michael@0: /* -*- Mode: C; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
michael@0: /* This Source Code Form is subject to the terms of the Mozilla Public
michael@0:  * License, v. 2.0. If a copy of the MPL was not distributed with this
michael@0:  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
michael@0: 
michael@0: #ifdef IS_LITTLE_ENDIAN
michael@0: 
michael@0: #define PREFIX(ident) little2_ ## ident
michael@0: #define BYTE_TYPE(p) LITTLE2_BYTE_TYPE(XmlGetUtf16InternalEncodingNS(), p)
michael@0: #define IS_NAME_CHAR_MINBPC(p) LITTLE2_IS_NAME_CHAR_MINBPC(0, p)
michael@0: #define IS_NMSTRT_CHAR_MINBPC(p) LITTLE2_IS_NMSTRT_CHAR_MINBPC(0, p)
michael@0: 
michael@0: #else
michael@0: 
michael@0: #define PREFIX(ident) big2_ ## ident
michael@0: #define BYTE_TYPE(p) BIG2_BYTE_TYPE(XmlGetUtf16InternalEncodingNS(), p)
michael@0: #define IS_NAME_CHAR_MINBPC(p) BIG2_IS_NAME_CHAR_MINBPC(0, p)
michael@0: #define IS_NMSTRT_CHAR_MINBPC(p) BIG2_IS_NMSTRT_CHAR_MINBPC(0, p)
michael@0: 
michael@0: #endif
michael@0: 
michael@0: #define MOZ_EXPAT_VALID_QNAME       (0)
michael@0: #define MOZ_EXPAT_EMPTY_QNAME       (1 << 0)
michael@0: #define MOZ_EXPAT_INVALID_CHARACTER (1 << 1)
michael@0: #define MOZ_EXPAT_MALFORMED         (1 << 2)
michael@0: 
michael@0: int MOZ_XMLCheckQName(const char* ptr, const char* end, int ns_aware,
michael@0:                       const char** colon)
michael@0: {
michael@0:   int result = MOZ_EXPAT_VALID_QNAME;
michael@0:   int nmstrt = 1;
michael@0:   *colon = 0;
michael@0:   if (ptr == end) {
michael@0:     return MOZ_EXPAT_EMPTY_QNAME;
michael@0:   }
michael@0:   do {
michael@0:     switch (BYTE_TYPE(ptr)) {
michael@0:     case BT_COLON:
michael@0:        /* We're namespace-aware and either first or last character is a colon
michael@0:           or we've already seen a colon. */
michael@0:       if (ns_aware && (nmstrt || *colon || ptr + 2 == end)) {
michael@0:         return MOZ_EXPAT_MALFORMED;
michael@0:       }
michael@0:       *colon = ptr;
michael@0:       nmstrt = ns_aware; /* e.g. "a:0" should be valid if !ns_aware */
michael@0:       break;
michael@0:     case BT_NONASCII:
michael@0:       if (nmstrt && !IS_NMSTRT_CHAR_MINBPC(ptr)) {
michael@0:         /* If this is a valid name character and we're namespace-aware, the
michael@0:            QName is malformed.  Otherwise, this character's invalid at the
michael@0:            start of a name (or, if we're namespace-aware, at the start of a
michael@0:            localpart). */
michael@0:         return (IS_NAME_CHAR_MINBPC(ptr) && ns_aware) ?
michael@0:                MOZ_EXPAT_MALFORMED :
michael@0:                MOZ_EXPAT_INVALID_CHARACTER;
michael@0:       }
michael@0:       if (!IS_NAME_CHAR_MINBPC(ptr)) {
michael@0:         return MOZ_EXPAT_INVALID_CHARACTER;
michael@0:       }
michael@0:       nmstrt = 0;
michael@0:       break;
michael@0:     case BT_NMSTRT:
michael@0:     case BT_HEX:
michael@0:       nmstrt = 0;
michael@0:       break;
michael@0:     case BT_DIGIT:
michael@0:     case BT_NAME:
michael@0:     case BT_MINUS:
michael@0:       if (nmstrt) {
michael@0:         return MOZ_EXPAT_INVALID_CHARACTER;
michael@0:       }
michael@0:       break;
michael@0:     default:
michael@0:       return MOZ_EXPAT_INVALID_CHARACTER;
michael@0:     }
michael@0:     ptr += 2;
michael@0:   } while (ptr != end);
michael@0:   return result;
michael@0: }
michael@0: 
michael@0: int MOZ_XMLIsLetter(const char* ptr)
michael@0: {
michael@0:   switch (BYTE_TYPE(ptr)) {
michael@0:   case BT_NONASCII:
michael@0:     if (!IS_NMSTRT_CHAR_MINBPC(ptr)) {
michael@0:       return 0;
michael@0:     }
michael@0:     /* fall through */
michael@0:   case BT_NMSTRT:
michael@0:   case BT_HEX:
michael@0:     return 1;
michael@0:   default:
michael@0:     return 0;
michael@0:   }
michael@0: }
michael@0: 
michael@0: int MOZ_XMLIsNCNameChar(const char* ptr)
michael@0: {
michael@0:   switch (BYTE_TYPE(ptr)) {
michael@0:   case BT_NONASCII:
michael@0:     if (!IS_NAME_CHAR_MINBPC(ptr)) {
michael@0:       return 0;
michael@0:     }
michael@0:     /* fall through */
michael@0:   case BT_NMSTRT:
michael@0:   case BT_HEX:
michael@0:   case BT_DIGIT:
michael@0:   case BT_NAME:
michael@0:   case BT_MINUS:
michael@0:     return 1;
michael@0:   default:
michael@0:     return 0;
michael@0:   }
michael@0: }
michael@0: 
michael@0: int MOZ_XMLTranslateEntity(const char* ptr, const char* end, const char** next,
michael@0:                            XML_Char* result)
michael@0: {
michael@0:   const ENCODING* enc = XmlGetUtf16InternalEncodingNS();
michael@0:   int tok = PREFIX(scanRef)(enc, ptr, end, next);
michael@0:   if (tok <= XML_TOK_INVALID) {
michael@0:     return 0;
michael@0:   }
michael@0: 
michael@0:   if (tok == XML_TOK_CHAR_REF) {
michael@0:     int n = XmlCharRefNumber(enc, ptr);
michael@0: 
michael@0:     /* We could get away with just < 0, but better safe than sorry. */
michael@0:     if (n <= 0) {
michael@0:       return 0;
michael@0:     }
michael@0: 
michael@0:     return XmlUtf16Encode(n, (unsigned short*)result);
michael@0:   }
michael@0: 
michael@0:   if (tok == XML_TOK_ENTITY_REF) {
michael@0:     /* *next points to after the semicolon, so the entity ends at
michael@0:        *next - enc->minBytesPerChar. */
michael@0:     XML_Char ch =
michael@0:       (XML_Char)XmlPredefinedEntityName(enc, ptr, *next - enc->minBytesPerChar);
michael@0:     if (!ch) {
michael@0:       return 0;
michael@0:     }
michael@0: 
michael@0:     *result = ch;
michael@0:     return 1;
michael@0:   }
michael@0: 
michael@0:   return 0;
michael@0: }
michael@0: 
michael@0: #undef PREFIX
michael@0: #undef BYTE_TYPE
michael@0: #undef IS_NAME_CHAR_MINBPC
michael@0: #undef IS_NMSTRT_CHAR_MINBPC