- Launch DOM checker in the subdirectory 'dom_checker/', as specified by the...

- Launch DOM checker in the subdirectory 'dom_checker/', as specified by its
  configuration.
- Use the cleaned up version of the error message that DOM checker generates.
- Remove <!-- NOP --> from test failure messages reported by DOM checker.
- Update the expected failures files as a result of these changes.
- Disable the test temporarily as this will require a Buildbot change and
  restart.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=12538

[Patrick Johnson, 2009-03-25 19:16:37]

I realized this morning that I made a mistake here.  DOM checker is configured
to run under the subdirectory 'dom_checker/', and I was running it at the
document root of the web server.

I'm fixing the issue on the Buildbot side in a partner change:
http://codereview.chromium.org/53060

Unfortunately, this change is going to break with the current parameters that
Buildbot is using, so I'll need to wait for a Buildbot restart.  This is why
I've disabled the test (temporarily).  I'll try to minimize the time the test is
actually disabled.

With this fix, there are some new failures.  Please let me know if any of them
look interesting to you.

[abarth-chromium, 2009-03-25 20:50:46]

LGTM.  Hopefully we can re-enable this soon.  :)

http://codereview.chromium.org/53064/diff/3006/3008
File chrome/test/data/dom_checker/expected_failures-file.txt (right):

http://codereview.chromium.org/53064/diff/3006/3008#newcode8
Line 8: (third-party).frames<!-- NOP -->[0] probe [value: [object global]]
Why is there this NOP here?  It probably doesn't matter, but the output would be
cleaner without it.

[Patrick Johnson, 2009-03-25 21:37:59]

http://codereview.chromium.org/53064/diff/3006/3008
File chrome/test/data/dom_checker/expected_failures-file.txt (right):

http://codereview.chromium.org/53064/diff/3006/3008#newcode8
Line 8: (third-party).frames<!-- NOP -->[0] probe [value: [object global]]
On 2009/03/25 20:50:46, abarth wrote:
> Why is there this NOP here?  It probably doesn't matter, but the output would
be
> cleaner without it.

Agreed, it's weird.  It's inserted by DOM checker, I believe it must be for HTML
formatting reasons (I'm really not sure, actually).  Removed.