I TestRunner(24082) at junit.framework.TestResult$1.protect(TestResult.java:115)
I TestRunner(24082) at junit.framework.TestResult.runProtected(TestResult.java:133)
I TestRunner(24082) at junit.framework.TestResult.run(TestResult.java:118)
I TestRunner(24082) at junit.framework.TestCase.run(TestCase.java:124)
I TestRunner(24082) at android.test.AndroidTestRunner.runTest(AndroidTestRunner.java:191)
I TestRunner(24082) at android.test.AndroidTestRunner.runTest(AndroidTestRunner.java:176)
I TestRunner(24082) at android.test.InstrumentationTestRunner.onStart(InstrumentationTestRunner.java:554)
I TestRunner(24082) at android.app.Instrumentation$InstrumentationThread.run(Instrumentation.java:1738)
I TestRunner(24082) ----- end exception -----
I TestRunner(24082) finished: testInitialLoad(org.mozilla.gecko.background.fxa.TestAccountLoader)
========
ce21e7f19a
Author: Nick Alexander <nalexander@mozilla.com>
Date: Wed Jun 4 14:15:09 2014 -0700
Bug 1019793 - Always upload in TestHealthReportUploadService.
========
e4c7393c6f
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Jun 3 15:47:33 2014 -0700
Bug 1019793 - Don't create storage in TestHealthReportPruneService.
Creating real storage requires a valid Context, and in this test, we
only have MockContext.
========
872d926e32
Author: Nick Alexander <nalexander@mozilla.com>
Date: Wed Jun 4 14:21:17 2014 -0700
Bug 993134 - Part 5: Work around Android intermittent issue.
========
c76bb53901
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Jun 3 17:02:15 2014 -0700
Bug 993134 - Part 4: Make fxa.TestAccountPickler robust.
This tested extracting data from an AndroidFxAccount before, and another
after, backed by different Android Accounts. Unfortunately, one Account
replaced the other, but had the same name, leading to a race
condition (between reading and deleting the first account). To address
this, we extract all the data before deleting, and then compare it to
the data from the newly pickled account.
========
bd32978f7b
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Jun 3 15:34:09 2014 -0700
Bug 993134 - Part 3: Use AndroidSyncTestCaseWithAccounts to make fxa.TestAccountPickler when accounts exist.
========
ec7d910187
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Jun 3 15:33:29 2014 -0700
Bug 993134 - Part 2: Extract AndroidSyncTestCaseWithAccounts.
========
6563ad7bd8
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Jun 3 15:30:15 2014 -0700
Bug 993134 - Part 1: Make getApplicationContext return getTargetContext.
I was finding that repeated instrumentation-test runs were returning a
null application context, indicating a race between test start-up and
target instantiation. This works well for me on two test devices.
This is the expedient thing to do. The right thing to do is to
deprecate Sync's clients database in favour of Fennec's clients table,
but that's a good chunk of work for a small gain.
========
8c7b253142
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Jun 3 15:48:40 2014 -0700
Bug 788688 - Review comment: Include timestamp in setClientName.
========
b53b9092c2
Author: Nick Alexander <nalexander@mozilla.com>
Date: Thu May 29 16:05:25 2014 -0700
Bug 788688 - Post: PII client data to ease debugging.
========
fd59f3c984
Author: Nick Alexander <nalexander@mozilla.com>
Date: Wed Jun 4 15:01:12 2014 -0700
Bug 788688 - Part 4: Work around Android DialogPreference caching bug.
========
87d10bc16a
Author: Nick Alexander <nalexander@mozilla.com>
Date: Thu May 29 14:44:24 2014 -0700
Bug 788688 - Part 3: Add "Device name" pref to Status activity.
In the edge case where what the user has entered (empty text) and what
is persisted (default client name) differ, Android does not update the
contents of the dialog's EditText correctly. Removing and re-creating
all preferences is the only way I found to work around this; that's in
the next commit.
========
7af72f6c2f
Author: Nick Alexander <nalexander@mozilla.com>
Date: Thu May 29 16:05:59 2014 -0700
Bug 788688 - Part 2: Upload clients and tabs records when client name changes.
========
0e99eae1b5
Author: Nick Alexander <nalexander@mozilla.com>
Date: Thu May 29 15:28:36 2014 -0700
Bug 788688 - Part 1: Add setClientName with timestamp to ClientsDataDelegate.
========
1999e263db
Author: Nick Alexander <nalexander@mozilla.com>
Date: Thu May 29 14:10:40 2014 -0700
Bug 788688 - Pre: Clean some imports.
========
6ced7b05cd
Author: Nick Alexander <nalexander@mozilla.com>
Bug 996749 - Part 3: Finally kill PrefsSource.
========
b37547fbd8
Author: Nick Alexander <nalexander@mozilla.com>
Date: Tue Apr 15 10:47:11 2014 -0700
Bug 996749 - Part 2: Remove storageEndpoint from FxAccountGlobalSession constructor.
This is just a separation of concerns: let's inject our dependency as
configured as possible.
========
42e8fe761f
Author: Nick Alexander <nalexander@mozilla.com>
Date: Mon Apr 14 12:30:21 2014 -0700
Bug 996749 - Part 1: Remove debugAssertion.
This is no longer valuable, since we're not getting timestamp HAWK
errors from the server.
I also took the opportunity to move a constant into
AndroidManifest.xml.in; it removes a layer of indirection when trying to
determine what Android package gets built.
The issue is that the preprocessed files need to be referenced as
$(CURDIR)/file to have Make handle the dependencies correctly. To
allow this, this patch adds ANDROID_MANIFEST_FILE (defaulting to
AndroidManifest.xml) and uses it in the appropriate places.
This is a cautionary move: when uploaded to infrastructure,
background.apk will not look so bad; but browser.apk is likely to
confuse folks. (These files will be in the same directory as
fennec-*apk, and sort earlier than that file.)
I considered including "tests" in the name, but decided not to make a
long name even longer.
This also renames the Android package. This is purely cosmetic, since
the package has no consumers (infrastructure, mach, etc). But this looks
nicer in IDEs: the generated classes fit into the class hierarchy.
