811734 - on android 4.0.4/pandaboard we fail test_bug640321.html with 140+ failures

Status: RESOLVED WORKSFORME

(Core :: DOM: Editor, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

We are trying to get tests up and running on the panda boards with ice cream sandwich.  This test runs fine on Android 2.3 and the tegra reference platform, but it fails miserably on the new panda boards.  Here is an example from the mochitest.log:
5411 ERROR TEST-UNEXPECTED-FAIL | /tests/editor/libeditor/html/tests/test_bug640321.html | The width should be increased by -10 pixels - got 100, expected 90
5416 ERROR TEST-UNEXPECTED-FAIL | /tests/editor/libeditor/html/tests/test_bug640321.html | The width should be increased by -10 pixels - got 100, expected 90
5421 ERROR TEST-UNEXPECTED-FAIL | /tests/editor/libeditor/html/tests/test_bug640321.html | The width should be increased by -10 pixels - got 100, expected 90
5441 ERROR TEST-UNEXPECTED-FAIL | /tests/editor/libeditor/html/tests/test_bug640321.html | The width should be increased by 10 pixels - got 100, expected 110
5446 ERROR TEST-UNEXPECTED-FAIL | /tests/editor/libeditor/html/tests/test_bug640321.html | The width should be increased by 10 pixels - got 100, expected 110
5451 ERROR TEST-UNEXPECTED-FAIL | /tests/editor/libeditor/html/tests/test_bug640321.html | The width should be increased by 10 pixels - got 100, expected 110


I also see this in the adb logcat:
11-14 15:41:41.562 W/System.err( 2234): java.lang.IndexOutOfBoundsException: replace (4 ... 3) has end before start
11-14 15:41:41.562 W/System.err( 2234): 	at android.text.SpannableStringBuilder.checkRange(SpannableStringBuilder.java:937)
11-14 15:41:41.562 W/System.err( 2234): 	at android.text.SpannableStringBuilder.change(SpannableStringBuilder.java:271)
11-14 15:41:41.562 W/System.err( 2234): 	at android.text.SpannableStringBuilder.change(SpannableStringBuilder.java:266)
11-14 15:41:41.562 W/System.err( 2234): 	at android.text.SpannableStringBuilder.replace(SpannableStringBuilder.java:443)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoEditable.onTextChange(GeckoEditable.java:552)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.notifyIMEChange(GeckoAppShell.java:597)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.nativeRun(Native Method)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.nativeRun(Native Method)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.runGecko(GeckoAppShell.java:532)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoThread.run(GeckoThread.java:77)
11-14 15:41:41.562 W/System.err( 2234): java.lang.IllegalArgumentException: invalid selection notification range
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoEditable.onSelectionChange(GeckoEditable.java:500)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.notifyIMEChange(GeckoAppShell.java:595)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.nativeRun(Native Method)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.nativeRun(Native Method)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoAppShell.runGecko(GeckoAppShell.java:532)
11-14 15:41:41.562 W/System.err( 2234): 	at org.mozilla.gecko.GeckoThread.run(GeckoThread.java:77)


These error messages are repeated hundreds of times, I suspect for each failed check.

Comment 1

[Boris Zbarsky [:bzbarsky]]

The test uses synthetic mouse events to simulate drag of a resizer.

I would not be surprised if this totally fails on Android, and it failing might even be right.  Maybe.

Comment 2

[Olli Pettay [:smaug][[e-mail]]]

Do we even have dragservice on Android?

Comment 3

[Joel Maher ( :jmaher ) (UTC -8)]

gbrown, can you answer the above question about dragservice on Android?

Comment 4

[(no longer active)]

(In reply to comment #1)
> I would not be surprised if this totally fails on Android, and it failing might
> even be right.  Maybe.

Why is that?  This basically means that resizing stuff is broken on Android.  It would be interesting to see why this was not failing before...

Comment 5

[Boris Zbarsky [:bzbarsky]]

> Why is that?

Depends on whether resizing on Android uses mouse events, basically.

Comment 6

[Joel Maher ( :jmaher ) (UTC -8)]

I updated my build (since Monday) and these tests are green.  Also green on Cedar, and were not yesterday.  I assume something was fixed, I did verify these tests did run.