We now require eSpeak 1.46.02, which contains bug fixes.

Fixes #1991.

readme.txt

@@ -13,7 +13,7 @@ General dependencies:
 	* comtypes, version 0.6.2 or later: http://www.sourceforge.net/projects/comtypes/
 	* wxPython 2.8 unicode (for Python 2.7), version 2.8.11.0 or later: http://www.wxpython.org/
 	* Python Windows Extensions (for Python 2.7), build 214 or later: http://www.sourceforge.net/projects/pywin32/ 
-	* eSpeak, version 1.46.01 or later:
+	* eSpeak, version 1.46.02 or later:
 		* Official web site: http://espeak.sourceforge.net/
 		* Download the espeak source archive. Note that it must be an official source archive from the espeak website containing already compiled phoneme data, not straight from svn. 
 		* Extract it in to include/espeak so that include/espeak/src, include/espeak/dictsource, include/espeak/platforms and include/espeak/espeak-data all exist.

user_docs/en/changes.t2t

@@ -27,6 +27,7 @@
 
 == Changes ==
 - Commands to describe the current character or to spell the current word or line now  will spell in the appropriate language according to the text, if auto language switching is turned on and the appropriate language information is available.
+- Updated eSpeak speech synthesizer to 1.46.02.
 - NVDA will now filter out extremely long (30 characters or greater) names guessed from graphic and link URLs as they are most likely garbage that gets in the way of reading. (#1989)
 - Some information displayed in braille has been abbreviated. (#1955)