X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=utf8.c;h=25cd0fdc4184135aa4e324e2ff8753a337202c27;hb=d57eded6852763e3753285b692dd66ed278e268d;hp=01afa010be29afecfebf929dc950ebe93f699b1a;hpb=c4d5f83add3e03ac76c328ed8a29701d939174ce;p=p5sagit%2Fp5-mst-13.2.git diff --git a/utf8.c b/utf8.c index 01afa01..25cd0fd 100644 --- a/utf8.c +++ b/utf8.c @@ -131,28 +131,6 @@ Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv) #endif /* Loop style */ } -/* -=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 -*/ - -U8 * -Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) -{ - return Perl_uvuni_to_utf8(aTHX_ d, NATIVE_TO_UNI(uv)); -} /* @@ -461,25 +439,6 @@ malformed: } /* -=for apidoc A|U8* s|utf8n_to_uvchr|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 -*/ - -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|U8* s|utf8_to_uvchr|STRLEN *retlen Returns the native character value of the first character in the string C @@ -835,7 +794,7 @@ bool Perl_is_uni_alnum(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alnum(tmpbuf); } @@ -843,7 +802,7 @@ bool Perl_is_uni_alnumc(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alnumc(tmpbuf); } @@ -851,7 +810,7 @@ bool Perl_is_uni_idfirst(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_idfirst(tmpbuf); } @@ -859,7 +818,7 @@ bool Perl_is_uni_alpha(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_alpha(tmpbuf); } @@ -867,7 +826,7 @@ bool Perl_is_uni_ascii(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_ascii(tmpbuf); } @@ -875,7 +834,7 @@ bool Perl_is_uni_space(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_space(tmpbuf); } @@ -883,7 +842,7 @@ bool Perl_is_uni_digit(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_digit(tmpbuf); } @@ -891,7 +850,7 @@ bool Perl_is_uni_upper(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_upper(tmpbuf); } @@ -899,7 +858,7 @@ bool Perl_is_uni_lower(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_lower(tmpbuf); } @@ -907,7 +866,7 @@ bool Perl_is_uni_cntrl(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_cntrl(tmpbuf); } @@ -915,7 +874,7 @@ bool Perl_is_uni_graph(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_graph(tmpbuf); } @@ -923,7 +882,7 @@ bool Perl_is_uni_print(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_print(tmpbuf); } @@ -931,7 +890,7 @@ bool Perl_is_uni_punct(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_punct(tmpbuf); } @@ -939,7 +898,7 @@ bool Perl_is_uni_xdigit(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return is_utf8_xdigit(tmpbuf); } @@ -947,7 +906,7 @@ U32 Perl_to_uni_upper(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return to_utf8_upper(tmpbuf); } @@ -955,7 +914,7 @@ U32 Perl_to_uni_title(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return to_utf8_title(tmpbuf); } @@ -963,7 +922,7 @@ U32 Perl_to_uni_lower(pTHX_ U32 c) { U8 tmpbuf[UTF8_MAXLEN+1]; - uvuni_to_utf8(tmpbuf, (UV)c); + uvchr_to_utf8(tmpbuf, (UV)c); return to_utf8_lower(tmpbuf); } @@ -1321,14 +1280,34 @@ UV Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) { HV* hv = (HV*)SvRV(sv); + /* 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 is length of encoded char -1 + */ U32 klen = UTF8SKIP(ptr) - 1; - U32 off = ptr[klen] & 127; /* NB: 64 bit always 0 when len > 1 */ + U32 off = ptr[klen]; STRLEN slen; - STRLEN needents = (klen ? 64 : 128); + STRLEN needents; U8 *tmps; U32 bit; SV *retval; + 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, @@ -1352,6 +1331,10 @@ 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); ENTER; SAVETMPS; save_re_context(); @@ -1359,10 +1342,8 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr) PUSHMARK(SP); EXTEND(SP,3); PUSHs((SV*)sv); - /* We call utf8_to_uni as we want and index into Unicode tables, - not a native character number. - */ - PUSHs(sv_2mortal(newSViv(utf8_to_uvuni(ptr, 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; if (call_method("SWASHGET", G_SCALAR)) @@ -1377,7 +1358,7 @@ 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"); } @@ -1406,3 +1387,56 @@ 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|U8* s|utf8n_to_uvchr|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); +} + +