You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@xerces.apache.org by tn...@apache.org on 2001/11/09 22:27:21 UTC
cvs commit: xml-xerces/c/src/util/regx BlockRangeFactory.cpp RegxDefs.hpp
tng 01/11/09 13:27:21
Modified: c/src/util/regx BlockRangeFactory.cpp RegxDefs.hpp
Log:
Schema Regx Fix: manually add the extra specials and private use block range.
Revision Changes Path
1.6 +24 -15 xml-xerces/c/src/util/regx/BlockRangeFactory.cpp
Index: BlockRangeFactory.cpp
===================================================================
RCS file: /home/cvs/xml-xerces/c/src/util/regx/BlockRangeFactory.cpp,v
retrieving revision 1.5
retrieving revision 1.6
diff -u -r1.5 -r1.6
--- BlockRangeFactory.cpp 2001/11/09 18:10:09 1.5
+++ BlockRangeFactory.cpp 2001/11/09 21:27:21 1.6
@@ -56,6 +56,9 @@
/*
* $Log: BlockRangeFactory.cpp,v $
+ * Revision 1.6 2001/11/09 21:27:21 tng
+ * Schema Regx Fix: manually add the extra specials and private use block range.
+ *
* Revision 1.5 2001/11/09 18:10:09 tng
* Schema Regx: Block name 'X' should begin with 'IsX'. And add missing block names and range.
*
@@ -105,9 +108,12 @@
// ---------------------------------------------------------------------------
// Local static data
// ---------------------------------------------------------------------------
-const int BLOCKNAMESIZE = 68;
+const int BLOCKNAMESIZE = 96;
// Block Names IsX
+// only define Specials as FEFF..FEFF, missing Specials as FFF0..FFFD, add manually
+// only define private use as E000..F8FF,
+// missing 2 private use (F0000..FFFFD and 100000..10FFFD), add manually
const XMLCh fgBlockNames[][BLOCKNAMESIZE] =
{
{ chLatin_I, chLatin_s, chLatin_B, chLatin_a, chLatin_s, chLatin_i, chLatin_c, chLatin_L, chLatin_a,
@@ -320,8 +326,6 @@
chLatin_h, chLatin_a, chLatin_n, chLatin_d, chLatin_F, chLatin_u,
chLatin_l, chLatin_l, chLatin_w, chLatin_i, chLatin_d, chLatin_t, chLatin_h,
chLatin_F, chLatin_o, chLatin_r, chLatin_m, chLatin_s, chNull },
- { chLatin_I, chLatin_s, chLatin_S, chLatin_p, chLatin_e, chLatin_c, chLatin_i, chLatin_a, chLatin_l, chLatin_s,
- chNull },
{ chLatin_I, chLatin_s, chLatin_O, chLatin_l, chLatin_d, chLatin_I, chLatin_t, chLatin_a, chLatin_l, chLatin_i,
chLatin_c, chNull },
{ chLatin_I, chLatin_s, chLatin_G, chLatin_o, chLatin_t, chLatin_h, chLatin_i, chLatin_c, chNull },
@@ -346,10 +350,6 @@
chLatin_I, chLatin_d, chLatin_e, chLatin_o, chLatin_g, chLatin_r, chLatin_a, chLatin_p, chLatin_h, chLatin_s,
chLatin_S, chLatin_u, chLatin_p, chLatin_p, chLatin_l, chLatin_e, chLatin_m, chLatin_e, chLatin_n, chLatin_t, chNull },
{ chLatin_I, chLatin_s, chLatin_T, chLatin_a, chLatin_g, chLatin_s, chNull },
- { chLatin_I, chLatin_s, chLatin_P, chLatin_r, chLatin_i, chLatin_v, chLatin_a, chLatin_t, chLatin_e,
- chLatin_U, chLatin_s, chLatin_e, chNull },
- { chLatin_I, chLatin_s, chLatin_P, chLatin_r, chLatin_i, chLatin_v, chLatin_a, chLatin_t, chLatin_e,
- chLatin_U, chLatin_s, chLatin_e, chNull }
};
const XMLInt32 blockRanges[] =
@@ -368,9 +368,9 @@
0x31A0,0x31BF,0x3200,0x32FF,0x3300,0x33FF,0x3400,0x4DB5,0x4E00,0x9FFF,0xA000,0xA48F,
0xA490,0xA4CF,0xAC00,0xD7A3,0xD800,0xDB7F,0xDB80,0xDBFF,0xDC00,0xDFFF,0xE000,0xF8FF,
0xF900,0xFAFF,0xFB00,0xFB4F,0xFB50,0xFDFF,0xFE20,0xFE2F,0xFE30,0xFE4F,0xFE50,0xFE6F,
- 0xFE70,0xFEFE,0xFEFF,0xFEFF,0xFF00,0xFFEF,0xFFF0,0xFFFD,0x10300,0x1032F,0x10330,0x1034F,
+ 0xFE70,0xFEFE,0xFEFF,0xFEFF,0xFF00,0xFFEF,0x10300,0x1032F,0x10330,0x1034F,
0x10400,0x1044F,0x1D000,0x1D0FF,0x1D100,0x1D1FF,0x1D400,0x1D7FF,0x20000,0x2A6D6,0x2F800,0x2FA1F,
- 0xE0000,0xE007F,0xF0000,0xFFFFD,0x100000,0x10FFFD, chNull
+ 0xE0000,0xE007F, chNull
};
// ---------------------------------------------------------------------------
@@ -399,16 +399,25 @@
RangeTokenMap* rangeTokMap = RangeTokenMap::instance();
TokenFactory* tokFactory = rangeTokMap->getTokenFactory();
- for (int i=0; i < BLOCKNAMESIZE; i++) {
+ //for performance, once the desired specials and private use are found
+ //don't need to compareString anymore
+ bool foundSpecial = false;
+ bool foundPrivate = false;
- RangeToken* tok = tokFactory->createRange();
+ for (int i=0; i < BLOCKNAMESIZE; i++) {
+ RangeToken* tok = tokFactory->createRange();
tok->addRange(blockRanges[i*2], blockRanges[(i*2)+1]);
- if (XMLString::compareString((XMLCh*)fgBlockNames[i] , (XMLCh*) fgBlockSpecials) == 0) {
- tok->addRange(0xFFF0, 0xFFFd);
+ if (!foundSpecial && !XMLString::compareString((XMLCh*)fgBlockNames[i] , (XMLCh*) fgBlockIsSpecials)) {
+ tok->addRange(0xFFF0, 0xFFFD);
+ foundSpecial = true;
}
-
- rangeTokMap->setRangeToken(fgBlockNames[i], tok);
+ if (!foundPrivate && !XMLString::compareString((XMLCh*)fgBlockNames[i] , (XMLCh*) fgBlockIsPrivateUse)) {
+ tok->addRange(0xF0000, 0xFFFFD);
+ tok->addRange(0x100000, 0x10FFFD);
+ foundPrivate = true;
+ }
+ rangeTokMap->setRangeToken(fgBlockNames[i], tok);
}
fRangesCreated = true;
1.4 +9 -3 xml-xerces/c/src/util/regx/RegxDefs.hpp
Index: RegxDefs.hpp
===================================================================
RCS file: /home/cvs/xml-xerces/c/src/util/regx/RegxDefs.hpp,v
retrieving revision 1.3
retrieving revision 1.4
diff -u -r1.3 -r1.4
--- RegxDefs.hpp 2001/05/11 13:26:47 1.3
+++ RegxDefs.hpp 2001/11/09 21:27:21 1.4
@@ -55,7 +55,7 @@
*/
/*
- * $Id: RegxDefs.hpp,v 1.3 2001/05/11 13:26:47 tng Exp $
+ * $Id: RegxDefs.hpp,v 1.4 2001/11/09 21:27:21 tng Exp $
*/
#if !defined(REGXDEFS_HPP)
@@ -219,10 +219,16 @@
chLatin_N, chLatin_d, chNull
};
-const XMLCh fgBlockSpecials[] =
+const XMLCh fgBlockIsSpecials[] =
{
- chLatin_S, chLatin_p, chLatin_e, chLatin_c, chLatin_i, chLatin_a,
+ chLatin_I, chLatin_s, chLatin_S, chLatin_p, chLatin_e, chLatin_c, chLatin_i, chLatin_a,
chLatin_l, chLatin_s, chNull
+};
+
+const XMLCh fgBlockIsPrivateUse[] =
+{
+ chLatin_I, chLatin_s, chLatin_P, chLatin_r, chLatin_i, chLatin_v, chLatin_a, chLatin_t, chLatin_e,
+ chLatin_U, chLatin_s, chLatin_e, chNull
};
const XMLCh fgUniLetter[] =
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-cvs-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-cvs-help@xml.apache.org