/* utf8.c * * 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. * */ /* * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever * heard of that we don't want to see any closer; and that's the one place * we're trying to get to! And that's just where we can't get, nohow.' * * 'Well do I understand your speech,' he answered in the same language; * 'yet few strangers do so. Why then do you not speak in the Common Tongue, * as is the custom in the West, if you wish to be answered?' * * ...the travellers perceived that the floor was paved with stones of many * hues; branching runes and strange devices intertwined beneath their feet. */ #include "EXTERN.h" #define PERL_IN_UTF8_C #include "perl.h" /* 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_uvuni_to_utf8(pTHX_ U8 *d, UV uv) { if (UNI_IS_INVARIANT(uv)) { *d++ = UTF_TO_NATIVE(uv); return d; } #if defined(EBCDIC) || 1 /* always for testing */ 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); return d; } if (uv < 0x10000) { *d++ = (( uv >> 12) | 0xe0); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); return d; } if (uv < 0x200000) { *d++ = (( uv >> 18) | 0xf0); *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); return d; } if (uv < 0x4000000) { *d++ = (( uv >> 24) | 0xf8); *d++ = (((uv >> 18) & 0x3f) | 0x80); *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); return d; } if (uv < 0x80000000) { *d++ = (( uv >> 30) | 0xfc); *d++ = (((uv >> 24) & 0x3f) | 0x80); *d++ = (((uv >> 18) & 0x3f) | 0x80); *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); return d; } #ifdef HAS_QUAD if (uv < UTF8_QUAD_MAX) #endif { *d++ = 0xfe; /* Can't match U+FEFF! */ *d++ = (((uv >> 30) & 0x3f) | 0x80); *d++ = (((uv >> 24) & 0x3f) | 0x80); *d++ = (((uv >> 18) & 0x3f) | 0x80); *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); return d; } #ifdef HAS_QUAD { *d++ = 0xff; /* Can't match U+FFFE! */ *d++ = 0x80; /* 6 Reserved bits */ *d++ = (((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ *d++ = (((uv >> 54) & 0x3f) | 0x80); *d++ = (((uv >> 48) & 0x3f) | 0x80); *d++ = (((uv >> 42) & 0x3f) | 0x80); *d++ = (((uv >> 36) & 0x3f) | 0x80); *d++ = (((uv >> 30) & 0x3f) | 0x80); *d++ = (((uv >> 24) & 0x3f) | 0x80); *d++ = (((uv >> 18) & 0x3f) | 0x80); *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); return d; } #endif #endif /* Loop style */ } /* =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) { U8 u = *s; STRLEN slen, len; UV uv, ouv; if (UTF8_IS_INVARIANT(u)) return 1; if (!UTF8_IS_START(u)) return 0; len = UTF8SKIP(s); if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) return 0; slen = len - 1; s++; u &= UTF_START_MASK(len); uv = u; ouv = uv; while (slen--) { if (!UTF8_IS_CONTINUATION(*s)) return 0; uv = UTF8_ACCUMULATE(uv, *s); if (uv < ouv) return 0; ouv = uv; s++; } if (UNISKIP(uv) < len) return 0; return len; } /* =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len 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 */ bool Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len) { U8* x = s; 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; return TRUE; } /* =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags 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. 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 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. 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_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) { UV uv = *s, ouv = 0; STRLEN len = 1; bool dowarn = ckWARN_d(WARN_UTF8); STRLEN expectlen = 0; 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 (UTF8_IS_INVARIANT(uv)) { if (retlen) *retlen = 1; return (UV) (NATIVE_TO_UTF(*s)); } if (UTF8_IS_CONTINUATION(uv) && !(flags & UTF8_ALLOW_CONTINUATION)) { warning = UTF8_WARN_CONTINUATION; goto malformed; } if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && !(flags & UTF8_ALLOW_NON_CONTINUATION)) { 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)) { 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! */ #endif if (retlen) *retlen = len; expectlen = len; if ((curlen < expectlen) && !(flags & UTF8_ALLOW_SHORT)) { warning = UTF8_WARN_SHORT; goto malformed; } len--; s++; ouv = uv; while (len--) { if (!UTF8_IS_CONTINUATION(*s) && !(flags & UTF8_ALLOW_NON_CONTINUATION)) { s--; warning = UTF8_WARN_NON_CONTINUATION; goto malformed; } else 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 (UNICODE_IS_SURROGATE(uv) && !(flags & UTF8_ALLOW_SURROGATE)) { warning = UTF8_WARN_SURROGATE; goto malformed; } else if (UNICODE_IS_BYTE_ORDER_MARK(uv) && !(flags & UTF8_ALLOW_BOM)) { warning = UTF8_WARN_BOM; goto malformed; } else if ((expectlen > UNISKIP(uv)) && !(flags & UTF8_ALLOW_LONG)) { warning = UTF8_WARN_LONG; goto malformed; } else if (UNICODE_IS_ILLEGAL(uv) && !(flags & UTF8_ALLOW_FFFF)) { warning = UTF8_WARN_FFFF; goto malformed; } return uv; malformed: if (flags & UTF8_CHECK_ONLY) { if (retlen) *retlen = -1; 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); 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 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 A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen 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. 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. =cut */ UV Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen) { /* Call the low level routine asking for checks */ return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen, 0); } /* =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 up past C, croaks. =cut */ STRLEN 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) { U8 t = UTF8SKIP(s); if (e - s < t) Perl_croak(aTHX_ "panic: utf8_length: unaligned end"); s += t; len++; } return len; } /* =for apidoc A|IV|utf8_distance|U8 *a|U8 *b Returns the number of UTF8 characters between the UTF-8 pointers C and C. WARNING: use only if you *know* that the pointers point inside the same UTF-8 buffer. =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); if (b - a < c) Perl_croak(aTHX_ "panic: utf8_distance: unaligned end"); a += c; off--; } } else { while (b < a) { U8 c = UTF8SKIP(b); if (a - b < c) Perl_croak(aTHX_ "panic: utf8_distance: unaligned end"); b += c; off++; } } return off; } /* =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off 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 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 */ 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); } else { while (off++) { s--; while (UTF8_IS_CONTINUATION(*s)) s--; } } return s; } /* =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 updates len to contain the new length. Returns zero on failure, setting C to -1. =cut */ U8 * Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) { U8 *send; U8 *d; U8 *save = s; /* ensure valid UTF8 and chars < 256 before updating string */ for (send = s + *len; s < send; ) { U8 c = *s++; if (!UTF8_IS_INVARIANT(c) && (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { *len = -1; return 0; } } d = s = save; while (s < send) { STRLEN ulen; *d++ = (U8)utf8_to_uvchr(s, &ulen); s += ulen; } *d = '\0'; *len = d - save; return save; } /* =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 reflect the new length. =cut */ U8* Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len) { U8 *send; U8 *d; U8 *dst; send = s + (*len); Newz(801, d, (*len) * 2 + 1, U8); dst = d; while (s < send) { UV uv = NATIVE_TO_ASCII(*s++); if (UNI_IS_INVARIANT(uv)) *d++ = UTF_TO_NATIVE(uv); else { *d++ = UTF8_EIGHT_BIT_HI(uv); *d++ = UTF8_EIGHT_BIT_LO(uv); } } *d = '\0'; *len = d-dst; return dst; } /* * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. * * Destination must be pre-extended to 3/2 source. Do not use in-place. * We optimize for native, for obvious reasons. */ U8* Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) { U8* pend; U8* dstart = d; if (bytelen & 1) Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen"); pend = p + bytelen; while (p < pend) { UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ p += 2; if (uv < 0x80) { *d++ = uv; continue; } if (uv < 0x800) { *d++ = (( uv >> 6) | 0xc0); *d++ = (( uv & 0x3f) | 0x80); continue; } if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */ UV low = *p++; if (low < 0xdc00 || low >= 0xdfff) Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; } if (uv < 0x10000) { *d++ = (( uv >> 12) | 0xe0); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); continue; } else { *d++ = (( uv >> 18) | 0xf0); *d++ = (((uv >> 12) & 0x3f) | 0x80); *d++ = (((uv >> 6) & 0x3f) | 0x80); *d++ = (( uv & 0x3f) | 0x80); continue; } } *newlen = d - dstart; return d; } /* Note: this one is slightly destructive of the source. */ U8* Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) { U8* s = (U8*)p; U8* send = s + bytelen; while (s < send) { U8 tmp = s[0]; s[0] = s[1]; s[1] = tmp; s += 2; } return utf16_to_utf8(p, d, bytelen, newlen); } /* for now these are all defined (inefficiently) in terms of the utf8 versions */ bool Perl_is_uni_alnum(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alnum(tmpbuf); } bool Perl_is_uni_alnumc(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alnumc(tmpbuf); } bool Perl_is_uni_idfirst(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_idfirst(tmpbuf); } bool Perl_is_uni_alpha(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alpha(tmpbuf); } bool Perl_is_uni_ascii(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_ascii(tmpbuf); } bool Perl_is_uni_space(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_space(tmpbuf); } bool Perl_is_uni_digit(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_digit(tmpbuf); } bool Perl_is_uni_upper(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_upper(tmpbuf); } bool Perl_is_uni_lower(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_lower(tmpbuf); } bool Perl_is_uni_cntrl(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_cntrl(tmpbuf); } bool Perl_is_uni_graph(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_graph(tmpbuf); } bool Perl_is_uni_print(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_print(tmpbuf); } bool Perl_is_uni_punct(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_punct(tmpbuf); } bool Perl_is_uni_xdigit(pTHX_ U32 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, U8* p, STRLEN *lenp) { U8 tmpbuf[UTF8_MAXLEN*2+1]; uvchr_to_utf8(tmpbuf, (UV)c); return to_utf8_upper(tmpbuf, p, lenp); } U32 Perl_to_uni_title(pTHX_ U32 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_lower(pTHX_ U32 c, U8* p, STRLEN *lenp) { U8 tmpbuf[UTF8_MAXLEN+1]; uvchr_to_utf8(tmpbuf, (UV)c); return to_utf8_lower(tmpbuf, p, lenp); } /* for now these all assume no locale info available for Unicode > 255 */ bool Perl_is_uni_alnum_lc(pTHX_ U32 c) { return is_uni_alnum(c); /* XXX no locale support yet */ } bool Perl_is_uni_alnumc_lc(pTHX_ U32 c) { return is_uni_alnumc(c); /* XXX no locale support yet */ } bool Perl_is_uni_idfirst_lc(pTHX_ U32 c) { return is_uni_idfirst(c); /* XXX no locale support yet */ } bool Perl_is_uni_alpha_lc(pTHX_ U32 c) { return is_uni_alpha(c); /* XXX no locale support yet */ } bool Perl_is_uni_ascii_lc(pTHX_ U32 c) { return is_uni_ascii(c); /* XXX no locale support yet */ } bool Perl_is_uni_space_lc(pTHX_ U32 c) { return is_uni_space(c); /* XXX no locale support yet */ } bool Perl_is_uni_digit_lc(pTHX_ U32 c) { return is_uni_digit(c); /* XXX no locale support yet */ } bool Perl_is_uni_upper_lc(pTHX_ U32 c) { return is_uni_upper(c); /* XXX no locale support yet */ } bool Perl_is_uni_lower_lc(pTHX_ U32 c) { return is_uni_lower(c); /* XXX no locale support yet */ } bool Perl_is_uni_cntrl_lc(pTHX_ U32 c) { return is_uni_cntrl(c); /* XXX no locale support yet */ } bool Perl_is_uni_graph_lc(pTHX_ U32 c) { return is_uni_graph(c); /* XXX no locale support yet */ } bool Perl_is_uni_print_lc(pTHX_ U32 c) { return is_uni_print(c); /* XXX no locale support yet */ } bool Perl_is_uni_punct_lc(pTHX_ U32 c) { return is_uni_punct(c); /* XXX no locale support yet */ } bool Perl_is_uni_xdigit_lc(pTHX_ U32 c) { return is_uni_xdigit(c); /* XXX no locale support yet */ } bool Perl_is_utf8_alnum(pTHX_ U8 *p) { if (!is_utf8_char(p)) return FALSE; if (!PL_utf8_alnum) /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true * 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, 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, TRUE); #endif } bool Perl_is_utf8_alnumc(pTHX_ U8 *p) { if (!is_utf8_char(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, 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, TRUE); #endif } bool Perl_is_utf8_idfirst(pTHX_ U8 *p) { return *p == '_' || is_utf8_alpha(p); } bool Perl_is_utf8_alpha(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_ascii(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_space(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_digit(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_upper(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_lower(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_cntrl(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_graph(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_print(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_punct(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_xdigit(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } bool Perl_is_utf8_mark(pTHX_ U8 *p) { if (!is_utf8_char(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, TRUE); } UV Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp,char *normal, char *special) { UV uv; 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_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) { return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); } 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, U8* ustrp, STRLEN *lenp) { return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); } /* a "swash" is a swatch hash */ SV* Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) { SV* retval; SV* tokenbufsv = sv_2mortal(NEWSV(0,0)); dSP; HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE); SV* errsv_save; 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; PUSHSTACKi(PERLSI_MAGIC); PUSHMARK(SP); EXTEND(SP,5); PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg)))); PUSHs(sv_2mortal(newSVpvn(name, strlen(name)))); PUSHs(listsv); PUSHs(sv_2mortal(newSViv(minbits))); PUSHs(sv_2mortal(newSViv(none))); PUTBACK; ENTER; SAVEI32(PL_hints); PL_hints = 0; save_re_context(); 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) { 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) Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); 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, bool do_utf8) { HV* hv = (HV*)SvRV(sv); U32 klen; U32 off; STRLEN slen; STRLEN needents; U8 *tmps; 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 * suite. (That is, only 7-8% overall over just a hash cache. Still, * it's nothing to sniff at.) Pity we usually come through at least * two function calls to get here... * * NB: this code assumes that swatches are never modified, once generated! */ if (hv == PL_last_swash_hv && klen == PL_last_swash_klen && (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) { tmps = PL_last_swash_tmps; slen = PL_last_swash_slen; } else { /* Try our second-level swatch cache, kept in a hash. */ SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE); /* 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(); PUSHSTACKi(PERLSI_MAGIC); PUSHMARK(SP); EXTEND(SP,3); PUSHs((SV*)sv); /* 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; if (PL_curcop == &PL_compiling) PL_curcop->op_private = PL_hints; svp = hv_store(hv, (char*)ptr, klen, retval, 0); if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents) Perl_croak(aTHX_ "SWASHGET didn't return result of proper length"); } PL_last_swash_hv = hv; PL_last_swash_klen = klen; PL_last_swash_tmps = tmps; PL_last_swash_slen = slen; if (klen) Copy(ptr, PL_last_swash_key, klen, U8); } switch ((int)((slen << 3) / needents)) { case 1: bit = 1 << (off & 7); off >>= 3; return (tmps[off] & bit) != 0; case 8: return tmps[off]; case 16: off <<= 1; return (tmps[off] << 8) + tmps[off + 1] ; case 32: off <<= 2; return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; } 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); }