eSpeak synthDriver: when speaking characters in character mode, use the correct...

diff --git a/source/synthDrivers/espeak.py b/source/synthDrivers/espeak.py
index 37ffde2..d7ebac5 100644 (file)
--- a/source/synthDrivers/espeak.py
+++ b/source/synthDrivers/espeak.py
@@ -64,7 +64,7 @@ class SynthDriver(SynthDriver):
                        elif isinstance(item,speech.IndexCommand):\r
                                textList.append("<mark name=\"%d\" />"%item.index)\r
                        elif isinstance(item,speech.CharacterModeCommand):\r
-                               textList.append("<say-as type=\"spell-out\">" if item.state else "</say-as>")\r
+                               textList.append("<say-as interpret-as=\"characters\">" if item.state else "</say-as>")\r
                        elif isinstance(item,speech.LangChangeCommand):\r
                                if langChanged:\r
                                        textList.append("</voice>")\r

diff --git a/user_docs/en/changes.t2t b/user_docs/en/changes.t2t
index 59de60c..1c7bb46 100644 (file)
--- a/user_docs/en/changes.t2t
+++ b/user_docs/en/changes.t2t
@@ -78,6 +78,7 @@ Please see the [Commands Quick Reference keyCommands.html] document for the new
 - Fixed inability to use certain Microsoft Speech API version 5 voices such as Koba Speech 2 voices. (#2629)\r
 - In applications using the Java Access Bridge, braille displays are now updated correctly when the caret moves in editable text fields . (#3107)\r
 - Support the form landmark in browse mode documents that support landmarks. (#2997) \r
+- The eSpeak synth driver now handles reading by character more appropriately (e.g. announcing a letter's name or value rather than just its sound or generic name). (#3106)\r
 \r
 \r
 == Changes for Developers ==\r