A quick fix for Issue 23745....

A quick fix for Issue 23745.
This issue is caused by my another mistake that I forgot changing test_shell to apply my input-code refactoring. (Since my refactoring, EventSendingController::keyDown() doesn't need platform-dependent code.)
This change just removes the platform-dependent code in EventSendingController::keyDown() and updates 'test_expectations.txt'.

BUG=23745
TEST=run-webkit-tests.sh LayoutTests/fast/events/node-event-anchor-lock.html

[Avi (use Gerrit), 2009-10-06 13:41:22]

LG

CCing Nate since IIRC this was his test

[Evan Martin, 2009-10-06 15:46:48]

LGTM

[Hironori Bono, 2009-10-07 11:09:08]

Avi and Evan,

Thank you so much for your quick reviews.
Today, I notice this change probably makes another layout test fail on Mac, and
I have been fixing this issue. I will update this change when I fix this issue.

Regards,

Hironori Bono

[Hironori Bono, 2009-10-21 10:19:34]

Avi and Evan,

Sorry for my slow update.
I have finally fixed the regressions on Mac caused by my previous change. Some
layout tests checks the platform and send option-key events on Mac. (On the
other and, these tests send control-key events on Linux and Windows.) 
Since my change also removed "#if defined(OS_WIN)" macros from the code where we
set "event_down.isSystemKey" to true, the above option-key events are ignored by
our EditorClientImpl class and caused the regressions.
As a workaround, I added code that resets "event_down.isSystemKey" for Mac. (We
need to set "event_char.isSystemKey" and "event_up.isSystemKey" to true for
access keys.)

Would it be possible to review the updated change?

Regards,

Hironori Bono

[Evan Martin, 2009-10-21 15:12:08]

darin: this is just the sort of thing we were discussing in that thread;
hironori here made the platforms unified but then has to special-case OS X to
make a layout test pass on OS X because the test expects *different* behavior on
OS X.

The change looks fine to me (fewer ifdefs is always good) but I am concerned
about the long term impact of this.

[Avi (use Gerrit), 2009-10-21 15:19:15]

LG

http://codereview.chromium.org/258032/diff/15001/16002
File webkit/tools/test_shell/event_sending_controller.cc (right):

http://codereview.chromium.org/258032/diff/15001/16002#newcode518
Line 518: event_down.isSystemKey = false;
I thought the system_key param on events was always false on the Mac...