Imported Upstream version 1.1~hg20130304upstream/1.1_hg20130304

4 files changed, 116 insertions, 33 deletions

diff --git a/src/lib9/utf/Makefile b/src/lib9/utf/Makefile
index bbb2da6a9..5c9cdf051 100644
--- a/src/lib9/utf/Makefile
+++ b/src/lib9/utf/Makefile
@@ -15,13 +15,13 @@ UnicodeData-%.txt:
 mkrunetype: mkrunetype.c
 	cc -I../../../include -o mkrunetype -L$(GOROOT)/pkg/obj/$(GOOS)_$(GOARCH)/ mkrunetype.c -l9
 
-runetypebody-%.c: mkrunetype UnicodeData-%.txt
+runetypebody-%.h: mkrunetype UnicodeData-%.txt
 	mkrunetype -p UnicodeData-$*.txt >_$@
 	mv _$@ $@
 
 CLEANFILES+=UnicodeData.txt
 
-UNICODE_VERSION=6.0.0
+UNICODE_VERSION=6.2.0
 
 test: mkrunetype UnicodeData-$(UNICODE_VERSION).txt
 	mkrunetype -c UnicodeData-$(UNICODE_VERSION).txt
diff --git a/src/lib9/utf/rune.c b/src/lib9/utf/rune.c
index cf98bab15..818771cfd 100644
--- a/src/lib9/utf/rune.c
+++ b/src/lib9/utf/rune.c
@@ -36,12 +36,14 @@ enum
 	Rune1	= (1<<(Bit1+0*Bitx))-1,		/* 0000 0000 0111 1111 */
 	Rune2	= (1<<(Bit2+1*Bitx))-1,		/* 0000 0111 1111 1111 */
 	Rune3	= (1<<(Bit3+2*Bitx))-1,		/* 1111 1111 1111 1111 */
-	Rune4	= (1<<(Bit4+3*Bitx))-1,
-                                        /* 0001 1111 1111 1111 1111 1111 */
+	Rune4	= (1<<(Bit4+3*Bitx))-1,		/* 0001 1111 1111 1111 1111 1111 */
 
 	Maskx	= (1<<Bitx)-1,			/* 0011 1111 */
 	Testx	= Maskx ^ 0xFF,			/* 1100 0000 */
 
+	SurrogateMin	= 0xD800,
+	SurrogateMax	= 0xDFFF,
+
 	Bad	= Runeerror,
 };
 
@@ -122,6 +124,8 @@ charntorune(Rune *rune, const char *str, int length)
 		l = ((((c << Bitx) | c1) << Bitx) | c2) & Rune3;
 		if(l <= Rune2)
 			goto bad;
+		if (SurrogateMin <= l && l <= SurrogateMax)
+			goto bad;
 		*rune = l;
 		return 3;
 	}
@@ -138,7 +142,7 @@ charntorune(Rune *rune, const char *str, int length)
 		goto bad;
 	if (c < T5) {
 		l = ((((((c << Bitx) | c1) << Bitx) | c2) << Bitx) | c3) & Rune4;
-		if (l <= Rune3)
+		if (l <= Rune3 || l > Runemax)
 			goto bad;
 		*rune = l;
 		return 4;
@@ -208,6 +212,8 @@ chartorune(Rune *rune, const char *str)
 		l = ((((c << Bitx) | c1) << Bitx) | c2) & Rune3;
 		if(l <= Rune2)
 			goto bad;
+		if (SurrogateMin <= l && l <= SurrogateMax)
+			goto bad;
 		*rune = l;
 		return 3;
 	}
@@ -221,7 +227,7 @@ chartorune(Rune *rune, const char *str)
 		goto bad;
 	if (c < T5) {
 		l = ((((((c << Bitx) | c1) << Bitx) | c2) << Bitx) | c3) & Rune4;
-		if (l <= Rune3)
+		if (l <= Rune3 || l > Runemax)
 			goto bad;
 		*rune = l;
 		return 4;
@@ -241,7 +247,8 @@ bad:
 }
 
 int
-isvalidcharntorune(const char* str, int length, Rune* rune, int* consumed) {
+isvalidcharntorune(const char* str, int length, Rune* rune, int* consumed)
+{
 	*consumed = charntorune(rune, str, length);
 	return *rune != Runeerror || *consumed == 3;
 }
@@ -273,13 +280,15 @@ runetochar(char *str, const Rune *rune)
 	}
 
 	/*
-	 * If the Rune is out of range, convert it to the error rune.
+	 * If the Rune is out of range or a surrogate half, convert it to the error rune.
 	 * Do this test here because the error rune encodes to three bytes.
 	 * Doing it earlier would duplicate work, since an out of range
 	 * Rune wouldn't have fit in one or two bytes.
 	 */
 	if (c > Runemax)
 		c = Runeerror;
+	if (SurrogateMin <= c && c <= SurrogateMax)
+		c = Runeerror;
 
 	/*
 	 * three character sequence
diff --git a/src/lib9/utf/runetype.c b/src/lib9/utf/runetype.c
index 51729fb01..b3634965f 100644
--- a/src/lib9/utf/runetype.c
+++ b/src/lib9/utf/runetype.c
@@ -35,4 +35,4 @@ rbsearch(Rune c, Rune *t, int n, int ne)
 	return 0;
 }
 
-#include "runetypebody-6.0.0.h"
+#include "runetypebody-6.2.0.h"
diff --git a/src/lib9/utf/runetypebody-6.0.0.h b/src/lib9/utf/runetypebody-6.2.0.h
index 47c0faf73..a603af0df 100644
--- a/src/lib9/utf/runetypebody-6.0.0.h
+++ b/src/lib9/utf/runetypebody-6.2.0.h
@@ -1,4 +1,4 @@
-/* generated automatically by mkrunetype.c from UnicodeData-6.0.0.txt */
+/* generated automatically by mkrunetype.c from UnicodeData-6.2.0.txt */
 
 static Rune __isspacer[] = {
 	0x0009, 0x000d,
@@ -64,6 +64,10 @@ static Rune __isdigitr[] = {
 	0xff10, 0xff19,
 	0x104a0, 0x104a9,
 	0x11066, 0x1106f,
+	0x110f0, 0x110f9,
+	0x11136, 0x1113f,
+	0x111d0, 0x111d9,
+	0x116c0, 0x116c9,
 	0x1d7ce, 0x1d7ff,
 };
 
@@ -110,6 +114,7 @@ static Rune __isalphar[] = {
 	0x07f4, 0x07f5,
 	0x0800, 0x0815,
 	0x0840, 0x0858,
+	0x08a2, 0x08ac,
 	0x0904, 0x0939,
 	0x0958, 0x0961,
 	0x0971, 0x0977,
@@ -189,7 +194,7 @@ static Rune __isalphar[] = {
 	0x0ead, 0x0eb0,
 	0x0eb2, 0x0eb3,
 	0x0ec0, 0x0ec4,
-	0x0edc, 0x0edd,
+	0x0edc, 0x0edf,
 	0x0f40, 0x0f47,
 	0x0f49, 0x0f6c,
 	0x0f88, 0x0f8c,
@@ -201,7 +206,7 @@ static Rune __isalphar[] = {
 	0x1075, 0x1081,
 	0x10a0, 0x10c5,
 	0x10d0, 0x10fa,
-	0x1100, 0x1248,
+	0x10fc, 0x1248,
 	0x124a, 0x124d,
 	0x1250, 0x1256,
 	0x125a, 0x125d,
@@ -242,12 +247,13 @@ static Rune __isalphar[] = {
 	0x1b45, 0x1b4b,
 	0x1b83, 0x1ba0,
 	0x1bae, 0x1baf,
-	0x1bc0, 0x1be5,
+	0x1bba, 0x1be5,
 	0x1c00, 0x1c23,
 	0x1c4d, 0x1c4f,
 	0x1c5a, 0x1c7d,
 	0x1ce9, 0x1cec,
 	0x1cee, 0x1cf1,
+	0x1cf5, 0x1cf6,
 	0x1d00, 0x1dbf,
 	0x1e00, 0x1f15,
 	0x1f18, 0x1f1d,
@@ -276,8 +282,9 @@ static Rune __isalphar[] = {
 	0x2c30, 0x2c5e,
 	0x2c60, 0x2ce4,
 	0x2ceb, 0x2cee,
+	0x2cf2, 0x2cf3,
 	0x2d00, 0x2d25,
-	0x2d30, 0x2d65,
+	0x2d30, 0x2d67,
 	0x2d80, 0x2d96,
 	0x2da0, 0x2da6,
 	0x2da8, 0x2dae,
@@ -299,7 +306,7 @@ static Rune __isalphar[] = {
 	0x31a0, 0x31ba,
 	0x31f0, 0x31ff,
 	0x3400, 0x4db5,
-	0x4e00, 0x9fcb,
+	0x4e00, 0x9fcc,
 	0xa000, 0xa48c,
 	0xa4d0, 0xa4fd,
 	0xa500, 0xa60c,
@@ -311,9 +318,9 @@ static Rune __isalphar[] = {
 	0xa717, 0xa71f,
 	0xa722, 0xa788,
 	0xa78b, 0xa78e,
-	0xa790, 0xa791,
-	0xa7a0, 0xa7a9,
-	0xa7fa, 0xa801,
+	0xa790, 0xa793,
+	0xa7a0, 0xa7aa,
+	0xa7f8, 0xa801,
 	0xa803, 0xa805,
 	0xa807, 0xa80a,
 	0xa80c, 0xa822,
@@ -332,6 +339,8 @@ static Rune __isalphar[] = {
 	0xaab5, 0xaab6,
 	0xaab9, 0xaabd,
 	0xaadb, 0xaadd,
+	0xaae0, 0xaaea,
+	0xaaf2, 0xaaf4,
 	0xab01, 0xab06,
 	0xab09, 0xab0e,
 	0xab11, 0xab16,
@@ -341,8 +350,7 @@ static Rune __isalphar[] = {
 	0xac00, 0xd7a3,
 	0xd7b0, 0xd7c6,
 	0xd7cb, 0xd7fb,
-	0xf900, 0xfa2d,
-	0xfa30, 0xfa6d,
+	0xf900, 0xfa6d,
 	0xfa70, 0xfad9,
 	0xfb00, 0xfb06,
 	0xfb13, 0xfb17,
@@ -387,6 +395,8 @@ static Rune __isalphar[] = {
 	0x1083f, 0x10855,
 	0x10900, 0x10915,
 	0x10920, 0x10939,
+	0x10980, 0x109b7,
+	0x109be, 0x109bf,
 	0x10a10, 0x10a13,
 	0x10a15, 0x10a17,
 	0x10a19, 0x10a33,
@@ -397,9 +407,16 @@ static Rune __isalphar[] = {
 	0x10c00, 0x10c48,
 	0x11003, 0x11037,
 	0x11083, 0x110af,
+	0x110d0, 0x110e8,
+	0x11103, 0x11126,
+	0x11183, 0x111b2,
+	0x111c1, 0x111c4,
+	0x11680, 0x116aa,
 	0x12000, 0x1236e,
 	0x13000, 0x1342e,
 	0x16800, 0x16a38,
+	0x16f00, 0x16f44,
+	0x16f93, 0x16f9f,
 	0x1b000, 0x1b001,
 	0x1d400, 0x1d454,
 	0x1d456, 0x1d49c,
@@ -428,6 +445,23 @@ static Rune __isalphar[] = {
 	0x1d78a, 0x1d7a8,
 	0x1d7aa, 0x1d7c2,
 	0x1d7c4, 0x1d7cb,
+	0x1ee00, 0x1ee03,
+	0x1ee05, 0x1ee1f,
+	0x1ee21, 0x1ee22,
+	0x1ee29, 0x1ee32,
+	0x1ee34, 0x1ee37,
+	0x1ee4d, 0x1ee4f,
+	0x1ee51, 0x1ee52,
+	0x1ee61, 0x1ee62,
+	0x1ee67, 0x1ee6a,
+	0x1ee6c, 0x1ee72,
+	0x1ee74, 0x1ee77,
+	0x1ee79, 0x1ee7c,
+	0x1ee80, 0x1ee89,
+	0x1ee8b, 0x1ee9b,
+	0x1eea1, 0x1eea3,
+	0x1eea5, 0x1eea9,
+	0x1eeab, 0x1eebb,
 	0x20000, 0x2a6d6,
 	0x2a700, 0x2b734,
 	0x2b740, 0x2b81d,
@@ -451,6 +485,7 @@ static Rune __isalphas[] = {
 	0x081a,
 	0x0824,
 	0x0828,
+	0x08a0,
 	0x093d,
 	0x0950,
 	0x09b2,
@@ -481,7 +516,8 @@ static Rune __isalphas[] = {
 	0x103f,
 	0x1061,
 	0x108e,
-	0x10fc,
+	0x10c7,
+	0x10cd,
 	0x1258,
 	0x12c0,
 	0x17d7,
@@ -501,6 +537,8 @@ static Rune __isalphas[] = {
 	0x2126,
 	0x2128,
 	0x214e,
+	0x2d27,
+	0x2d2d,
 	0x2d6f,
 	0x2e2f,
 	0xa8fb,
@@ -514,9 +552,26 @@ static Rune __isalphas[] = {
 	0x10808,
 	0x1083c,
 	0x10a00,
+	0x16f50,
 	0x1d4a2,
 	0x1d4bb,
 	0x1d546,
+	0x1ee24,
+	0x1ee27,
+	0x1ee39,
+	0x1ee3b,
+	0x1ee42,
+	0x1ee47,
+	0x1ee49,
+	0x1ee4b,
+	0x1ee54,
+	0x1ee57,
+	0x1ee59,
+	0x1ee5b,
+	0x1ee5d,
+	0x1ee5f,
+	0x1ee64,
+	0x1ee7e,
 };
 
 int
@@ -652,7 +707,8 @@ static Rune __isupperp[] = {
 	0xa779, 0xa77b,
 	0xa780, 0xa786,
 	0xa78b, 0xa78d,
-	0xa7a0, 0xa7a8,
+	0xa790, 0xa792,
+	0xa7a0, 0xa7aa,
 };
 
 static Rune __isuppers[] = {
@@ -673,6 +729,8 @@ static Rune __isuppers[] = {
 	0x03cf,
 	0x03f4,
 	0x03f7,
+	0x10c7,
+	0x10cd,
 	0x2102,
 	0x2107,
 	0x2115,
@@ -681,7 +739,7 @@ static Rune __isuppers[] = {
 	0x2c60,
 	0x2c72,
 	0x2c75,
-	0xa790,
+	0x2cf2,
 	0x1d49c,
 	0x1d4a2,
 	0x1d546,
@@ -733,7 +791,7 @@ static Rune __islowerr[] = {
 	0x04ce, 0x04cf,
 	0x0561, 0x0587,
 	0x1d00, 0x1d2b,
-	0x1d62, 0x1d77,
+	0x1d6b, 0x1d77,
 	0x1d79, 0x1d9a,
 	0x1e95, 0x1e9d,
 	0x1eff, 0x1f07,
@@ -764,7 +822,7 @@ static Rune __islowerr[] = {
 	0x2c30, 0x2c5e,
 	0x2c65, 0x2c66,
 	0x2c73, 0x2c74,
-	0x2c76, 0x2c7c,
+	0x2c76, 0x2c7b,
 	0x2ce3, 0x2ce4,
 	0x2d00, 0x2d25,
 	0xa72f, 0xa731,
@@ -832,13 +890,12 @@ static Rune __islowerp[] = {
 	0xa77a, 0xa77c,
 	0xa77f, 0xa787,
 	0xa78c, 0xa78e,
+	0xa791, 0xa793,
 	0xa7a1, 0xa7a9,
 };
 
 static Rune __islowers[] = {
-	0x00aa,
 	0x00b5,
-	0x00ba,
 	0x0188,
 	0x0192,
 	0x0195,
@@ -864,7 +921,9 @@ static Rune __islowers[] = {
 	0x2184,
 	0x2c61,
 	0x2c71,
-	0xa791,
+	0x2cf3,
+	0x2d27,
+	0x2d2d,
 	0xa7fa,
 	0x1d4bb,
 	0x1d7cb,
@@ -973,7 +1032,8 @@ static Rune __istitlep[] = {
 	0xa779, 0xa77b,
 	0xa780, 0xa786,
 	0xa78b, 0xa78d,
-	0xa7a0, 0xa7a8,
+	0xa790, 0xa792,
+	0xa7a0, 0xa7aa,
 };
 
 static Rune __istitles[] = {
@@ -990,12 +1050,14 @@ static Rune __istitles[] = {
 	0x038c,
 	0x03cf,
 	0x03f7,
+	0x10c7,
+	0x10cd,
 	0x2132,
 	0x2183,
 	0x2c60,
 	0x2c72,
 	0x2c75,
-	0xa790,
+	0x2cf2,
 };
 
 int
@@ -1088,6 +1150,7 @@ static Rune __toupperp[] = {
 	0xa733, 0xa76f, 1048575,
 	0xa77a, 0xa77c, 1048575,
 	0xa77f, 0xa787, 1048575,
+	0xa791, 0xa793, 1048575,
 	0xa7a1, 0xa7a9, 1048575,
 };
 
@@ -1132,6 +1195,7 @@ static Rune __touppers[] = {
 	0x0260, 1048371,
 	0x0263, 1048369,
 	0x0265, 1090856,
+	0x0266, 1090884,
 	0x0268, 1048367,
 	0x0269, 1048365,
 	0x026b, 1059319,
@@ -1178,8 +1242,10 @@ static Rune __touppers[] = {
 	0x2c66, 1037784,
 	0x2c73, 1048575,
 	0x2c76, 1048575,
+	0x2cf3, 1048575,
+	0x2d27, 1041312,
+	0x2d2d, 1041312,
 	0xa78c, 1048575,
-	0xa791, 1048575,
 };
 
 Rune
@@ -1271,6 +1337,7 @@ static Rune __tolowerp[] = {
 	0xa732, 0xa76e, 1048577,
 	0xa779, 0xa77b, 1048577,
 	0xa780, 0xa786, 1048577,
+	0xa790, 0xa792, 1048577,
 	0xa7a0, 0xa7a8, 1048577,
 };
 
@@ -1337,6 +1404,8 @@ static Rune __tolowers[] = {
 	0x03fa, 1048577,
 	0x04c0, 1048591,
 	0x04c1, 1048577,
+	0x10c7, 1055840,
+	0x10cd, 1055840,
 	0x1e9e, 1040961,
 	0x1fbc, 1048567,
 	0x1fcc, 1048567,
@@ -1357,11 +1426,12 @@ static Rune __tolowers[] = {
 	0x2c70, 1037794,
 	0x2c72, 1048577,
 	0x2c75, 1048577,
+	0x2cf2, 1048577,
 	0xa77d, 1013244,
 	0xa77e, 1048577,
 	0xa78b, 1048577,
 	0xa78d, 1006296,
-	0xa790, 1048577,
+	0xa7aa, 1006268,
 };
 
 Rune
@@ -1455,6 +1525,7 @@ static Rune __totitlep[] = {
 	0xa733, 0xa76f, 1048575,
 	0xa77a, 0xa77c, 1048575,
 	0xa77f, 0xa787, 1048575,
+	0xa791, 0xa793, 1048575,
 	0xa7a1, 0xa7a9, 1048575,
 };
 
@@ -1496,6 +1567,7 @@ static Rune __totitles[] = {
 	0x0260, 1048371,
 	0x0263, 1048369,
 	0x0265, 1090856,
+	0x0266, 1090884,
 	0x0268, 1048367,
 	0x0269, 1048365,
 	0x026b, 1059319,
@@ -1542,8 +1614,10 @@ static Rune __totitles[] = {
 	0x2c66, 1037784,
 	0x2c73, 1048575,
 	0x2c76, 1048575,
+	0x2cf3, 1048575,
+	0x2d27, 1041312,
+	0x2d2d, 1041312,
 	0xa78c, 1048575,
-	0xa791, 1048575,
 };
 
 Rune