X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=doop.c;h=22160aff631c24011fda495109059caa84666beb;hb=d6802e4345360e741b212b52b6af4a704cfce207;hp=dc64c454badc5e46b66e063f31aea8ef7ca41011;hpb=a77e643ad0ddc577b22643ae4113c389b2dab048;p=p5sagit%2Fp5-mst-13.2.git diff --git a/doop.c b/doop.c index dc64c45..22160af 100644 --- a/doop.c +++ b/doop.c @@ -1,7 +1,7 @@ /* doop.c * * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, - * 2000, 2001, 2002, by Larry Wall and others + * 2000, 2001, 2002, 2004, 2005, by Larry Wall and others * * You may distribute under the terms of either the GNU General Public * License or the Artistic License, as specified in the README file. @@ -12,6 +12,11 @@ * "'So that was the job I felt I had to do when I started,' thought Sam." */ +/* This file contains some common functions needed to carry out certain + * ops. For example both pp_schomp() and pp_chomp() - scalar and array + * chomp operations - call the function do_chomp() found in this file. + */ + #include "EXTERN.h" #define PERL_IN_DOOP_C #include "perl.h" @@ -25,15 +30,13 @@ S_do_trans_simple(pTHX_ SV *sv) { U8 *s; U8 *d; - U8 *send; + const U8 *send; U8 *dstart; I32 matches = 0; - I32 grows = PL_op->op_private & OPpTRANS_GROWS; + const I32 grows = PL_op->op_private & OPpTRANS_GROWS; STRLEN len; - short *tbl; - I32 ch; - tbl = (short*)cPVOP->op_pv; + const short *tbl = (short*)cPVOP->op_pv; if (!tbl) Perl_croak(aTHX_ "panic: do_trans_simple line %d",__LINE__); @@ -43,7 +46,8 @@ S_do_trans_simple(pTHX_ SV *sv) /* First, take care of non-UTF-8 input strings, because they're easy */ if (!SvUTF8(sv)) { while (s < send) { - if ((ch = tbl[*s]) >= 0) { + const I32 ch = tbl[*s]; + if (ch >= 0) { matches++; *s++ = (U8)ch; } @@ -62,10 +66,10 @@ S_do_trans_simple(pTHX_ SV *sv) dstart = d; while (s < send) { STRLEN ulen; - UV c; + I32 ch; /* Need to check this, otherwise 128..255 won't match */ - c = utf8n_to_uvchr(s, send - s, &ulen, 0); + const UV c = utf8n_to_uvchr(s, send - s, &ulen, 0); if (c < 0x100 && (ch = tbl[c]) >= 0) { matches++; d = uvchr_to_utf8(d, ch); @@ -97,10 +101,9 @@ S_do_trans_count(pTHX_ SV *sv) U8 *send; I32 matches = 0; STRLEN len; - short *tbl; - I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT; + const I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT; - tbl = (short*)cPVOP->op_pv; + const short * const tbl = (short*)cPVOP->op_pv; if (!tbl) Perl_croak(aTHX_ "panic: do_trans_count line %d",__LINE__); @@ -114,9 +117,8 @@ S_do_trans_count(pTHX_ SV *sv) } else while (s < send) { - UV c; STRLEN ulen; - c = utf8n_to_uvchr(s, send - s, &ulen, 0); + const UV c = utf8n_to_uvchr(s, send - s, &ulen, 0); if (c < 0x100) { if (tbl[c] >= 0) matches++; @@ -137,14 +139,12 @@ S_do_trans_complex(pTHX_ SV *sv) U8 *dstart; I32 isutf8; I32 matches = 0; - I32 grows = PL_op->op_private & OPpTRANS_GROWS; - I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT; - I32 del = PL_op->op_private & OPpTRANS_DELETE; + const I32 grows = PL_op->op_private & OPpTRANS_GROWS; + const I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT; + const I32 del = PL_op->op_private & OPpTRANS_DELETE; STRLEN len, rlen = 0; - short *tbl; - I32 ch; - tbl = (short*)cPVOP->op_pv; + const short * const tbl = (short*)cPVOP->op_pv; if (!tbl) Perl_croak(aTHX_ "panic: do_trans_complex line %d",__LINE__); @@ -155,9 +155,10 @@ S_do_trans_complex(pTHX_ SV *sv) if (!isutf8) { dstart = d = s; if (PL_op->op_private & OPpTRANS_SQUASH) { - U8* p = send; + const U8* p = send; while (s < send) { - if ((ch = tbl[*s]) >= 0) { + const I32 ch = tbl[*s]; + if (ch >= 0) { *d = (U8)ch; matches++; if (p != d - 1 || *p != *d) @@ -172,7 +173,8 @@ S_do_trans_complex(pTHX_ SV *sv) } else { while (s < send) { - if ((ch = tbl[*s]) >= 0) { + const I32 ch = tbl[*s]; + if (ch >= 0) { matches++; *d++ = (U8)ch; } @@ -203,7 +205,8 @@ S_do_trans_complex(pTHX_ SV *sv) UV pch = 0xfeedface; while (s < send) { STRLEN len; - UV comp = utf8_to_uvchr(s, &len); + const UV comp = utf8_to_uvchr(s, &len); + I32 ch; if (comp > 0xff) { if (!complement) { @@ -247,7 +250,8 @@ S_do_trans_complex(pTHX_ SV *sv) else { while (s < send) { STRLEN len; - UV comp = utf8_to_uvchr(s, &len); + const UV comp = utf8_to_uvchr(s, &len); + I32 ch; if (comp > 0xff) { if (!complement) { Move(s, d, len, U8); @@ -299,14 +303,14 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) U8 *start; U8 *dstart, *dend; I32 matches = 0; - I32 grows = PL_op->op_private & OPpTRANS_GROWS; + const I32 grows = PL_op->op_private & OPpTRANS_GROWS; STRLEN len; SV* rv = (SV*)cSVOP->op_sv; HV* hv = (HV*)SvRV(rv); SV** svp = hv_fetch(hv, "NONE", 4, FALSE); - UV none = svp ? SvUV(*svp) : 0x7fffffff; - UV extra = none + 1; + const UV none = svp ? SvUV(*svp) : 0x7fffffff; + const UV extra = none + 1; UV final = 0; UV uv; I32 isutf8; @@ -315,9 +319,9 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) s = (U8*)SvPV(sv, len); isutf8 = SvUTF8(sv); if (!isutf8) { - U8 *t = s, *e = s + len; + const U8 *t = s, *e = s + len; while (t < e) { - U8 ch = *t++; + const U8 ch = *t++; if ((hibit = !NATIVE_IS_INVARIANT(ch))) break; } @@ -333,7 +337,7 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) if (grows) { /* d needs to be bigger than s, in case e.g. upgrading is required */ - New(0, d, len*3+UTF8_MAXLEN, U8); + New(0, d, len * 3 + UTF8_MAXBYTES, U8); dend = d + len * 3; dstart = d; } @@ -349,14 +353,13 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) d = uvuni_to_utf8(d, uv); } else if (uv == none) { - int i = UTF8SKIP(s); + const int i = UTF8SKIP(s); Move(s, d, i, U8); d += i; s += i; } else if (uv == extra) { - int i = UTF8SKIP(s); - s += i; + s += UTF8SKIP(s); matches++; d = uvuni_to_utf8(d, final); } @@ -364,11 +367,11 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) s += UTF8SKIP(s); if (d > dend) { - STRLEN clen = d - dstart; - STRLEN nlen = dend - dstart + len + UTF8_MAXLEN; + const STRLEN clen = d - dstart; + const STRLEN nlen = dend - dstart + len + UTF8_MAXBYTES; if (!grows) Perl_croak(aTHX_ "panic: do_trans_simple_utf8 line %d",__LINE__); - Renew(dstart, nlen+UTF8_MAXLEN, U8); + Renew(dstart, nlen + UTF8_MAXBYTES, U8); d = dstart + clen; dend = dstart + nlen; } @@ -400,16 +403,16 @@ S_do_trans_count_utf8(pTHX_ SV *sv) SV* rv = (SV*)cSVOP->op_sv; HV* hv = (HV*)SvRV(rv); SV** svp = hv_fetch(hv, "NONE", 4, FALSE); - UV none = svp ? SvUV(*svp) : 0x7fffffff; - UV extra = none + 1; - UV uv; + const UV none = svp ? SvUV(*svp) : 0x7fffffff; + const UV extra = none + 1; U8 hibit = 0; s = (U8*)SvPV(sv, len); if (!SvUTF8(sv)) { - U8 *t = s, *e = s + len; + const U8 *t = s; + const U8 *e = s + len; while (t < e) { - U8 ch = *t++; + const U8 ch = *t++; if ((hibit = !NATIVE_IS_INVARIANT(ch))) break; } @@ -419,6 +422,7 @@ S_do_trans_count_utf8(pTHX_ SV *sv) send = s + len; while (s < send) { + UV uv; if ((uv = swash_fetch(rv, s, TRUE)) < none || uv == extra) matches++; s += UTF8SKIP(s); @@ -432,32 +436,29 @@ S_do_trans_count_utf8(pTHX_ SV *sv) STATIC I32 S_do_trans_complex_utf8(pTHX_ SV *sv) { - U8 *s; U8 *start, *send; U8 *d; I32 matches = 0; - I32 squash = PL_op->op_private & OPpTRANS_SQUASH; - I32 del = PL_op->op_private & OPpTRANS_DELETE; - I32 grows = PL_op->op_private & OPpTRANS_GROWS; + const I32 squash = PL_op->op_private & OPpTRANS_SQUASH; + const I32 del = PL_op->op_private & OPpTRANS_DELETE; + const I32 grows = PL_op->op_private & OPpTRANS_GROWS; SV* rv = (SV*)cSVOP->op_sv; HV* hv = (HV*)SvRV(rv); SV** svp = hv_fetch(hv, "NONE", 4, FALSE); - UV none = svp ? SvUV(*svp) : 0x7fffffff; - UV extra = none + 1; + const UV none = svp ? SvUV(*svp) : 0x7fffffff; + const UV extra = none + 1; UV final = 0; bool havefinal = FALSE; - UV uv; STRLEN len; U8 *dstart, *dend; - I32 isutf8; U8 hibit = 0; - s = (U8*)SvPV(sv, len); - isutf8 = SvUTF8(sv); + U8 *s = (U8*)SvPV(sv, len); + const I32 isutf8 = SvUTF8(sv); if (!isutf8) { - U8 *t = s, *e = s + len; + const U8 *t = s, *e = s + len; while (t < e) { - U8 ch = *t++; + const U8 ch = *t++; if ((hibit = !NATIVE_IS_INVARIANT(ch))) break; } @@ -475,7 +476,7 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) if (grows) { /* d needs to be bigger than s, in case e.g. upgrading is required */ - New(0, d, len*3+UTF8_MAXLEN, U8); + New(0, d, len * 3 + UTF8_MAXBYTES, U8); dend = d + len * 3; dstart = d; } @@ -487,14 +488,14 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) if (squash) { UV puv = 0xfeedface; while (s < send) { - uv = swash_fetch(rv, s, TRUE); + UV uv = swash_fetch(rv, s, TRUE); if (d > dend) { - STRLEN clen = d - dstart; - STRLEN nlen = dend - dstart + len + UTF8_MAXLEN; + const STRLEN clen = d - dstart; + const STRLEN nlen = dend - dstart + len + UTF8_MAXBYTES; if (!grows) Perl_croak(aTHX_ "panic: do_trans_complex_utf8 line %d",__LINE__); - Renew(dstart, nlen+UTF8_MAXLEN, U8); + Renew(dstart, nlen + UTF8_MAXBYTES, U8); d = dstart + clen; dend = dstart + nlen; } @@ -508,7 +509,7 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) continue; } else if (uv == none) { /* "none" is unmapped character */ - int i = UTF8SKIP(s); + const int i = UTF8SKIP(s); Move(s, d, i, U8); d += i; s += i; @@ -542,13 +543,13 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) } else { while (s < send) { - uv = swash_fetch(rv, s, TRUE); + const UV uv = swash_fetch(rv, s, TRUE); if (d > dend) { - STRLEN clen = d - dstart; - STRLEN nlen = dend - dstart + len + UTF8_MAXLEN; + const STRLEN clen = d - dstart; + const STRLEN nlen = dend - dstart + len + UTF8_MAXBYTES; if (!grows) Perl_croak(aTHX_ "panic: do_trans_complex_utf8 line %d",__LINE__); - Renew(dstart, nlen+UTF8_MAXLEN, U8); + Renew(dstart, nlen + UTF8_MAXBYTES, U8); d = dstart + clen; dend = dstart + nlen; } @@ -559,7 +560,7 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) continue; } else if (uv == none) { /* "none" is unmapped character */ - int i = UTF8SKIP(s); + const int i = UTF8SKIP(s); Move(s, d, i, U8); d += i; s += i; @@ -595,7 +596,7 @@ I32 Perl_do_trans(pTHX_ SV *sv) { STRLEN len; - I32 hasutf = (PL_op->op_private & + const I32 hasutf = (PL_op->op_private & (OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF)); if (SvREADONLY(sv)) { @@ -646,7 +647,6 @@ Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **s register I32 items = sp - mark; register STRLEN len; STRLEN delimlen; - STRLEN tmplen; (void) SvPV(del, delimlen); /* stringify and get the delimlen */ /* SvCUR assumes it's SvPOK() and woe betide you if it's not. */ @@ -657,6 +657,7 @@ Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **s if (SvLEN(sv) < len + items) { /* current length is way too short */ while (items-- > 0) { if (*mark && !SvGAMAGIC(*mark) && SvOK(*mark)) { + STRLEN tmplen; SvPV(*mark, tmplen); len += tmplen; } @@ -669,7 +670,10 @@ Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **s ++mark; } - sv_setpv(sv, ""); + sv_setpvn(sv, "", 0); + /* sv_setpv retains old UTF8ness [perl #24846] */ + SvUTF8_off(sv); + if (PL_tainting && SvMAGICAL(sv)) SvTAINTED_off(sv); @@ -696,7 +700,7 @@ void Perl_do_sprintf(pTHX_ SV *sv, I32 len, SV **sarg) { STRLEN patlen; - char *pat = SvPV(*sarg, patlen); + const char *pat = SvPV(*sarg, patlen); bool do_taint = FALSE; SvUTF8_off(sv); @@ -936,10 +940,10 @@ Perl_do_chop(pTHX_ register SV *astr, register SV *sv) if (SvTYPE(sv) == SVt_PVAV) { register I32 i; - I32 max; AV* av = (AV*)sv; - max = AvFILL(av); - for (i = 0; i <= max; i++) { + const I32 max = AvFILL(av); + + for (i = 0; i <= max; i++) { sv = (SV*)av_fetch(av, i, FALSE); if (sv && ((sv = *(SV**)sv), sv != &PL_sv_undef)) do_chop(astr, sv); @@ -1004,6 +1008,8 @@ Perl_do_chomp(pTHX_ register SV *sv) STRLEN len; STRLEN n_a; char *s; + char *temp_buffer = NULL; + SV* svrecode = Nullsv; if (RsSNARF(PL_rs)) return 0; @@ -1012,10 +1018,10 @@ Perl_do_chomp(pTHX_ register SV *sv) count = 0; if (SvTYPE(sv) == SVt_PVAV) { register I32 i; - I32 max; AV* av = (AV*)sv; - max = AvFILL(av); - for (i = 0; i <= max; i++) { + const I32 max = AvFILL(av); + + for (i = 0; i <= max; i++) { sv = (SV*)av_fetch(av, i, FALSE); if (sv && ((sv = *(SV**)sv), sv != &PL_sv_undef)) count += do_chomp(sv); @@ -1039,6 +1045,18 @@ Perl_do_chomp(pTHX_ register SV *sv) if (SvREADONLY(sv)) Perl_croak(aTHX_ PL_no_modify); } + + if (PL_encoding) { + if (!SvUTF8(sv)) { + /* XXX, here sv is utf8-ized as a side-effect! + If encoding.pm is used properly, almost string-generating + operations, including literal strings, chr(), input data, etc. + should have been utf8-ized already, right? + */ + sv_recode_to_utf8(sv, PL_encoding); + } + } + s = SvPV(sv, len); if (s && len) { s += --len; @@ -1053,8 +1071,43 @@ Perl_do_chomp(pTHX_ register SV *sv) } } else { - STRLEN rslen; + STRLEN rslen, rs_charlen; char *rsptr = SvPV(PL_rs, rslen); + + rs_charlen = SvUTF8(PL_rs) + ? sv_len_utf8(PL_rs) + : rslen; + + if (SvUTF8(PL_rs) != SvUTF8(sv)) { + /* Assumption is that rs is shorter than the scalar. */ + if (SvUTF8(PL_rs)) { + /* RS is utf8, scalar is 8 bit. */ + bool is_utf8 = TRUE; + temp_buffer = (char*)bytes_from_utf8((U8*)rsptr, + &rslen, &is_utf8); + if (is_utf8) { + /* Cannot downgrade, therefore cannot possibly match + */ + assert (temp_buffer == rsptr); + temp_buffer = NULL; + goto nope; + } + rsptr = temp_buffer; + } + else if (PL_encoding) { + /* RS is 8 bit, encoding.pm is used. + * Do not recode PL_rs as a side-effect. */ + svrecode = newSVpvn(rsptr, rslen); + sv_recode_to_utf8(svrecode, PL_encoding); + rsptr = SvPV(svrecode, rslen); + rs_charlen = sv_len_utf8(svrecode); + } + else { + /* RS is 8 bit, scalar is utf8. */ + temp_buffer = (char*)bytes_to_utf8((U8*)rsptr, &rslen); + rsptr = temp_buffer; + } + } if (rslen == 1) { if (*s != *rsptr) goto nope; @@ -1067,7 +1120,7 @@ Perl_do_chomp(pTHX_ register SV *sv) s -= rslen - 1; if (memNE(s, rsptr, rslen)) goto nope; - count += rslen; + count += rs_charlen; } } s = SvPV_force(sv, n_a); @@ -1077,6 +1130,11 @@ Perl_do_chomp(pTHX_ register SV *sv) SvSETMAGIC(sv); } nope: + + if (svrecode) + SvREFCNT_dec(svrecode); + + Safefree(temp_buffer); return count; } @@ -1097,8 +1155,8 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) I32 lensave; char *lsave; char *rsave; - bool left_utf = DO_UTF8(left); - bool right_utf = DO_UTF8(right); + const bool left_utf = DO_UTF8(left); + const bool right_utf = DO_UTF8(right); I32 needlen = 0; if (left_utf && !right_utf) @@ -1108,8 +1166,8 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) if (sv != left || (optype != OP_BIT_AND && !SvOK(sv) && !SvGMAGICAL(sv))) sv_setpvn(sv, "", 0); /* avoid undef warning on |= and ^= */ - lsave = lc = SvPV(left, leftlen); - rsave = rc = SvPV(right, rightlen); + lsave = lc = SvPV_nomg(left, leftlen); + rsave = rc = SvPV_nomg(right, rightlen); len = leftlen < rightlen ? leftlen : rightlen; lensave = len; if ((left_utf || right_utf) && (sv == left || sv == right)) { @@ -1118,7 +1176,7 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) } else if (SvOK(sv) || SvTYPE(sv) > SVt_PVMG) { STRLEN n_a; - dc = SvPV_force(sv, n_a); + dc = SvPV_force_nomg(sv, n_a); if (SvCUR(sv) < (STRLEN)len) { dc = SvGROW(sv, (STRLEN)(len + 1)); (void)memzero(dc + SvCUR(sv), len - SvCUR(sv) + 1); @@ -1203,7 +1261,7 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) !((long)lc % sizeof(long)) && !((long)rc % sizeof(long))) /* It's almost always aligned... */ { - I32 remainder = len % (sizeof(long)*4); + const I32 remainder = len % (sizeof(long)*4); len /= (sizeof(long)*4); dl = (long*)dc; @@ -1278,13 +1336,10 @@ Perl_do_kv(pTHX) HV *hv = (HV*)POPs; HV *keys; register HE *entry; - SV *tmpstr; - I32 gimme = GIMME_V; - I32 dokeys = (PL_op->op_type == OP_KEYS); - I32 dovalues = (PL_op->op_type == OP_VALUES); - - if (PL_op->op_type == OP_RV2HV || PL_op->op_type == OP_PADHV) - dokeys = dovalues = TRUE; + const I32 gimme = GIMME_V; + const I32 dokv = (PL_op->op_type == OP_RV2HV || PL_op->op_type == OP_PADHV); + const I32 dokeys = dokv || (PL_op->op_type == OP_KEYS); + const I32 dovalues = dokv || (PL_op->op_type == OP_VALUES); if (!hv) { if (PL_op->op_flags & OPf_MOD || LVRET) { /* lvalue */ @@ -1342,6 +1397,7 @@ Perl_do_kv(pTHX) XPUSHs(sv); /* won't clobber stack_sp */ } if (dovalues) { + SV *tmpstr; PUTBACK; tmpstr = hv_iterval(hv,entry); DEBUG_H(Perl_sv_setpvf(aTHX_ tmpstr, "%lu%%%d=%lu", @@ -1356,3 +1412,12 @@ Perl_do_kv(pTHX) return NORMAL; } +/* + * Local variables: + * c-indentation-style: bsd + * c-basic-offset: 4 + * indent-tabs-mode: t + * End: + * + * ex: set ts=8 sts=4 sw=4 noet: + */