Update UCDN.

1 files changed, 49 insertions, 47 deletions

diff --git a/source/fitz/ucdn.c b/source/fitz/ucdn.c
index d6b4135e..4e000246 100644
--- a/source/fitz/ucdn.c
+++ b/source/fitz/ucdn.c
@@ -17,13 +17,13 @@
 #include "mupdf/fitz.h"
 #include "mupdf/ucdn.h"
 
+#include <stdio.h>
 #include <stdlib.h>
 
 typedef struct {
 	unsigned char category;
 	unsigned char combining;
 	unsigned char bidi_class;
-	unsigned char mirrored;
 	unsigned char east_asian_width;
 	unsigned char script;
 	unsigned char linebreak_class;
@@ -43,7 +43,7 @@ typedef struct {
 	short count, index;
 } Reindex;
 
-#include "unicodedata_db.h"
+#include "ucdn_db.h"
 
 /* constants required for Hangul (de)composition */
 #define SBASE 0xAC00
@@ -91,20 +91,30 @@ static const unsigned short *get_decomp_record(uint32_t code)
 	return &decomp_data[index];
 }
 
-static int get_comp_index(uint32_t code, const Reindex *idx)
+static int compare_reindex(const void *a, const void *b)
 {
-	int i;
-
-	for (i = 0; idx[i].start; i++) {
-		const Reindex *cur = &idx[i];
-		if (code < cur->start)
-			return -1;
-		if (code <= cur->start + cur->count) {
-			return cur->index + (code - cur->start);
-		}
-	}
+	Reindex *ra = (Reindex *)a;
+	Reindex *rb = (Reindex *)b;
 
-	return -1;
+	if (ra->start < rb->start)
+		return -1;
+	else if (ra->start > (rb->start + rb->count))
+		return 1;
+	else
+		return 0;
+}
+
+static int get_comp_index(uint32_t code, const Reindex *idx, size_t len)
+{
+	Reindex *res;
+	Reindex r = {0, 0, 0};
+	r.start = code;
+	res = (Reindex *) bsearch(&r, idx, len, sizeof(Reindex), compare_reindex);
+
+	if (res != NULL)
+		return res->index + (code - res->start);
+	else
+		return -1;
 }
 
 static int compare_mp(const void *a, const void *b)
@@ -154,23 +164,18 @@ static int hangul_pair_decompose(uint32_t code, uint32_t *a, uint32_t *b)
 
 static int hangul_pair_compose(uint32_t *code, uint32_t a, uint32_t b)
 {
-	if (b < VBASE || b >= (TBASE + TCOUNT))
-		return 0;
-
-	if ((a < LBASE || a >= (LBASE + LCOUNT))
-			&& (a < SBASE || a >= (SBASE + SCOUNT)))
-		return 0;
-
-	if (a >= SBASE) {
+	if (a >= SBASE && a < (SBASE + SCOUNT) && b >= TBASE && b < (TBASE + TCOUNT)) {
 		/* LV,T */
 		*code = a + (b - TBASE);
 		return 3;
-	} else {
+	} else if (a >= LBASE && a < (LBASE + LCOUNT) && b >= VBASE && b < (VBASE + VCOUNT)) {
 		/* L,V */
 		int li = a - LBASE;
 		int vi = b - VBASE;
 		*code = SBASE + li * NCOUNT + vi * TCOUNT;
 		return 2;
+	} else {
+		return 0;
 	}
 }
 
@@ -178,7 +183,7 @@ static uint32_t decode_utf16(const unsigned short **code_ptr)
 {
 	const unsigned short *code = *code_ptr;
 
-	if ((code[0] & 0xd800) != 0xd800) {
+	if (code[0] < 0xd800 || code[0] > 0xdc00) {
 		*code_ptr += 1;
 		return (uint32_t)code[0];
 	} else {
@@ -215,7 +220,7 @@ int ucdn_get_bidi_class(uint32_t code)
 
 int ucdn_get_mirrored(uint32_t code)
 {
-	return get_ucd_record(code)->mirrored;
+	return ucdn_mirror(code) != code;
 }
 
 int ucdn_get_script(uint32_t code)
@@ -234,28 +239,28 @@ int ucdn_get_resolved_linebreak_class(uint32_t code)
 
 	switch (record->linebreak_class)
 	{
-		case UCDN_LINEBREAK_CLASS_AI:
-		case UCDN_LINEBREAK_CLASS_SG:
-		case UCDN_LINEBREAK_CLASS_XX:
-			return UCDN_LINEBREAK_CLASS_AL;
+	case UCDN_LINEBREAK_CLASS_AI:
+	case UCDN_LINEBREAK_CLASS_SG:
+	case UCDN_LINEBREAK_CLASS_XX:
+		return UCDN_LINEBREAK_CLASS_AL;
 
-		case UCDN_LINEBREAK_CLASS_SA:
-			if (record->category == UCDN_GENERAL_CATEGORY_MC ||
-					record->category == UCDN_GENERAL_CATEGORY_MN)
-				return UCDN_LINEBREAK_CLASS_CM;
-			return UCDN_LINEBREAK_CLASS_AL;
+	case UCDN_LINEBREAK_CLASS_SA:
+		if (record->category == UCDN_GENERAL_CATEGORY_MC ||
+				record->category == UCDN_GENERAL_CATEGORY_MN)
+			return UCDN_LINEBREAK_CLASS_CM;
+		return UCDN_LINEBREAK_CLASS_AL;
 
-		case UCDN_LINEBREAK_CLASS_CJ:
-			return UCDN_LINEBREAK_CLASS_NS;
+	case UCDN_LINEBREAK_CLASS_CJ:
+		return UCDN_LINEBREAK_CLASS_NS;
 
-		case UCDN_LINEBREAK_CLASS_CB:
-			return UCDN_LINEBREAK_CLASS_B2;
+	case UCDN_LINEBREAK_CLASS_CB:
+		return UCDN_LINEBREAK_CLASS_B2;
 
-		case UCDN_LINEBREAK_CLASS_NL:
-			return UCDN_LINEBREAK_CLASS_BK;
+	case UCDN_LINEBREAK_CLASS_NL:
+		return UCDN_LINEBREAK_CLASS_BK;
 
-		default:
-			return record->linebreak_class;
+	default:
+		return record->linebreak_class;
 	}
 }
 
@@ -264,9 +269,6 @@ uint32_t ucdn_mirror(uint32_t code)
 	MirrorPair mp = {0};
 	MirrorPair *res;
 
-	if (get_ucd_record(code)->mirrored == 0)
-		return code;
-
 	mp.from = code;
 	res = (MirrorPair *) bsearch(&mp, mirror_pairs, BIDI_MIRROR_LEN,
 			sizeof(MirrorPair), compare_mp);
@@ -326,8 +328,8 @@ int ucdn_compose(uint32_t *code, uint32_t a, uint32_t b)
 	if (hangul_pair_compose(code, a, b))
 		return 1;
 
-	l = get_comp_index(a, nfc_first);
-	r = get_comp_index(b, nfc_last);
+	l = get_comp_index(a, nfc_first, sizeof(nfc_first) / sizeof(Reindex));
+	r = get_comp_index(b, nfc_last, sizeof(nfc_last) / sizeof(Reindex));
 
 	if (l < 0 || r < 0)
 		return 0;