Regards

Quentin.

On Mon, Jun 19, 2017 at 5:17 PM, Brian's Mail list account <bglists@...> wrote:

Hmm, I'm sure they know about this error condition, but I'm always a little sceptical  when I see errors that in the end it might be a problem that could cause issues if you don't trap it somehow.



Brian

bglists@...
Sent via blueyonder.
Please address personal email to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
----- Original Message ----- From: "Quentin Christensen" <quentin@...>
To: <nvda@nvda.groups.io>
Sent: Monday, June 19, 2017 12:31 AM
Subject: Re: [nvda] A question about non destructive errors

Hi Brian,

Without being specifically familiar with this particular code, a couple of
points off the top of my head:

- Often different programs (and sometimes even the same program) will have
controls which may look and behaves exactly the same to the end user, but
underneath they work differently.  This can be why some controls throw up
errors when others do not.
- Partly for the same reason, if we try and find out information a certain
way and it fails, there may be another (backup, if you like) way that we
can try which may get the same information.  In that case, the control may
throw up an error, but appear to work as normal.
- My lack of familiarity with the code will show here, but the error about
CoInitialize not being called is likely because it would be usual to
initialize that prior to making a certain request.  For whatever reason
that wasn't done, which may in some situations mean that the value returned
by that request may be unexpected.  Whether that's a problem or not I'd
need Mick or Jamie to answer, but all that might give you a bit of an idea
to start with.  Again, that's all just general off the top of my head and
without being specifically familiar with the code.

Regards

Quentin.

On Sun, Jun 18, 2017 at 11:28 PM, Brian's Mail list account <
bglists@...> wrote:

As quite a lot of us do use snapshot builds we do hear errors from time to
time. one of the most annoying ones I see in windows 7 is when some of the
search fields are used. There can be multiple errors as one types in and
even when the edit area is opened, yet to all  intents and purposes it
seems like the control itself has worked.
Here is a log to demonstrate what I mean.

IO - inputCore.InputManager.executeGesture (14:18:51.645):
Input: kb(desktop):shift+tab
IO - speech.speak (14:18:51.730):
Speaking [LangChangeCommand ('en_GB'), u'Search Box  edit   Search Windows
(C:)']
DEBUG - NVDAObjects.UIA.UIATextInfo._getTextWithFieldsForUIARange
(14:18:51.732):
_getTextWithFieldsForUIARange
DEBUG - NVDAObjects.UIA.UIATextInfo._getTextWithFieldsForUIARange
(14:18:51.733):
rootElement: edit
ERROR - eventHandler.executeEvent (14:18:51.733):
error executing event: gainFocus on <NVDAObjects.Dynamic_EditableT
extWithoutAutoSelectDetectionUIA object at 0x05C24D90> with extra args of
{}
Traceback (most recent call last):
 File "eventHandler.pyc", line 143, in executeEvent
 File "eventHandler.pyc", line 91, in __init__
 File "eventHandler.pyc", line 98, in next
 File "C:\Users\Brian\AppData\Roaming\nvda\addons\easyTableNavigat
or\globalPlugins\easyTableNavigator.py", line 82, in event_gainFocus
 File "eventHandler.pyc", line 98, in next
 File "C:\Users\Brian\AppData\Roaming\nvda\addons\focusHighlight\
globalPlugins\focusHighlight.py", line 427, in event_gainFocus
 File "eventHandler.pyc", line 98, in next
 File "appModules\explorer.pyc", line 280, in event_gainFocus
 File "eventHandler.pyc", line 98, in next
 File "NVDAObjects\__init__.pyc", line 907, in event_gainFocus
 File "NVDAObjects\__init__.pyc", line 819, in reportFocus
 File "speech.pyc", line 367, in speakObject
 File "speech.pyc", line 689, in speakTextInfo
 File "NVDAObjects\UIA\__init__.pyc", line 584, in getTextWithFields
 File "NVDAObjects\UIA\__init__.pyc", line 444, in
_getTextWithFieldsForUIARange
COMError: (-2147221008, 'CoInitialize has not been called.', (None, None,
None, 0, None))
ERROR - eventHandler.executeEvent (14:18:51.740):
error executing event: caret on <NVDAObjects.Dynamic_EditableT
extWithoutAutoSelectDetectionUIA object at 0x05C24D90> with extra args of
{}
Traceback (most recent call last):
 File "eventHandler.pyc", line 143, in executeEvent
 File "eventHandler.pyc", line 91, in __init__
 File "eventHandler.pyc", line 98, in next
 File "NVDAObjects\__init__.pyc", line 940, in event_caret
 File "review.pyc", line 167, in handleCaretMove
 File "NVDAObjects\__init__.pyc", line 973, in makeTextInfo
 File "NVDAObjects\UIA\__init__.pyc", line 244, in __init__
COMError: (-2147221008, 'CoInitialize has not been called.', (None, None,
None, 0, None))
ERROR - eventHandler.executeEvent (14:18:51.752):
error executing event: caret on <NVDAObjects.Dynamic_EditableT
extWithoutAutoSelectDetectionUIA object at 0x05C24D90> with extra args of
{}
Traceback (most recent call last):
 File "eventHandler.pyc", line 143, in executeEvent
 File "eventHandler.pyc", line 91, in __init__
 File "eventHandler.pyc", line 98, in next
 File "NVDAObjects\__init__.pyc", line 940, in event_caret
 File "review.pyc", line 167, in handleCaretMove
 File "NVDAObjects\__init__.pyc", line 973, in makeTextInfo
 File "NVDAObjects\UIA\__init__.pyc", line 244, in __init__
COMError: (-2147221008, 'CoInitialize has not been called.', (None, None,
None, 0, None))


This is in the master branch and has been around for quite some time now
but of course one does not hear it on the final builds.
Also quit why it refers to two add-ons which would seem to have nothing to
do with it is a mystery as well.
Just wondered why this c_inisialise or whatever is not called and why it
seems that some edit areas work while others do not.

Brian

bglists@...
Sent via blueyonder.
Please address personal email to:-
briang1@..., putting 'Brian Gaff'
in the display name field.

--
Quentin Christensen
Training and Support Manager

Basic Training for NVDA & Microsoft Word with NVDA E-Books now available:
http://www.nvaccess.org/shop/

Ph +61 7 3149 3306
www.nvaccess.org
Facebook: http://www.facebook.com/NVAccess
Twitter: @NVAccess