Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

NVDA causes errors on exit of some windows software #2254

Closed
nvaccessAuto opened this issue Apr 20, 2012 · 26 comments
Closed

NVDA causes errors on exit of some windows software #2254

nvaccessAuto opened this issue Apr 20, 2012 · 26 comments
Labels
Milestone

Comments

@nvaccessAuto
Copy link

Reported by briang1 on 2012-04-20 13:39
This is probably related to my previous ticket. It seems as if nvda is doing nasty things when certain windows applications try to exit. Here is the log from task manager.

IO - inputCore.InputManager.executeGesture (14:29:43):
Input: kb(desktop):control+shift+escape
DEBUGWARNING - RPC process 2752 (taskmgr.exe) (14:29:45):
Thread 2756, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 2752 (taskmgr.exe) (14:29:45):
Thread 2756, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 2752 (taskmgr.exe) (14:29:45):
Thread 2756, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 2752 (taskmgr.exe) (14:29:45):
Thread 2756, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 2752 (taskmgr.exe) (14:29:45):
Thread 2756, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

IO - speech.speak (14:29:45):
Speaking ('en_UK'), u'Windows Task Manager dialog'
IO - speech.speak (14:29:45):
Speaking ('en_UK'), u'Processes list'
IO - speech.speak (14:29:45):
Speaking ('en_UK'), u'alg.exe User Name: LOCAL SERVICE, CPU: 00 , Mem Usage: 3,584 K 1 of 40'
IO - inputCore.InputManager.executeGesture (14:29:53):
Input: kb(desktop):alt+f4
DEBUGWARNING - IAccessibleHandler.accessibleObjectFromEvent (14:29:53):
oleacc.AccessibleObjectFromEvent with window 66592, objectID -4 and childID 0: -2147467259 Unspecified error
IO - speech.speak (14:29:54):
Speaking ('en_UK'), u'Taskbar'
IO - speech.speak (14:29:54):
Speaking ('en_UK'), u'taskmgr.exe - Application Error dialog The instruction at "0x00961f6e" referenced memory at "0x00000004". The memory could not be "read".\r\n\nClick on OK to terminate the program'
IO - speech.speak (14:29:54):
Speaking ('en_UK'), u'OK button'
IO - inputCore.InputManager.executeGesture (14:30:07):
Input: kb(desktop):enter
DEBUG - appModuleHandler.update (14:30:07):
application taskmgr closed
IO - speech.speak (14:30:07):
Speaking ('en_UK'), u'Taskbar'
IO - inputCore.InputManager.executeGesture (14:30:21):
Input: kb(desktop):control+alt+r
INFO - core.main (14:30:21):
Exiting

Notepad and probably others are affected also.

@nvaccessAuto
Copy link
Author

Comment 1 by briang1 on 2012-04-20 13:42
After such an error nvda is virtually unusable particularly on Outlook Express as some items are not spoken any more. It has to be rebooted to fix it.

@nvaccessAuto
Copy link
Author

Comment 2 by briang1 on 2012-04-20 14:09
Also affected, Jarte, IE8 All versions of Firefox, Winamp, malwarebytes. Not all scew up nvda after the error but all the errors look the same to me, as far as the windows message is concerned.

@nvaccessAuto
Copy link
Author

Comment 3 by mdcurran on 2012-04-20 15:35
According to discussions on nvda-support its clear this only affects XP. I'm hoping this won't be too hard to debug -- I know exactly where to look.
Changes:
Milestone changed from None to 2012.2

@nvaccessAuto
Copy link
Author

Comment 4 by briang1 on 2012-04-20 15:49
Hi well, I also tried to navigate the error alert with flat review when it happened in notepad. this is the log of that.

IO - speech.speak (16:35:51):
Speaking ('en_UK'), u'NOTEPAD 39 of 60'
DEBUG - queueHandler.pumpAll (16:35:51):
generator 3 finished
IO - inputCore.InputManager.executeGesture (16:35:54):
Input: kb(desktop):enter
DEBUGWARNING - RPC process 296 (notepad.exe) (16:35:54):
Thread 3824, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 296 (notepad.exe) (16:35:54):
Thread 3824, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 296 (notepad.exe) (16:35:54):
Thread 3824, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

IO - speech.speak (16:35:54):
Speaking ('en_UK'), u'Untitled - Notepad'
IO - speech.speak (16:35:54):
Speaking ('en_UK'), u'edit multi line'
IO - speech.speak (16:35:54):
Speaking ('en_UK'), u'blank'
IO - inputCore.InputManager.executeGesture (16:35:56):
Input: kb(desktop):alt+f4
IO - speech.speak (16:35:56):
Speaking ('en_UK'), u'Taskbar'
IO - speech.speak (16:35:57):
Speaking ('en_UK'), u'notepad.exe - Application Error dialog The instruction at "0x00921f6e" referenced memory at "0x00000004". The memory could not be "read".\r\n\nClick on OK to terminate the program'
IO - speech.speak (16:35:57):
Speaking ('en_UK'), u'OK button'
IO - inputCore.InputManager.executeGesture (16:36:14):
Input: kb(desktop):NVDA+numpad7
ERROR - scriptHandler.executeScript (16:36:14):
error executing script: <bound method GlobalCommands.script_navigatorObject_moveToFlatReviewAtObjectPosition of <globalCommands.GlobalCommands object at 0x01D598F0>> with gesture u'NVDA+numpad 7'
Traceback (most recent call last):
File "scriptHandler.pyc", line 165, in executeScript
File "globalCommands.pyc", line 247, in script_navigatorObject_moveToFlatReviewAtObjectPosition
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in getPropertyViaCache
File "NVDAObjects__init
_.pyc", line 830, in get_flatReviewPosition
File "NVDAObjects__init
_.pyc", line 848, in makeTextInfo
File "textInfos\offsets.pyc", line 266, in init
File "displayModel.pyc", line 176, in _getOffsetsFromNVDAObject
File "displayModel.pyc", line 148, in _getClosestOffsetFromPoint
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in getPropertyViaCache
File "displayModel.pyc", line 71, in get__textAndRects
File "displayModel.pyc", line 25, in getWindowTextInRect
File "watchdog.pyc", line 274, in cancellableExecute
File "watchdog.pyc", line 246, in run
ArgumentError: argument 1: <type 'exceptions.TypeError'>: wrong type
IO - inputCore.InputManager.executeGesture (16:36:16):
Input: kb(desktop):NVDA+numpad7
ERROR - scriptHandler.executeScript (16:36:16):
error executing script: <bound method GlobalCommands.script_navigatorObject_moveToFlatReviewAtObjectPosition of <globalCommands.GlobalCommands object at 0x01D598F0>> with gesture u'NVDA+numpad 7'
Traceback (most recent call last):
File "scriptHandler.pyc", line 165, in executeScript
File "globalCommands.pyc", line 247, in script_navigatorObject_moveToFlatReviewAtObjectPosition
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in getPropertyViaCache
File "NVDAObjects__init
.pyc", line 830, in get_flatReviewPosition
File "NVDAObjects__init
.pyc", line 848, in makeTextInfo
File "textInfos\offsets.pyc", line 266, in init
File "displayModel.pyc", line 176, in _getOffsetsFromNVDAObject
File "displayModel.pyc", line 148, in _getClosestOffsetFromPoint
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in _getPropertyViaCache
File "displayModel.pyc", line 71, in _get__textAndRects
File "displayModel.pyc", line 25, in getWindowTextInRect
File "watchdog.pyc", line 274, in cancellableExecute
File "watchdog.pyc", line 246, in run
ArgumentError: argument 1: <type 'exceptions.TypeError'>: wrong type
IO - inputCore.InputManager.executeGesture (16:36:20):
Input: kb(desktop):control+alt+r
INFO - core.main (16:36:21):
Exiting

@nvaccessAuto
Copy link
Author

Comment 5 by briang1 on 2012-04-20 16:07
After a while the effects seem to grow, even some explorer folders then get affected and trying to run Outlook express well...

IO - speech.speak (17:00:43):
Speaking ('en_UK'), u'outlook express 45 of 60'
DEBUG - queueHandler.pumpAll (17:00:43):
generator 1 finished
IO - inputCore.InputManager.executeGesture (17:00:43):
Input: kb(desktop):enter
ERROR - eventHandler.executeEvent (17:00:44):
error executing event: gainFocus on <NVDAObjects.IAccessible.ContentGenericClient object at 0x03EF5F30> with extra args of {}
Traceback (most recent call last):
File "eventHandler.pyc", line 136, in executeEvent
File "eventHandler.pyc", line 84, in init
File "eventHandler.pyc", line 91, in next
File "appModules\msimn.pyc", line 70, in event_gainFocus
File "eventHandler.pyc", line 91, in next
File "NVDAObjects__init__.pyc", line 778, in event_gainFocus
File "NVDAObjects__init__.pyc", line 722, in reportFocus
File "speech.pyc", line 300, in speakObject
File "speech.pyc", line 232, in speakObjectProperties
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in getPropertyViaCache
File "NVDAObjects\IAccessible__init
_.pyc", line 1303, in get_value
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in getPropertyViaCache
File "NVDAObjects\window__init
.pyc", line 192, in get_displayText
File "displayModel.pyc", line 25, in getWindowTextInRect
File "watchdog.pyc", line 274, in cancellableExecute
File "watchdog.pyc", line 246, in run
ArgumentError: argument 1: <type 'exceptions.TypeError'>: wrong type
IO - speech.speak (17:00:44):
Speaking ('en_UK'), u'Inbox - Outlook Express - Brian Gaff Main'
IO - speech.speak (17:00:44):
Speaking ('en_UK'), u'Outlook Express Message List list'
ERROR - eventHandler.executeEvent (17:00:44):
error executing event: gainFocus on <appModules.msimn.MessageListItem object at 0x03F38F10> with extra args of {}
Traceback (most recent call last):
File "eventHandler.pyc", line 136, in executeEvent
File "eventHandler.pyc", line 84, in init
File "eventHandler.pyc", line 91, in next
File "appModules\msimn.pyc", line 70, in event_gainFocus
File "eventHandler.pyc", line 91, in next
File "NVDAObjects__init
_.pyc", line 778, in event_gainFocus
File "NVDAObjects__init__.pyc", line 722, in reportFocus
File "speech.pyc", line 300, in speakObject
File "speech.pyc", line 232, in speakObjectProperties
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in _getPropertyViaCache
File "appModules\msimn.pyc", line 107, in get_name
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in getPropertyViaCache
File "appModules\msimn.pyc", line 93, in get_isUnread
File "displayModel.pyc", line 87, in getTextWithFields
File "displayModel.pyc", line 71, in get__textAndRects
File "displayModel.pyc", line 25, in getWindowTextInRect
File "watchdog.pyc", line 274, in cancellableExecute
File "watchdog.pyc", line 246, in run
ArgumentError: argument 1: <type 'exceptions.TypeError'>: wrong type
ERROR - eventHandler.executeEvent (17:00:44):
error executing event: gainFocus on <appModules.msimn.MessageListItem object at 0x03F38FD0> with extra args of {}
Traceback (most recent call last):
File "eventHandler.pyc", line 131, in executeEvent
File "eventHandler.pyc", line 144, in doPreGainFocus
File "api.pyc", line 86, in setFocusObject
File "NVDAObjects__init
.pyc", line 288, in eq
File "NVDAObjects\IAccessible__init
.pyc", line 642, in _isEqual
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in _getPropertyViaCache
File "appModules\msimn.pyc", line 107, in _get_name
File "baseObject.pyc", line 34, in get
File "baseObject.pyc", line 110, in _getPropertyViaCache
File "appModules\msimn.pyc", line 93, in _get_isUnread
File "displayModel.pyc", line 87, in getTextWithFields
File "displayModel.pyc", line 71, in _get__textAndRects
File "displayModel.pyc", line 25, in getWindowTextInRect
File "watchdog.pyc", line 274, in cancellableExecute
File "watchdog.pyc", line 246, in run
ArgumentError: argument 1: <type 'exceptions.TypeError'>: wrong type
IO - inputCore.InputManager.executeGesture (17:00:53):
Input: kb(desktop):control+alt+r
INFO - core.main (17:00:53):
Exiting

@nvaccessAuto
Copy link
Author

Comment 6 by mdcurran on 2012-04-20 17:29
Hopefully fixed in 67ec0cd. Please close if so.

@nvaccessAuto
Copy link
Author

Comment 7 by elliott94 on 2012-04-20 18:22
I've just been speaking to another user who has just updated to 5121, and the issue still exists. Here's a section from their log:

DEBUGWARNING - RPC process 1412 (taskmgr.exe) (19:18:16):
Thread 5636, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

@nvaccessAuto
Copy link
Author

Comment 8 by Palacee_hun on 2012-04-20 19:23
Interestingly I don't have either any of these crashes and odd behaviours or any errors when exiting apps on my XP system with snap 5120.
[portable copy from main-5120 package to "c:\nvda_test" folder. Ran NVDA.EXE from there with Total Commander after quitting running version of NVDA. Using Espeak synth, no braille.
[[BR]([BR]]
Particulars:
[[BR]]
Created)]
System: Windows XP Home SP3 (Hungarian lang.) 32-bit, AMD Sempron 3200+ CPU 1.81 Ghz, 1 Gb of RAM, bunch of NVIDIA Forceware and network drivers preinstalled on Windows+Eset NOD32 av
[[BR]]
Tested apps: Total Commander, Firefox 10.0.2, VLC Media Player, Notepad, Utorrent, Adobe Reader, Explorer. Great stability, smooth and responsive operation, as responsive and slick as earlier snaps or even a bit better.

@nvaccessAuto
Copy link
Author

Comment 9 by mdcurran (in reply to comment 7) on 2012-04-20 20:34
Replying to elliott94:

I've just been speaking to another user who has just updated to 5121, and the issue still exists.

Did the user still definitely still experience the same crash? (i.e. a dialog saying the memory could not be read at blahblahblah)?
Those log lines may still continue to appear once the crash is fixed.
It also may be useful to reboot the PC as an old dll may still be in use.

@nvaccessAuto
Copy link
Author

Comment 10 by PZajda on 2012-04-20 20:58
Hi,

I have the same bug with the snapshot 5120 and 5121 even after rebooting.

@nvaccessAuto
Copy link
Author

Comment 11 by mdcurran (in reply to comment 10) on 2012-04-20 21:06
Replying to PZajda:

I have the same bug with the snapshot 5120 and 5121 even after rebooting.

What service pack of XP?
Could you please provide some exact steps to reproduce? E.g. started notepad. Typed particular text. Closed notepad, said no to save. Dialog saing the memory could not be read.
Annoyingly on XP SP3 I was only ever able to reproduce in notepad when pasting several lines in to a new document and then closing without saving. But it went away with my changes in 5121.

@nvaccessAuto
Copy link
Author

Comment 12 by mdcurran on 2012-04-21 00:25
Its now looking like the cause is minhook.dll being marked for delay loading. As if I revert this change completely the issue goes away even in Taskmgr.
The errors can be quite different, but it seems as though nvdahelperRemote's dllmain is either never being called on process exit, or its being called much too late. This means that lots of our code is running while the process is exiting. Have tried protecting it in different ways but each time I do one thing another error is uncovered.
If delay loading cannot be solved we'll have to just load minhook manually and getProcAddress for each needed function.
All this for Windows 8.

@nvaccessAuto
Copy link
Author

Comment 13 by briang1 (in reply to comment 12) on 2012-04-21 07:08
Replying to mdcurran:

Its now looking like the cause is minhook.dll being marked for delay loading. As if I revert this change completely the issue goes away even in Taskmgr.

The errors can be quite different, but it seems as though nvdahelperRemote's dllmain is either never being called on process exit, or its being called much too late. This means that lots of our code is running while the process is exiting. Have tried protecting it in different ways but each time I do one thing another error is uncovered.

If delay loading cannot be solved we'll have to just load minhook manually and getProcAddress for each needed function.

All this for Windows 8.

Actually I'd more or less figured out that it was code holding the routines while stuff wanted to exit.
Below I'll paste most of the stuff from the non ticket messages for posterity, so to speak. However I wonder if its also a problem more for those of us using older non multi threaded processors?
Sounds to me like Microsoft need to sort out their own code.

The complete reboot of the computer with the snaps 20 or 21 does not solve the issue..., but snap21 is no better
than snap 20 here.
INFO - nvda (23:20:13):
Starting NVDA
INFO - core.main (23:20:14):
Config dir: C:\Documents and Settings\brian\Application Data\nvda
DEBUG - core.main (23:20:14):
setting language to Windows
INFO - core.main (23:20:14):
NVDA version main-5121
INFO - core.main (23:20:14):
Using Windows version sys.getwindowsversion(major=5, minor=1, build=2600,
platform=2, service_pack='Service Pack 3')
INFO - core.main (23:20:14):
Using Python version 2.7.2 (default, Jun 12 2011, 15:08:59) v.1500 32
bit (Intel)

INFO - core.main (23:20:14):
Using comtypes version 0.6.2
DEBUG - core.main (23:20:14):
Creating wx application instance
DEBUG - core.main (23:20:15):
Initializing appModule Handler
DEBUG - core.main (23:20:15):
Initializing NVDAHelper
DEBUG - core.main (23:20:15):
Speech Dictionary processing
DEBUG - speechDictHandler.SpeechDict.load (23:20:15):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\default.dic'...
DEBUG - speechDictHandler.SpeechDict.load (23:20:15):
8 loaded records.
DEBUG - speechDictHandler.SpeechDict.load (23:20:15):
Loading speech dictionary 'builtin.dic'...
DEBUG - speechDictHandler.SpeechDict.load (23:20:15):
3 loaded records.
DEBUG - core.main (23:20:15):
Initializing speech
INFO - synthDrivers.espeak.SynthDriver.init (23:20:15):
Using eSpeak version 1.46.02 06.Jan.12
DEBUG - speechDictHandler.SpeechDict.load (23:20:15):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\espeak-english.dic'...
DEBUG - speechDictHandler.SpeechDict.load (23:20:15):
1 loaded records.
INFO - synthDriverHandler.setSynth (23:20:15):
Loaded synthDriver espeak
INFO - core.main (23:20:15):
Using wx version 2.8.12.1 (msw-unicode)
DEBUG - core.main (23:20:15):
Initializing braille
INFO - braille.initialize (23:20:15):
Using liblouis version 2.4.1
INFO - braille.BrailleHandler.setDisplayByName (23:20:15):
Loaded braille display driver noBraille, current display has 0 cells.
DEBUG - core.main (23:20:15):
Initializing displayModel
DEBUG - core.main (23:20:15):
Initializing GUI
DEBUG - core.main (23:20:16):
initializing Java Access Bridge support
DEBUG - core.main (23:20:16):
Initializing winConsole support
DEBUG - core.main (23:20:16):
Initializing UIA support
WARNING - core.main (23:20:16):
UIA not available
DEBUG - core.main (23:20:16):
Initializing IAccessible support
DEBUG - core.main (23:20:16):
Initializing input core
DEBUGWARNING - inputCore.InputManager.loadLocaleGestureMap (23:20:16):
No locale gesture map for language en
DEBUGWARNING - inputCore.InputManager.loadUserGestureMap (23:20:16):
No user gesture map
DEBUG - core.main (23:20:16):
Initializing keyboard handler
DEBUG - core.main (23:20:16):
initializing mouse handler
DEBUG - core.main (23:20:16):
Initializing global plugin handler
DEBUG - core.main (23:20:16):
starting core pump
DEBUG - core.CorePump.init (23:20:16):
Core pump starting
DEBUG - core.main (23:20:16):
Initializing watchdog
DEBUG - core.main (23:20:16):
initializing updateCheck
INFO - core.main (23:20:16):
NVDA initialized
DEBUG - core.main (23:20:16):
entering wx application main loop
IO - speech.speak (23:20:16):
Speaking list'
IO - speech.speak (23:20:16):
Speaking Explorer 5 of 58'
IO - inputCore.InputManager.executeGesture (23:20:24):
Input: kb(desktop):control+shift+escape
DEBUGWARNING - RPC process 1416 (taskmgr.exe) (23:20:24):
Thread 3884, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1416 (taskmgr.exe) (23:20:24):
Thread 3884, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1416 (taskmgr.exe) (23:20:24):
Thread 3884, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1416 (taskmgr.exe) (23:20:24):
Thread 3884, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1416 (taskmgr.exe) (23:20:24):
Thread 3884, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

IO - speech.speak (23:20:24):
Speaking Task Manager dialog'
IO - speech.speak (23:20:24):
Speaking list'
IO - speech.speak (23:20:24):
Speaking User Name: SYSTEM, CPU: 00 , CPU Time: 0:00:00,
Mem Usage: 3,512 K, Peak Mem Usage: 3,520 K, Threads: 8 1 of 38'

IO - inputCore.InputManager.executeGesture (23:20:30):
Input: kb(desktop):alt+f4
IO - speech.speak (23:20:30):
Speaking [- speech.speak (23:20:30):
Speaking [u'taskmgr.exe - Application Error dialog The instruction at
"0x00961f3e" referenced memory at "0x00000004". The memory could not be
"read".\r\n\nClick on OK to terminate the program'](u'Taskbar']
IO)
IO - speech.speak (23:20:30):
Speaking button'
IO - inputCore.InputManager.executeGesture (23:20:57):
Input: kb(desktop):control+alt+i
IO - speech.speak (23:20:57):
Speaking nvda with log terminal'
IO - speech.speak (23:20:57):

Speaking [- core.main (23:20:57):

21 is really little better than 20, just the same apps.

INFO - nvda (00:00:58):
Starting NVDA
INFO - core.main (00:00:58):
Config dir: C:\Documents and Settings\brian\Application Data\nvda
DEBUG - core.main (00:00:58):
setting language to Windows
INFO - core.main (00:00:58):
NVDA version main-5121
INFO - core.main (00:00:58):
Using Windows version sys.getwindowsversion(major=5, minor=1, build=2600,
platform=2, service_pack='Service Pack 3')
INFO - core.main (00:00:58):
Using Python version 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32
bit (Intel)](u'blank']
INFO)
INFO - core.main (00:00:58):
Using comtypes version 0.6.2
DEBUG - core.main (00:00:58):
Creating wx application instance
DEBUG - core.main (00:00:58):
Initializing appModule Handler
DEBUG - core.main (00:00:58):
Initializing NVDAHelper
DEBUG - core.main (00:00:59):
Speech Dictionary processing
DEBUG - speechDictHandler.SpeechDict.load (00:00:59):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\default.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:00:59):
8 loaded records.
DEBUG - speechDictHandler.SpeechDict.load (00:00:59):
Loading speech dictionary 'builtin.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:00:59):
3 loaded records.
DEBUG - core.main (00:00:59):
Initializing speech
INFO - synthDrivers.espeak.SynthDriver.init (00:00:59):
Using eSpeak version 1.46.02 06.Jan.12
DEBUG - speechDictHandler.SpeechDict.load (00:00:59):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\espeak-english.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:00:59):
1 loaded records.
INFO - synthDriverHandler.setSynth (00:00:59):
Loaded synthDriver espeak
INFO - core.main (00:00:59):
Using wx version 2.8.12.1 (msw-unicode)
DEBUG - core.main (00:00:59):
Initializing braille
INFO - braille.initialize (00:00:59):
Using liblouis version 2.4.1
INFO - braille.BrailleHandler.setDisplayByName (00:00:59):
Loaded braille display driver noBraille, current display has 0 cells.
DEBUG - core.main (00:00:59):
Initializing displayModel
DEBUG - core.main (00:00:59):
Initializing GUI
DEBUG - core.main (00:00:59):
initializing Java Access Bridge support
DEBUG - core.main (00:00:59):
Initializing winConsole support
DEBUG - core.main (00:00:59):
Initializing UIA support
WARNING - core.main (00:00:59):
UIA not available
DEBUG - core.main (00:00:59):
Initializing IAccessible support
DEBUG - core.main (00:00:59):
Initializing input core
DEBUGWARNING - inputCore.InputManager.loadLocaleGestureMap (00:00:59):
No locale gesture map for language en
DEBUGWARNING - inputCore.InputManager.loadUserGestureMap (00:00:59):
No user gesture map
DEBUG - core.main (00:00:59):
Initializing keyboard handler
DEBUG - core.main (00:00:59):
initializing mouse handler
DEBUG - core.main (00:00:59):
Initializing global plugin handler
DEBUG - core.main (00:00:59):
starting core pump
DEBUG - core.CorePump.init (00:00:59):
Core pump starting
DEBUG - core.main (00:00:59):
Initializing watchdog
DEBUG - core.main (00:00:59):
initializing updateCheck
INFO - core.main (00:00:59):
NVDA initialized
DEBUG - core.main (00:00:59):
entering wx application main loop
IO - speech.speak (00:00:59):
Speaking list'
IO - speech.speak (00:00:59):
Speaking Checker 55 of 58'
IO - inputCore.InputManager.executeGesture (00:01:03):
Input: kb(desktop):control+alt+a
DEBUGWARNING - RPC process 3404 (notepad.exe) (00:01:03):
Thread 1436, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 3404 (notepad.exe) (00:01:03):
Thread 1436, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 3404 (notepad.exe) (00:01:03):
Thread 1436, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support,
81:
Error registering class object, code -2147221008

IO - speech.speak (00:01:04):
Speaking - Notepad'
IO - speech.speak (00:01:04):
Speaking multi line'
IO - speech.speak (00:01:04):
Speaking [- inputCore.InputManager.executeGesture (00:01:08):
Input: kb(desktop):alt+f4
IO - speech.speak (00:01:08):
Speaking [u'Desktop list'](u'blank']
IO)
IO - speech.speak (00:01:08):
Speaking Checker 55 of 58'
IO - speech.speak (00:01:08):
Speaking - Application Error dialog The instruction at
"0x00921f3e" referenced memory at "0x00000004". The memory could not be
"read".\r\n\nClick on OK to terminate the program'

IO - speech.speak (00:01:08):
Speaking button'
IO - inputCore.InputManager.executeGesture (00:01:22):
Input: kb(desktop):enter
IO - speech.speak (00:01:22):
Speaking [- inputCore.InputManager.executeGesture (00:01:33):
Input: kb(desktop):control+alt+i
DEBUG - appModuleHandler.update (00:01:33):
application notepad closed
IO - speech.speak (00:01:33):
Speaking [u'reboot nvda with log terminal'](u'Taskbar']
IO)
IO - speech.speak (00:01:33):
Speaking [- core.main (00:01:33):
Exiting
and.
INFO - nvda (00:04:17):
Starting NVDA
INFO - core.main (00:04:17):
Config dir: C:\Documents and Settings\brian\Application Data\nvda
DEBUG - core.main (00:04:17):
setting language to Windows
INFO - core.main (00:04:17):
NVDA version main-5121
INFO - core.main (00:04:17):
Using Windows version sys.getwindowsversion(major=5, minor=1, build=2600,
platform=2, service_pack='Service Pack 3')
INFO - core.main (00:04:17):
Using Python version 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32
bit (Intel)](u'blank']
INFO)
INFO - core.main (00:04:17):
Using comtypes version 0.6.2
DEBUG - core.main (00:04:17):
Creating wx application instance
DEBUG - core.main (00:04:17):
Initializing appModule Handler
DEBUG - core.main (00:04:17):
Initializing NVDAHelper
DEBUG - core.main (00:04:17):
Speech Dictionary processing
DEBUG - speechDictHandler.SpeechDict.load (00:04:17):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\default.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:04:17):
8 loaded records.
DEBUG - speechDictHandler.SpeechDict.load (00:04:17):
Loading speech dictionary 'builtin.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:04:17):
3 loaded records.
DEBUG - core.main (00:04:17):
Initializing speech
INFO - synthDrivers.espeak.SynthDriver.init (00:04:18):
Using eSpeak version 1.46.02 06.Jan.12
DEBUG - speechDictHandler.SpeechDict.load (00:04:18):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\espeak-english.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:04:18):
1 loaded records.
INFO - synthDriverHandler.setSynth (00:04:18):
Loaded synthDriver espeak
INFO - core.main (00:04:18):
Using wx version 2.8.12.1 (msw-unicode)
DEBUG - core.main (00:04:18):
Initializing braille
INFO - braille.initialize (00:04:18):
Using liblouis version 2.4.1
INFO - braille.BrailleHandler.setDisplayByName (00:04:18):
Loaded braille display driver noBraille, current display has 0 cells.
DEBUG - core.main (00:04:18):
Initializing displayModel
DEBUG - core.main (00:04:18):
Initializing GUI
DEBUG - core.main (00:04:18):
initializing Java Access Bridge support
DEBUG - core.main (00:04:18):
Initializing winConsole support
DEBUG - core.main (00:04:18):
Initializing UIA support
WARNING - core.main (00:04:18):
UIA not available
DEBUG - core.main (00:04:18):
Initializing IAccessible support
DEBUG - core.main (00:04:18):
Initializing input core
DEBUGWARNING - inputCore.InputManager.loadLocaleGestureMap (00:04:18):
No locale gesture map for language en
DEBUGWARNING - inputCore.InputManager.loadUserGestureMap (00:04:18):
No user gesture map
DEBUG - core.main (00:04:18):
Initializing keyboard handler
DEBUG - core.main (00:04:18):
initializing mouse handler
DEBUG - core.main (00:04:18):
Initializing global plugin handler
DEBUG - core.main (00:04:18):
starting core pump
DEBUG - core.CorePump.init (00:04:18):
Core pump starting
DEBUG - core.main (00:04:18):
Initializing watchdog
DEBUG - core.main (00:04:18):
initializing updateCheck
INFO - core.main (00:04:18):
NVDA initialized
DEBUG - core.main (00:04:18):
entering wx application main loop
IO - speech.speak (00:04:18):
Speaking - Outlook Express - Brians mail lists'
IO - inputCore.InputManager.executeGesture (00:04:24):
Input: kb(desktop):windows+m
IO - speech.speak (00:04:25):
Speaking list'
IO - speech.speak (00:04:25):
Speaking 12 of 58'
IO - inputCore.InputManager.executeGesture (00:04:26):
Input: kb(desktop):enter
DEBUGWARNING - watchdog.watcher (00:04:28):
Trying to recover from freeze, core stack:
File "nvda.pyw", line 155, in
File "core.pyc", line 330, in main
File "wx_core.pyc", line 8010, in MainLoop
File "wx_core.pyc", line 7306, in MainLoop
File "core.pyc", line 307, in Notify
File "IAccessibleHandler.pyc", line 837, in pumpAll
File "IAccessibleHandler.pyc", line 746, in processForegroundWinEvent
File "IAccessibleHandler.pyc", line 532, in winEventToNVDAEvent
File "NVDAObjects\IAccessible__init
_.pyc", line 38, in
getNVDAObjectFromEvent
File "IAccessibleHandler.pyc", line 358, in accessibleObjectFromEvent
File "oleacc.pyc", line 240, in AccessibleObjectFromEvent

DEBUGWARNING - IAccessibleHandler.accessibleObjectFromEvent (00:04:28):
oleacc.AccessibleObjectFromEvent with window 525096, objectID -4 and childID
0:
DEBUGWARNING - IAccessibleHandler.accessibleObjectFromEvent (00:04:29):
oleacc.AccessibleObjectFromEvent with window 525096, objectID -4 and childID
0:
DEBUGWARNING - IAccessibleHandler.accessibleObjectFromEvent (00:04:30):
oleacc.AccessibleObjectFromEvent with window 65666, objectID -4 and childID
6:
DEBUGWARNING - watchdog.watcher (00:04:30):
Trying to recover from freeze, core stack:
File "nvda.pyw", line 155, in
File "core.pyc", line 330, in main
File "wx_core.pyc", line 8010, in MainLoop
File "wx_core.pyc", line 7306, in MainLoop
File "core.pyc", line 307, in Notify
File "IAccessibleHandler.pyc", line 851, in pumpAll
File "IAccessibleHandler.pyc", line 610, in processGenericWinEvent
File "IAccessibleHandler.pyc", line 532, in winEventToNVDAEvent
File "NVDAObjects\IAccessible__init
_.pyc", line 38, in
getNVDAObjectFromEvent
File "IAccessibleHandler.pyc", line 358, in accessibleObjectFromEvent
File "oleacc.pyc", line 240, in AccessibleObjectFromEvent

IO - speech.speak (00:04:31):
Speaking (screen reader mode)'
IO - speech.speak (00:04:31):
Speaking multi line'
IO - speech.speak (00:04:31):
Speaking [- inputCore.InputManager.executeGesture (00:04:34):
Input: kb(desktop):alt+f4
DEBUGWARNING - watchdog.watcher (00:04:35):
Trying to recover from freeze, core stack:
File "nvda.pyw", line 155, in
File "core.pyc", line 330, in main
File "wx_core.pyc", line 8010, in MainLoop
File "wx_core.pyc", line 7306, in MainLoop
File "core.pyc", line 308, in Notify
File "queueHandler.pyc", line 76, in pumpAll
File "queueHandler.pyc", line 47, in flushQueue
File "eventHandler.pyc", line 54, in queueEventCallback
File "eventHandler.pyc", line 131, in executeEvent
File "core.pyc", line 68, in newCpbDel
File "comtypes__init
.pyc", line 841, in del
File "comtypes__init__.pyc", line 1092, in Release

IO - speech.speak (00:04:35):
Speaking [u'Inbox - Outlook Express - Brians mail lists'](u'blank']
IO)
IO - speech.speak (00:04:35):
Speaking - Application Error dialog The instruction at
"0x01fe1f3e" referenced memory at "0x00000004". The memory could not be
"read".\r\n\nClick on OK to terminate the program'

IO - speech.speak (00:04:35):
Speaking button'
IO - inputCore.InputManager.executeGesture (00:04:44):
Input: kb(desktop):control+alt+i
IO - speech.speak (00:04:44):
Speaking nvda with log terminal'
IO - speech.speak (00:04:44):
Speaking [- core.main (00:04:45):
Exiting

and..
INFO - nvda (00:06:42):
Starting NVDA
INFO - core.main (00:06:42):
Config dir: C:\Documents and Settings\brian\Application Data\nvda
DEBUG - core.main (00:06:42):
setting language to Windows
INFO - core.main (00:06:42):
NVDA version main-5121
INFO - core.main (00:06:42):
Using Windows version sys.getwindowsversion(major=5, minor=1, build=2600,
platform=2, service_pack='Service Pack 3')
INFO - core.main (00:06:42):
Using Python version 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32
bit (Intel)](u'blank']
INFO)
INFO - core.main (00:06:42):
Using comtypes version 0.6.2
DEBUG - core.main (00:06:42):
Creating wx application instance
DEBUG - core.main (00:06:43):
Initializing appModule Handler
DEBUG - core.main (00:06:43):
Initializing NVDAHelper
DEBUG - core.main (00:06:43):
Speech Dictionary processing
DEBUG - speechDictHandler.SpeechDict.load (00:06:43):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\default.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:06:43):
8 loaded records.
DEBUG - speechDictHandler.SpeechDict.load (00:06:43):
Loading speech dictionary 'builtin.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:06:43):
3 loaded records.
DEBUG - core.main (00:06:43):
Initializing speech
INFO - synthDrivers.espeak.SynthDriver.init (00:06:43):
Using eSpeak version 1.46.02 06.Jan.12
DEBUG - speechDictHandler.SpeechDict.load (00:06:43):
Loading speech dictionary 'C:\Documents and Settings\brian\Application
Data\nvda\speechDicts\espeak-english.dic'...
DEBUG - speechDictHandler.SpeechDict.load (00:06:43):
1 loaded records.
INFO - synthDriverHandler.setSynth (00:06:43):
Loaded synthDriver espeak
INFO - core.main (00:06:43):
Using wx version 2.8.12.1 (msw-unicode)
DEBUG - core.main (00:06:43):
Initializing braille
INFO - braille.initialize (00:06:43):
Using liblouis version 2.4.1
INFO - braille.BrailleHandler.setDisplayByName (00:06:43):
Loaded braille display driver noBraille, current display has 0 cells.
DEBUG - core.main (00:06:43):
Initializing displayModel
DEBUG - core.main (00:06:43):
Initializing GUI
DEBUG - core.main (00:06:43):
initializing Java Access Bridge support
DEBUG - core.main (00:06:43):
Initializing winConsole support
DEBUG - core.main (00:06:43):
Initializing UIA support
WARNING - core.main (00:06:43):
UIA not available
DEBUG - core.main (00:06:43):
Initializing IAccessible support
DEBUG - core.main (00:06:43):
Initializing input core
DEBUGWARNING - inputCore.InputManager.loadLocaleGestureMap (00:06:43):
No locale gesture map for language en
DEBUGWARNING - inputCore.InputManager.loadUserGestureMap (00:06:43):
No user gesture map
DEBUG - core.main (00:06:43):
Initializing keyboard handler
DEBUG - core.main (00:06:43):
initializing mouse handler
DEBUG - core.main (00:06:43):
Initializing global plugin handler
DEBUG - core.main (00:06:43):
starting core pump
DEBUG - core.CorePump.init (00:06:43):
Core pump starting
DEBUG - core.main (00:06:43):
Initializing watchdog
DEBUG - core.main (00:06:43):
initializing updateCheck
INFO - core.main (00:06:43):
NVDA initialized
DEBUG - core.main (00:06:43):
entering wx application main loop
IO - speech.speak (00:06:43):
Speaking - Outlook Express - Brians mail lists'
IO - inputCore.InputManager.executeGesture (00:06:46):
Input: kb(desktop):windows+m
IO - speech.speak (00:06:47):
Speaking list'
IO - speech.speak (00:06:47):
Speaking 12 of 58'
IO - inputCore.InputManager.executeGesture (00:06:50):
Input: kb(desktop):w
IO - speech._speakSpellingGen (00:06:50):
Speaking character u'w'
DEBUG - queueHandler.registerGeneratorObject (00:06:50):
Adding generator 1
IO - speech.speak (00:06:50):
Speaking 18 of 58'
DEBUG - queueHandler.pumpAll (00:06:50):
generator 1 finished
IO - inputCore.InputManager.executeGesture (00:06:53):
Input: kb(desktop):enter
IO - speech.speak (00:06:59):
Speaking view'
IO - speech.speak (00:06:59):
Speaking 1 Audio expanded 1 of 7'
IO - speech.speak (00:07:00):
Speaking 5.623'
IO - inputCore.InputManager.executeGesture (00:07:06):
Input: kb(desktop):alt+f4
IO - speech.speak (00:07:07):
Speaking - Outlook Express - Brians mail lists'
IO - speech.speak (00:07:09):
Speaking - Application Error dialog The instruction at
"0x02ab1f3e" referenced memory at "0x00000004". The memory could not be
"read".\r\n\nClick on OK to terminate the program'

IO - speech.speak (00:07:09):
Speaking button'
IO - inputCore.InputManager.executeGesture (00:07:20):
Input: kb(desktop):control+alt+i
IO - speech.speak (00:07:21):
Speaking nvda with log terminal'
IO - speech.speak (00:07:21):
Speaking [u'blank']
INFO - core.main (00:07:21):
Exiting

It seems that anything that exits has the error. The exceptions in snap 21 are IE and Explorer now, though if you continue long enough funnies start to happen as nvda is obviously confused.
Live mail does it, but Outlook express seems not to.
Msse does not probably due to being already running.
Interestingly, at shut down its only the dolphin service I see get into trouble.
I hav also tried a new version froma downloaded snap as a portable with default user configs and this is also doing it, so the idea I saw of rogue app modules is not the cause. Back in the mists of time I'm sure I recall a problem very much like this, but its a long while ago.

Is this memory location its claiming not to be able to read real or is it just something locking up a file that is using it?

@nvaccessAuto
Copy link
Author

Comment 14 by briang1 on 2012-04-21 07:27
Is there any way that the already detected operating system as used to mask out the problems in xp trying to use the new routines, could be used to branch to the relevent routines for each OS in this case also?
Or would having two of everything make it a lot bigger?

@nvaccessAuto
Copy link
Author

Comment 15 by mdcurran on 2012-04-21 07:56
We'd have to have two copies of nvdaHelperRemote.dll and each of the vbufBackend dlls probably in order to do that.
However, further testing has shown that even when completely loading minHook.dll dynamically (no delay loading, no linking, just LoadLibrary and GetProcAddress) the issue still is happening with taskmgr.
Its now really looking like some kind of memory corruption.
I'll test a few more things this weekend but its possible I might just have to back out the merge and keep it as a separate branch... and perhaps get some willing XP users to try some specific snapshots for that branch.

@nvaccessAuto
Copy link
Author

Comment 16 by mdcurran on 2012-04-21 08:06
Brian, do you only see those log lines about error registering class object in taskmgr.exe since 5120? I understand the crash did not happen before these, but what about those debug warnings?
If they did not, then I'm really barking up the wrong tree hear and I should be looking in a rather different place.
These log lines are appearing when ever the focus moves in taskmgr, not just when it exits.

@nvaccessAuto
Copy link
Author

Comment 17 by briang1 on 2012-04-21 08:25
The lines have been there for a long time as far as I'm aware, here is 2012.1.
IO - inputCore.InputManager.executeGesture (09:17:56):
Input: kb(desktop):control+shift+escape
DEBUGWARNING - RPC process 1084 (taskmgr.exe) (09:17:56):
Thread 1828, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1084 (taskmgr.exe) (09:17:56):
Thread 1828, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1084 (taskmgr.exe) (09:17:56):
Thread 1828, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1084 (taskmgr.exe) (09:17:56):
Thread 1828, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

DEBUGWARNING - RPC process 1084 (taskmgr.exe) (09:17:56):
Thread 1828, nvdaHelper\build\x86\remote\ia2Support.cpp, installIA2Support, 81:
Error registering class object, code -2147221008

IO - speech.speak (09:17:56):
Speaking Task Manager dialog'
IO - speech.speak (09:17:56):
Speaking list'
IO - speech.speak (09:17:56):
Speaking User Name: brian gaff, CPU: 01 , CPU Time: 0:00:03, Mem Usage: 28,980 K, Peak Mem Usage: 31,544 K, Page Faults: 18,225, Threads: 19, I/O Reads: 2,690, I/O Writes: 22 1 of 36'
IO - inputCore.InputManager.executeGesture (09:18:08):
Input: kb(desktop):alt+f4
IO - speech.speak (09:18:08):
Speaking messages - Outlook Express - Brian's main account"
IO - speech.speak (09:18:08):
Speaking Express Message List list'
IO - speech.speak (09:18:08):
Speaking [Trac Subject: Re: NVDA #2254: NVDA causes errors on exit of some windows software, Received: 21/04/2012 09:06 1257 of 1257']
IO - inputCore.InputManager.executeGesture (09:18:13):
Input: kb(desktop):control+alt+i
INFO - core.main (09:18:14):
Exiting

I seem to recall Jamie saying they are not an issue, but as its hard to do a detailed log without them, not much I can do to preserve context otherwise.

@nvaccessAuto
Copy link
Author

Comment 18 by mdcurran on 2012-04-21 11:09
Right, now I believe its the fact that as nvdaHelperRemote no longer statically depends on minHook, and minHook is being loaded after nvdaHelperRemote, then (on XP anyway) at process exit minHook is being unloaded before nvdaHelperRemote. In deed if I compile with a debug CRT, on process exits I get a runtime error saying that the CRT is not initialized, I'd say that's probably minHook's crt, not NVDAHelperRemote's.

I guess, minHook is going to have to need its own dllmain, which on process exit can clean up all hooks and uninitialize. But, it seems that Win 7/8 must still be unloading minHook later as we don't see the error. But not sure if that is trust worthy. So have no idea yet how to work out at runtime which one unloaded first.

@nvaccessAuto
Copy link
Author

Comment 19 by mdcurran on 2012-04-21 12:22
More debugging shows me that on both XP and win8 (therefore I hope Vista and 7) that if minhook is loaded later, it is always unloaded first on process exit. No idea though why on 7/8 this does not actually cause exceptions r memory corruption. Perhaps 7/8 does not bother uninitializing crts on proces exit or something.
8484214 now tries to handle this by making sure that minhook itself disables hooks and uninitializes in its own dllmain on process exit, and nvdaHelperRemote's apiHooks_terminate does not bother doing anything if the process is exiting.
This fixes the taskmgr crash on XP for me. Lets hope it does for others.
A new snapshot is building now.

@nvaccessAuto
Copy link
Author

Comment 20 by Palacee_hun on 2012-04-21 13:19
Reading all this I got more and more puzzled why my XP system seems totally immune to all these issues mentioned in this ticket. I am using Main-5120 for more than a day and not a single crash, nor any error messages in my log similar to those cited here. TaskMgr works totally flawlessly too, exiting processes from there works fine, not a trace of errors. Furthermore it struck me that some users got error messages when only navigating TaskMgr even with 2012.1 which never happened to me.
So I found it was high time to investigate my WinXP SP3 from a little closer. I queried Windows version with winver.exe built-in utility and this is what I got:
[(build: 2600.xpsp.080413-2111)
[[BR]([BR]]
5.1)]
So my system is indeed a post SP3 WinXP Home with updates applied to the date 13 April 2008. I was unaware of this as my Windows had been preinstalled on my computer when I bought it and NVDA just reported it to be SP3. Other system details are in Comment 8.
So there are indeed updates to official service pack 3 in which Microsoft sorted out some things that lie behind this mess described in this ticket.
[[BR]]
I hope this might shed some light on this issue and furthermore I suggest that everybody who comments on that ticket with log excerts also mention their exact Windows XP version from winver. It can easily be copy-pasted from winver dialog with NVDA+F9/F10.

@nvaccessAuto
Copy link
Author

Comment 21 by briang1 on 2012-04-21 14:16
OK for the record then..

Version 5.1 (Build 2600.xpsp_sp3_gdr.111025-1629 : Service Pack 3)
Copyright © 2007 Microsoft Corporation
Copyright © 2007 Microsoft Corporation
Physical memory available to Windows:
1,572,336 K
However, as I have all the security and suggested updates on my machine, some saying they are sp4, I'd assume that in fact the copies are the same in the end. all winver seems to put is the date of the last update either on the cd or the service pack, ignoring the updates thereafter.

I still strongly suspect that you have perhaps a multithreading processor or a faster machine. I'll go look at the new snap in a while.

@nvaccessAuto
Copy link
Author

Comment 22 by briang1 (in reply to comment 19) on 2012-04-21 14:37
Replying to mdcurran:

More debugging shows me that on both XP and win8 (therefore I hope Vista and 7) that if minhook is loaded later, it is always unloaded first on process exit. No idea though why on 7/8 this does not actually cause exceptions r memory corruption. Perhaps 7/8 does not bother uninitializing crts on proces exit or something.

8484214 now tries to handle this by making sure that minhook itself disables hooks and uninitializes in its own dllmain on process exit, and nvdaHelperRemote's apiHooks_terminate does not bother doing anything if the process is exiting.

This fixes the taskmgr crash on XP for me. Lets hope it does for others.

A new snapshot is building now.

OK, I have just downloaded snap 22 on this machine and so far no funny errors on any programs. Also the error on terminating the dolserv service has stopped as well. I'll now go and install it on other machines and see what happens.

@nvaccessAuto
Copy link
Author

Comment 23 by PZajda on 2012-04-21 17:53
After snapshot 22 I have not crash on exiting programmes, but NVDA start slowly and is sometimes slower. I'll try to find something I could reproduce because it is very randomness.

It seems restarting NVDA allows to make him quicker, strange...

@nvaccessAuto
Copy link
Author

Comment 24 by briang1 on 2012-04-21 19:05
Hmm, not sure about this one. I've certainly been having issues with the fresh copy losing bits of its capabilities, for example the menus don't work or edit areas seem to not speak when you cursor through them,but both of these oddities were around before as I recall, they are just more common now. As you say, a reboot of nvda normally fixes things.

I think this may well have been happening previously but the current situation has made us notice it more?

@nvaccessAuto
Copy link
Author

Comment 25 by Palacee_hun (in reply to comment 21) on 2012-04-21 21:27
No, single-core, single-threaded CPU. Tested with CPU-Z external utility v1.58. 1.81 Ghz stock frequency, not overclocked, not overclockable at all anyway.
[to [comment:21 briang1]([BR]]
Replying):

OK for the record then..

Version 5.1 (Build 2600.xpsp_sp3_gdr.111025-1629 : Service Pack 3)

Copyright © 2007 Microsoft Corporation

Copyright © 2007 Microsoft Corporation

Physical memory available to Windows:

1,572,336 K

However, as I have all the security and suggested updates on my machine, some saying they are sp4, I'd assume that in fact the copies are the same in the end. all winver seems to put is the date of the last update either on the cd or the service pack, ignoring the updates thereafter.

I still strongly suspect that you have perhaps a multithreading processor or a faster machine. I'll go look at the new snap in a while.

@nvaccessAuto
Copy link
Author

Comment 26 by briang1 on 2012-04-23 08:43
I think, having seen no other reports of this error all weekend on any group or list, this was fixed by the snap 22 and this could be closed. However several are now getting the errors I refer to in ticket 58, but its very intermittent, so is probably only under certain specific conditions under which watchdog throws it up as a ding.

Changes:
State: closed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant