X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=utf8.c;h=1a024b719cf9da8a7d0939cb0536232037f2ec55;hb=564bd1e9882f812e2096523b4d5d7a449b2ae3ea;hp=244bb639bb048f992fac1f3fda5b05685169d820;hpb=b06226ff370ef661c3ff28e6f65e1ba0ef078609;p=p5sagit%2Fp5-mst-13.2.git diff --git a/utf8.c b/utf8.c index 244bb63..1a024b7 100644 --- a/utf8.c +++ b/utf8.c @@ -1,6 +1,6 @@ /* utf8.c * - * Copyright (c) 1998-2000, Larry Wall + * Copyright (c) 1998-2001, Larry Wall * * You may distribute under the terms of either the GNU General Public * License or the Artistic License, as specified in the README file. @@ -26,25 +26,51 @@ /* Unicode support */ +/* +=for apidoc A|U8 *|uvuni_to_utf8|U8 *d|UV uv + +Adds the UTF8 representation of the Unicode codepoint C to the end +of the string C; C should be have at least C free +bytes available. The return value is the pointer to the byte after the +end of the new character. In other words, + + d = uvuni_to_utf8(d, uv); + +is the recommended Unicode-aware way of saying + + *(d++) = uv; + +=cut +*/ + U8 * -Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) /* the d must be UTF8_MAXLEN+1 deep */ +Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv) { - if (uv < 0x80) { - *d++ = uv; - *d = 0; + if (UNI_IS_INVARIANT(uv)) { + *d++ = UTF_TO_NATIVE(uv); return d; } +#if defined(EBCDIC) + else { + STRLEN len = UNISKIP(uv); + U8 *p = d+len-1; + while (p > d) { + *p-- = UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); + uv >>= UTF_ACCUMULATION_SHIFT; + } + *p = UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); + return d+len; + } +#else /* Non loop style */ if (uv < 0x800) { *d++ = (( uv >> 6) | 0xc0); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } if (uv < 0x10000) { *d++ = (( uv >> 12) | 0xe0); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } if (uv < 0x200000) { @@ -52,7 +78,6 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) /* the d must be UTF8_MAXLEN+1 deep */ *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } if (uv < 0x4000000) { @@ -61,7 +86,6 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) /* the d must be UTF8_MAXLEN+1 deep */ *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } if (uv < 0x80000000) { @@ -71,7 +95,6 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) /* the d must be UTF8_MAXLEN+1 deep */ *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } #ifdef HAS_QUAD @@ -85,7 +108,6 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) /* the d must be UTF8_MAXLEN+1 deep */ *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } #ifdef HAS_QUAD @@ -103,15 +125,24 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) /* the d must be UTF8_MAXLEN+1 deep */ *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); - *d = 0; return d; } #endif +#endif /* Loop style */ } -/* Tests if some arbitrary number of bytes begins in a valid UTF-8 character. - * The actual number of bytes in the UTF-8 character will be returned if it - * is valid, otherwise 0. */ + + +/* +=for apidoc A|STRLEN|is_utf8_char|U8 *s + +Tests if some arbitrary number of bytes begins in a valid UTF-8 +character. Note that an INVARIANT (i.e. ASCII) character is a valid UTF-8 character. +The actual number of bytes in the UTF-8 character will be returned if +it is valid, otherwise 0. + +=cut +*/ STRLEN Perl_is_utf8_char(pTHX_ U8 *s) { @@ -119,26 +150,27 @@ Perl_is_utf8_char(pTHX_ U8 *s) STRLEN slen, len; UV uv, ouv; - if (u <= 0x7f) + if (UTF8_IS_INVARIANT(u)) return 1; - if (u >= 0x80 && u <= 0xbf) + if (!UTF8_IS_START(u)) return 0; len = UTF8SKIP(s); - if (len < 2 || (u >= 0xc0 && u <= 0xfd && s[1] < 0x80)) + if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) return 0; slen = len - 1; s++; - uv = u; + u &= UTF_START_MASK(len); + uv = u; ouv = uv; while (slen--) { - if ((*s & 0xc0) != 0x80) + if (!UTF8_IS_CONTINUATION(*s)) return 0; - uv = (uv << 6) | (*s & 0x3f); - if (uv < ouv) + uv = UTF8_ACCUMULATE(uv, *s); + if (uv < ouv) return 0; ouv = uv; s++; @@ -151,10 +183,12 @@ Perl_is_utf8_char(pTHX_ U8 *s) } /* -=for apidoc Am|is_utf8_string|U8 *s|STRLEN len +=for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len -Returns true if first C bytes of the given string form valid a UTF8 -string, false otherwise. +Returns true if first C bytes of the given string form a valid UTF8 +string, false otherwise. Note that 'a valid UTF8 string' does not mean +'a string that contains UTF8' because a valid ASCII string is a valid +UTF8 string. =cut */ @@ -163,109 +197,126 @@ bool Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len) { U8* x = s; - U8* send = s + len; + U8* send; STRLEN c; + if (!len) + len = strlen((char *)s); + send = s + len; + while (x < send) { c = is_utf8_char(x); if (!c) return FALSE; x += c; - if (x > send) - return FALSE; } + if (x != send) + return FALSE; return TRUE; } /* -=for apidoc Am|U8* s|utf8_to_uv|STRLEN curlen|STRLEN *retlen|U32 flags +=for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags -Returns the character value of the first character in the string C +Bottom level UTF-8 decode routine. +Returns the unicode code point value of the first character in the string C which is assumed to be in UTF8 encoding and no longer than C; -C will be set to the length, in bytes, of that character, -and the pointer C will be advanced to the end of the character. +C will be set to the length, in bytes, of that character. If C does not point to a well-formed UTF8 character, the behaviour is dependent on the value of C: if it contains UTF8_CHECK_ONLY, it is assumed that the caller will raise a warning, and this function -will set C to C<-1> and return. The C can also contain -various flags to allow deviations from the strict UTF-8 encoding -(see F). +will silently just set C to C<-1> and return zero. If the +C does not contain UTF8_CHECK_ONLY, warnings about +malformations will be given, C will be set to the expected +length of the UTF-8 character in bytes, and zero will be returned. -=cut */ +The C can also contain various flags to allow deviations from +the strict UTF-8 encoding (see F). + +Most code should use utf8_to_uvchr() rather than call this directly. + +=cut +*/ UV -Perl_utf8_to_uv(pTHX_ U8* s, STRLEN curlen, STRLEN* retlen, U32 flags) +Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) { - UV uv = *s, ouv; + UV uv = *s, ouv = 0; STRLEN len = 1; -#ifdef EBCDIC - bool dowarn = 0; -#else bool dowarn = ckWARN_d(WARN_UTF8); -#endif STRLEN expectlen = 0; - - if (curlen == 0) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (an empty string)"); + U32 warning = 0; + +/* This list is a superset of the UTF8_ALLOW_XXX. */ + +#define UTF8_WARN_EMPTY 1 +#define UTF8_WARN_CONTINUATION 2 +#define UTF8_WARN_NON_CONTINUATION 3 +#define UTF8_WARN_FE_FF 4 +#define UTF8_WARN_SHORT 5 +#define UTF8_WARN_OVERFLOW 6 +#define UTF8_WARN_SURROGATE 7 +#define UTF8_WARN_BOM 8 +#define UTF8_WARN_LONG 9 +#define UTF8_WARN_FFFF 10 + + if (curlen == 0 && + !(flags & UTF8_ALLOW_EMPTY)) { + warning = UTF8_WARN_EMPTY; goto malformed; } - if (uv <= 0x7f) { /* Pure ASCII. */ + if (UTF8_IS_INVARIANT(uv)) { if (retlen) *retlen = 1; - return *s; + return (UV) (NATIVE_TO_UTF(*s)); } - if ((uv >= 0x80 && uv <= 0xbf) && + if (UTF8_IS_CONTINUATION(uv) && !(flags & UTF8_ALLOW_CONTINUATION)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (unexpected continuation byte 0x%02"UVxf")", - uv); + warning = UTF8_WARN_CONTINUATION; goto malformed; } - if ((uv >= 0xc0 && uv <= 0xfd && curlen > 1 && s[1] < 0x80) && + if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && !(flags & UTF8_ALLOW_NON_CONTINUATION)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (unexpected non-continuation byte 0x%02"UVxf" after byte 0x%02"UVxf")", - (UV)s[1], uv); + warning = UTF8_WARN_NON_CONTINUATION; goto malformed; } - + +#ifdef EBCDIC + uv = NATIVE_TO_UTF(uv); +#else if ((uv == 0xfe || uv == 0xff) && !(flags & UTF8_ALLOW_FE_FF)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (byte 0x%02"UVxf")", - uv); + warning = UTF8_WARN_FE_FF; goto malformed; } - +#endif + if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } +#ifdef EBCDIC + else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } + else { len = 7; uv &= 0x01; } +#else else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } else if (!(uv & 0x01)) { len = 7; uv = 0; } - else { len = 13; uv = 0; } /* whoa! */ - + else { len = 13; uv = 0; } /* whoa! */ +#endif + if (retlen) *retlen = len; - + expectlen = len; if ((curlen < expectlen) && !(flags & UTF8_ALLOW_SHORT)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (%d byte%s, need %d)", - curlen, curlen == 1 ? "" : "s", expectlen); + warning = UTF8_WARN_SHORT; goto malformed; } @@ -274,54 +325,47 @@ Perl_utf8_to_uv(pTHX_ U8* s, STRLEN curlen, STRLEN* retlen, U32 flags) ouv = uv; while (len--) { - if ((*s & 0xc0) != 0x80) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (unexpected continuation byte 0x%02x)", - *s); + if (!UTF8_IS_CONTINUATION(*s) && + !(flags & UTF8_ALLOW_NON_CONTINUATION)) { + s--; + warning = UTF8_WARN_NON_CONTINUATION; goto malformed; } else - uv = (uv << 6) | (*s & 0x3f); - if (uv < ouv) { - /* This cannot be allowed. */ - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (overflow at 0x%"UVxf", byte 0x%02x)", - ouv, *s); - goto malformed; + uv = UTF8_ACCUMULATE(uv, *s); + if (!(uv > ouv)) { + /* These cannot be allowed. */ + if (uv == ouv) { + if (!(flags & UTF8_ALLOW_LONG)) { + warning = UTF8_WARN_LONG; + goto malformed; + } + } + else { /* uv < ouv */ + /* This cannot be allowed. */ + warning = UTF8_WARN_OVERFLOW; + goto malformed; + } } s++; ouv = uv; } - if ((uv >= 0xd800 && uv <= 0xdfff) && + if (UNICODE_IS_SURROGATE(uv) && !(flags & UTF8_ALLOW_SURROGATE)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (UTF-16 surrogate 0x%04"UVxf")", - uv); + warning = UTF8_WARN_SURROGATE; goto malformed; - } else if ((uv == 0xfffe) && + } else if (UNICODE_IS_BYTE_ORDER_MARK(uv) && !(flags & UTF8_ALLOW_BOM)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (byte order mark 0x%04"UVxf")", - uv); + warning = UTF8_WARN_BOM; goto malformed; } else if ((expectlen > UNISKIP(uv)) && !(flags & UTF8_ALLOW_LONG)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (%d byte%s, need %d)", - expectlen, expectlen == 1 ? "": "s", UNISKIP(uv)); + warning = UTF8_WARN_LONG; goto malformed; - } else if ((uv == 0xffff) && + } else if (UNICODE_IS_ILLEGAL(uv) && !(flags & UTF8_ALLOW_FFFF)) { - if (dowarn) - Perl_warner(aTHX_ WARN_UTF8, - "Malformed UTF-8 character (character 0x%04"UVxf")", - uv); + warning = UTF8_WARN_FFFF; goto malformed; } @@ -335,19 +379,96 @@ malformed: return 0; } + if (dowarn) { + SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0)); + + switch (warning) { + case 0: /* Intentionally empty. */ break; + case UTF8_WARN_EMPTY: + Perl_sv_catpvf(aTHX_ sv, "(empty string)"); + break; + case UTF8_WARN_CONTINUATION: + Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf")", uv); + break; + case UTF8_WARN_NON_CONTINUATION: + Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf" after start byte 0x%02"UVxf")", + (UV)s[1], uv); + break; + case UTF8_WARN_FE_FF: + Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); + break; + case UTF8_WARN_SHORT: + Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d)", + curlen, curlen == 1 ? "" : "s", expectlen); + expectlen = curlen; /* distance for caller to skip */ + break; + case UTF8_WARN_OVERFLOW: + Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x)", + ouv, *s); + break; + case UTF8_WARN_SURROGATE: + Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); + break; + case UTF8_WARN_BOM: + Perl_sv_catpvf(aTHX_ sv, "(byte order mark 0x%04"UVxf")", uv); + break; + case UTF8_WARN_LONG: + Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d)", + expectlen, expectlen == 1 ? "": "s", UNISKIP(uv)); + break; + case UTF8_WARN_FFFF: + Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv); + break; + default: + Perl_sv_catpvf(aTHX_ sv, "(unknown reason)"); + break; + } + + if (warning) { + char *s = SvPVX(sv); + + if (PL_op) + Perl_warner(aTHX_ WARN_UTF8, + "%s in %s", s, OP_DESC(PL_op)); + else + Perl_warner(aTHX_ WARN_UTF8, "%s", s); + } + } + if (retlen) *retlen = expectlen ? expectlen : len; - return UNICODE_REPLACEMENT_CHARACTER; + return 0; +} + +/* +=for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen + +Returns the native character value of the first character in the string C +which is assumed to be in UTF8 encoding; C will be set to the +length, in bytes, of that character. + +If C does not point to a well-formed UTF8 character, zero is +returned and retlen is set, if possible, to -1. + +=cut +*/ + +UV +Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen) +{ + return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen, 0); } /* -=for apidoc Am|U8* s|utf8_to_uv_simple|STRLEN *retlen +=for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen -Returns the character value of the first character in the string C +Returns the Unicode code point of the first character in the string C which is assumed to be in UTF8 encoding; C will be set to the -length, in bytes, of that character, and the pointer C will be -advanced to the end of the character. +length, in bytes, of that character. + +This function should only be used when returned UV is considered +an index into the Unicode semantic tables (e.g. swashes). If C does not point to a well-formed UTF8 character, zero is returned and retlen is set, if possible, to -1. @@ -356,13 +477,14 @@ returned and retlen is set, if possible, to -1. */ UV -Perl_utf8_to_uv_simple(pTHX_ U8* s, STRLEN* retlen) +Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen) { - return Perl_utf8_to_uv(aTHX_ s, UTF8_MAXLEN, retlen, 0); + /* Call the low level routine asking for checks */ + return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen, 0); } /* -=for apidoc Am|STRLEN|utf8_length|U8* s|U8 *e +=for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e Return the length of the UTF-8 char encoded string C in characters. Stops at C (inclusive). If C s> or if the scan would end @@ -372,10 +494,14 @@ up past C, croaks. */ STRLEN -Perl_utf8_length(pTHX_ U8* s, U8* e) +Perl_utf8_length(pTHX_ U8 *s, U8 *e) { STRLEN len = 0; + /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. + * the bitops (especially ~) can create illegal UTF-8. + * In other words: in Perl UTF-8 is not just for Unicode. */ + if (e < s) Perl_croak(aTHX_ "panic: utf8_length: unexpected end"); while (s < e) { @@ -391,7 +517,7 @@ Perl_utf8_length(pTHX_ U8* s, U8* e) } /* -=for apidoc Am|IV|utf8_distance|U8 *a|U8 *b +=for apidoc A|IV|utf8_distance|U8 *a|U8 *b Returns the number of UTF8 characters between the UTF-8 pointers C and C. @@ -399,13 +525,18 @@ and C. WARNING: use only if you *know* that the pointers point inside the same UTF-8 buffer. -=cut */ +=cut +*/ IV Perl_utf8_distance(pTHX_ U8 *a, U8 *b) { IV off = 0; + /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. + * the bitops (especially ~) can create illegal UTF-8. + * In other words: in Perl UTF-8 is not just for Unicode. */ + if (a < b) { while (a < b) { U8 c = UTF8SKIP(a); @@ -431,19 +562,25 @@ Perl_utf8_distance(pTHX_ U8 *a, U8 *b) } /* -=for apidoc Am|U8*|utf8_hop|U8 *s|I32 off +=for apidoc A|U8 *|utf8_hop|U8 *s|I32 off -Move the C pointing to UTF-8 data by C characters, either forward -or backward. +Return the UTF-8 pointer C displaced by C characters, either +forward or backward. WARNING: do not use the following unless you *know* C is within -the UTF-8 buffer pointed to by C. +the UTF-8 data pointed to by C *and* that on entry C is aligned +on the first byte of character or just after the last byte of a character. -=cut */ +=cut +*/ U8 * Perl_utf8_hop(pTHX_ U8 *s, I32 off) { + /* Note: cannot use UTF8_IS_...() too eagerly here since e.g + * the bitops (especially ~) can create illegal UTF-8. + * In other words: in Perl UTF-8 is not just for Unicode. */ + if (off >= 0) { while (off--) s += UTF8SKIP(s); @@ -451,17 +588,15 @@ Perl_utf8_hop(pTHX_ U8 *s, I32 off) else { while (off++) { s--; - if (*s & 0x80) { - while ((*s & 0xc0) == 0x80) - s--; - } + while (UTF8_IS_CONTINUATION(*s)) + s--; } } return s; } /* -=for apidoc Am|U8 *|utf8_to_bytes|U8 *s|STRLEN *len +=for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len Converts a string C of length C from UTF8 into byte encoding. Unlike C, this over-writes the original string, and @@ -472,7 +607,7 @@ Returns zero on failure, setting C to -1. */ U8 * -Perl_utf8_to_bytes(pTHX_ U8* s, STRLEN *len) +Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) { U8 *send; U8 *d; @@ -482,9 +617,9 @@ Perl_utf8_to_bytes(pTHX_ U8* s, STRLEN *len) for (send = s + *len; s < send; ) { U8 c = *s++; - if (c >= 0x80 && - ((s >= send) || - ((*s++ & 0xc0) != 0x80) || ((c & 0xfe) != 0xc2))) { + if (!UTF8_IS_INVARIANT(c) && + (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) + || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { *len = -1; return 0; } @@ -492,14 +627,9 @@ Perl_utf8_to_bytes(pTHX_ U8* s, STRLEN *len) d = s = save; while (s < send) { - if (*s < 0x80) { - *d++ = *s++; - } - else { - STRLEN ulen; - *d++ = (U8)utf8_to_uv_simple(s, &ulen); - s += ulen; - } + STRLEN ulen; + *d++ = (U8)utf8_to_uvchr(s, &ulen); + s += ulen; } *d = '\0'; *len = d - save; @@ -507,7 +637,61 @@ Perl_utf8_to_bytes(pTHX_ U8* s, STRLEN *len) } /* -=for apidoc Am|U8 *|bytes_to_utf8|U8 *s|STRLEN *len +=for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8 + +Converts a string C of length C from UTF8 into byte encoding. +Unlike but like C, returns a pointer to +the newly-created string, and updates C to contain the new +length. Returns the original string if no conversion occurs, C +is unchanged. Do nothing if C points to 0. Sets C to +0 if C is converted or contains all 7bit characters. + +=cut +*/ + +U8 * +Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8) +{ + U8 *d; + U8 *start = s; + U8 *send; + I32 count = 0; + + if (!*is_utf8) + return start; + + /* ensure valid UTF8 and chars < 256 before converting string */ + for (send = s + *len; s < send;) { + U8 c = *s++; + if (!UTF8_IS_INVARIANT(c)) { + if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && + (c = *s++) && UTF8_IS_CONTINUATION(c)) + count++; + else + return start; + } + } + + *is_utf8 = 0; + + Newz(801, d, (*len) - count + 1, U8); + s = start; start = d; + while (s < send) { + U8 c = *s++; + if (!UTF8_IS_INVARIANT(c)) { + /* Then it is two-byte encoded */ + c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); + c = ASCII_TO_NATIVE(c); + } + *d++ = c; + } + *d = '\0'; + *len = d - start; + return start; +} + +/* +=for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len Converts a string C of length C from ASCII into UTF8 encoding. Returns a pointer to the newly-created string, and sets C to @@ -517,7 +701,7 @@ reflect the new length. */ U8* -Perl_bytes_to_utf8(pTHX_ U8* s, STRLEN *len) +Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len) { U8 *send; U8 *d; @@ -528,12 +712,12 @@ Perl_bytes_to_utf8(pTHX_ U8* s, STRLEN *len) dst = d; while (s < send) { - if (*s < 0x80) - *d++ = *s++; + UV uv = NATIVE_TO_ASCII(*s++); + if (UNI_IS_INVARIANT(uv)) + *d++ = UTF_TO_NATIVE(uv); else { - UV uv = *s++; - *d++ = (( uv >> 6) | 0xc0); - *d++ = (( uv & 0x3f) | 0x80); + *d++ = UTF8_EIGHT_BIT_HI(uv); + *d++ = UTF8_EIGHT_BIT_LO(uv); } } *d = '\0'; @@ -613,245 +797,235 @@ Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) /* for now these are all defined (inefficiently) in terms of the utf8 versions */ bool -Perl_is_uni_alnum(pTHX_ U32 c) +Perl_is_uni_alnum(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alnum(tmpbuf); } bool -Perl_is_uni_alnumc(pTHX_ U32 c) +Perl_is_uni_alnumc(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alnumc(tmpbuf); } bool -Perl_is_uni_idfirst(pTHX_ U32 c) +Perl_is_uni_idfirst(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_idfirst(tmpbuf); } bool -Perl_is_uni_alpha(pTHX_ U32 c) +Perl_is_uni_alpha(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alpha(tmpbuf); } bool -Perl_is_uni_ascii(pTHX_ U32 c) +Perl_is_uni_ascii(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_ascii(tmpbuf); } bool -Perl_is_uni_space(pTHX_ U32 c) +Perl_is_uni_space(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_space(tmpbuf); } bool -Perl_is_uni_digit(pTHX_ U32 c) +Perl_is_uni_digit(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_digit(tmpbuf); } bool -Perl_is_uni_upper(pTHX_ U32 c) +Perl_is_uni_upper(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_upper(tmpbuf); } bool -Perl_is_uni_lower(pTHX_ U32 c) +Perl_is_uni_lower(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_lower(tmpbuf); } bool -Perl_is_uni_cntrl(pTHX_ U32 c) +Perl_is_uni_cntrl(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_cntrl(tmpbuf); } bool -Perl_is_uni_graph(pTHX_ U32 c) +Perl_is_uni_graph(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_graph(tmpbuf); } bool -Perl_is_uni_print(pTHX_ U32 c) +Perl_is_uni_print(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_print(tmpbuf); } bool -Perl_is_uni_punct(pTHX_ U32 c) +Perl_is_uni_punct(pTHX_ UV c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_punct(tmpbuf); } bool -Perl_is_uni_xdigit(pTHX_ U32 c) +Perl_is_uni_xdigit(pTHX_ UV c) { - U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); + U8 tmpbuf[UTF8_MAXLEN*2+1]; + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_xdigit(tmpbuf); } -U32 -Perl_to_uni_upper(pTHX_ U32 c) +UV +Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) { - U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); - return to_utf8_upper(tmpbuf); + U8 tmpbuf[UTF8_MAXLEN*2+1]; + uvchr_to_utf8(tmpbuf, (UV)c); + return to_utf8_upper(tmpbuf, p, lenp); +} + +UV +Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) +{ + U8 tmpbuf[UTF8_MAXLEN*2+1]; + uvchr_to_utf8(tmpbuf, (UV)c); + return to_utf8_title(tmpbuf, p, lenp); } -U32 -Perl_to_uni_title(pTHX_ U32 c) +UV +Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); - return to_utf8_title(tmpbuf); + uvchr_to_utf8(tmpbuf, (UV)c); + return to_utf8_lower(tmpbuf, p, lenp); } -U32 -Perl_to_uni_lower(pTHX_ U32 c) +UV +Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) { U8 tmpbuf[UTF8_MAXLEN+1]; - uv_to_utf8(tmpbuf, (UV)c); - return to_utf8_lower(tmpbuf); + uvchr_to_utf8(tmpbuf, (UV)c); + return to_utf8_fold(tmpbuf, p, lenp); } /* for now these all assume no locale info available for Unicode > 255 */ bool -Perl_is_uni_alnum_lc(pTHX_ U32 c) +Perl_is_uni_alnum_lc(pTHX_ UV c) { return is_uni_alnum(c); /* XXX no locale support yet */ } bool -Perl_is_uni_alnumc_lc(pTHX_ U32 c) +Perl_is_uni_alnumc_lc(pTHX_ UV c) { return is_uni_alnumc(c); /* XXX no locale support yet */ } bool -Perl_is_uni_idfirst_lc(pTHX_ U32 c) +Perl_is_uni_idfirst_lc(pTHX_ UV c) { return is_uni_idfirst(c); /* XXX no locale support yet */ } bool -Perl_is_uni_alpha_lc(pTHX_ U32 c) +Perl_is_uni_alpha_lc(pTHX_ UV c) { return is_uni_alpha(c); /* XXX no locale support yet */ } bool -Perl_is_uni_ascii_lc(pTHX_ U32 c) +Perl_is_uni_ascii_lc(pTHX_ UV c) { return is_uni_ascii(c); /* XXX no locale support yet */ } bool -Perl_is_uni_space_lc(pTHX_ U32 c) +Perl_is_uni_space_lc(pTHX_ UV c) { return is_uni_space(c); /* XXX no locale support yet */ } bool -Perl_is_uni_digit_lc(pTHX_ U32 c) +Perl_is_uni_digit_lc(pTHX_ UV c) { return is_uni_digit(c); /* XXX no locale support yet */ } bool -Perl_is_uni_upper_lc(pTHX_ U32 c) +Perl_is_uni_upper_lc(pTHX_ UV c) { return is_uni_upper(c); /* XXX no locale support yet */ } bool -Perl_is_uni_lower_lc(pTHX_ U32 c) +Perl_is_uni_lower_lc(pTHX_ UV c) { return is_uni_lower(c); /* XXX no locale support yet */ } bool -Perl_is_uni_cntrl_lc(pTHX_ U32 c) +Perl_is_uni_cntrl_lc(pTHX_ UV c) { return is_uni_cntrl(c); /* XXX no locale support yet */ } bool -Perl_is_uni_graph_lc(pTHX_ U32 c) +Perl_is_uni_graph_lc(pTHX_ UV c) { return is_uni_graph(c); /* XXX no locale support yet */ } bool -Perl_is_uni_print_lc(pTHX_ U32 c) +Perl_is_uni_print_lc(pTHX_ UV c) { return is_uni_print(c); /* XXX no locale support yet */ } bool -Perl_is_uni_punct_lc(pTHX_ U32 c) +Perl_is_uni_punct_lc(pTHX_ UV c) { return is_uni_punct(c); /* XXX no locale support yet */ } bool -Perl_is_uni_xdigit_lc(pTHX_ U32 c) +Perl_is_uni_xdigit_lc(pTHX_ UV c) { return is_uni_xdigit(c); /* XXX no locale support yet */ } -U32 -Perl_to_uni_upper_lc(pTHX_ U32 c) -{ - return to_uni_upper(c); /* XXX no locale support yet */ -} - -U32 -Perl_to_uni_title_lc(pTHX_ U32 c) -{ - return to_uni_title(c); /* XXX no locale support yet */ -} - -U32 -Perl_to_uni_lower_lc(pTHX_ U32 c) -{ - return to_uni_lower(c); /* XXX no locale support yet */ -} - bool Perl_is_utf8_alnum(pTHX_ U8 *p) { @@ -862,13 +1036,13 @@ Perl_is_utf8_alnum(pTHX_ U8 *p) * descendant of isalnum(3), in other words, it doesn't * contain the '_'. --jhi */ PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_alnum, p); + return swash_fetch(PL_utf8_alnum, p, TRUE); /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */ #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ if (!PL_utf8_alnum) PL_utf8_alnum = swash_init("utf8", "", sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0); - return swash_fetch(PL_utf8_alnum, p); + return swash_fetch(PL_utf8_alnum, p, TRUE); #endif } @@ -879,13 +1053,13 @@ Perl_is_utf8_alnumc(pTHX_ U8 *p) return FALSE; if (!PL_utf8_alnum) PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_alnum, p); + return swash_fetch(PL_utf8_alnum, p, TRUE); /* return is_utf8_alpha(p) || is_utf8_digit(p); */ #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ if (!PL_utf8_alnum) PL_utf8_alnum = swash_init("utf8", "", sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0); - return swash_fetch(PL_utf8_alnum, p); + return swash_fetch(PL_utf8_alnum, p, TRUE); #endif } @@ -902,7 +1076,7 @@ Perl_is_utf8_alpha(pTHX_ U8 *p) return FALSE; if (!PL_utf8_alpha) PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_alpha, p); + return swash_fetch(PL_utf8_alpha, p, TRUE); } bool @@ -912,7 +1086,7 @@ Perl_is_utf8_ascii(pTHX_ U8 *p) return FALSE; if (!PL_utf8_ascii) PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_ascii, p); + return swash_fetch(PL_utf8_ascii, p, TRUE); } bool @@ -922,7 +1096,7 @@ Perl_is_utf8_space(pTHX_ U8 *p) return FALSE; if (!PL_utf8_space) PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_space, p); + return swash_fetch(PL_utf8_space, p, TRUE); } bool @@ -932,7 +1106,7 @@ Perl_is_utf8_digit(pTHX_ U8 *p) return FALSE; if (!PL_utf8_digit) PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_digit, p); + return swash_fetch(PL_utf8_digit, p, TRUE); } bool @@ -942,7 +1116,7 @@ Perl_is_utf8_upper(pTHX_ U8 *p) return FALSE; if (!PL_utf8_upper) PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_upper, p); + return swash_fetch(PL_utf8_upper, p, TRUE); } bool @@ -952,7 +1126,7 @@ Perl_is_utf8_lower(pTHX_ U8 *p) return FALSE; if (!PL_utf8_lower) PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_lower, p); + return swash_fetch(PL_utf8_lower, p, TRUE); } bool @@ -962,7 +1136,7 @@ Perl_is_utf8_cntrl(pTHX_ U8 *p) return FALSE; if (!PL_utf8_cntrl) PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_cntrl, p); + return swash_fetch(PL_utf8_cntrl, p, TRUE); } bool @@ -972,7 +1146,7 @@ Perl_is_utf8_graph(pTHX_ U8 *p) return FALSE; if (!PL_utf8_graph) PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_graph, p); + return swash_fetch(PL_utf8_graph, p, TRUE); } bool @@ -982,7 +1156,7 @@ Perl_is_utf8_print(pTHX_ U8 *p) return FALSE; if (!PL_utf8_print) PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_print, p); + return swash_fetch(PL_utf8_print, p, TRUE); } bool @@ -992,7 +1166,7 @@ Perl_is_utf8_punct(pTHX_ U8 *p) return FALSE; if (!PL_utf8_punct) PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_punct, p); + return swash_fetch(PL_utf8_punct, p, TRUE); } bool @@ -1002,7 +1176,7 @@ Perl_is_utf8_xdigit(pTHX_ U8 *p) return FALSE; if (!PL_utf8_xdigit) PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_xdigit, p); + return swash_fetch(PL_utf8_xdigit, p, TRUE); } bool @@ -1012,40 +1186,98 @@ Perl_is_utf8_mark(pTHX_ U8 *p) return FALSE; if (!PL_utf8_mark) PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0); - return swash_fetch(PL_utf8_mark, p); + return swash_fetch(PL_utf8_mark, p, TRUE); } +/* +=for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special + +The "p" contains the pointer to the UTF-8 string encoding +the character that is being converted. + +The "ustrp" is a pointer to the character buffer to put the +conversion result to. The "lenp" is a pointer to the length +of the result. + +The "swash" is a pointer to the swash to use. + +The "normal" is a string like "ToLower" which means the swash +$utf8::ToLower, which is stored in lib/unicore/To/Lower.pl, +and loaded by SWASHGET, using lib/utf8_heavy.pl. + +The "special" is a string like "utf8::ToSpecLower", which means +the hash %utf8::ToSpecLower, which is stored in the same file, +lib/unicore/To/Lower.pl, and also loaded by SWASHGET. The access +to the hash is by Perl_to_utf8_case(). + +=cut + */ + UV -Perl_to_utf8_upper(pTHX_ U8 *p) +Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp,char *normal, char *special) { UV uv; - if (!PL_utf8_toupper) - PL_utf8_toupper = swash_init("utf8", "ToUpper", &PL_sv_undef, 4, 0); - uv = swash_fetch(PL_utf8_toupper, p); - return uv ? uv : utf8_to_uv(p,UTF8_MAXLEN,0,0); + if (!*swashp) + *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); + uv = swash_fetch(*swashp, p, TRUE); + if (uv) + uv = UNI_TO_NATIVE(uv); + else { + HV *hv; + SV *keysv; + HE *he; + + uv = utf8_to_uvchr(p, 0); + + if ((hv = get_hv(special, FALSE)) && + (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv))) && + (he = hv_fetch_ent(hv, keysv, FALSE, 0))) { + SV *val = HeVAL(he); + char *s = SvPV(val, *lenp); + U8 c = *(U8*)s; + if (*lenp > 1 || UNI_IS_INVARIANT(c)) + Copy(s, ustrp, *lenp, U8); + else { + /* something in the 0x80..0xFF range */ + ustrp[0] = UTF8_EIGHT_BIT_HI(c); + ustrp[1] = UTF8_EIGHT_BIT_LO(c); + *lenp = 2; + } + return 0; + } + } + *lenp = UNISKIP(uv); + uvuni_to_utf8(ustrp, uv); + return uv; } UV -Perl_to_utf8_title(pTHX_ U8 *p) +Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) { - UV uv; + return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, + &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); +} - if (!PL_utf8_totitle) - PL_utf8_totitle = swash_init("utf8", "ToTitle", &PL_sv_undef, 4, 0); - uv = swash_fetch(PL_utf8_totitle, p); - return uv ? uv : utf8_to_uv(p,UTF8_MAXLEN,0,0); +UV +Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) +{ + return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, + &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); } UV -Perl_to_utf8_lower(pTHX_ U8 *p) +Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) { - UV uv; + return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, + &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); +} - if (!PL_utf8_tolower) - PL_utf8_tolower = swash_init("utf8", "ToLower", &PL_sv_undef, 4, 0); - uv = swash_fetch(PL_utf8_tolower, p); - return uv ? uv : utf8_to_uv(p,UTF8_MAXLEN,0,0); +UV +Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) +{ + return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, + &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); } /* a "swash" is a swatch hash */ @@ -1054,12 +1286,18 @@ SV* Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) { SV* retval; - char tmpbuf[256]; + SV* tokenbufsv = sv_2mortal(NEWSV(0,0)); dSP; + HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE); + SV* errsv_save; - if (!gv_stashpv(pkg, 0)) { /* demand load utf8 */ + if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ ENTER; + errsv_save = newSVsv(ERRSV); Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv); + if (!SvTRUE(ERRSV)) + sv_setsv(ERRSV, errsv_save); + SvREFCNT_dec(errsv_save); LEAVE; } SPAGAIN; @@ -1076,16 +1314,24 @@ Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) SAVEI32(PL_hints); PL_hints = 0; save_re_context(); - if (PL_curcop == &PL_compiling) /* XXX ought to be handled by lex_start */ - strncpy(tmpbuf, PL_tokenbuf, sizeof tmpbuf); + if (PL_curcop == &PL_compiling) + /* XXX ought to be handled by lex_start */ + sv_setpv(tokenbufsv, PL_tokenbuf); + errsv_save = newSVsv(ERRSV); if (call_method("SWASHNEW", G_SCALAR)) retval = newSVsv(*PL_stack_sp--); else retval = &PL_sv_undef; + if (!SvTRUE(ERRSV)) + sv_setsv(ERRSV, errsv_save); + SvREFCNT_dec(errsv_save); LEAVE; POPSTACK; if (PL_curcop == &PL_compiling) { - strncpy(PL_tokenbuf, tmpbuf, sizeof tmpbuf); + STRLEN len; + char* pv = SvPV(tokenbufsv, len); + + Copy(pv, PL_tokenbuf, len+1, char); PL_curcop->op_private = PL_hints; } if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) @@ -1093,17 +1339,53 @@ Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) return retval; } + +/* This API is wrong for special case conversions since we may need to + * return several Unicode characters for a single Unicode character + * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is + * the lower-level routine, and it is similarly broken for returning + * multiple values. --jhi */ UV -Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) +Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8) { HV* hv = (HV*)SvRV(sv); - U32 klen = UTF8SKIP(ptr) - 1; - U32 off = ptr[klen] & 127; /* NB: 64 bit always 0 when len > 1 */ + U32 klen; + U32 off; STRLEN slen; - STRLEN needents = (klen ? 64 : 128); - U8 *tmps; + STRLEN needents; + U8 *tmps = NULL; U32 bit; SV *retval; + U8 tmputf8[2]; + UV c = NATIVE_TO_ASCII(*ptr); + + if (!do_utf8 && !UNI_IS_INVARIANT(c)) { + tmputf8[0] = UTF8_EIGHT_BIT_HI(c); + tmputf8[1] = UTF8_EIGHT_BIT_LO(c); + ptr = tmputf8; + } + /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ + * then the "swatch" is a vec() for al the chars which start + * with 0xAA..0xYY + * So the key in the hash (klen) is length of encoded char -1 + */ + klen = UTF8SKIP(ptr) - 1; + off = ptr[klen]; + + if (klen == 0) + { + /* If char in invariant then swatch is for all the invariant chars + * In both UTF-8 and UTF8-MOD that happens to be UTF_CONTINUATION_MARK + */ + needents = UTF_CONTINUATION_MARK; + off = NATIVE_TO_UTF(ptr[klen]); + } + else + { + /* If char is encoded then swatch is for the prefix */ + needents = (1 << UTF_ACCUMULATION_SHIFT); + off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; + } /* * This single-entry cache saves about 1/3 of the utf8 overhead in test @@ -1114,9 +1396,9 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) * NB: this code assumes that swatches are never modified, once generated! */ - if (hv == PL_last_swash_hv && + if (hv == PL_last_swash_hv && klen == PL_last_swash_klen && - (!klen || memEQ((char *)ptr,(char *)PL_last_swash_key,klen)) ) + (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) { tmps = PL_last_swash_tmps; slen = PL_last_swash_slen; @@ -1128,6 +1410,11 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) /* If not cached, generate it via utf8::SWASHGET */ if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) { dSP; + /* We use utf8n_to_uvuni() as we want an index into + Unicode tables, not a native character number. + */ + UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, NULL, 0); + SV *errsv_save; ENTER; SAVETMPS; save_re_context(); @@ -1135,13 +1422,19 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) PUSHMARK(SP); EXTEND(SP,3); PUSHs((SV*)sv); - PUSHs(sv_2mortal(newSViv(utf8_to_uv(ptr, UTF8_MAXLEN, 0, 0) & ~(needents - 1)))); + /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ + PUSHs(sv_2mortal(newSViv((klen) ? + (code_point & ~(needents - 1)) : 0))); PUSHs(sv_2mortal(newSViv(needents))); PUTBACK; + errsv_save = newSVsv(ERRSV); if (call_method("SWASHGET", G_SCALAR)) retval = newSVsv(*PL_stack_sp--); else retval = &PL_sv_undef; + if (!SvTRUE(ERRSV)) + sv_setsv(ERRSV, errsv_save); + SvREFCNT_dec(errsv_save); POPSTACK; FREETMPS; LEAVE; @@ -1150,7 +1443,7 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) svp = hv_store(hv, (char*)ptr, klen, retval, 0); - if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || slen < 8) + if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents) Perl_croak(aTHX_ "SWASHGET didn't return result of proper length"); } @@ -1179,3 +1472,166 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) Perl_croak(aTHX_ "panic: swash_fetch"); return 0; } + + +/* +=for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv + +Adds the UTF8 representation of the Native codepoint C to the end +of the string C; C should be have at least C free +bytes available. The return value is the pointer to the byte after the +end of the new character. In other words, + + d = uvchr_to_utf8(d, uv); + +is the recommended wide native character-aware way of saying + + *(d++) = uv; + +=cut +*/ + +/* On ASCII machines this is normally a macro but we want a + real function in case XS code wants it +*/ +#undef Perl_uvchr_to_utf8 +U8 * +Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) +{ + return Perl_uvuni_to_utf8(aTHX_ d, NATIVE_TO_UNI(uv)); +} + + +/* +=for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags + +Returns the native character value of the first character in the string C +which is assumed to be in UTF8 encoding; C will be set to the +length, in bytes, of that character. + +Allows length and flags to be passed to low level routine. + +=cut +*/ +/* On ASCII machines this is normally a macro but we want a + real function in case XS code wants it +*/ +#undef Perl_utf8n_to_uvchr +UV +Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) +{ + UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); + return UNI_TO_NATIVE(uv); +} + +/* +=for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags + +Build to the scalar dsv a displayable version of the string spv, +length len, the displayable version being at most pvlim bytes long +(if longer, the rest is truncated and "..." will be appended). +The flags argument is currently unused but available for future extensions. +The pointer to the PV of the dsv is returned. + +=cut */ +char * +Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags) +{ + int truncated = 0; + char *s, *e; + + sv_setpvn(dsv, "", 0); + for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { + UV u; + if (pvlim && SvCUR(dsv) >= pvlim) { + truncated++; + break; + } + u = utf8_to_uvchr((U8*)s, 0); + Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); + } + if (truncated) + sv_catpvn(dsv, "...", 3); + + return SvPVX(dsv); +} + +/* +=for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags + +Build to the scalar dsv a displayable version of the scalar sv, +he displayable version being at most pvlim bytes long +(if longer, the rest is truncated and "..." will be appended). +The flags argument is currently unused but available for future extensions. +The pointer to the PV of the dsv is returned. + +=cut */ +char * +Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) +{ + return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv), + pvlim, flags); +} + +/* +=for apidoc A|I32|ibcmp_utf8|const char *s1|bool u1|const char *s2|bool u2|register I32 len + +Return true if the strings s1 and s2 differ case-insensitively, false +if not (if they are equal case-insensitively). If u1 is true, the +string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true, +the string s2 is assumed to be in UTF-8-encoded Unicode. (If both u1 +and u2 are false, ibcmp() is called.) + +For case-insensitiveness, the "casefolding" of Unicode is used +instead of upper/lowercasing both the characters, see +http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). + +=cut */ +I32 +Perl_ibcmp_utf8(pTHX_ const char *s1, bool u1, const char *s2, bool u2, register I32 len) +{ + if (u1 || u2) { + register U8 *a = (U8*)s1; + register U8 *b = (U8*)s2; + STRLEN la, lb; + UV ca, cb; + STRLEN ulen1, ulen2; + U8 tmpbuf1[UTF8_MAXLEN*3+1]; + U8 tmpbuf2[UTF8_MAXLEN*3+1]; + + while (len) { + if (u1) + ca = utf8_to_uvchr((U8*)a, &la); + else { + ca = *a; + la = 1; + } + if (u2) + cb = utf8_to_uvchr((U8*)b, &lb); + else { + cb = *b; + lb = 1; + } + if (ca != cb) { + if (u1) + to_uni_fold(NATIVE_TO_UNI(ca), tmpbuf1, &ulen1); + else + ulen1 = 1; + if (u2) + to_uni_fold(NATIVE_TO_UNI(cb), tmpbuf2, &ulen2); + else + ulen2 = 1; + if (ulen1 != ulen2 + || (ulen1 == 1 && PL_fold[ca] != PL_fold[cb]) + || memNE((char *)tmpbuf1, (char *)tmpbuf2, ulen1)) + return 1; + } + a += la; + b += lb; + } + return 0; + } + else + return ibcmp(s1, s2, len); +} +