Android test runner: Update cctest status

Android test runner: Update cctest status

Committed: https://code.google.com/p/v8/source/detail?r=12154

[haitao.feng, 2012-07-19 09:39:33]

Update cctest status.

The following cases passed for both ia32 and ARM
-test-debug/DebuggerAgent:
-test-debug/DebuggerAgentProtocolOverflowHeader:
-test-sockets/Socket:

There is only one fail: cctest/test-threads/ThreadJoinSelf. It seems that
thread.join does not stop in Android. This is the only fail when running make
android_ia.release.check. I might spend some time on it after the Mac build
done.

[Jakob Kummerow, 2012-07-19 13:26:09]

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.status
File test/cctest/cctest.status (left):

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.sta...
test/cctest/cctest.status:68: # BUG(945): Socket connect fails on ARM
This comment is misleading, the bug has more information: these tests fail when
they're run in parallel. We need to keep SKIPping them, at least on ARM (we can
run them on Android as long as we assume that the test driver runs with -j1
there, which might not be a safe assumption).

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.status
File test/cctest/cctest.status (right):

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.sta...
test/cctest/cctest.status:66: test-log/LogAccessorCallbacks: SKIP
No reason to SKIP this when $arch == arm, is there? Please put these new entries
into an Android specific section.

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.sta...
test/cctest/cctest.status:72: # BUG(355): Test crashes on ARM and IA32.
Really? Have you seen this crash? The bug is pretty inconclusive and several
years old; maybe this has long since been fixed.

[haitao.feng, 2012-07-20 02:09:20]

Added an android section including android_arm and android_ia32. The ARM section
is untouched as I do not have ARM device to test the cross compiled binary. Some
of them passed in ARM simulator mode in my Linux machine as you have mentioned.

You should really set up a buildbot for all those kinds of testing!

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.status
File test/cctest/cctest.status (left):

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.sta...
test/cctest/cctest.status:68: # BUG(945): Socket connect fails on ARM
Fix the comment and add them in Android section. Actually
test-debug/DebuggerAgent passed with -j8. But from your comments in the bug
entry, I skipped it.

On 2012/07/19 13:26:09, Jakob wrote:
> This comment is misleading, the bug has more information: these tests fail
when
> they're run in parallel. We need to keep SKIPping them, at least on ARM (we
can
> run them on Android as long as we assume that the test driver runs with -j1
> there, which might not be a safe assumption).

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.status
File test/cctest/cctest.status (right):

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.sta...
test/cctest/cctest.status:66: test-log/LogAccessorCallbacks: SKIP
On 2012/07/19 13:26:09, Jakob wrote:
> No reason to SKIP this when $arch == arm, is there? Please put these new
entries
> into an Android specific section.

Done.

https://chromiumcodereview.appspot.com/10802030/diff/1/test/cctest/cctest.sta...
test/cctest/cctest.status:72: # BUG(355): Test crashes on ARM and IA32.
Yes, it crashes as there is no /tmp in Android.

On 2012/07/19 13:26:09, Jakob wrote:
> Really? Have you seen this crash? The bug is pretty inconclusive and several
> years old; maybe this has long since been fixed.

[haitao.feng, 2012-07-20 02:11:03]

Added an android section including android_arm and android_ia32. The ARM section
is untouched as I do not have ARM device to test the cross compiled binary. Some
of them passed in ARM simulator mode in my Linux machine as you have mentioned.

You should really set up a buildbot for all those kinds of testing!

[Jakob Kummerow, 2012-07-20 10:01:07]

LGTM, landed.

One minor nit for next time: Please end comments with a full stop.