X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/c825ef8c59b8c249cbbb50f46c9bbce54e5a42e9..4c1d9526da221c268d1fbbdcdbf89a424894c74c:/utf8.h diff --git a/utf8.h b/utf8.h index e0fc6b6..f60202b 100644 --- a/utf8.h +++ b/utf8.h @@ -34,11 +34,6 @@ #define FOLD_FLAGS_FULL 0x2 #define FOLD_FLAGS_NOMIX_ASCII 0x4 -/* For _core_swash_init(), internal core use only */ -#define _CORE_SWASH_INIT_USER_DEFINED_PROPERTY 0x1 -#define _CORE_SWASH_INIT_RETURN_IF_UNDEF 0x2 -#define _CORE_SWASH_INIT_ACCEPT_INVLIST 0x4 - /* =head1 Unicode Support L has an introduction to this API. @@ -74,8 +69,7 @@ the string is invariant. #define uvchr_to_utf8_flags_msgs(d,uv,flags,msgs) \ uvoffuni_to_utf8_flags_msgs(d,NATIVE_TO_UNI(uv),flags, msgs) #define utf8_to_uvchr_buf(s, e, lenp) \ - utf8n_to_uvchr(s, (U8*)(e) - (U8*)(s), lenp, \ - ckWARN_d(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY) + utf8_to_uvchr_buf_helper((const U8 *) (s), (const U8 *) e, lenp) #define utf8n_to_uvchr(s, len, lenp, flags) \ utf8n_to_uvchr_error(s, len, lenp, flags, 0) #define utf8n_to_uvchr_error(s, len, lenp, flags, errors) \ @@ -83,15 +77,6 @@ the string is invariant. #define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, FOLD_FLAGS_FULL) -#define to_utf8_fold(s, r, lenr) \ - _to_utf8_fold_flags (s, NULL, r, lenr, FOLD_FLAGS_FULL, __FILE__, __LINE__) -#define to_utf8_lower(s, r, lenr) \ - _to_utf8_lower_flags(s, NULL, r ,lenr, 0, __FILE__, __LINE__) -#define to_utf8_upper(s, r, lenr) \ - _to_utf8_upper_flags(s, NULL, r, lenr, 0, __FILE__, __LINE__) -#define to_utf8_title(s, r, lenr) \ - _to_utf8_title_flags(s, NULL, r, lenr ,0, __FILE__, __LINE__) - #define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \ foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0) #define FOLDEQ_UTF8_NOMIX_ASCII (1 << 0) @@ -114,11 +99,19 @@ the string is invariant. #else /* ! EBCDIC */ START_EXTERN_C -/* How wide can a single UTF-8 encoded character become in bytes. */ -/* NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 since UTF-8 - * is an encoding of Unicode, and Unicode's upper limit, 0x10FFFF, can be - * expressed with 4 bytes. However, Perl thinks of UTF-8 as a way to encode - * non-negative integers in a binary format, even those above Unicode */ +/* + +=for apidoc AmnU|STRLEN|UTF8_MAXBYTES + +The maximum width of a single UTF-8 encoded character, in bytes. + +NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 since UTF-8 +is an encoding of Unicode, and Unicode's upper limit, 0x10FFFF, can be +expressed with 4 bytes. However, Perl thinks of UTF-8 as a way to encode +non-negative integers in a binary format, even those above Unicode. + +=cut + */ #define UTF8_MAXBYTES 13 #ifdef DOINIT @@ -151,20 +144,53 @@ EXTCONST unsigned char PL_utf8skip[]; END_EXTERN_C -#if defined(_MSC_VER) && _MSC_VER < 1400 -/* older MSVC versions have a smallish macro buffer */ -#define PERL_SMALL_MACRO_BUFFER -#endif +/* -/* Native character to/from iso-8859-1. Are the identity functions on ASCII - * platforms */ -#ifdef PERL_SMALL_MACRO_BUFFER -#define NATIVE_TO_LATIN1(ch) ((U8)(ch)) -#define LATIN1_TO_NATIVE(ch) ((U8)(ch)) -#else -#define NATIVE_TO_LATIN1(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch))) -#define LATIN1_TO_NATIVE(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch))) -#endif +=for apidoc Am|U8|NATIVE_TO_LATIN1|U8 ch + +Returns the Latin-1 (including ASCII and control characters) equivalent of the +input native code point given by C. Thus, C on +EBCDIC platforms returns 65. These each represent the character C<"A"> on +their respective platforms. On ASCII platforms no conversion is needed, so +this macro expands to just its input, adding no time nor space requirements to +the implementation. + +For conversion of code points potentially larger than will fit in a character, +use L. + +=for apidoc Am|U8|LATIN1_TO_NATIVE|U8 ch + +Returns the native equivalent of the input Latin-1 code point (including ASCII +and control characters) given by C. Thus, C on +EBCDIC platforms returns 194. These each represent the character C<"B"> on +their respective platforms. On ASCII platforms no conversion is needed, so +this macro expands to just its input, adding no time nor space requirements to +the implementation. + +For conversion of code points potentially larger than will fit in a character, +use L. + +=for apidoc Am|UV|NATIVE_TO_UNI|UV ch + +Returns the Unicode equivalent of the input native code point given by C. +Thus, C on EBCDIC platforms returns 67. These each +represent the character C<"C"> on their respective platforms. On ASCII +platforms no conversion is needed, so this macro expands to just its input, +adding no time nor space requirements to the implementation. + +=for apidoc Am|UV|UNI_TO_NATIVE|UV ch + +Returns the native equivalent of the input Unicode code point given by C. +Thus, C on EBCDIC platforms returns 196. These each +represent the character C<"D"> on their respective platforms. On ASCII +platforms no conversion is needed, so this macro expands to just its input, +adding no time nor space requirements to the implementation. + +=cut +*/ + +#define NATIVE_TO_LATIN1(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) ((ch) | 0))) +#define LATIN1_TO_NATIVE(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) ((ch) | 0))) /* I8 is an intermediate version of UTF-8 used only in UTF-EBCDIC. We thus * consider it to be identical to UTF-8 on ASCII platforms. Strictly speaking @@ -172,17 +198,11 @@ END_EXTERN_C * because they are 8-bit encodings that serve the same purpose in Perl, and * rarely do we need to distinguish them. The term "NATIVE_UTF8" applies to * whichever one is applicable on the current platform */ -#ifdef PERL_SMALL_MACRO_BUFFER -#define NATIVE_UTF8_TO_I8(ch) ((U8) (ch)) -#define I8_TO_NATIVE_UTF8(ch) ((U8) (ch)) -#else -#define NATIVE_UTF8_TO_I8(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch))) -#define I8_TO_NATIVE_UTF8(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch))) -#endif +#define NATIVE_UTF8_TO_I8(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) ((ch) | 0))) +#define I8_TO_NATIVE_UTF8(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) ((ch) | 0))) -/* Transforms in wide UV chars */ -#define UNI_TO_NATIVE(ch) ((UV) (ch)) -#define NATIVE_TO_UNI(ch) ((UV) (ch)) +#define UNI_TO_NATIVE(ch) ((UV) ((ch) | 0)) +#define NATIVE_TO_UNI(ch) ((UV) ((ch) | 0)) /* @@ -237,66 +257,7 @@ Perl's extended UTF-8 means we can have start bytes up through FF, though any beginning with FF yields a code point that is too large for 32-bit ASCII platforms. FF signals to use 13 bytes for the encoded character. This breaks the paradigm that the number of leading bits gives how many total bytes there -are in the character. - -*/ - -/* Is the representation of the Unicode code point 'cp' the same regardless of - * being encoded in UTF-8 or not? */ -#define OFFUNI_IS_INVARIANT(cp) isASCII(cp) - -/* -=for apidoc Am|bool|UVCHR_IS_INVARIANT|UV cp - -Evaluates to 1 if the representation of code point C is the same whether or -not it is encoded in UTF-8; otherwise evaluates to 0. UTF-8 invariant -characters can be copied as-is when converting to/from UTF-8, saving time. -C is Unicode if above 255; otherwise is platform-native. - -=cut - */ - -#define UVCHR_IS_INVARIANT(cp) OFFUNI_IS_INVARIANT(cp) - -/* This defines the bits that are to be in the continuation bytes of a multi-byte - * UTF-8 encoded character that mark it is a continuation byte. */ -#define UTF_CONTINUATION_MARK 0x80 - -/* Misleadingly named: is the UTF8-encoded byte 'c' part of a variant sequence - * in UTF-8? This is the inverse of UTF8_IS_INVARIANT. The |0 makes sure this - * isn't mistakenly called with a ptr argument */ -#define UTF8_IS_CONTINUED(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ - ((U8)((c) | 0)) & UTF_CONTINUATION_MARK) - -/* Is the byte 'c' the first byte of a multi-byte UTF8-8 encoded sequence? - * This doesn't catch invariants (they are single-byte). It also excludes the - * illegal overlong sequences that begin with C0 and C1. The |0 makes sure - * this isn't mistakenly called with a ptr argument */ -#define UTF8_IS_START(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ - ((U8)((c) | 0)) >= 0xc2) - -/* For use in UTF8_IS_CONTINUATION() below */ -#define UTF_IS_CONTINUATION_MASK 0xC0 - -/* Is the byte 'c' part of a multi-byte UTF8-8 encoded sequence, and not the - * first byte thereof? The |0 makes sure this isn't mistakenly called with a - * ptr argument */ -#define UTF8_IS_CONTINUATION(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ - (((U8)((c) | 0)) & UTF_IS_CONTINUATION_MASK) == UTF_CONTINUATION_MARK) - -/* Is the UTF8-encoded byte 'c' the first byte of a two byte sequence? Use - * UTF8_IS_NEXT_CHAR_DOWNGRADEABLE() instead if the input isn't known to - * be well-formed. Masking with 0xfe allows the low bit to be 0 or 1; thus - * this matches 0xc[23]. The |0 makes sure this isn't mistakenly called with a - * ptr argument */ -#define UTF8_IS_DOWNGRADEABLE_START(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ - (((U8)((c) | 0)) & 0xfe) == 0xc2) - -/* Is the UTF8-encoded byte 'c' the first byte of a sequence of bytes that - * represent a code point > 255? The |0 makes sure this isn't mistakenly - * called with a ptr argument */ -#define UTF8_IS_ABOVE_LATIN1(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ - ((U8)((c) | 0)) >= 0xc4) +are in the character. */ /* This is the number of low-order bits a continuation byte in a UTF-8 encoded * sequence contributes to the specification of the code point. In the bit @@ -315,111 +276,48 @@ C is Unicode if above 255; otherwise is platform-native. #define isUTF8_POSSIBLY_PROBLEMATIC(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ (U8) c >= 0xED) -/* A helper macro for isUTF8_CHAR, so use that one instead of this. This was - * generated by regen/regcharclass.pl, and then moved here. Then it was - * hand-edited to add some LIKELY() calls, presuming that malformations are - * unlikely. The lines that generated it were then commented out. This was - * done because it takes on the order of 10 minutes to generate, and is never - * going to change, unless the generated code is improved, and figuring out - * the LIKELYs there would be hard. - * - UTF8_CHAR: Matches legal UTF-8 variant code points up through 0x1FFFFFF - - 0x80 - 0x1FFFFF -*/ -/*** GENERATED CODE ***/ -#define is_UTF8_CHAR_utf8_no_length_checks(s) \ -( ( 0xC2 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xDF ) ? \ - ( LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 ) \ -: ( 0xE0 == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( 0xE1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xEF ) ? \ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( 0xF0 == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( 0x90 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBF ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\ -: ( ( ( ( 0xF1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xF7 ) && LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ) && LIKELY( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 ) - -/* The above macro handles UTF-8 that has this start byte as the maximum */ -#define _IS_UTF8_CHAR_HIGHEST_START_BYTE 0xF7 - -/* A helper macro for isSTRICT_UTF8_CHAR, so use that one instead of this. - * Like is_UTF8_CHAR_utf8_no_length_checks(), this was moved here and LIKELYs - * added manually. - * - STRICT_UTF8_CHAR: Matches legal Unicode UTF-8 variant code points, no - surrrogates nor non-character code points -*/ -/*** GENERATED CODE ***/ -#define is_STRICT_UTF8_CHAR_utf8_no_length_checks(s) \ -( ( 0xC2 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xDF ) ? \ - ( LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 ) \ -: ( 0xE0 == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( ( 0xE1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xEC ) || 0xEE == ((const U8*)s)[0] ) ?\ - ( ( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( 0xED == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( 0xEF == ((const U8*)s)[0] ) ? \ - ( ( ( 0x80 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xB6 ) || ( 0xB8 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBE ) ) ?\ - ( LIKELY( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ? 3 : 0 ) \ - : ( 0xB7 == ((const U8*)s)[1] ) ? \ - ( LIKELY( ( ((const U8*)s)[2] & 0xF0 ) == 0x80 || ( ((const U8*)s)[2] & 0xF0 ) == 0xB0 ) ? 3 : 0 )\ - : ( ( 0xBF == ((const U8*)s)[1] ) && ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBD ) ) ? 3 : 0 )\ -: ( 0xF0 == ((const U8*)s)[0] ) ? \ - ( ( ( 0x90 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0x9E ) || ( 0xA0 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xAE ) || ( 0xB0 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBE ) ) ?\ - ( LIKELY( ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\ - : ( ((const U8*)s)[1] == 0x9F || ( ( ((const U8*)s)[1] & 0xEF ) == 0xAF ) ) ? \ - ( ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBE ) ? \ - ( LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ? 4 : 0 ) \ - : LIKELY( ( 0xBF == ((const U8*)s)[2] ) && ( 0x80 <= ((const U8*)s)[3] && ((const U8*)s)[3] <= 0xBD ) ) ? 4 : 0 )\ - : 0 ) \ -: ( 0xF1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xF3 ) ? \ - ( ( ( ( ((const U8*)s)[1] & 0xC8 ) == 0x80 ) || ( ( ((const U8*)s)[1] & 0xCC ) == 0x88 ) || ( ( ((const U8*)s)[1] & 0xCE ) == 0x8C ) || ( ( ((const U8*)s)[1] & 0xCF ) == 0x8E ) ) ?\ - ( LIKELY( ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\ - : ( ( ((const U8*)s)[1] & 0xCF ) == 0x8F ) ? \ - ( ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBE ) ? \ - ( LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ? 4 : 0 ) \ - : LIKELY( ( 0xBF == ((const U8*)s)[2] ) && ( 0x80 <= ((const U8*)s)[3] && ((const U8*)s)[3] <= 0xBD ) ) ? 4 : 0 )\ - : 0 ) \ -: ( 0xF4 == ((const U8*)s)[0] ) ? \ - ( ( 0x80 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0x8E ) ? \ - ( LIKELY( ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\ - : ( 0x8F == ((const U8*)s)[1] ) ? \ - ( ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBE ) ? \ - ( LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ? 4 : 0 ) \ - : LIKELY( ( 0xBF == ((const U8*)s)[2] ) && ( 0x80 <= ((const U8*)s)[3] && ((const U8*)s)[3] <= 0xBD ) ) ? 4 : 0 )\ - : 0 ) \ -: 0 ) - -/* Similarly, - C9_STRICT_UTF8_CHAR: Matches legal Unicode UTF-8 variant code - points, no surrogates - 0x0080 - 0xD7FF - 0xE000 - 0x10FFFF -*/ -/*** GENERATED CODE ***/ -#define is_C9_STRICT_UTF8_CHAR_utf8_no_length_checks(s) \ -( ( 0xC2 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xDF ) ? \ - ( LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 ) \ -: ( 0xE0 == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( ( 0xE1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xEC ) || ( ((const U8*)s)[0] & 0xFE ) == 0xEE ) ?\ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( 0xED == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\ -: ( 0xF0 == ((const U8*)s)[0] ) ? \ - ( LIKELY( ( ( 0x90 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBF ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\ -: ( 0xF1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xF3 ) ? \ - ( LIKELY( ( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\ -: LIKELY( ( ( ( 0xF4 == ((const U8*)s)[0] ) && ( ( ((const U8*)s)[1] & 0xF0 ) == 0x80 ) ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 ) - #define UNICODE_IS_PERL_EXTENDED(uv) UNLIKELY((UV) (uv) > 0x7FFFFFFF) #endif /* EBCDIC vs ASCII */ -/* 2**UTF_ACCUMULATION_SHIFT - 1 */ +/* 2**UTF_ACCUMULATION_SHIFT - 1. This masks out all but the bits that carry + * real information in a continuation byte. This turns out to be 0x3F in + * UTF-8, 0x1F in UTF-EBCDIC. */ #define UTF_CONTINUATION_MASK ((U8) ((1U << UTF_ACCUMULATION_SHIFT) - 1)) +/* For use in UTF8_IS_CONTINUATION(). This turns out to be 0xC0 in UTF-8, + * E0 in UTF-EBCDIC */ +#define UTF_IS_CONTINUATION_MASK ((U8) (0xFF << UTF_ACCUMULATION_SHIFT)) + +/* This defines the bits that are to be in the continuation bytes of a + * multi-byte UTF-8 encoded character that mark it is a continuation byte. + * This turns out to be 0x80 in UTF-8, 0xA0 in UTF-EBCDIC. (khw doesn't know + * the underlying reason that B0 works here) */ +#define UTF_CONTINUATION_MARK (UTF_IS_CONTINUATION_MASK & 0xB0) + +/* Is the byte 'c' part of a multi-byte UTF8-8 encoded sequence, and not the + * first byte thereof? */ +#define UTF8_IS_CONTINUATION(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ + (((NATIVE_UTF8_TO_I8(c) & UTF_IS_CONTINUATION_MASK) \ + == UTF_CONTINUATION_MARK))) + +/* Is the representation of the Unicode code point 'cp' the same regardless of + * being encoded in UTF-8 or not? This is a fundamental property of + * UTF-8,EBCDIC */ +#define OFFUNI_IS_INVARIANT(c) (((WIDEST_UTYPE)(c)) < UTF_CONTINUATION_MARK) + +/* +=for apidoc Am|bool|UVCHR_IS_INVARIANT|UV cp + +Evaluates to 1 if the representation of code point C is the same whether or +not it is encoded in UTF-8; otherwise evaluates to 0. UTF-8 invariant +characters can be copied as-is when converting to/from UTF-8, saving time. +C is Unicode if above 255; otherwise is platform-native. + +=cut + */ +#define UVCHR_IS_INVARIANT(cp) (OFFUNI_IS_INVARIANT(NATIVE_TO_UNI(cp))) + /* Internal macro to be used only in this file to aid in constructing other * publicly accessible macros. * The number of bytes required to express this uv in UTF-8, for just those @@ -478,6 +376,31 @@ encoded as UTF-8. C is a native (ASCII or EBCDIC) code point if less than */ #define UVCHR_SKIP(uv) ( UVCHR_IS_INVARIANT(uv) ? 1 : __BASE_UNI_SKIP(uv)) +#define UTF_MIN_START_BYTE \ + ((UTF_CONTINUATION_MARK >> UTF_ACCUMULATION_SHIFT) | UTF_START_MARK(2)) + +/* Is the byte 'c' the first byte of a multi-byte UTF8-8 encoded sequence? + * This doesn't catch invariants (they are single-byte). It also excludes the + * illegal overlong sequences that begin with C0 and C1 on ASCII platforms, and + * C0-C4 I8 start bytes on EBCDIC ones */ +#define UTF8_IS_START(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ + (NATIVE_UTF8_TO_I8(c) >= UTF_MIN_START_BYTE)) + +#define UTF_MIN_ABOVE_LATIN1_BYTE \ + ((0x100 >> UTF_ACCUMULATION_SHIFT) | UTF_START_MARK(2)) + +/* Is the UTF8-encoded byte 'c' the first byte of a sequence of bytes that + * represent a code point > 255? */ +#define UTF8_IS_ABOVE_LATIN1(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ + (NATIVE_UTF8_TO_I8(c) >= UTF_MIN_ABOVE_LATIN1_BYTE)) + +/* Is the UTF8-encoded byte 'c' the first byte of a two byte sequence? Use + * UTF8_IS_NEXT_CHAR_DOWNGRADEABLE() instead if the input isn't known to + * be well-formed. */ +#define UTF8_IS_DOWNGRADEABLE_START(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ + inRANGE(NATIVE_UTF8_TO_I8(c), \ + UTF_MIN_START_BYTE, UTF_MIN_ABOVE_LATIN1_BYTE - 1)) + /* The largest code point representable by two UTF-8 bytes on this platform. * As explained in the comments for __COMMON_UNI_SKIP, 32 start bytes with * UTF_ACCUMULATION_SHIFT bits of information each */ @@ -488,18 +411,26 @@ encoded as UTF-8. C is a native (ASCII or EBCDIC) code point if less than * continuation byte */ #define MAX_PORTABLE_UTF8_TWO_BYTE (32 * (1U << 5) - 1) -/* The maximum number of UTF-8 bytes a single Unicode character can - * uppercase/lowercase/fold into. Unicode guarantees that the maximum - * expansion is UTF8_MAX_FOLD_CHAR_EXPAND characters, but any above-Unicode - * code point will fold to itself, so we only have to look at the expansion of - * the maximum Unicode code point. But this number may be less than the space - * occupied by a very large code point under Perl's extended UTF-8. We have to - * make it large enough to fit any single character. (It turns out that ASCII - * and EBCDIC differ in which is larger) */ +/* + +=for apidoc AmnU|STRLEN|UTF8_MAXBYTES_CASE + +The maximum number of UTF-8 bytes a single Unicode character can +uppercase/lowercase/titlecase/fold into. + +=cut + + * Unicode guarantees that the maximum expansion is UTF8_MAX_FOLD_CHAR_EXPAND + * characters, but any above-Unicode code point will fold to itself, so we only + * have to look at the expansion of the maximum Unicode code point. But this + * number may be less than the space occupied by a very large code point under + * Perl's extended UTF-8. We have to make it large enough to fit any single + * character. (It turns out that ASCII and EBCDIC differ in which is larger) + * +=cut +*/ #define UTF8_MAXBYTES_CASE \ - (UTF8_MAXBYTES >= (UTF8_MAX_FOLD_CHAR_EXPAND * OFFUNISKIP(0x10FFFF)) \ - ? UTF8_MAXBYTES \ - : (UTF8_MAX_FOLD_CHAR_EXPAND * OFFUNISKIP(0x10FFFF))) + MAX(UTF8_MAXBYTES, UTF8_MAX_FOLD_CHAR_EXPAND * OFFUNISKIP(0x10FFFF)) /* Rest of these are attributes of Unicode and perl's internals rather than the * encoding, or happen to be the same in both ASCII and EBCDIC (at least at @@ -522,7 +453,7 @@ encoded as UTF-8. C is a native (ASCII or EBCDIC) code point if less than * UTF-8 encoded character that mark it as a start byte and give the number of * bytes that comprise the character. 'len' is the number of bytes in the * multi-byte sequence. */ -#define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFF & (0xFE << (7-(len))))) +#define UTF_START_MARK(len) (((len) > 7) ? 0xFF : ((U8) (0xFE << (7-(len))))) /* Masks out the initial one bits in a start byte, leaving the real data ones. * Doesn't work on an invariant byte. 'len' is the number of bytes in the @@ -537,7 +468,7 @@ encoded as UTF-8. C is a native (ASCII or EBCDIC) code point if less than * code point in process of being generated */ #define UTF8_ACCUMULATE(old, new) (__ASSERT_(FITS_IN_8_BITS(new)) \ ((old) << UTF_ACCUMULATION_SHIFT) \ - | ((NATIVE_UTF8_TO_I8((U8)new)) \ + | ((NATIVE_UTF8_TO_I8(new)) \ & UTF_CONTINUATION_MASK)) /* This works in the face of malformed UTF-8. */ @@ -592,14 +523,75 @@ encoded as UTF-8. C is a native (ASCII or EBCDIC) code point if less than /* =for apidoc Am|STRLEN|UTF8SKIP|char* s -returns the number of bytes in the UTF-8 encoded character whose first (perhaps -only) byte is pointed to by C. +returns the number of bytes a non-malformed UTF-8 encoded character whose first +(perhaps only) byte is pointed to by C. + +If there is a possibility of malformed input, use instead: + +=over + +=item L> if you know the maximum ending pointer in the +buffer pointed to by C; or + +=item L> if you don't know it. + +=back + +It is better to restructure your code so the end pointer is passed down so that +you know what it actually is at the point of this call, but if that isn't +possible, L> can minimize the chance of accessing beyond the end +of the input buffer. =cut */ #define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)] + +/* +=for apidoc Am|STRLEN|UTF8_SKIP|char* s +This is a synonym for L> + +=cut +*/ + #define UTF8_SKIP(s) UTF8SKIP(s) +/* +=for apidoc Am|STRLEN|UTF8_CHK_SKIP|char* s + +This is a safer version of L>, but still not as safe as +L>. This version doesn't blindly assume that the input +string pointed to by C is well-formed, but verifies that there isn't a NUL +terminating character before the expected end of the next character in C. +The length C returns stops just before any such NUL. + +Perl tends to add NULs, as an insurance policy, after the end of strings in +SV's, so it is likely that using this macro will prevent inadvertent reading +beyond the end of the input buffer, even if it is malformed UTF-8. + +This macro is intended to be used by XS modules where the inputs could be +malformed, and it isn't feasible to restructure to use the safer +L>, for example when interfacing with a C library. + +=cut +*/ + +#define UTF8_CHK_SKIP(s) \ + (s[0] == '\0' ? 1 : MIN(UTF8SKIP(s), \ + my_strnlen((char *) (s), UTF8SKIP(s)))) +/* + +=for apidoc Am|STRLEN|UTF8_SAFE_SKIP|char* s|char* e +returns 0 if S= e>>; otherwise returns the number of bytes in the +UTF-8 encoded character whose first byte is pointed to by C. But it never +returns beyond C. On DEBUGGING builds, it asserts that S= e>>. + +=cut + */ +#define UTF8_SAFE_SKIP(s, e) (__ASSERT_((e) >= (s)) \ + ((e) - (s)) <= 0 \ + ? 0 \ + : MIN(((e) - (s)), UTF8_SKIP(s))) + /* Most code that says 'UNI_' really means the native value for code points up * through 255 */ #define UNI_IS_INVARIANT(cp) UVCHR_IS_INVARIANT(cp) @@ -631,6 +623,11 @@ with a ptr argument. * above show, doesn't matter as to its implementation */ #define NATIVE_BYTE_IS_INVARIANT(c) UVCHR_IS_INVARIANT(c) +/* Misleadingly named: is the UTF8-encoded byte 'c' part of a variant sequence + * in UTF-8? This is the inverse of UTF8_IS_INVARIANT. */ +#define UTF8_IS_CONTINUED(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ + (! UTF8_IS_INVARIANT(c))) + /* The macros in the next 4 sets are used to generate the two utf8 or utfebcdic * bytes from an ordinal that is known to fit into exactly two (not one) bytes; * it must be less than 0x3FF to work across both encodings. */ @@ -706,6 +703,7 @@ with a ptr argument. ((IN_BYTES || !UTF) \ ? isWORDCHAR(*(p)) \ : isWORDCHAR_utf8_safe((U8 *) p, (U8 *) e)) +#define isALNUM_lazy_if_safe(p, e, UTF) isWORDCHAR_lazy_if_safe(p, e, UTF) #define isALNUM_lazy_if(p,UTF) \ _is_utf8_FOO(_CC_IDFIRST, (const U8 *) p, "isALNUM_lazy_if", \ @@ -751,7 +749,7 @@ case any call to string overloading updates the internal UTF-8 encoding flag. #define UTF8_ALLOW_CONTINUATION 0x0002 #define UTF8_GOT_CONTINUATION UTF8_ALLOW_CONTINUATION -/* Unexpected continuation byte */ +/* Unexpected non-continuation byte */ #define UTF8_ALLOW_NON_CONTINUATION 0x0004 #define UTF8_GOT_NON_CONTINUATION UTF8_ALLOW_NON_CONTINUATION @@ -835,8 +833,7 @@ case any call to string overloading updates the internal UTF-8 encoding flag. |UTF8_ALLOW_OVERFLOW) /* Accept any Perl-extended UTF-8 that evaluates to any UV on the platform, but - * not any malformed. This is the default. (Note that UVs above IV_MAX are - * deprecated. */ + * not any malformed. This is the default. */ #define UTF8_ALLOW_ANYUV 0 #define UTF8_ALLOW_DEFAULT UTF8_ALLOW_ANYUV @@ -856,6 +853,8 @@ point's representation. #define UTF8_IS_REPLACEMENT(s, send) is_REPLACEMENT_utf8_safe(s,send) +#define MAX_LEGAL_CP ((UV)IV_MAX) + /* =for apidoc Am|bool|UTF8_IS_SUPER|const U8 *s|const U8 *e @@ -872,7 +871,7 @@ represent a code point that cannot fit in a UV on the current platform. Hence this macro can give different results when run on a 64-bit word machine than on one with a 32-bit word size. -Note that it is deprecated to have code points that are larger than what can +Note that it is illegal to have code points that are larger than what can fit in an IV on the current machine. =cut @@ -889,14 +888,14 @@ fit in an IV on the current machine. && ( NATIVE_UTF8_TO_I8(*(s)) > 0xF9 \ || (NATIVE_UTF8_TO_I8(*((s) + 1)) >= 0xA2)) \ && LIKELY((s) + UTF8SKIP(s) <= (e))) \ - ? _is_utf8_char_helper(s, s + UTF8SKIP(s), 0) : 0) + ? is_utf8_char_helper(s, s + UTF8SKIP(s), 0) : 0) #else # define UTF8_IS_SUPER(s, e) \ (( LIKELY((e) > (s) + 3) \ && (*(U8*) (s)) >= 0xF4 \ && ((*(U8*) (s)) > 0xF4 || (*((U8*) (s) + 1) >= 0x90))\ && LIKELY((s) + UTF8SKIP(s) <= (e))) \ - ? _is_utf8_char_helper(s, s + UTF8SKIP(s), 0) : 0) + ? is_utf8_char_helper(s, s + UTF8SKIP(s), 0) : 0) #endif /* These are now machine generated, and the 'given' clause is no longer @@ -913,6 +912,10 @@ of the Unicode non-character code points; otherwise it evaluates to 0. If non-zero, the value gives how many bytes starting at C comprise the code point's representation. +=for apidoc AmnU|UV|UNICODE_REPLACEMENT + +Evaluates to 0xFFFD, the code point of the Unicode REPLACEMENT CHARACTER + =cut */ #define UTF8_IS_NONCHAR(s, e) \ @@ -936,6 +939,10 @@ point's representation. #define UNICODE_DISALLOW_NONCHAR 0x0020 #define UNICODE_DISALLOW_SUPER 0x0040 #define UNICODE_DISALLOW_PERL_EXTENDED 0x0080 + +#ifdef PERL_CORE +# define UNICODE_ALLOW_ABOVE_IV_MAX 0x0100 +#endif #define UNICODE_DISALLOW_ABOVE_31_BIT UNICODE_DISALLOW_PERL_EXTENDED #define UNICODE_GOT_SURROGATE UNICODE_DISALLOW_SURROGATE @@ -1022,134 +1029,12 @@ point's representation. #define SHARP_S_SKIP 2 -/* - -=for apidoc Am|STRLEN|isUTF8_CHAR|const U8 *s|const U8 *e - -Evaluates to non-zero if the first few bytes of the string starting at C and -looking no further than S> are well-formed UTF-8, as extended by Perl, -that represents some code point; otherwise it evaluates to 0. If non-zero, the -value gives how many bytes starting at C comprise the code point's -representation. Any bytes remaining before C, but beyond the ones needed to -form the first code point in C, are not examined. - -The code point can be any that will fit in a UV on this machine, using Perl's -extension to official UTF-8 to represent those higher than the Unicode maximum -of 0x10FFFF. That means that this macro is used to efficiently decide if the -next few bytes in C is legal UTF-8 for a single character. - -Use C> to restrict the acceptable code points to those -defined by Unicode to be fully interchangeable across applications; -C> to use the L definition of allowable -code points; and C> for a more customized definition. - -Use C>, C>, and -C> to check entire strings. - -Note that it is deprecated to use code points higher than what will fit in an -IV. This macro does not raise any warnings for such code points, treating them -as valid. - -Note also that a UTF-8 INVARIANT character (i.e. ASCII on non-EBCDIC machines) -is a valid UTF-8 character. - -=cut -*/ - -#define isUTF8_CHAR(s, e) \ - (UNLIKELY((e) <= (s)) \ - ? 0 \ - : (UTF8_IS_INVARIANT(*s)) \ - ? 1 \ - : UNLIKELY(((e) - (s)) < UTF8SKIP(s)) \ - ? 0 \ - : LIKELY(NATIVE_UTF8_TO_I8(*s) <= _IS_UTF8_CHAR_HIGHEST_START_BYTE) \ - ? is_UTF8_CHAR_utf8_no_length_checks(s) \ - : _is_utf8_char_helper(s, e, 0)) - #define is_utf8_char_buf(buf, buf_end) isUTF8_CHAR(buf, buf_end) #define bytes_from_utf8(s, lenp, is_utf8p) \ bytes_from_utf8_loc(s, lenp, is_utf8p, 0) /* -=for apidoc Am|STRLEN|isSTRICT_UTF8_CHAR|const U8 *s|const U8 *e - -Evaluates to non-zero if the first few bytes of the string starting at C and -looking no further than S> are well-formed UTF-8 that represents some -Unicode code point completely acceptable for open interchange between all -applications; otherwise it evaluates to 0. If non-zero, the value gives how -many bytes starting at C comprise the code point's representation. Any -bytes remaining before C, but beyond the ones needed to form the first code -point in C, are not examined. - -The largest acceptable code point is the Unicode maximum 0x10FFFF, and must not -be a surrogate nor a non-character code point. Thus this excludes any code -point from Perl's extended UTF-8. - -This is used to efficiently decide if the next few bytes in C is -legal Unicode-acceptable UTF-8 for a single character. - -Use C> to use the L definition of allowable -code points; C> to check for Perl's extended UTF-8; -and C> for a more customized definition. - -Use C>, C>, and -C> to check entire strings. - -=cut -*/ - -#define isSTRICT_UTF8_CHAR(s, e) \ - (UNLIKELY((e) <= (s)) \ - ? 0 \ - : (UTF8_IS_INVARIANT(*s)) \ - ? 1 \ - : UNLIKELY(((e) - (s)) < UTF8SKIP(s)) \ - ? 0 \ - : is_STRICT_UTF8_CHAR_utf8_no_length_checks(s)) - -/* - -=for apidoc Am|STRLEN|isC9_STRICT_UTF8_CHAR|const U8 *s|const U8 *e - -Evaluates to non-zero if the first few bytes of the string starting at C and -looking no further than S> are well-formed UTF-8 that represents some -Unicode non-surrogate code point; otherwise it evaluates to 0. If non-zero, -the value gives how many bytes starting at C comprise the code point's -representation. Any bytes remaining before C, but beyond the ones needed to -form the first code point in C, are not examined. - -The largest acceptable code point is the Unicode maximum 0x10FFFF. This -differs from C> only in that it accepts non-character -code points. This corresponds to -L. -which said that non-character code points are merely discouraged rather than -completely forbidden in open interchange. See -L. - -Use C> to check for Perl's extended UTF-8; and -C> for a more customized definition. - -Use C>, C>, and -C> to check entire strings. - -=cut -*/ - -#define isC9_STRICT_UTF8_CHAR(s, e) \ - (UNLIKELY((e) <= (s)) \ - ? 0 \ - : (UTF8_IS_INVARIANT(*s)) \ - ? 1 \ - : UNLIKELY(((e) - (s)) < UTF8SKIP(s)) \ - ? 0 \ - : is_C9_STRICT_UTF8_CHAR_utf8_no_length_checks(s)) - -/* - =for apidoc Am|STRLEN|isUTF8_CHAR_flags|const U8 *s|const U8 *e| const U32 flags Evaluates to non-zero if the first few bytes of the string starting at C and @@ -1185,7 +1070,7 @@ L to check entire strings. ? 1 \ : UNLIKELY(((e) - (s)) < UTF8SKIP(s)) \ ? 0 \ - : _is_utf8_char_helper(s, e, flags)) + : is_utf8_char_helper(s, e, flags)) /* Do not use; should be deprecated. Use isUTF8_CHAR() instead; this is * retained solely for backwards compatibility */