usp10: Windows breaks punctuation into 2 different script types.

dlls/usp10/shape.c

@@ -583,6 +583,7 @@ static const ScriptShapeData ShapingData[] =
     {{ devanagari_features, 6}, required_telugu_features, "mlym", "mlm2", ContextualShape_Malayalam, ShapeCharGlyphProp_Malayalam},
     {{ devanagari_features, 6}, required_telugu_features, "mlym", "mlm2", ContextualShape_Malayalam, ShapeCharGlyphProp_Malayalam},
     {{ standard_features, 2}, NULL, "" , "", NULL, NULL},
+    {{ standard_features, 2}, NULL, "latn" , "", NULL, NULL},
 };
 
 static INT GSUB_is_glyph_covered(LPCVOID table , UINT glyph)

dlls/usp10/tests/usp10.c

@@ -285,6 +285,13 @@ static void test_ScriptItemize( void )
     static const WCHAR test21[] = {0x0710, 0x0712, 0x0308, 0x0712, 0x0714,0};
     static const itemTest t211[2] = {{{0,0,0,0,0},0,1,1,1,syrc_tag,FALSE},{{0,0,0,0,0},5,0,0,0,-1,FALSE}};
 
+    /* Latin Punctuation */
+    static const WCHAR test22[] = {'#','$',',','!','\"','*',0};
+    static const itemTest t221[3] = {{{0,0,0,0,0},0,0,0,0,latn_tag,FALSE},{{0,0,0,0,0},3,0,0,0,0,FALSE},{{0,0,0,0,0},6,0,0,0,-1,FALSE}};
+    static const itemTest t222[3] = {{{0,0,0,0,0},0,1,1,1,latn_tag,FALSE},{{0,0,0,0,0},3,1,1,1,0,FALSE},{{0,0,0,0,0},6,0,0,0,-1,FALSE}};
+    static const itemTest t223[2] = {{{0,0,0,0,0},0,1,1,1,latn_tag,FALSE},{{0,0,0,0,0},6,0,0,0,-1,FALSE}};
+
+
     SCRIPT_ITEM items[15];
     SCRIPT_CONTROL  Control;
     SCRIPT_STATE    State;
@@ -339,6 +346,7 @@ static void test_ScriptItemize( void )
     test_items_ok(test19,6,NULL,NULL,1,t191,FALSE,0);
     test_items_ok(test20,5,NULL,NULL,2,t201,FALSE,0);
     test_items_ok(test21,5,NULL,NULL,1,t211,FALSE,0);
+    test_items_ok(test22,6,NULL,NULL,2,t221,FALSE,0);
 
     State.uBidiLevel = 0;
     test_items_ok(test1,4,&Control,&State,1,t11,FALSE,0);
@@ -367,6 +375,7 @@ static void test_ScriptItemize( void )
     test_items_ok(test19,6,&Control,&State,1,t191,FALSE,0);
     test_items_ok(test20,5,&Control,&State,2,t201,FALSE,0);
     test_items_ok(test21,5,&Control,&State,1,t211,FALSE,0);
+    test_items_ok(test22,6,&Control,&State,2,t221,FALSE,0);
 
     State.uBidiLevel = 1;
     test_items_ok(test1,4,&Control,&State,1,t12,FALSE,0);
@@ -395,6 +404,7 @@ static void test_ScriptItemize( void )
     test_items_ok(test19,6,&Control,&State,1,t192,FALSE,0);
     test_items_ok(test20,5,&Control,&State,2,t202,FALSE,0);
     test_items_ok(test21,5,&Control,&State,1,t211,FALSE,0);
+    test_items_ok(test22,6,&Control,&State,2,t222,FALSE,1);
 
     State.uBidiLevel = 1;
     Control.fMergeNeutralItems = TRUE;
@@ -424,6 +434,7 @@ static void test_ScriptItemize( void )
     test_items_ok(test19,6,&Control,&State,1,t192,FALSE,0);
     test_items_ok(test20,5,&Control,&State,2,t202,FALSE,0);
     test_items_ok(test21,5,&Control,&State,1,t211,FALSE,0);
+    test_items_ok(test22,6,&Control,&State,1,t223,FALSE,2);
 }
 
 static inline void _test_shape_ok(int valid, HDC hdc, LPCWSTR string,

dlls/usp10/usp10.c

@@ -362,6 +362,10 @@ static const scriptData scriptInformation[] = {
      {LANG_ENGLISH, 0, 1, 0, 1, ANSI_CHARSET, 0, 0, 0, 0, 0, 1, 1, 0, 0},
      0x00000000,
      {0}},
+    {{Script_Punctuation2, 0, 0, 0, 0, 0, 0, { 0,0,0,0,0,0,0,0,0,0,0}},
+     {LANG_ENGLISH, 0, 0, 0, 0, ANSI_CHARSET, 0, 0, 0, 0, 0, 0, 0, 0, 0},
+     MS_MAKE_TAG('l','a','t','n'),
+     {0}},
 };
 
 static const SCRIPT_PROPERTIES *script_props[] =
@@ -522,14 +526,19 @@ static WCHAR mirror_char( WCHAR ch )
 
 static WORD get_char_script( WCHAR ch)
 {
+    static const WCHAR latin_punc[] = {'#','$','&','\'',',',';','<','>','?','@','\\','^','_','`','{','|','}','~', 0x00a0, 0};
     WORD type = 0;
     int i;
 
     if (ch == 0xc || ch == 0x20 || ch == 0x202f)
         return Script_CR;
 
+    /* These punctuation are seperated out as Latin punctuation */
+    if (strchrW(latin_punc,ch))
+        return Script_Punctuation2;
+
     /* These chars are itemized as Punctuation by Windows */
-    if (ch == 0x2212 || ch == 0x2044 || ch == 0x00a0)
+    if (ch == 0x2212 || ch == 0x2044)
         return Script_Punctuation;
 
     GetStringTypeW(CT_CTYPE1, &ch, 1, &type);

dlls/usp10/usp10_internal.h

@@ -73,6 +73,7 @@
 #define Script_Malayalam_Numeric 43
 /* More supplemental */
 #define Script_Diacritical 44
+#define Script_Punctuation2 45
 
 #define GLYPH_BLOCK_SHIFT 8
 #define GLYPH_BLOCK_SIZE  (1UL << GLYPH_BLOCK_SHIFT)