Resolves: fdo#38983 allow extra word boundary characters

i.e. word overrides emdash and endash to be word boundary characters
for the purposes of counting words. And there are some who want
to treat =,- etc similarly.

Default to a configuration that gives the same results as Word for
word counting.

Change-Id: Ia8ce6ac12011a1d6e547f11644c76163c4c993c5

1 files changed, 5 insertions, 1 deletions

diff --git a/sw/inc/swscanner.hxx b/sw/inc/swscanner.hxx
index f657f238c2ca..7a71c0a1e9e9 100644
--- a/sw/inc/swscanner.hxx
+++ b/sw/inc/swscanner.hxx
@@ -43,13 +43,15 @@ class SwScanner
 {
     rtl::OUString aWord;
     const SwTxtNode& rNode;
-    const rtl::OUString aText;
+    const rtl::OUString aPreDashReplacementText;
+    rtl::OUString aText;
     const LanguageType* pLanguage;
     const ModelToViewHelper& rConversionMap;
     sal_Int32 nStartPos;
     sal_Int32 nEndPos;
     sal_Int32 nBegin;
     sal_Int32 nLen;
+    sal_Int32 nOverriddenDashCount;
     LanguageType aCurrLang;
     sal_uInt16 nWordType;
     sal_Bool bClip;
@@ -74,6 +76,8 @@ public:
     sal_Int32 GetLen() const           { return nLen; }
 
     LanguageType GetCurrentLanguage() const {return aCurrLang;}
+
+    sal_Int32 getOverriddenDashCount() const {return nOverriddenDashCount; }
 };
 
 #endif