X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/81714fb9c03d91d66b66cab6e899e81bf64a2ca7..73e91d5ab26c8e505f47d63fdd32c83a208f60d9:/doop.c diff --git a/doop.c b/doop.c index 1620465..c11555f 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, 2004, 2005, 2006, by Larry Wall and others + * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, + * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2009 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. @@ -9,7 +9,9 @@ */ /* - * "'So that was the job I felt I had to do when I started,' thought Sam." + * 'So that was the job I felt I had to do when I started,' thought Sam. + * + * [p.934 of _The Lord of the Rings_, VI/iii: "Mount Doom"] */ /* This file contains some common functions needed to carry out certain @@ -31,10 +33,12 @@ S_do_trans_simple(pTHX_ SV * const sv) dVAR; I32 matches = 0; STRLEN len; - U8 *s = (U8*)SvPV(sv,len); + U8 *s = (U8*)SvPV_nomg(sv,len); U8 * const send = s+len; - const short * const tbl = (short*)cPVOP->op_pv; + + PERL_ARGS_ASSERT_DO_TRANS_SIMPLE; + if (!tbl) Perl_croak(aTHX_ "panic: do_trans_simple line %d",__LINE__); @@ -97,11 +101,13 @@ S_do_trans_count(pTHX_ SV * const sv) { dVAR; STRLEN len; - const U8 *s = (const U8*)SvPV_const(sv, len); + const U8 *s = (const U8*)SvPV_nomg_const(sv, len); const U8 * const send = s + len; I32 matches = 0; - const short * const tbl = (short*)cPVOP->op_pv; + + PERL_ARGS_ASSERT_DO_TRANS_COUNT; + if (!tbl) Perl_croak(aTHX_ "panic: do_trans_count line %d",__LINE__); @@ -133,11 +139,13 @@ S_do_trans_complex(pTHX_ SV * const sv) { dVAR; STRLEN len; - U8 *s = (U8*)SvPV(sv, len); + U8 *s = (U8*)SvPV_nomg(sv, len); U8 * const send = s+len; I32 matches = 0; - const short * const tbl = (short*)cPVOP->op_pv; + + PERL_ARGS_ASSERT_DO_TRANS_COMPLEX; + if (!tbl) Perl_croak(aTHX_ "panic: do_trans_complex line %d",__LINE__); @@ -195,10 +203,6 @@ S_do_trans_complex(pTHX_ SV * const sv) if (complement && !del) rlen = tbl[0x100]; -#ifdef MACOS_TRADITIONAL -#define comp CoMP /* "comp" is a keyword in some compilers ... */ -#endif - if (PL_op->op_private & OPpTRANS_SQUASH) { UV pch = 0xfeedface; while (s < send) { @@ -306,16 +310,22 @@ S_do_trans_simple_utf8(pTHX_ SV * const sv) I32 matches = 0; const I32 grows = PL_op->op_private & OPpTRANS_GROWS; STRLEN len; - - SV* const rv = (SV*)cSVOP->op_sv; - HV* const hv = (HV*)SvRV(rv); + SV* const rv = +#ifdef USE_ITHREADS + PAD_SVl(cPADOP->op_padix); +#else + MUTABLE_SV(cSVOP->op_sv); +#endif + HV* const hv = MUTABLE_HV(SvRV(rv)); SV* const * svp = hv_fetchs(hv, "NONE", FALSE); const UV none = svp ? SvUV(*svp) : 0x7fffffff; const UV extra = none + 1; UV final = 0; U8 hibit = 0; - s = (U8*)SvPV(sv, len); + PERL_ARGS_ASSERT_DO_TRANS_SIMPLE_UTF8; + + s = (U8*)SvPV_nomg(sv, len); if (!SvUTF8(sv)) { const U8 *t = s; const U8 * const e = s + len; @@ -402,15 +412,21 @@ S_do_trans_count_utf8(pTHX_ SV * const sv) const U8 *send; I32 matches = 0; STRLEN len; - - SV* const rv = (SV*)cSVOP->op_sv; - HV* const hv = (HV*)SvRV(rv); + SV* const rv = +#ifdef USE_ITHREADS + PAD_SVl(cPADOP->op_padix); +#else + MUTABLE_SV(cSVOP->op_sv); +#endif + HV* const hv = MUTABLE_HV(SvRV(rv)); SV* const * const svp = hv_fetchs(hv, "NONE", FALSE); const UV none = svp ? SvUV(*svp) : 0x7fffffff; const UV extra = none + 1; U8 hibit = 0; - s = (const U8*)SvPV_const(sv, len); + PERL_ARGS_ASSERT_DO_TRANS_COUNT_UTF8; + + s = (const U8*)SvPV_nomg_const(sv, len); if (!SvUTF8(sv)) { const U8 *t = s; const U8 * const e = s + len; @@ -447,8 +463,13 @@ S_do_trans_complex_utf8(pTHX_ SV * const sv) 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 * const rv = (SV*)cSVOP->op_sv; - HV * const hv = (HV*)SvRV(rv); + SV* const rv = +#ifdef USE_ITHREADS + PAD_SVl(cPADOP->op_padix); +#else + MUTABLE_SV(cSVOP->op_sv); +#endif + HV * const hv = MUTABLE_HV(SvRV(rv)); SV * const *svp = hv_fetchs(hv, "NONE", FALSE); const UV none = svp ? SvUV(*svp) : 0x7fffffff; const UV extra = none + 1; @@ -457,8 +478,10 @@ S_do_trans_complex_utf8(pTHX_ SV * const sv) STRLEN len; U8 *dstart, *dend; U8 hibit = 0; + U8 *s = (U8*)SvPV_nomg(sv, len); + + PERL_ARGS_ASSERT_DO_TRANS_COMPLEX_UTF8; - U8 *s = (U8*)SvPV(sv, len); if (!SvUTF8(sv)) { const U8 *t = s; const U8 * const e = s + len; @@ -606,11 +629,13 @@ Perl_do_trans(pTHX_ SV *sv) const I32 hasutf = (PL_op->op_private & (OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF)); - if (SvREADONLY(sv)) { + PERL_ARGS_ASSERT_DO_TRANS; + + if (SvREADONLY(sv) && !(PL_op->op_private & OPpTRANS_IDENTICAL)) { if (SvIsCOW(sv)) sv_force_normal_flags(sv, 0); - if (SvREADONLY(sv) && !(PL_op->op_private & OPpTRANS_IDENTICAL)) - Perl_croak(aTHX_ PL_no_modify); + if (SvREADONLY(sv)) + Perl_croak_no_modify(aTHX); } (void)SvPV_const(sv, len); if (!len) @@ -656,6 +681,8 @@ Perl_do_join(pTHX_ register SV *sv, SV *delim, register SV **mark, register SV * register STRLEN len; STRLEN delimlen; + PERL_ARGS_ASSERT_DO_JOIN; + (void) SvPV_const(delim, delimlen); /* stringify and get the delimlen */ /* SvCUR assumes it's SvPOK() and woe betide you if it's not. */ @@ -678,7 +705,7 @@ Perl_do_join(pTHX_ register SV *sv, SV *delim, register SV **mark, register SV * ++mark; } - sv_setpvn(sv, "", 0); + sv_setpvs(sv, ""); /* sv_setpv retains old UTF8ness [perl #24846] */ SvUTF8_off(sv); @@ -712,6 +739,16 @@ Perl_do_sprintf(pTHX_ SV *sv, I32 len, SV **sarg) const char * const pat = SvPV_const(*sarg, patlen); bool do_taint = FALSE; + PERL_ARGS_ASSERT_DO_SPRINTF; + + if (SvTAINTED(*sarg)) + TAINT_PROPER( + (PL_op && PL_op->op_type < OP_max) + ? (PL_op->op_type == OP_PRTF) + ? "printf" + : PL_op_name[PL_op->op_type] + : "(unknown)" + ); SvUTF8_off(sv); if (DO_UTF8(*sarg)) SvUTF8_on(sv); @@ -726,25 +763,34 @@ UV Perl_do_vecget(pTHX_ SV *sv, I32 offset, I32 size) { dVAR; - STRLEN srclen, len, uoffset; + STRLEN srclen, len, uoffset, bitoffs = 0; const unsigned char *s = (const unsigned char *) SvPV_const(sv, srclen); UV retnum = 0; + PERL_ARGS_ASSERT_DO_VECGET; + if (offset < 0) - return retnum; + return 0; if (size < 1 || (size & (size-1))) /* size < 1 or not a power of two */ Perl_croak(aTHX_ "Illegal number of bits in vec"); if (SvUTF8(sv)) (void) Perl_sv_utf8_downgrade(aTHX_ sv, TRUE); - uoffset = offset*size; /* turn into bit offset */ - len = (uoffset + size + 7) / 8; /* required number of bytes */ + if (size < 8) { + bitoffs = ((offset%8)*size)%8; + uoffset = offset/(8/size); + } + else if (size > 8) + uoffset = offset*(size/8); + else + uoffset = offset; + + len = uoffset + (bitoffs + size + 7)/8; /* required number of bytes */ if (len > srclen) { if (size <= 8) retnum = 0; else { - uoffset >>= 3; /* turn into byte offset */ if (size == 16) { if (uoffset >= srclen) retnum = 0; @@ -769,9 +815,8 @@ Perl_do_vecget(pTHX_ SV *sv, I32 offset, I32 size) } #ifdef UV_IS_QUAD else if (size == 64) { - if (ckWARN(WARN_PORTABLE)) - Perl_warner(aTHX_ packWARN(WARN_PORTABLE), - "Bit vector size > 32 non-portable"); + Perl_ck_warner(aTHX_ packWARN(WARN_PORTABLE), + "Bit vector size > 32 non-portable"); if (uoffset >= srclen) retnum = 0; else if (uoffset + 1 >= srclen) @@ -821,9 +866,8 @@ Perl_do_vecget(pTHX_ SV *sv, I32 offset, I32 size) } } else if (size < 8) - retnum = (s[uoffset >> 3] >> (uoffset & 7)) & ((1 << size) - 1); + retnum = (s[uoffset] >> bitoffs) & ((1 << size) - 1); else { - uoffset >>= 3; /* turn into byte offset */ if (size == 8) retnum = s[uoffset]; else if (size == 16) @@ -838,9 +882,8 @@ Perl_do_vecget(pTHX_ SV *sv, I32 offset, I32 size) s[uoffset + 3]; #ifdef UV_IS_QUAD else if (size == 64) { - if (ckWARN(WARN_PORTABLE)) - Perl_warner(aTHX_ packWARN(WARN_PORTABLE), - "Bit vector size > 32 non-portable"); + Perl_ck_warner(aTHX_ packWARN(WARN_PORTABLE), + "Bit vector size > 32 non-portable"); retnum = ((UV) s[uoffset ] << 56) + ((UV) s[uoffset + 1] << 48) + @@ -865,7 +908,7 @@ void Perl_do_vecset(pTHX_ SV *sv) { dVAR; - register I32 offset; + register I32 offset, bitoffs = 0; register I32 size; register unsigned char *s; register UV lval; @@ -874,6 +917,8 @@ Perl_do_vecset(pTHX_ SV *sv) STRLEN len; SV * const targ = LvTARG(sv); + PERL_ARGS_ASSERT_DO_VECSET; + if (!targ) return; s = (unsigned char*)SvPV_force(targ, targlen); @@ -894,8 +939,14 @@ Perl_do_vecset(pTHX_ SV *sv) if (size < 1 || (size & (size-1))) /* size < 1 or not a power of two */ Perl_croak(aTHX_ "Illegal number of bits in vec"); - offset *= size; /* turn into bit offset */ - len = (offset + size + 7) / 8; /* required number of bytes */ + if (size < 8) { + bitoffs = ((offset%8)*size)%8; + offset /= 8/size; + } + else if (size > 8) + offset *= size/8; + + len = offset + (bitoffs + size + 7)/8; /* required number of bytes */ if (len > targlen) { s = (unsigned char*)SvGROW(targ, len + 1); (void)memzero((char *)(s + targlen), len - targlen + 1); @@ -904,14 +955,11 @@ Perl_do_vecset(pTHX_ SV *sv) if (size < 8) { mask = (1 << size) - 1; - size = offset & 7; lval &= mask; - offset >>= 3; /* turn into byte offset */ - s[offset] &= ~(mask << size); - s[offset] |= lval << size; + s[offset] &= ~(mask << bitoffs); + s[offset] |= lval << bitoffs; } else { - offset >>= 3; /* turn into byte offset */ if (size == 8) s[offset ] = (U8)( lval & 0xff); else if (size == 16) { @@ -926,9 +974,8 @@ Perl_do_vecset(pTHX_ SV *sv) } #ifdef UV_IS_QUAD else if (size == 64) { - if (ckWARN(WARN_PORTABLE)) - Perl_warner(aTHX_ packWARN(WARN_PORTABLE), - "Bit vector size > 32 non-portable"); + Perl_ck_warner(aTHX_ packWARN(WARN_PORTABLE), + "Bit vector size > 32 non-portable"); s[offset ] = (U8)((lval >> 56) & 0xff); s[offset+1] = (U8)((lval >> 48) & 0xff); s[offset+2] = (U8)((lval >> 40) & 0xff); @@ -944,218 +991,6 @@ Perl_do_vecset(pTHX_ SV *sv) } void -Perl_do_chop(pTHX_ register SV *astr, register SV *sv) -{ - dVAR; - STRLEN len; - char *s; - - if (SvTYPE(sv) == SVt_PVAV) { - register I32 i; - AV* const av = (AV*)sv; - 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); - } - return; - } - else if (SvTYPE(sv) == SVt_PVHV) { - HV* const hv = (HV*)sv; - HE* entry; - (void)hv_iterinit(hv); - while ((entry = hv_iternext(hv))) - do_chop(astr,hv_iterval(hv,entry)); - return; - } - else if (SvREADONLY(sv)) { - if (SvFAKE(sv)) { - /* SV is copy-on-write */ - sv_force_normal_flags(sv, 0); - } - if (SvREADONLY(sv)) - Perl_croak(aTHX_ PL_no_modify); - } - - if (PL_encoding && !SvUTF8(sv)) { - /* like in do_chomp(), utf8-ize the sv as a side-effect - * if we're using encoding. */ - sv_recode_to_utf8(sv, PL_encoding); - } - - s = SvPV(sv, len); - if (len && !SvPOK(sv)) - s = SvPV_force(sv, len); - if (DO_UTF8(sv)) { - if (s && len) { - char * const send = s + len; - char * const start = s; - s = send - 1; - while (s > start && UTF8_IS_CONTINUATION(*s)) - s--; - if (is_utf8_string((U8*)s, send - s)) { - sv_setpvn(astr, s, send - s); - *s = '\0'; - SvCUR_set(sv, s - start); - SvNIOK_off(sv); - SvUTF8_on(astr); - } - } - else - sv_setpvn(astr, "", 0); - } - else if (s && len) { - s += --len; - sv_setpvn(astr, s, 1); - *s = '\0'; - SvCUR_set(sv, len); - SvUTF8_off(sv); - SvNIOK_off(sv); - } - else - sv_setpvn(astr, "", 0); - SvSETMAGIC(sv); -} - -I32 -Perl_do_chomp(pTHX_ register SV *sv) -{ - dVAR; - register I32 count; - STRLEN len; - char *s; - char *temp_buffer = NULL; - SV* svrecode = NULL; - - if (RsSNARF(PL_rs)) - return 0; - if (RsRECORD(PL_rs)) - return 0; - count = 0; - if (SvTYPE(sv) == SVt_PVAV) { - register I32 i; - AV* const av = (AV*)sv; - 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); - } - return count; - } - else if (SvTYPE(sv) == SVt_PVHV) { - HV* const hv = (HV*)sv; - HE* entry; - (void)hv_iterinit(hv); - while ((entry = hv_iternext(hv))) - count += do_chomp(hv_iterval(hv,entry)); - return count; - } - else if (SvREADONLY(sv)) { - if (SvFAKE(sv)) { - /* SV is copy-on-write */ - sv_force_normal_flags(sv, 0); - } - 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; - if (RsPARA(PL_rs)) { - if (*s != '\n') - goto nope; - ++count; - while (len && s[-1] == '\n') { - --len; - --s; - ++count; - } - } - else { - STRLEN rslen, rs_charlen; - const char *rsptr = SvPV_const(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_const(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; - ++count; - } - else { - if (len < rslen - 1) - goto nope; - len -= rslen - 1; - s -= rslen - 1; - if (memNE(s, rsptr, rslen)) - goto nope; - count += rs_charlen; - } - } - s = SvPV_force_nolen(sv); - SvCUR_set(sv, len); - *SvEND(sv) = '\0'; - SvNIOK_off(sv); - SvSETMAGIC(sv); - } - nope: - - if (svrecode) - SvREFCNT_dec(svrecode); - - Safefree(temp_buffer); - return count; -} - -void Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) { dVAR; @@ -1177,10 +1012,17 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) bool right_utf; STRLEN needlen = 0; + PERL_ARGS_ASSERT_DO_VOP; if (sv != left || (optype != OP_BIT_AND && !SvOK(sv) && !SvGMAGICAL(sv))) - sv_setpvn(sv, "", 0); /* avoid undef warning on |= and ^= */ - lsave = lc = SvPV_nomg_const(left, leftlen); + sv_setpvs(sv, ""); /* avoid undef warning on |= and ^= */ + if (sv == left) { + lsave = lc = SvPV_force_nomg(left, leftlen); + } + else { + lsave = lc = SvPV_nomg_const(left, leftlen); + SvPV_force_nomg_nolen(sv); + } rsave = rc = SvPV_nomg_const(right, rightlen); /* This need to come after SvPV to ensure that string overloading has @@ -1193,13 +1035,13 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) /* Avoid triggering overloading again by using temporaries. Maybe there should be a variant of sv_utf8_upgrade that takes pvn */ - right = sv_2mortal(newSVpvn(rsave, rightlen)); + right = newSVpvn_flags(rsave, rightlen, SVs_TEMP); sv_utf8_upgrade(right); rsave = rc = SvPV_nomg_const(right, rightlen); right_utf = TRUE; } else if (!left_utf && right_utf) { - left = sv_2mortal(newSVpvn(lsave, leftlen)); + left = newSVpvn_flags(lsave, leftlen, SVs_TEMP); sv_utf8_upgrade(left); lsave = lc = SvPV_nomg_const(left, leftlen); left_utf = TRUE; @@ -1382,13 +1224,14 @@ Perl_do_kv(pTHX) { dVAR; dSP; - HV * const hv = (HV*)POPs; + HV * const hv = MUTABLE_HV(POPs); HV *keys; register HE *entry; 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); + /* op_type is OP_RKEYS/OP_RVALUES if pp_rkeys delegated to here */ + const I32 dokeys = dokv || (PL_op->op_type == OP_KEYS || PL_op->op_type == OP_RKEYS); + const I32 dovalues = dokv || (PL_op->op_type == OP_VALUES || PL_op->op_type == OP_RVALUES); if (!hv) { if (PL_op->op_flags & OPf_MOD || LVRET) { /* lvalue */ @@ -1407,38 +1250,30 @@ Perl_do_kv(pTHX) RETURN; if (gimme == G_SCALAR) { - IV i; - dTARGET; - if (PL_op->op_flags & OPf_MOD || LVRET) { /* lvalue */ - if (SvTYPE(TARG) < SVt_PVLV) { - sv_upgrade(TARG, SVt_PVLV); - sv_magic(TARG, NULL, PERL_MAGIC_nkeys, NULL, 0); - } - LvTYPE(TARG) = 'k'; - if (LvTARG(TARG) != (SV*)keys) { - if (LvTARG(TARG)) - SvREFCNT_dec(LvTARG(TARG)); - LvTARG(TARG) = SvREFCNT_inc_simple(keys); - } - PUSHs(TARG); - RETURN; - } - - if (! SvTIED_mg((SV*)keys, PERL_MAGIC_tied) - && ! SvTIED_mg((SV*)keys, PERL_MAGIC_regdata_names)) - { - i = HvKEYS(keys); + SV * const ret = sv_2mortal(newSV_type(SVt_PVLV)); /* Not TARG RT#67838 */ + sv_magic(ret, NULL, PERL_MAGIC_nkeys, NULL, 0); + LvTYPE(ret) = 'k'; + LvTARG(ret) = SvREFCNT_inc_simple(keys); + PUSHs(ret); } else { - i = 0; - while (hv_iternext(keys)) i++; + IV i; + dTARGET; + + if (! SvTIED_mg((const SV *)keys, PERL_MAGIC_tied) ) { + i = HvUSEDKEYS(keys); + } + else { + i = 0; + while (hv_iternext(keys)) i++; + } + PUSHi( i ); } - PUSHi( i ); RETURN; } - EXTEND(SP, HvKEYS(keys) * (dokeys + dovalues)); + EXTEND(SP, HvUSEDKEYS(keys) * (dokeys + dovalues)); PUTBACK; /* hv_iternext and hv_iterval might clobber stack_sp */ while ((entry = hv_iternext(keys))) {