OSDN Git Service
Michael Curran [Sun, 4 Dec 2011 22:10:20 +0000 (08:10 +1000)]
WordDocumentTextInfo: remove _getFormatFieldAtRange and _expandToFormatRange methods as these are no longer used.
Michael Curran [Sun, 4 Dec 2011 22:01:51 +0000 (08:01 +1000)]
Merged main
Dinakar T.D. [Sun, 4 Dec 2011 21:57:59 +0000 (21:57 +0000)]
Tamil (ta):
userGuide.t2t matching 4820
nvda.po
Jani Kinnunen [Sun, 4 Dec 2011 21:57:29 +0000 (21:57 +0000)]
Finnish (fi):
userGuide.t2t matching 4820
changes.t2t matching 4816
nvda.po
Mesar Hameed [Sun, 4 Dec 2011 21:50:08 +0000 (21:50 +0000)]
Brazilian Portuguese (pt_BR):
userGuide.t2t matching 4817
changes.t2t matching 4816
nvda.po
Authors: Marlin Rodrigues <marlincgrodrigues@yahoo.com.br>, Cleverson Casarin Uliana <clever92000@yahoo.com.br>
Mesar Hameed [Sun, 4 Dec 2011 21:30:57 +0000 (21:30 +0000)]
Corrected simple formatting mistakes in po files for the following languages
No errors outstanding:
af_ZA, fr, he, hr, ka, pl and th.
hi: fixed some mistakes, many remain. Contributer needs to look and fix.
Michael Curran [Sun, 4 Dec 2011 11:32:35 +0000 (21:32 +1000)]
Use DisplayModelTextInfo for UIA objects with a role of window, that have a valid nativeWindowHandle, and do not support the UIA text pattern. This allows flat review of many Windows 8 apps (Wordpad, Windows explorer etc).
Peter Vágner [Fri, 2 Dec 2011 13:54:25 +0000 (14:54 +0100)]
Updated slovak language file, user guide and changes.
James Teh [Thu, 1 Dec 2011 21:35:45 +0000 (07:35 +1000)]
Remove more extraneous whitespace. Translators can ignore this commit.
Sebastian Kruber [Thu, 1 Dec 2011 10:44:34 +0000 (20:44 +1000)]
Add hedo MobilLine USB braille display driver.
Fixes #1897.
James Teh [Thu, 1 Dec 2011 06:48:13 +0000 (16:48 +1000)]
English User Guide: More capitalisation and spacing fixes.
James Teh [Thu, 1 Dec 2011 06:38:41 +0000 (16:38 +1000)]
English User Guide: Cosmetic fixes: capitalisation, spacing and pucntuation.
James Teh [Thu, 1 Dec 2011 05:24:48 +0000 (15:24 +1000)]
User Guide: Fix punctuation.
James Teh [Thu, 1 Dec 2011 05:23:48 +0000 (15:23 +1000)]
User Guide: Clarification and examples for Object Navigation, Text Review and Browse Mode sections.
James Teh [Thu, 1 Dec 2011 04:52:54 +0000 (14:52 +1000)]
object navigation commands: Change the "No children" and "No parents" messages to be more user friendly. I missed this when I was changing the terminology for 2011.2.
Michael Curran [Thu, 1 Dec 2011 03:02:53 +0000 (13:02 +1000)]
_UIAHandler.UIAHandler.isUIAWindow: normalize the window class name before checking for it in badUIAWindowClassList. Also add all the possible RichEdit window classes to badUIAWindowClassList so that UIA is no longer used for them. Windows 8 introduces UIA support for RichEdit which is wonderful, but for now causes some problems in NVDA -- more our fault than Microsofts. For now we disable support and fall back to our older win32 RichEdit window message code. this will also allow the poedit appModule to work on Windows 8. It also stops double selection announcement in RichEdit documents in Windows 8
James Teh [Thu, 1 Dec 2011 01:12:00 +0000 (11:12 +1000)]
Remove unnecessary import.
Michael Curran [Wed, 30 Nov 2011 09:28:07 +0000 (19:28 +1000)]
Fix a bug that could very possibly cause a crash in a 64 bit application if that application provided glyphs to an ExtTextOut call and the OS was Windows 8 or the memory consumption was rather high (forcing memory allocations above the 32 bit boundary).
Specifically: some pointers were being cast to DWORD before being cast to their correct pointer type, which in a 64 bit app would cause the pointer to loose bits.
James Teh [Wed, 30 Nov 2011 07:34:10 +0000 (17:34 +1000)]
logHandler.Logger.exception: Don't output debugWarnings if the configured log level excludes them.
Michael Curran [Wed, 30 Nov 2011 01:04:59 +0000 (11:04 +1000)]
Merged main
James Teh [Tue, 29 Nov 2011 23:45:51 +0000 (09:45 +1000)]
sconstruct: Add output\devDocs\nvda target to generate documentation for the NVDA Python code using Epydoc.
Michael Curran [Tue, 29 Nov 2011 09:35:14 +0000 (19:35 +1000)]
Another charBaseline was missing for displayModel's textInRect
Michael Curran [Mon, 28 Nov 2011 10:20:11 +0000 (20:20 +1000)]
Merged main
Michael Curran [Mon, 28 Nov 2011 10:19:29 +0000 (20:19 +1000)]
Fix typo
Michael Curran [Mon, 28 Nov 2011 09:57:37 +0000 (19:57 +1000)]
WordDocumentTextInfo.getTextwithFields: return an empty list if the range is collapsed. Should fix #1949 (multiple bullets shown on braille display)
Michael Curran [Mon, 28 Nov 2011 09:55:41 +0000 (19:55 +1000)]
Merged main
Michael Curran [Mon, 28 Nov 2011 09:31:53 +0000 (19:31 +1000)]
speech.getCurrentLanguage: catch NotImplementeError from a synth's language property. _getAvailableVoices, and therefore language, is not implemented on the noSpeech synthDriver. Fixes #1963
Michael Curran [Mon, 28 Nov 2011 07:21:19 +0000 (17:21 +1000)]
Slightly reimplemented the way NVDA tracks the caret when using display model, in order to fix #1658 (tracking the caret while editing a cell in Excel with a Chinese font), and it could possibly allow access to a few other picky applications. Specific change:
* displayModelChunk_t now has a baseline member, rather than a baselineFromTop member. Its possible that previously baselineFromTop could have become incorrect if the chunk was ever clipped above the baseline. Now as baseline is no longer relative, this should not be a problem. Of course though copyRect needed to be changed to make sure that it tweeked baseline via deltaY, just like it does for rect->top and rect->bottom.
* GDIHooks's ExtTextOut_helper: tmAscent is in logical units not physical units, so make sure it gets converted to physical units when being used to calculate baseline. Previously baselineFromTop was most likely always way off what it should have been.
* displayModel_t::renderText: rather than just providing a rectangle for a character's location, provide left, top, right, bottom and baseline as well (we call it charLocation_t). However, for vertical whitespace added between lines and or at the top or bottom of a bounding rectangle, set that representing character's baseline to -1 so that its easily identified as added vertical whitespace.
* displayModel_getWindowTextInRect: also output the baseline of each character, rather than just left, top, right and bottom.
* EditableTextDisplayModelTextInfo._getCaretOffset: rewrite the matching rules to hopefully be more accurate and less picky. Firstly it tries to match on the first real character that is positioned so that the caret's left is anywhere to the left of that character's horizontal center (including before the character entirely), and where the caret's bottom sits on or below the character's baseline. Only if it does not find a character like this does it then reiterate over the characters this time searching for added vertical whitespace (blank lines) and uses the first character where the caret overlaps vertically and is entirely contained horizontally.
Please report if this change breaks any previously working text editors using displayModel. Or in deed if it allows access to otherones.
Michael Curran [Sun, 27 Nov 2011 12:01:27 +0000 (22:01 +1000)]
Reimplement table navigation in MS Word to allow navigation in tables with merged cells, and to improve over all speed while navigating. Fixes #1944. Specific changes:
* winword_getTextInRange: use table.rows.count and table.columns.count to calculate rowCount and columnCount respectively, rather than range.information as range.information only gave row and column counts within the current part of the table. Now the table reports the real maximum row and column count even if there are merged cells in the table.
* Winword document NVDAObject: implement a new _moveInTable method which handles the movement within a table, including announcing once it has moved. The scripts now just call this, rather than using their own custom code plus the old WorddocumentTextInfo._moveInTable which as been removed. This new _moveInTable method uses getTextInFields and also table.cell(row,column) which makes navigation much faster, and also allows it to work where merged cells exist.
Marlin Rodrigues [Sun, 27 Nov 2011 00:06:48 +0000 (00:06 +0000)]
Brazilian Portuguese (pt_BR):
userGuide.t2t matching 4799
changes.t2t
Peter Vágner [Sat, 26 Nov 2011 09:39:55 +0000 (10:39 +0100)]
Updated slovak language file and changes.
Michael Curran [Sat, 26 Nov 2011 07:11:54 +0000 (17:11 +1000)]
Updated what's new to mention MS Word improvements
Michael Curran [Sat, 26 Nov 2011 06:46:45 +0000 (16:46 +1000)]
Do not report comments in MS word while inside the comments story itself.
Michael Curran [Sat, 26 Nov 2011 06:39:46 +0000 (16:39 +1000)]
Report text color in Microsoft Word
James Teh [Fri, 25 Nov 2011 12:17:31 +0000 (22:17 +1000)]
Merge 2011.3.
James Teh [Fri, 25 Nov 2011 12:07:28 +0000 (22:07 +1000)]
Add comment about SAPI5 initialisation change so we remember why it was done. :)
Bernd Dorer [Fri, 25 Nov 2011 08:11:55 +0000 (08:11 +0000)]
German (de):
symbols.dic, fixing tick
James Teh [Fri, 25 Nov 2011 02:28:49 +0000 (12:28 +1000)]
versionInfo: Set version to 2011.3 (final).
Dinakar T.D. [Thu, 24 Nov 2011 14:09:06 +0000 (14:09 +0000)]
Tamil (ta):
userGuide.t2t matching 4784 (correcting terms)
symbols.dic (correcting terms)
Michael Curran [Thu, 24 Nov 2011 13:42:32 +0000 (23:42 +1000)]
Updated Chinese Taiwan po file
James Teh [Thu, 24 Nov 2011 11:32:49 +0000 (21:32 +1000)]
globalCommands: reportFormatting script: Add entry for reportFormatting to the format config, as all possible config entries must be present. Fixes exceptions.
Peter Vágner [Thu, 24 Nov 2011 10:09:59 +0000 (11:09 +0100)]
Updated croatian language file.
James Teh [Thu, 24 Nov 2011 08:25:59 +0000 (18:25 +1000)]
sapi5 synth driver: Set the voice before setting the audio output instead of the other way around.
It seems that SAPI 5 doesn't reset the audio parameters when you change the voice. Setting the audio output after setting the voice seems to fix this.
Fixes poor audio quality in some cases when using Microsoft Speech API version 5 voices with the audio output device set to something other than the default (Microsoft Sound Mapper).
Fixes #749.
Michael Curran [Thu, 24 Nov 2011 04:47:06 +0000 (14:47 +1000)]
Report heading levels 1 through 6 for paragraphs with a respective outlineLevel in MS Word.
Michael Curran [Thu, 24 Nov 2011 04:08:02 +0000 (14:08 +1000)]
Reimplement the way we expose MS Word tables so that they are control fields in the XML rather than format fields. Provide _startOfNode and _endOfNode attributes to both table and tabelCell control fields, and also now a level attribute for tables. Supporting level in MS Word tables provides a clearer idea of when you are in a nested table, although currently its still a bit more verbose than it should be. Plus converting the table implementation to control fields is much nicer (hopefully) for braille, though this should not affect speech at all.
Peter Vágner [Wed, 23 Nov 2011 11:01:34 +0000 (12:01 +0100)]
Merged 2011.3
Peter Vágner [Wed, 23 Nov 2011 10:58:46 +0000 (11:58 +0100)]
updated portuguese and czech language files
Michael Curran [Wed, 23 Nov 2011 09:49:40 +0000 (19:49 +1000)]
Support announcement of the existance of comments in Microsoft Word. (Says 'has comment' at the end of the text that has the comment). No automated way yet of reading the comment, except for going context menu -> edit comment. Added a reportComments config option to nvda.ini, but not bothered yet to add a GUI option.
Michael Curran [Wed, 23 Nov 2011 09:23:34 +0000 (19:23 +1000)]
Support reporting of links in MS Word
Michael Curran [Wed, 23 Nov 2011 08:00:37 +0000 (18:00 +1000)]
NVDAHelper's appendCharToXML: rather than just representing all invalid XML unicode characters with the unicode replacement character, instead output a speciall "unich" collapsed tag with a 'value' attribute containing the really unicode value as an unsigned int. XMLFormatting in NVDA now watches for these tags and converts the values back in to real unicode characters in the text. This again allows NVDA to announce page break characters in Microsoft Word.
Michael Curran [Wed, 23 Nov 2011 01:58:56 +0000 (11:58 +1000)]
Updated Chinese Taiwan userguide
Simone Dal Maso [Tue, 22 Nov 2011 10:58:49 +0000 (10:58 +0000)]
Italian (it):
nvda.po correction to progress bars.
Cagri Dogan [Tue, 22 Nov 2011 08:26:47 +0000 (08:26 +0000)]
Turkish (tr):
symbols.dic, localized negative numbers.
Michael Curran [Tue, 22 Nov 2011 08:26:29 +0000 (18:26 +1000)]
winword_getTextInRange: only report page numbers greater than 0. Stops NVDA announcing things like page -1 when in the comments story.
Michael Curran [Tue, 22 Nov 2011 08:22:53 +0000 (18:22 +1000)]
speech.speakObject: when speaking the line of text at the caret, and or from the start of the control if the caret can't be found, use a reason of REASON_CARET rather than the reason given to speakObject. This allows control fields etc to be spoken more appropriately. Most importantly, this now allows NVDA to announce the comments frame etc when focus lands in a word document while in the comments sotry.
Michael Curran [Tue, 22 Nov 2011 05:32:47 +0000 (15:32 +1000)]
WordDocumentTextInfo's getTextWithFields: check for the situation where winword_getTextInRange may give back no text (some strange error occured in early versions of word) and log a debug warning and return self.text instead.
Michael Curran [Tue, 22 Nov 2011 04:59:18 +0000 (14:59 +1000)]
winword_getTextInRange: add extra checks to make sure that IDispatch objects are not null even if a method that fetched it returned s_OK.
Michael Curran [Tue, 22 Nov 2011 03:33:03 +0000 (13:33 +1000)]
Use _com_dispatch_raw_method etc rather than _com_dispatch_method etc in our winword inproc support as the "raw" functions always return hResults where as the non-raw ones sometimes try to be smart and throw _com_error exceptions for various hresults. This may stop some of the crashes people are seeing with older versions of MS Word.
Michael Curran [Tue, 22 Nov 2011 03:08:38 +0000 (13:08 +1000)]
Support for announcing footnote and endnote numbers in MS Word documents. Currently some strange COM exceptions occuring when checking for an endnote when on a footnote or vice-versa so getfootnoteIndex and getEndnoteIndex bodies are surrounded by try catch(_com_error).
Mesar Hameed [Mon, 21 Nov 2011 23:54:16 +0000 (23:54 +0000)]
merge 2011.3
Jani Kinnunen [Mon, 21 Nov 2011 23:47:47 +0000 (23:47 +0000)]
Finnish (fi):
userGuide.t2t matching 4784
nvda.po
Dinakar T.D. [Mon, 21 Nov 2011 23:40:07 +0000 (23:40 +0000)]
Tamil (ta):
userGuide.t2t matching 4784
nvda.po
Bernd Dorer [Mon, 21 Nov 2011 23:35:00 +0000 (23:35 +0000)]
German (de):
changes.t2t matching 4761
symbols.dic using the inheritance from english.
James Teh [Sun, 20 Nov 2011 23:37:14 +0000 (09:37 +1000)]
Merge 2011.3.
Cleverson Casarin Uliana [Sun, 20 Nov 2011 16:35:00 +0000 (16:35 +0000)]
Brazilian Portuguese (pt_BR):
changes.t2t matching 4783
Michael Curran [Sun, 20 Nov 2011 03:20:52 +0000 (13:20 +1000)]
A first try at supporting annoucement of special parts of the document ssuch as the comments, header, footer, footnotes and endnotes stories. When entering one of these sections NVDA will now announce the section as a frame with an appropriate name reflecting the section type.
Bernd Dorer [Sat, 19 Nov 2011 22:15:03 +0000 (22:15 +0000)]
de:
symbols.dic
Hubert Meyer [Sat, 19 Nov 2011 15:55:42 +0000 (15:55 +0000)]
pl:
userGuide.t2t matching 4784
changes.t2t matching 4783
nvda.po
Jani Kinnunen [Sat, 19 Nov 2011 15:47:27 +0000 (15:47 +0000)]
fi:
userGuide.t2t small corrections
changes.t2t matching 4783
nvda.po
Michael Curran [Fri, 18 Nov 2011 08:49:40 +0000 (18:49 +1000)]
winword_getTextInRange: Make the range from the given offsets by fetching the current selection, copying it and moving the range to the offsets. This is rather than fetching the active document and creating a range with the offsets. This again allows NVDA to correctly read text when inside headers, footers, footnotes, endnotes and comments etc. (Note that this is not support for these sections but just making behaviour the same as the older outOfProc implementation.... support for announcing these will come later).
James Teh [Fri, 18 Nov 2011 01:15:00 +0000 (11:15 +1000)]
English User Guide: Fix typo. Translators should ignore this.
James Teh [Fri, 18 Nov 2011 00:58:45 +0000 (10:58 +1000)]
Merge 2011.3.
James Teh [Fri, 18 Nov 2011 00:57:58 +0000 (10:57 +1000)]
Add What's New entry for Tamil language, which was missed when it was added.
James Teh [Fri, 18 Nov 2011 00:49:41 +0000 (10:49 +1000)]
Use browse mode for full screen Adobe Flash content.
Quetzatl [Thu, 17 Nov 2011 23:28:19 +0000 (23:28 +0000)]
es:
symbols.dic fixing typo.
Author notified that complex section should idealy be empty so that the regexps can be inherited.
Only new rules should be added, instead of redifining rules.
Mesar Hameed [Thu, 17 Nov 2011 10:44:29 +0000 (10:44 +0000)]
it:
nvda.po fixing small problem with input help message.
Michael Curran [Thu, 17 Nov 2011 07:43:16 +0000 (17:43 +1000)]
Merged main
Michael Curran [Thu, 17 Nov 2011 07:42:12 +0000 (17:42 +1000)]
winword_getTextInRange: make sure never to allow formatconfig_reportLists to generateXMLAttribsForformatting if not on the first loop. reportLists is also now just a fomatConfig not an initialFormatconfig. Stops line-prefix from being repeated over and orver again for each word or format change. Also make all text xml-friendly.
Michael Curran [Thu, 17 Nov 2011 07:00:54 +0000 (17:00 +1000)]
winword_getTextWithFields: support tables.
Michael Curran [Thu, 17 Nov 2011 06:29:24 +0000 (16:29 +1000)]
winword_getTextInRange: support line-prefix (listFormat) and allow it to be enabled/disabled with formatconfig reportLists. Change WordDocumentTextInfo.getTextWithFields to use a static formatconfig flags map rather than calculating at runtime as we don't use all formatconfig keys and therefore things would be now out of order.
James Teh [Thu, 17 Nov 2011 02:58:06 +0000 (12:58 +1000)]
Pointless whitespace.
James Teh [Thu, 17 Nov 2011 02:52:51 +0000 (12:52 +1000)]
In browse mode, you can activate the object under the cursor using a braille display. This is done by pressing the cursor routing key where the cursor is located (which means pressing it twice if the cursor is not already there).
To facilitate this, a new TextInfo.activate() method was introduced.
Fixes #1893.
James Teh [Thu, 17 Nov 2011 01:14:18 +0000 (11:14 +1000)]
It is now possible to have braille displayed by paragraphs instead of lines, which may allow for more fluent reading of large amounts of text. This is configurable using the Read by paragraphs option in the Braille Settings dialog.
Fixes #1891.
James Teh [Thu, 17 Nov 2011 01:12:47 +0000 (11:12 +1000)]
Update documentation.
James Teh [Thu, 17 Nov 2011 00:12:46 +0000 (10:12 +1000)]
BrailleSettingsDialog: Add read by paragraph option.
James Teh [Thu, 17 Nov 2011 00:04:39 +0000 (10:04 +1000)]
braille.TextInfoRegion: Use UNIT_PARAGRAPH if readByParagraph is enabled. Also, change comments and variable names to refer to reading unit instead of line now that the reading unit might be paragraph.
James Teh [Thu, 17 Nov 2011 00:03:07 +0000 (10:03 +1000)]
config: Add braille readByParagraph option.
Cagri Dogan [Wed, 16 Nov 2011 19:01:02 +0000 (19:01 +0000)]
tr:
symbols.dic added some missing symbols
nvda.po corrected some translations.
Jani Kinnunen [Wed, 16 Nov 2011 18:38:32 +0000 (18:38 +0000)]
fi:
symbols.dic removed obsoleted symbols.
fatma mehanna [Wed, 16 Nov 2011 18:34:48 +0000 (18:34 +0000)]
ar:
changes.t2t matching 2011.3/4761
userGuide.t2t matching 2011.3/4758
Katsutoshi Tsuji [Wed, 16 Nov 2011 18:26:04 +0000 (18:26 +0000)]
ja:
nvda.po corrected typos.
changes.t2t corrected typos and updated for 2011.3
Simone Dal Maso [Wed, 16 Nov 2011 18:20:34 +0000 (18:20 +0000)]
it:
langstrings.txt corrected copyright years.
changes.t2t completed for 2011.3.
userGuide.t2t formatting corrections.
DINAKAR T.D. [Wed, 16 Nov 2011 16:42:00 +0000 (16:42 +0000)]
ta:
nvda.po no outstanding messages.
userGuide.t2t Complete Tamil userGuide.
A Special Thanks to Dinakar for his excellent work and for completing
the Tamil userGuide in this short time.
Bram Duvigneau [Wed, 16 Nov 2011 16:29:58 +0000 (16:29 +0000)]
nl:
nvda.po no fuzzy or untranslated messages.
Bernd Dorer [Wed, 16 Nov 2011 16:24:35 +0000 (16:24 +0000)]
de:
symbols.dic fixes #1895
changes.t2t textual improvements, matching previous rev.
userGuide.t2t textual improvements, matching previous rev.
Michel Such [Wed, 16 Nov 2011 15:52:54 +0000 (15:52 +0000)]
fr:
changes.t2t, fixing a small typo.
Michael Curran [Wed, 16 Nov 2011 08:58:52 +0000 (18:58 +1000)]
winword_getTextInRange: if spellingErrors is a requested format then always move by word, even if the chracterFormat unit may be supported.
Michael Curran [Wed, 16 Nov 2011 08:44:35 +0000 (18:44 +1000)]
winword_getTextInRange: allow the fetching of each format attribute to be configurable through a new formatConfig argument which replaces the (unused) flags argument. This speeds things up some more as particular formatting is no longer fetched unless its actually needed.
Michael Curran [Wed, 16 Nov 2011 06:29:40 +0000 (16:29 +1000)]
winword_getTextInRange: The xml now contains font name, font size, bold, italic, underline, subscript, superscript, paragraph alignment, style, line number and page number. However, line number, page number and paragraph alignment are only fetched for the first format field (to speed things up). wordDocumentTextInfo: _get_text now just calls MS Word's text property again, but getTextWithFields now calls winword_getTextInRange, and processes the xml in to formatFields. Therefore NVDA can now report formatting after the cursor many many times faster than before. Tables are not supported yet, and displabling/enabling of specific formatting still most be implemented.
James Teh [Wed, 16 Nov 2011 01:51:31 +0000 (11:51 +1000)]
Update SynthDriver code doc.