Catch LookupError when getting NVDAObject from offset in virtual buffers

[LeonarddeR]

Link to issue number:

Fixes #8667, regression caused by #8479. Originally reported by @ruifontes via mail

Summary of the issue (steps to reproduce):

1. Open the following in Firefox:

   data:text/html, <p></p>

2. Navigate through the empty buffer.

Actual result

An error sound is played. The following is printed in the log:

Log snippet

ERROR - scriptHandler.executeScript (07:13:10.999):
error executing script: > with gesture u'down arrow'
Traceback (most recent call last):
  File "scriptHandler.pyc", line 187, in executeScript
  File "cursorManager.pyc", line 210, in script_moveByLine_forward
  File "cursorManager.pyc", line 140, in _caretMovementScriptHelper
  File "browseMode.pyc", line 1210, in _set_selection
  File "baseObject.pyc", line 21, in __get__
  File "textInfos\offsets.pyc", line 311, in _get_NVDAObjectAtStart
  File "virtualBuffers\__init__.pyc", line 181, in _getNVDAObjectFromOffset
  File "virtualBuffers\__init__.pyc", line 160, in _getFieldIdentifierFromOffset
LookupError: Neither docHandle nor ID found for offset 0

Expected result

NVDA just says blank.

Description of how this pull request fixes the issue:

With #8479, I introduced raising a LookupError for cases where both the vbuf handle and id of for an offset are 0 when executing _getFieldIdentifierFromOffset, since this is an invalid case. However, LookupErrors wheren't caught in _getNVDAObjectFromOffset. With this pr, they are, and an additional debugWarning is logged.

Note that the old path (pre #8479) was as follows:

1. Call _getNVDAObjectFromOffset with an offset
2. This calls _getFieldIdentifierFromOffset, which returns (0, 0)
3. _getNVDAObjectFromOffset calls getNVDAObjectFromIdentifier on the virtual buffer with parameters (0, 0). This returns None.
4. Therefore, we now return None early in _getNVDAObjectFromOffset.

Testing performed:

Tested the case as described above. Also made sure that _get_fieldIdentifierFromOffset isn't called elsewhere where this could cause an issue due to a not caught LookupError.

Known issues with pull request:

None

Change log entry:

None

[LeonarddeR]

@michaelDCurran: Just a reminder to make sure you're not missing this one.