OSDN Git Service
Michael Curran [Fri, 8 Apr 2011 22:25:42 +0000 (08:25 +1000)]
Merged main
Michael Curran [Thu, 7 Apr 2011 07:54:13 +0000 (17:54 +1000)]
characterProcessing.getCharacterDescription: For now remove the code that uses unicodedata if there is no english character description available. Although this was very useful for English, it caused problems in other locales as it eclipsed data in characterSymbols. In future we may consider a better way to handle this, but for now its gone.
Peter Vágner [Wed, 6 Apr 2011 12:46:57 +0000 (14:46 +0200)]
Added finnish characters descriptions,
Updated finnish user guide and changes,
Added german characters descriptions,
Updated turkish language file,ä
Peter Vágner [Wed, 6 Apr 2011 09:21:48 +0000 (11:21 +0200)]
Added croatian characters descriptions.
Peter Vágner [Wed, 6 Apr 2011 09:06:20 +0000 (11:06 +0200)]
Added turkish characters descriptions.
Peter Vágner [Wed, 6 Apr 2011 07:44:32 +0000 (09:44 +0200)]
Added czech and slovak characters descriptions
Aleksey Sadovoy [Wed, 6 Apr 2011 07:10:00 +0000 (10:10 +0300)]
Added Russian character descriptions from wikipedia.
Michael Curran [Wed, 6 Apr 2011 06:01:00 +0000 (16:01 +1000)]
Added Japanese character descriptions
Michael Curran [Wed, 6 Apr 2011 04:50:14 +0000 (14:50 +1000)]
Merged character descriptions branch (#55). This implements support in NVDA to hear a description or examples of a given character. Pressing review current character twice will provide a description of the character under the review cursor. Pressing review current word or review current line three times will spell the word or line but using descriptions. So far descriptions for english, french and traditional Chinese (Taiwan) have been added. Descriptions are chosen based on NVDA's current locale and or the English locale. In future we may consider tying this to the synth's current language. This implementation is especially important to Asian languages as it now provides them with a way to have example phraises spoken for a given symbol, as for in Chinese, many symbols can sound the same but have different meanings.
Michael Curran [Wed, 6 Apr 2011 04:48:28 +0000 (14:48 +1000)]
Update english changes file
Michael Curran [Wed, 6 Apr 2011 04:41:53 +0000 (14:41 +1000)]
Update english changes and user guide files to mention character descriptions
Michael Curran [Wed, 6 Apr 2011 04:25:22 +0000 (14:25 +1000)]
characterProcessing.getCharacterDescription:
* If locale data cannot be found for a particular locale and its not english, fall back to English by calling getCharacterDescription this time with 'en'.
* If a character description can not be found in the locale data for a given locale, and its not english, fall back to English by calling getCharacterDescription with 'en'.
* If a character description can not be found in the english locale, instead look up the name in the unicode database. The unicode database is only available in english, however I think it is a useful fall back when there is nothing else to say.
* Finally if absolutely nothing can be found, then None is returned, which means that NVDA will just use the character itself.
Michael Curran [Tue, 5 Apr 2011 23:56:46 +0000 (09:56 +1000)]
setup.py: include *.dic files in locale directories.
Michael Curran [Tue, 5 Apr 2011 23:44:29 +0000 (09:44 +1000)]
Added characterDescriptions.dic for French
Michael Curran [Tue, 5 Apr 2011 23:42:48 +0000 (09:42 +1000)]
Merged main
Michael Curran [Tue, 5 Apr 2011 23:24:37 +0000 (09:24 +1000)]
Merged releases/2011.1 branch
Michael Curran [Tue, 5 Apr 2011 23:22:42 +0000 (09:22 +1000)]
Merged fix for #1425 as the user guide states it works this way.
Peter Vágner [Tue, 5 Apr 2011 06:48:44 +0000 (08:48 +0200)]
Updated czech changes.
Michael Curran [Tue, 5 Apr 2011 05:58:06 +0000 (15:58 +1000)]
NVDAObjects.window.edit.EditTextInfo._getWordOffsets: re-implement word offset support for standard edit controls (pre richEdit) by following the Microsoft documented rule of only breaking on spaces. This code replaces the extremely problematic implementation that used to move the caret by sending key presses, in order to calculate the word offsets.
This implementation of course is much faster, and should no longer cause problems for braille or other code that may want to expand to word during typing.
Microsoft KB article: http://support.microsoft.com/kb/109551
Michael Curran [Tue, 5 Apr 2011 00:19:02 +0000 (10:19 +1000)]
Fix for 1425 (NVDA should make it easier to discover unbound gestures): now when in input help mode, all gestures are logged whether or not they are bound to a script. If bound to a script, then that information is also appended to the log message.
Michael Curran [Mon, 4 Apr 2011 23:07:21 +0000 (09:07 +1000)]
Added a Traditional Chinese (Taiwan) characterDescriptions.dic which provides examples for many traditional Chinese characters. Based on data from the libchewingdata project: libchewingdata.openfoundry.org/
Michael Curran [Mon, 4 Apr 2011 23:06:01 +0000 (09:06 +1000)]
Added an English characterDescriptions.dic which provides the standard English phonetic alphabet.
Peter Vágner [Mon, 4 Apr 2011 09:28:40 +0000 (11:28 +0200)]
Updated czech changes,
Updated finnish user guide.
Michael Curran [Mon, 4 Apr 2011 07:27:14 +0000 (17:27 +1000)]
Merged main
Michael Curran [Mon, 4 Apr 2011 07:14:05 +0000 (17:14 +1000)]
characterProcessing.CharacterDescriptions.getCharacterDescription: better handle the case where there are no descriptions for the given character. Now returns None, rather than raising an exception.
Michael Curran [Mon, 4 Apr 2011 06:56:43 +0000 (16:56 +1000)]
characterProcessing: Added a LocaleDataMap class, which can handle the dynamic loading and storage and lookup of data objects for particular locales. The reason for this abstraction is that symbol Framework can hopefully make use of this also. Also added CharacterDescriptions, which is a class when instanciated with a locale, loads characters and their descriptions from a file and allows lookup. Based on code by Timothy Lee. Finally the module-level getCharacterDescriptions makes use of all this completing its implementation. Note though that LocaleDataMap could be improved to do timed caching (removing old objects to save memory).
Michael Curran [Mon, 4 Apr 2011 06:43:24 +0000 (16:43 +1000)]
speech: for now only use NVDA's locale for character descriptions, the synth language can still not be entirely trusted yet.
Michael Curran [Mon, 4 Apr 2011 06:38:10 +0000 (16:38 +1000)]
speech: No longer handle character descriptions in processSymbol, instead handle it directly in _speakSpellingGen. This means that processSymbol again only takes a character, no longer locale or useCharacterDescriptions. Fixes bugs where certain symbols were no longer being spoken when moving by word or selecting text etc.
Also, force characters to lowercase when looking up their description.
Michael Curran [Mon, 4 Apr 2011 00:54:26 +0000 (10:54 +1000)]
* speech.speakSpelling, speech._speakSpellingGen and speech.processSymbol now all take a locale argument and a useCharacterDescriptions argument. In the case of speakSpelling both of these are optional, but for the other internal functions these are manditory. SpeakSpelling if locale is not provided defaults to either the current synth's language if it has one, or instead NVDA's current locale. SpeakSpelling's useCharacterDescriptions by default is false.
* speech.processSymbol: if useCharacterDescriptions is True, then characterProcessing.getCharacterDescription is used to fetch a character description, and if it finds one, it is returned straight away. If it does not find one, or character descriptions were not requested, then the function stays as is.
* script_review_current{word|Line} scripts now pass useCharacterDescriptions as true to speech.speakSpelling if the script has been executed more than twice (i.e. pressed three times). So, once says the string as is, twice spells the string, three times or more spells with character descriptions.
* GlobalCommands.script_review_currentCharacter: pressing twice now announces the character with character descriptions, and three times now speaks the hex representation.
Michael Curran [Sun, 3 Apr 2011 23:48:47 +0000 (09:48 +1000)]
Fix syntax error
Michael Curran [Sun, 3 Apr 2011 23:45:18 +0000 (09:45 +1000)]
Added a characterProcessing module, which so far contains a getCharacterDescription function. This function takes alocale and a character, and returns a description for the given character in the given locale. For now the descritpion returned is hardcoded as the locale + the character, just for testing.
Peter Vágner [Fri, 1 Apr 2011 12:48:14 +0000 (14:48 +0200)]
Updated japanese language file,
Updated french changes.
Peter Vágner [Thu, 31 Mar 2011 19:24:43 +0000 (21:24 +0200)]
Updated italian language file and changes,
Updated finnish changes,
Updated czech user guide.
Peter Vágner [Thu, 31 Mar 2011 04:59:08 +0000 (06:59 +0200)]
arabic changes for 2011.1.1,
Peter Vágner [Thu, 31 Mar 2011 04:58:11 +0000 (06:58 +0200)]
Updated danish language file
James Teh [Wed, 30 Mar 2011 20:56:15 +0000 (06:56 +1000)]
Merge 2011.1.
Mesar Hameed [Wed, 30 Mar 2011 20:38:01 +0000 (21:38 +0100)]
arabic changes for 2011.1.1, small modifications to userguide, critical small changes to interface, and the new changes.
Peter Vágner [Wed, 30 Mar 2011 11:51:13 +0000 (13:51 +0200)]
Added turkish braille table into the braille settings dialog.
Peter Vágner [Wed, 30 Mar 2011 11:45:01 +0000 (13:45 +0200)]
Updated danish user guide
Peter Vágner [Wed, 30 Mar 2011 06:25:29 +0000 (08:25 +0200)]
Updated slovak changes.
Michael Curran [Wed, 30 Mar 2011 04:59:56 +0000 (14:59 +1000)]
No longer register for UIA textChange and textSelectionChange automation events, as this was causing major performance issues in WPF applications using .net 3.5 [apparently]. Certainly for one particular wpf app (#1205) no longer registering makes the app now quite usable. However, due to no longer receiving textSelectionChange, NVDA must use EditableTextWithoutAutoSelectDetection for UIA editable text controls. May partially fix #1205 and #1437, though no testing on XP as yet.
James Teh [Wed, 30 Mar 2011 01:15:09 +0000 (11:15 +1000)]
Update version string.
Michael Curran [Wed, 30 Mar 2011 00:57:31 +0000 (10:57 +1000)]
nvda.pyw: duplicate getInputDesktopName from nvda_service so that it does not have to be imporated (it pulls in too much unrelated code).
Michael Curran [Wed, 30 Mar 2011 00:36:40 +0000 (10:36 +1000)]
nvda.pyw: protect against multiple instances of NVDA being accidentilly started by using a mutex. The mutex name is specific to the desktop and session (hopefully). This new feature should be tested well with the service on secure desktops.
James Teh [Wed, 30 Mar 2011 00:31:25 +0000 (10:31 +1000)]
What's New: Add entry for moveToParentTreeInterceptor fix.
Peter Vágner [Tue, 29 Mar 2011 12:30:47 +0000 (14:30 +0200)]
Updated finnish language file, user guide, language strings for the NVDA installer and changes,
Updated slovak changes.
James Teh [Tue, 29 Mar 2011 11:38:16 +0000 (21:38 +1000)]
Merge 2011.1.
James Teh [Tue, 29 Mar 2011 08:56:12 +0000 (18:56 +1000)]
What's New: Minor tweak.
James Teh [Tue, 29 Mar 2011 08:54:37 +0000 (18:54 +1000)]
Disable logging on secure Windows screens, as this is a security risk.
Fixes #1435.
James Teh [Tue, 29 Mar 2011 07:51:39 +0000 (17:51 +1000)]
NVDAHelper.RemoteLoader64: The file descriptors returned by the fileno() method of Python file objects are not file handles. Therefore, when duplicating the nul file handle for the loader process, query the file handle with msvcrt.get_osfhandle().
This never caused a problem because there always seemed to be a file handle with the same number as the file descriptor we were supplying, though it's very possible it was the wrong one. However, once we have no log file, this fails with an invalid handle error.
In any case, this was *very* incorrect. I'm surprised it hasn't caused us problems before this.
Michael Curran [Tue, 29 Mar 2011 05:46:04 +0000 (15:46 +1000)]
developerGuide: remove double spacing at the beginning of each line via postProc as txt2tags seems to place two spaces before ach line of preformatted text (i.e. our code samples).
James Teh [Tue, 29 Mar 2011 05:25:38 +0000 (15:25 +1000)]
Report space pressed with modifiers other than shift as a command key.
Fixes #1424.
James Teh [Tue, 29 Mar 2011 05:23:47 +0000 (15:23 +1000)]
Don't report shifted characters as command keys.
Fixes #1422.
Michael Curran [Tue, 29 Mar 2011 03:05:32 +0000 (13:05 +1000)]
NVDAObjects.IAccessible.IAccessible.findOverlayClasses: when checking if GeckoPluginWindow should be used, also make sure the childID is 0.
Michael Curran [Tue, 29 Mar 2011 02:50:57 +0000 (12:50 +1000)]
Fix for #1429 (Firefox 4: Can't escape embedded objects using moveToParentTreeInterceptor script):
* NVDAObjects.IAccessible.adobeFlash.Root NVDAObject's parent property: ass accParent is broken on the Adobe Flash root client, instead override it to properly return its window root IAccessible.
* IAccessibleHandler: add NAVRELATION_EMBEDS Mozilla Gecko constant.
* NVDAObjects.IAccessible.mozilla: add a GeckoPluginWindowRoot overlay NVDAObject, which should be used on the window root IAccessible NVDAObject for GeckoPlugin windows (the newer plugin window class in Firefox). This NvDAObject overrides the parent property to appropriately return the document this plugin is embedded in. It does this by fetching the original parent, checking to see if it has a Gecko version, and in deed if its not version 1 (therefore 2 or higher) and then using Mozilla's NAVRELATION_EMBEDS to fetch the active document.
* NVDAObjects.IAccessible.IAccessible.findOverlayClasses: if the window class is GeckoPluginWindow and the object's MSAA objectID is 0 (window root) then use the mozilla.GeckoPluginWindowRoot overlay class.
* globalCommands.GlobalCommands.script_moveToParentTreeInterceptor: rather than just using eventHandler.executeEvent directly to fake the focus event, instead use wx.CallLater to run the executeEvent call 50 ms in to the future. This is needed for Firefox 4 to get around the issue where a focus event is fired on the top window after focus moves back to the document.
Michael Curran [Mon, 28 Mar 2011 03:51:13 +0000 (13:51 +1000)]
Fix for #837 (merge move caret script in to double press of move focus
Michael Curran [Fri, 25 Mar 2011 11:00:38 +0000 (21:00 +1000)]
Merged releases/2011.1 branch
Peter Vágner [Fri, 25 Mar 2011 10:02:51 +0000 (11:02 +0100)]
Updated czech language file and fixed a problem which caused czech quick command keys not to be generated for the official version.
James Teh [Fri, 25 Mar 2011 07:12:04 +0000 (17:12 +1000)]
api.copyToClip(): Catch exceptions from OpenClipboard and return False as callers expect.
Fixes exception when using review_copy script in secure mode.
Michael Curran [Fri, 25 Mar 2011 06:57:12 +0000 (16:57 +1000)]
nvdaHelper: allow nvdaHelperRemoteLoader to compile again by fixing the argument to injection_initialize... oops missed this.
Michael Curran [Fri, 25 Mar 2011 05:55:55 +0000 (15:55 +1000)]
Update changes file to mention clipboard fix.
Michael Curran [Fri, 25 Mar 2011 05:39:30 +0000 (15:39 +1000)]
Do not allow copy/paste at all inside NVDA's UI while in secure mode (e.g. when NVDA is running on the secure desktop -- lock screen -- windows logon). To do this, Only in the NVDA process itself, and only in secure mode, nvdaHelperRemote hooks the OpenClipboard Windows API function and forces it to return false, and does not call the original OpenClipboard. This simply denies access to the clipboard for NVDA's process. Fixes #1421.
James Teh [Wed, 23 Mar 2011 06:19:33 +0000 (16:19 +1000)]
Don't include the Donate menu item in secure mode, as this is a security risk.
Fixes #1419.
Michael Curran [Wed, 23 Mar 2011 01:52:22 +0000 (11:52 +1000)]
NVDAHelper gdiHooks.cpp: remove all comments from fake functions refering to keeping NVDAHelperRemote from unloading. This code was taken out a long time ago therefore the comments are not valid.
Michael Curran [Wed, 23 Mar 2011 01:46:15 +0000 (11:46 +1000)]
nvdaHelper fake_createCompatibleDc: don't create a new displayModel if the real_createCompatibleDc returns NULL. Very rare, but could happen.
James Teh [Tue, 22 Mar 2011 22:31:14 +0000 (08:31 +1000)]
Report space pressed with modifiers other than shift as a command key.
Returning early for space was causing any key press including space to be treated as a non-command key.
Fixes #1424.
James Teh [Tue, 22 Mar 2011 22:17:03 +0000 (08:17 +1000)]
Don't report shifted characters as command keys.
I seem to have forgotten that generalizedModifiers includes the extended flag as well as the VK code. Oops.
Fixes #1422.
James Teh [Tue, 22 Mar 2011 01:20:39 +0000 (11:20 +1000)]
Don't include the Donate menu item in secure mode, as this is a security risk.
Fixes #1419.
James Teh [Fri, 11 Mar 2011 00:10:55 +0000 (10:10 +1000)]
Merge main.
James Teh [Mon, 7 Mar 2011 01:49:41 +0000 (11:49 +1000)]
Update version string.
Aleksey Sadovoy [Mon, 7 Mar 2011 01:18:47 +0000 (03:18 +0200)]
Updated Russian language strings for the NVDA installer
Aleksey Sadovoy [Mon, 7 Mar 2011 01:06:14 +0000 (03:06 +0200)]
Update Russian language file and changes, update contributors.
Peter Vágner [Fri, 4 Mar 2011 21:27:56 +0000 (22:27 +0100)]
Updated dutch language file and added user guide in dutch.
Bram Duvigneau [Fri, 4 Mar 2011 19:15:45 +0000 (20:15 +0100)]
Merge
Bram Duvigneau [Fri, 4 Mar 2011 19:13:32 +0000 (20:13 +0100)]
Updated Dutch translation
Bram Duvigneau [Fri, 4 Mar 2011 15:15:12 +0000 (16:15 +0100)]
Added Dutch translation of the user guide
Peter Vágner [Fri, 4 Mar 2011 13:01:20 +0000 (14:01 +0100)]
Updated japanese language file.
Peter Vágner [Thu, 3 Mar 2011 17:47:07 +0000 (18:47 +0100)]
Updated simplified chinese user guide and the language strings for the NVDA installer.
Peter Vágner [Thu, 3 Mar 2011 14:33:27 +0000 (15:33 +0100)]
Updated traditional chinese language file and user guide,
Updated simplified chinese language file.
Peter Vágner [Wed, 2 Mar 2011 21:53:56 +0000 (22:53 +0100)]
Updated finnish user guide and changes.
Peter Vágner [Wed, 2 Mar 2011 06:11:24 +0000 (07:11 +0100)]
Updated croatian language file.
Michael Curran [Wed, 2 Mar 2011 05:41:41 +0000 (15:41 +1000)]
sconstruct: instead of copying nvdaHelper docs to the devDocs directory, move it. Also make sure that when cleaning devDocs, the output/devDocs directory is totally removed.
Michael Curran [Wed, 2 Mar 2011 05:37:34 +0000 (15:37 +1000)]
sconstruct: developerGuide.html is now placed in output/devDocs. scons devDocs is now an alias simply for output/devDocs, which means that scons devDocs will now build anything that ends up there, not just nvdaHelper docs.
Peter Vágner [Tue, 1 Mar 2011 19:30:01 +0000 (20:30 +0100)]
Updated contributers file,
Peter Vágner [Tue, 1 Mar 2011 19:28:22 +0000 (20:28 +0100)]
Updated italian language file, user guide and changes.
Mesar Hameed [Tue, 1 Mar 2011 11:59:55 +0000 (11:59 +0000)]
Updated contributers file, Arabic translation team 2011.1
James Teh [Tue, 1 Mar 2011 11:14:04 +0000 (21:14 +1000)]
Update What's New.
James Teh [Tue, 1 Mar 2011 11:08:29 +0000 (21:08 +1000)]
mshtml and gecko_ia2 vbuf backends: Treat inline-block elements as inline elements instead of block elements; i.e. make them appear on the same line instead of separate lines.
Partial fix for #1378.
James Teh [Tue, 1 Mar 2011 11:00:48 +0000 (21:00 +1000)]
Typo and linguistic fixes.
Michael Curran [Tue, 1 Mar 2011 09:20:59 +0000 (19:20 +1000)]
API documentation in html can now be generated for nvdaHelper. You will need doxygen: see the readme.txt for where to get it. To generate the documentation, run the command: scons devDocs
Specific changes:
* Added a doxygen scons tool, based on the tool from the SCons wiki, with various changes by me. Major changes were auto-detection of doxygen using Windows registry, and some fixes to the way it generates its scons target nodes -- they were relative to the wrong directory.
* Updated nvdaHelper/doxyfile, it still said Virtual Buffer Library 2.0!
* sconstruct: build documentation using nvdaHelper/doxyfile, and copy it to output/devDocs/nvdaHelper. The devDocs alias can be used to build this. More documentation such as epydoc for NVDA itself will also be added at a later stage.
Peter Vágner [Tue, 1 Mar 2011 07:34:28 +0000 (08:34 +0100)]
Updated hungarian user guide and changes,
Updated galician language file and user guide.
Peter Vágner [Mon, 28 Feb 2011 20:40:22 +0000 (21:40 +0100)]
userGuide: Fixed a typo (thx Aron)
Peter Vágner [Mon, 28 Feb 2011 07:24:46 +0000 (08:24 +0100)]
Updated Arabic locale, (userguide, changes ...
Mesar Hameed [Mon, 28 Feb 2011 06:26:12 +0000 (06:26 +0000)]
Updated Arabic locale, (userguide, changes and nvda.po).
Peter Vágner [Sun, 27 Feb 2011 23:26:17 +0000 (00:26 +0100)]
Added changes in polish,
Updated polish language file, language strings for the NVDA installer, user guide,
Updated german language file, user guide and changes,
Updated danish language file,
Updated czech language file, language strings for the NVDA installer and user guide,
Updated slovak user guide and changes.
Peter Vágner [Fri, 25 Feb 2011 12:00:41 +0000 (13:00 +0100)]
vkCodes: fixed a typo which in turn fixes anouncement of left control key press in the NVDA's user language.
Michael Curran [Fri, 25 Feb 2011 02:56:14 +0000 (12:56 +1000)]
Collapsing a combo box in a virtual buffer when focus mode has been forced with NVDA+space no longer auto-switches back to browse mode. Fixes #1386
James Teh [Fri, 25 Feb 2011 01:37:37 +0000 (11:37 +1000)]
English User Guide: Remapping Key Assignments...: Clarify that each entry key can only occur once per section. Note that gesture identifiers ending in a comma must be enclosed in quotes.
James Teh [Fri, 25 Feb 2011 01:25:51 +0000 (11:25 +1000)]
Merge 2011.1.
James Teh [Thu, 24 Feb 2011 22:19:52 +0000 (08:19 +1000)]
Revert recent change to User Guide, as we're well into string freeze now and it's a relatively minor clarification.