X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=utf8.c;h=e86c49fdcf63ab3aac316e0f63656274f4ddbf6b;hb=7bd161a15a083a0aed82ba8e1fa3d6df6f975c68;hp=66135bb6e63ccdc06ef06cbc178be0604fa709b3;hpb=dfe13c55d349c8cc782995becdedd62551082672;p=p5sagit%2Fp5-mst-13.2.git diff --git a/utf8.c b/utf8.c index 66135bb..e86c49f 100644 --- a/utf8.c +++ b/utf8.c @@ -1,6 +1,6 @@ /* utf8.c * - * Copyright (c) 1998, Larry Wall + * Copyright (c) 1998-2000, 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. @@ -21,12 +21,13 @@ */ #include "EXTERN.h" +#define PERL_IN_UTF8_C #include "perl.h" /* Unicode support */ U8 * -uv_to_utf8(U8 *d, UV uv) +Perl_uv_to_utf8(pTHX_ U8 *d, UV uv) { if (uv < 0x80) { *d++ = uv; @@ -67,8 +68,8 @@ uv_to_utf8(U8 *d, UV uv) *d++ = (( uv & 0x3f) | 0x80); return d; } -#ifdef Quad_t - if (uv < 0x2000000000) +#ifdef HAS_QUAD + if (uv < 0x1000000000LL) #endif { *d++ = 0xfe; /* Can't match U+FEFF! */ @@ -80,9 +81,14 @@ uv_to_utf8(U8 *d, UV uv) *d++ = (( uv & 0x3f) | 0x80); return d; } -#ifdef Quad_t +#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); @@ -95,8 +101,65 @@ uv_to_utf8(U8 *d, UV uv) #endif } +/* 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. */ +int +Perl_is_utf8_char(pTHX_ U8 *s) +{ + U8 u = *s; + int slen, len; + + if (!(u & 0x80)) + return 1; + + if (!(u & 0x40)) + return 0; + + if (!(u & 0x20)) { len = 2; } + else if (!(u & 0x10)) { len = 3; } + else if (!(u & 0x08)) { len = 4; } + else if (!(u & 0x04)) { len = 5; } + else if (!(u & 0x02)) { len = 6; } + else if (!(u & 0x01)) { len = 7; } + else { len = 13; } /* whoa! */ + + slen = len - 1; + s++; + while (slen--) { + if ((*s & 0xc0) != 0x80) + return 0; + s++; + } + return len; +} + +/* +=for apidoc Am|bool_utf8_string|U8 *s|STRLEN len + +Returns true if first C bytes of the given string form valid a UTF8 +string, false otherwise. + +=cut +*/ + +bool +Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len) +{ + U8* x=s; + U8* send=s+len; + int c; + while (x < send) { + c = is_utf8_char(x); + x += c; + if (!c || x > send) + return 0; + } + return 1; +} + UV -utf8_to_uv(U8* s, I32* retlen) +Perl_utf8_to_uv(pTHX_ U8* s, I32* retlen) { UV uv = *s; int len; @@ -106,7 +169,9 @@ utf8_to_uv(U8* s, I32* retlen) return *s; } if (!(uv & 0x40)) { - warn("Malformed UTF-8 character"); + dTHR; + if (ckWARN_d(WARN_UTF8)) + Perl_warner(aTHX_ WARN_UTF8, "Malformed UTF-8 character"); if (retlen) *retlen = 1; return *s; @@ -117,8 +182,8 @@ utf8_to_uv(U8* s, I32* retlen) else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } - else if (!(uv & 0x01)) { len = 7; uv &= 0x00; } - else len = 8; /* whoa! */ + else if (!(uv & 0x01)) { len = 7; uv = 0; } + else { len = 13; uv = 0; } /* whoa! */ if (retlen) *retlen = len; @@ -126,7 +191,9 @@ utf8_to_uv(U8* s, I32* retlen) s++; while (len--) { if ((*s & 0xc0) != 0x80) { - warn("Malformed UTF-8 character"); + dTHR; + if (ckWARN_d(WARN_UTF8)) + Perl_warner(aTHX_ WARN_UTF8, "Malformed UTF-8 character"); if (retlen) *retlen -= len + 1; return 0xfffd; @@ -140,7 +207,7 @@ utf8_to_uv(U8* s, I32* retlen) /* utf8_distance(a,b) is intended to be a - b in pointer arithmetic */ I32 -utf8_distance(U8 *a, U8 *b) +Perl_utf8_distance(pTHX_ U8 *a, U8 *b) { I32 off = 0; if (a < b) { @@ -161,7 +228,7 @@ utf8_distance(U8 *a, U8 *b) /* WARNING: do not use the following unless you *know* off is within bounds */ U8 * -utf8_hop(U8 *s, I32 off) +Perl_utf8_hop(pTHX_ U8 *s, I32 off) { if (off >= 0) { while (off--) @@ -179,19 +246,100 @@ utf8_hop(U8 *s, I32 off) return s; } -/* XXX NOTHING CALLS THE FOLLOWING TWO ROUTINES YET!!! */ /* - * Convert native or reversed UTF-16 to UTF-8. +=for apidoc Am|U8 *|utf8_to_bytes|U8 *s|STRLEN len + +Converts a string C of length C from UTF8 into ASCII encoding. +Unlike C, this over-writes the original string. +Returns zero on failure after converting as much as possible. + +=cut +*/ + +U8 * +Perl_utf8_to_bytes(pTHX_ U8* s, STRLEN len) +{ + dTHR; + U8 *send; + U8 *d; + U8 *save; + + send = s + len; + d = save = s; + while (s < send) { + if (*s < 0x80) + *d++ = *s++; + else { + I32 ulen; + UV uv = utf8_to_uv(s, &ulen); + if (uv > 255) { + *d = '\0'; + return 0; + } + s += ulen; + *d++ = (U8)uv; + } + } + *d = '\0'; + return save; +} + +/* +=for apidoc Am|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) +{ + dTHR; + U8 *send; + U8 *d; + U8 *dst; + send = s + (*len); + + Newz(801, d, (*len) * 2 + 1, U8); + dst = d; + + while (s < send) { + if (*s < 0x80) + *d++ = *s++; + else { + UV uv = *s++; + *d++ = (( uv >> 6) | 0xc0); + *d++ = (( uv & 0x3f) | 0x80); + } + } + *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* -utf16_to_utf8(U16* p, U8* d, I32 bytelen) +Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) { - U16* pend = p + bytelen / 2; + 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++; + UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ + p += 2; if (uv < 0x80) { *d++ = uv; continue; @@ -202,12 +350,10 @@ utf16_to_utf8(U16* p, U8* d, I32 bytelen) continue; } if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */ - int low = *p++; - if (low < 0xdc00 || low >= 0xdfff) { - warn("Malformed UTF-16 surrogate"); - p--; - uv = 0xfffd; - } + dTHR; + 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) { @@ -224,13 +370,14 @@ utf16_to_utf8(U16* p, U8* d, I32 bytelen) continue; } } + *newlen = d - dstart; return d; } /* Note: this one is slightly destructive of the source. */ U8* -utf16_to_utf8_reversed(U16* p, U8* d, I32 bytelen) +Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) { U8* s = (U8*)p; U8* send = s + bytelen; @@ -240,95 +387,143 @@ utf16_to_utf8_reversed(U16* p, U8* d, I32 bytelen) s[1] = tmp; s += 2; } - return utf16_to_utf8(p, d, bytelen); + return utf16_to_utf8(p, d, bytelen, newlen); } /* for now these are all defined (inefficiently) in terms of the utf8 versions */ bool -is_uni_alnum(U32 c) +Perl_is_uni_alnum(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_alnum(tmpbuf); } bool -is_uni_idfirst(U32 c) +Perl_is_uni_alnumc(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; + uv_to_utf8(tmpbuf, (UV)c); + return is_utf8_alnumc(tmpbuf); +} + +bool +Perl_is_uni_idfirst(pTHX_ U32 c) +{ + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_idfirst(tmpbuf); } bool -is_uni_alpha(U32 c) +Perl_is_uni_alpha(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_alpha(tmpbuf); } bool -is_uni_space(U32 c) +Perl_is_uni_ascii(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; + uv_to_utf8(tmpbuf, (UV)c); + return is_utf8_ascii(tmpbuf); +} + +bool +Perl_is_uni_space(pTHX_ U32 c) +{ + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_space(tmpbuf); } bool -is_uni_digit(U32 c) +Perl_is_uni_digit(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_digit(tmpbuf); } bool -is_uni_upper(U32 c) +Perl_is_uni_upper(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_upper(tmpbuf); } bool -is_uni_lower(U32 c) +Perl_is_uni_lower(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_lower(tmpbuf); } bool -is_uni_print(U32 c) +Perl_is_uni_cntrl(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; + uv_to_utf8(tmpbuf, (UV)c); + return is_utf8_cntrl(tmpbuf); +} + +bool +Perl_is_uni_graph(pTHX_ U32 c) +{ + U8 tmpbuf[UTF8_MAXLEN]; + uv_to_utf8(tmpbuf, (UV)c); + return is_utf8_graph(tmpbuf); +} + +bool +Perl_is_uni_print(pTHX_ U32 c) +{ + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return is_utf8_print(tmpbuf); } +bool +Perl_is_uni_punct(pTHX_ U32 c) +{ + U8 tmpbuf[UTF8_MAXLEN]; + uv_to_utf8(tmpbuf, (UV)c); + return is_utf8_punct(tmpbuf); +} + +bool +Perl_is_uni_xdigit(pTHX_ U32 c) +{ + U8 tmpbuf[UTF8_MAXLEN]; + uv_to_utf8(tmpbuf, (UV)c); + return is_utf8_xdigit(tmpbuf); +} + U32 -to_uni_upper(U32 c) +Perl_to_uni_upper(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return to_utf8_upper(tmpbuf); } U32 -to_uni_title(U32 c) +Perl_to_uni_title(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return to_utf8_title(tmpbuf); } U32 -to_uni_lower(U32 c) +Perl_to_uni_lower(pTHX_ U32 c) { - U8 tmpbuf[10]; + U8 tmpbuf[UTF8_MAXLEN]; uv_to_utf8(tmpbuf, (UV)c); return to_utf8_lower(tmpbuf); } @@ -336,77 +531,114 @@ to_uni_lower(U32 c) /* for now these all assume no locale info available for Unicode > 255 */ bool -is_uni_alnum_lc(U32 c) +Perl_is_uni_alnum_lc(pTHX_ U32 c) { return is_uni_alnum(c); /* XXX no locale support yet */ } bool -is_uni_idfirst_lc(U32 c) +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 -is_uni_alpha_lc(U32 c) +Perl_is_uni_alpha_lc(pTHX_ U32 c) { return is_uni_alpha(c); /* XXX no locale support yet */ } bool -is_uni_space_lc(U32 c) +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 -is_uni_digit_lc(U32 c) +Perl_is_uni_digit_lc(pTHX_ U32 c) { return is_uni_digit(c); /* XXX no locale support yet */ } bool -is_uni_upper_lc(U32 c) +Perl_is_uni_upper_lc(pTHX_ U32 c) { return is_uni_upper(c); /* XXX no locale support yet */ } bool -is_uni_lower_lc(U32 c) +Perl_is_uni_lower_lc(pTHX_ U32 c) { return is_uni_lower(c); /* XXX no locale support yet */ } bool -is_uni_print_lc(U32 c) +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 */ +} + U32 -to_uni_upper_lc(U32 c) +Perl_to_uni_upper_lc(pTHX_ U32 c) { return to_uni_upper(c); /* XXX no locale support yet */ } U32 -to_uni_title_lc(U32 c) +Perl_to_uni_title_lc(pTHX_ U32 c) { return to_uni_title(c); /* XXX no locale support yet */ } U32 -to_uni_lower_lc(U32 c) +Perl_to_uni_lower_lc(pTHX_ U32 c) { return to_uni_lower(c); /* XXX no locale support yet */ } - bool -is_utf8_alnum(U8 *p) +Perl_is_utf8_alnum(pTHX_ U8 *p) { + if (!is_utf8_char(p)) + return FALSE; if (!PL_utf8_alnum) - PL_utf8_alnum = swash_init("utf8", "IsAlnum", &sv_undef, 0, 0); + PL_utf8_alnum = swash_init("utf8", "IsAlnum", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_alnum, p); /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */ #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ @@ -418,96 +650,177 @@ is_utf8_alnum(U8 *p) } bool -is_utf8_idfirst(U8 *p) +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); +/* 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); +#endif +} + +bool +Perl_is_utf8_idfirst(pTHX_ U8 *p) { return *p == '_' || is_utf8_alpha(p); } bool -is_utf8_alpha(U8 *p) +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", &sv_undef, 0, 0); + PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_alpha, p); } bool -is_utf8_space(U8 *p) +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); +} + +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", "IsSpace", &sv_undef, 0, 0); + PL_utf8_space = swash_init("utf8", "IsSpace", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_space, p); } bool -is_utf8_digit(U8 *p) +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", &sv_undef, 0, 0); + PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_digit, p); } bool -is_utf8_upper(U8 *p) +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", &sv_undef, 0, 0); + PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_upper, p); } bool -is_utf8_lower(U8 *p) +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", &sv_undef, 0, 0); + PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_lower, p); } bool -is_utf8_print(U8 *p) +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); +} + +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); +} + +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", &sv_undef, 0, 0); + PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_print, p); } bool -is_utf8_mark(U8 *p) +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); +} + +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); +} + +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", &sv_undef, 0, 0); + PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0); return swash_fetch(PL_utf8_mark, p); } -U32 -to_utf8_upper(U8 *p) +UV +Perl_to_utf8_upper(pTHX_ U8 *p) { UV uv; if (!PL_utf8_toupper) - PL_utf8_toupper = swash_init("utf8", "ToUpper", &sv_undef, 4, 0); + 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,0); } -U32 -to_utf8_title(U8 *p) +UV +Perl_to_utf8_title(pTHX_ U8 *p) { UV uv; if (!PL_utf8_totitle) - PL_utf8_totitle = swash_init("utf8", "ToTitle", &sv_undef, 4, 0); + 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,0); } -U32 -to_utf8_lower(U8 *p) +UV +Perl_to_utf8_lower(pTHX_ U8 *p) { UV uv; if (!PL_utf8_tolower) - PL_utf8_tolower = swash_init("utf8", "ToLower", &sv_undef, 4, 0); + 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,0); } @@ -515,11 +828,18 @@ to_utf8_lower(U8 *p) /* a "swash" is a swatch hash */ SV* -swash_init(char* pkg, char* name, SV *listsv, I32 minbits, I32 none) +Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) { SV* retval; char tmpbuf[256]; dSP; + + if (!gv_stashpv(pkg, 0)) { /* demand load utf8 */ + ENTER; + Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv); + LEAVE; + } + SPAGAIN; PUSHSTACKi(PERLSI_MAGIC); PUSHMARK(SP); EXTEND(SP,5); @@ -533,25 +853,25 @@ swash_init(char* pkg, char* name, SV *listsv, I32 minbits, I32 none) SAVEI32(PL_hints); PL_hints = 0; save_re_context(); - if (curcop == &compiling) /* XXX ought to be handled by lex_start */ + if (PL_curcop == &PL_compiling) /* XXX ought to be handled by lex_start */ strncpy(tmpbuf, PL_tokenbuf, sizeof tmpbuf); - if (perl_call_method("SWASHNEW", G_SCALAR)) - retval = newSVsv(*stack_sp--); + if (call_method("SWASHNEW", G_SCALAR)) + retval = newSVsv(*PL_stack_sp--); else - retval = &sv_undef; + retval = &PL_sv_undef; LEAVE; POPSTACK; - if (curcop == &compiling) { + if (PL_curcop == &PL_compiling) { strncpy(PL_tokenbuf, tmpbuf, sizeof tmpbuf); - curcop->op_private = PL_hints; + PL_curcop->op_private = PL_hints; } if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) - croak("SWASHNEW didn't return an HV ref"); + Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); return retval; } UV -swash_fetch(SV *sv, U8 *ptr) +Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) { HV* hv = (HV*)SvRV(sv); U32 klen = UTF8SKIP(ptr) - 1; @@ -573,7 +893,7 @@ swash_fetch(SV *sv, U8 *ptr) if (hv == PL_last_swash_hv && klen == PL_last_swash_klen && - (!klen || memEQ(ptr,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; @@ -595,20 +915,20 @@ swash_fetch(SV *sv, U8 *ptr) PUSHs(sv_2mortal(newSViv(utf8_to_uv(ptr, 0) & ~(needents - 1)))); PUSHs(sv_2mortal(newSViv(needents))); PUTBACK; - if (perl_call_method("SWASHGET", G_SCALAR)) - retval = newSVsv(*stack_sp--); + if (call_method("SWASHGET", G_SCALAR)) + retval = newSVsv(*PL_stack_sp--); else - retval = &sv_undef; + retval = &PL_sv_undef; POPSTACK; FREETMPS; LEAVE; - if (curcop == &compiling) - curcop->op_private = PL_hints; + 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 < 8) - croak("SWASHGET didn't return result of proper length"); + Perl_croak(aTHX_ "SWASHGET didn't return result of proper length"); } PL_last_swash_hv = hv; @@ -633,6 +953,6 @@ swash_fetch(SV *sv, U8 *ptr) off <<= 2; return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; } - croak("panic: swash_fetch"); + Perl_croak(aTHX_ "panic: swash_fetch"); return 0; }