tdf#81144 Chinese full-width punctuation does not align properly

Ideographic fullstop and comma in most Chinese fonts are centered,
while those in Japanese fonts align closer to the left. Original
compression algorithm trimed right side of the punctuation, making
fullwidth fullstop or comma in Chinese font visually unbalanced.
In worst case, it crowds together with the followed compressed
punctuation.

This patch fix the situation in the folowing way
1) make compression less stronger.
2) Trim space according to glyph bearing to font height ratio.
3) fix a memory access violation issue

Change-Id: Icff215064e6c442fd36eac8e01b01fb6acb27594
Reviewed-on: https://gerrit.libreoffice.org/19517
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>

1 files changed, 2 insertions, 1 deletions

diff --git a/sw/source/core/inc/scriptinfo.hxx b/sw/source/core/inc/scriptinfo.hxx
index ee476ac983e4..2201343dc183 100644
--- a/sw/source/core/inc/scriptinfo.hxx
+++ b/sw/source/core/inc/scriptinfo.hxx
@@ -37,7 +37,7 @@ typedef std::list< sal_Int32 > PositionList;
 class SwScriptInfo
 {
 public:
-    enum CompType { KANA, SPECIAL_LEFT, SPECIAL_RIGHT, NONE };
+    enum CompType { KANA, SPECIAL_LEFT, SPECIAL_RIGHT, NONE, SPECIAL_MIDDLE};
 
 private:
     //! Records a single change in script type.
@@ -255,6 +255,7 @@ public:
     // modifies the kerning array according to a given compress value
     long Compress( long* pKernArray, sal_Int32 nIdx, sal_Int32 nLen,
                    const sal_uInt16 nCompress, const sal_uInt16 nFontHeight,
+                   const bool bCentered,
                    Point* pPoint = NULL ) const;
 
 /** Performs a kashida justification on the kerning array