Generate new tables for isalpha/toupper/tolower from UnicodeDate.txt

genucd.py

@@ -0,0 +1,77 @@
+# Create utfdata.h from UnicodeData.txt
+
+tolower = []
+toupper = []
+isalpha = []
+
+for line in open("UnicodeData.txt").readlines():
+	line = line.split(";")
+	code = int(line[0],16)
+	# if code > 65535: continue # skip non-BMP codepoints
+	if line[2][0] == 'L':
+		isalpha.append(code)
+	if line[12]:
+		toupper.append((code,int(line[12],16)))
+	if line[13]:
+		tolower.append((code,int(line[13],16)))
+
+def dumpalpha():
+	table = []
+	prev = 0
+	start = 0
+	for code in isalpha:
+		if code != prev+1:
+			if start:
+				table.append((start,prev))
+			start = code
+		prev = code
+	table.append((start,prev))
+
+	print("")
+	print("static const Rune ucd_alpha2[] = {")
+	for a, b in table:
+		if b - a > 0:
+			print(hex(a)+","+hex(b)+",")
+	print("};");
+
+	print("")
+	print("static const Rune ucd_alpha1[] = {")
+	for a, b in table:
+		if b - a == 0:
+			print(hex(a)+",")
+	print("};");
+
+def dumpmap(name, input):
+	table = []
+	prev_a = 0
+	prev_b = 0
+	start_a = 0
+	start_b = 0
+	for a, b in input:
+		if a != prev_a+1 or b != prev_b+1:
+			if start_a:
+				table.append((start_a,prev_a,start_b))
+			start_a = a
+			start_b = b
+		prev_a = a
+		prev_b = b
+	table.append((start_a,prev_a,start_b))
+
+	print("")
+	print("static const Rune " + name + "2[] = {")
+	for a, b, n in table:
+		if b - a > 0:
+			print(hex(a)+","+hex(b)+","+str(n-a)+",")
+	print("};");
+
+	print("")
+	print("static const Rune " + name + "1[] = {")
+	for a, b, n in table:
+		if b - a == 0:
+			print(hex(a)+","+str(n-a)+",")
+	print("};");
+
+print("/* This file was automatically created from UnicodeData.txt */")
+dumpalpha()
+dumpmap("ucd_tolower", tolower)
+dumpmap("ucd_toupper", toupper)

one.c

@@ -24,4 +24,3 @@
 #include "jsvalue.c"
 #include "regexp.c"
 #include "utf.c"
-#include "utftype.c"

utf.c

@@ -15,6 +15,9 @@
 #include <string.h>
 
 #include "utf.h"
+#include "utfdata.h"
+
+#define nelem(a) (int)(sizeof (a) / sizeof (a)[0])
 
 typedef unsigned char uchar;
 
@@ -210,3 +213,93 @@ utflen(const char *s)
 		n++;
 	}
 }
+
+static const Rune *
+ucd_bsearch(Rune c, const Rune *t, int n, int ne)
+{
+	const Rune *p;
+	int m;
+
+	while(n > 1) {
+		m = n/2;
+		p = t + m*ne;
+		if(c >= p[0]) {
+			t = p;
+			n = n-m;
+		} else
+			n = m;
+	}
+	if(n && c >= t[0])
+		return t;
+	return 0;
+}
+
+Rune
+tolowerrune(Rune c)
+{
+	const Rune *p;
+
+	p = ucd_bsearch(c, ucd_tolower2, nelem(ucd_tolower2)/3, 3);
+	if(p && c >= p[0] && c <= p[1])
+		return c + p[2];
+	p = ucd_bsearch(c, ucd_tolower1, nelem(ucd_tolower1)/2, 2);
+	if(p && c == p[0])
+		return c + p[1];
+	return c;
+}
+
+Rune
+toupperrune(Rune c)
+{
+	const Rune *p;
+
+	p = ucd_bsearch(c, ucd_toupper2, nelem(ucd_toupper2)/3, 3);
+	if(p && c >= p[0] && c <= p[1])
+		return c + p[2];
+	p = ucd_bsearch(c, ucd_toupper1, nelem(ucd_toupper1)/2, 2);
+	if(p && c == p[0])
+		return c + p[1];
+	return c;
+}
+
+int
+islowerrune(Rune c)
+{
+	const Rune *p;
+
+	p = ucd_bsearch(c, ucd_toupper2, nelem(ucd_toupper2)/3, 3);
+	if(p && c >= p[0] && c <= p[1])
+		return 1;
+	p = ucd_bsearch(c, ucd_toupper1, nelem(ucd_toupper1)/2, 2);
+	if(p && c == p[0])
+		return 1;
+	return 0;
+}
+
+int
+isupperrune(Rune c)
+{
+	const Rune *p;
+
+	p = ucd_bsearch(c, ucd_tolower2, nelem(ucd_tolower2)/3, 3);
+	if(p && c >= p[0] && c <= p[1])
+		return 1;
+	p = ucd_bsearch(c, ucd_tolower1, nelem(ucd_tolower1)/2, 2);
+	if(p && c == p[0])
+		return 1;
+	return 0;
+}
+
+int
+isalpharune(Rune c)
+{
+	const Rune *p;
+
+	p = ucd_bsearch(c, ucd_alpha2, nelem(ucd_alpha2)/2, 2);
+	if(p && c >= p[0] && c <= p[1])
+		return 1;
+	p = ucd_bsearch(c, ucd_alpha1, nelem(ucd_alpha1), 1);
+	if(p && c == p[0])
+		return 1;
+	return 0;
+}

utf.h

@@ -23,11 +23,8 @@ typedef int Rune;	/* 32 bits */
 
 #define isalpharune	jsU_isalpharune
 #define islowerrune	jsU_islowerrune
-#define isspacerune	jsU_isspacerune
-#define istitlerune	jsU_istitlerune
 #define isupperrune	jsU_isupperrune
 #define tolowerrune	jsU_tolowerrune
-#define totitlerune	jsU_totitlerune
 #define toupperrune	jsU_toupperrune
 
 enum
@@ -46,11 +43,8 @@ int	utflen(const char *s);
 
 int		isalpharune(Rune c);
 int		islowerrune(Rune c);
-int		isspacerune(Rune c);
-int		istitlerune(Rune c);
 int		isupperrune(Rune c);
 Rune		tolowerrune(Rune c);
-Rune		totitlerune(Rune c);
 Rune		toupperrune(Rune c);
 
 #endif