[orca] Always attempt broken-text recovery for Gecko sentences

[Joanmarie Diggs <joanied src gnome org>]

commit 988335ae6aee988f1c393cdc975783450d76ac7f
Author: Joanmarie Diggs <jdiggs igalia com>
Date:   Tue Jan 3 10:20:42 2017 -0500

    Always attempt broken-text recovery for Gecko sentences
    
    Because Gecko doesn't implement that support at all.

 .../scripts/toolkits/Gecko/script_utilities.py     |    8 +++++++-
 1 files changed, 7 insertions(+), 1 deletions(-)
---
diff --git a/src/orca/scripts/toolkits/Gecko/script_utilities.py 
b/src/orca/scripts/toolkits/Gecko/script_utilities.py
index 778b83f..fff8740 100644
--- a/src/orca/scripts/toolkits/Gecko/script_utilities.py
+++ b/src/orca/scripts/toolkits/Gecko/script_utilities.py
@@ -43,8 +43,14 @@ class Utilities(web.Utilities):
         super().__init__(script)
 
     def _attemptBrokenTextRecovery(self, obj, **args):
+        boundary = args.get('boundary')
+
+        # Gecko fails to implement this boundary type.
+        if boundary == pyatspi.TEXT_BOUNDARY_SENTENCE_START:
+            return True
+
         if self.isContentEditableWithEmbeddedObjects(obj):
-            return args.get('boundary') == pyatspi.TEXT_BOUNDARY_WORD_START
+            return boundary == pyatspi.TEXT_BOUNDARY_WORD_START
 
         return True