X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/cbbf89328f29f925842d180c1410e324b297744d..7049214e3eea022e77794ed9b9d9038f93695ef9:/doop.c diff --git a/doop.c b/doop.c index c163182..e31fae2 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, 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 @@ -26,23 +28,20 @@ #endif STATIC I32 -S_do_trans_simple(pTHX_ SV *sv) +S_do_trans_simple(pTHX_ SV * const sv) { - U8 *s; - U8 *d; - const U8 *send; - U8 *dstart; + dVAR; I32 matches = 0; - const I32 grows = PL_op->op_private & OPpTRANS_GROWS; STRLEN 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__); - s = (U8*)SvPV(sv, len); - send = s + len; - /* First, take care of non-UTF-8 input strings, because they're easy */ if (!SvUTF8(sv)) { while (s < send) { @@ -54,70 +53,75 @@ S_do_trans_simple(pTHX_ SV *sv) s++; } SvSETMAGIC(sv); - return matches; - } - - /* Allow for expansion: $_="a".chr(400); tr/a/\xFE/, FE needs encoding */ - if (grows) - Newx(d, len*2+1, U8); - else - d = s; - dstart = d; - while (s < send) { - STRLEN ulen; - I32 ch; - - /* Need to check this, otherwise 128..255 won't match */ - 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); - s += ulen; - } - else { /* No match -> copy */ - Move(s, d, ulen, U8); - d += ulen; - s += ulen; - } - } - if (grows) { - sv_setpvn(sv, (char*)dstart, d - dstart); - Safefree(dstart); } else { - *d = '\0'; - SvCUR_set(sv, d - dstart); + const I32 grows = PL_op->op_private & OPpTRANS_GROWS; + U8 *d; + U8 *dstart; + + /* Allow for expansion: $_="a".chr(400); tr/a/\xFE/, FE needs encoding */ + if (grows) + Newx(d, len*2+1, U8); + else + d = s; + dstart = d; + while (s < send) { + STRLEN ulen; + I32 ch; + + /* Need to check this, otherwise 128..255 won't match */ + const UV c = utf8n_to_uvchr(s, send - s, &ulen, UTF8_ALLOW_DEFAULT); + if (c < 0x100 && (ch = tbl[c]) >= 0) { + matches++; + d = uvchr_to_utf8(d, ch); + s += ulen; + } + else { /* No match -> copy */ + Move(s, d, ulen, U8); + d += ulen; + s += ulen; + } + } + if (grows) { + sv_setpvn(sv, (char*)dstart, d - dstart); + Safefree(dstart); + } + else { + *d = '\0'; + SvCUR_set(sv, d - dstart); + } + SvUTF8_on(sv); + SvSETMAGIC(sv); } - SvUTF8_on(sv); - SvSETMAGIC(sv); return matches; } STATIC I32 -S_do_trans_count(pTHX_ SV *sv) +S_do_trans_count(pTHX_ SV * const sv) { - const U8 *s; - const U8 *send; - I32 matches = 0; + dVAR; STRLEN 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__); - s = (const U8*)SvPV_const(sv, len); - send = s + len; - - if (!SvUTF8(sv)) + if (!SvUTF8(sv)) { while (s < send) { if (tbl[*s++] >= 0) matches++; } + } else { const I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT; while (s < send) { STRLEN ulen; - const UV c = utf8n_to_uvchr(s, send - s, &ulen, 0); + const UV c = utf8n_to_uvchr(s, send - s, &ulen, UTF8_ALLOW_DEFAULT); if (c < 0x100) { if (tbl[c] >= 0) matches++; @@ -131,26 +135,24 @@ S_do_trans_count(pTHX_ SV *sv) } STATIC I32 -S_do_trans_complex(pTHX_ SV *sv) +S_do_trans_complex(pTHX_ SV * const sv) { - U8 *s; - U8 *send; - U8 *d; - U8 *dstart; - I32 isutf8; + dVAR; + STRLEN len; + U8 *s = (U8*)SvPV_nomg(sv, len); + U8 * const send = s+len; I32 matches = 0; - STRLEN len, rlen = 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__); - s = (U8*)SvPV(sv, len); - isutf8 = SvUTF8(sv); - send = s + len; + if (!SvUTF8(sv)) { + U8 *d = s; + U8 * const dstart = d; - if (!isutf8) { - dstart = d = s; if (PL_op->op_private & OPpTRANS_SQUASH) { const U8* p = send; while (s < send) { @@ -185,10 +187,13 @@ S_do_trans_complex(pTHX_ SV *sv) *d = '\0'; SvCUR_set(sv, d - dstart); } - else { /* isutf8 */ + else { /* is utf8 */ const I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT; const I32 grows = PL_op->op_private & OPpTRANS_GROWS; const I32 del = PL_op->op_private & OPpTRANS_DELETE; + U8 *d; + U8 *dstart; + STRLEN rlen = 0; if (grows) Newx(d, len*2+1, U8); @@ -198,26 +203,23 @@ S_do_trans_complex(pTHX_ SV *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) { STRLEN len; - const UV comp = utf8_to_uvchr(s, &len); + const UV comp = utf8n_to_uvchr(s, send - s, &len, + UTF8_ALLOW_DEFAULT); I32 ch; if (comp > 0xff) { if (!complement) { - Copy(s, d, len, U8); + Move(s, d, len, U8); d += len; } else { matches++; if (!del) { - ch = (rlen == 0) ? comp : + ch = (rlen == 0) ? (I32)comp : (comp - 0x100 < rlen) ? tbl[comp+1] : tbl[0x100+rlen]; if ((UV)ch != pch) { @@ -239,7 +241,7 @@ S_do_trans_complex(pTHX_ SV *sv) continue; } else if (ch == -1) { /* -1 is unmapped character */ - Copy(s, d, len, U8); + Move(s, d, len, U8); d += len; } else if (ch == -2) /* -2 is delete character */ @@ -251,7 +253,8 @@ S_do_trans_complex(pTHX_ SV *sv) else { while (s < send) { STRLEN len; - const UV comp = utf8_to_uvchr(s, &len); + const UV comp = utf8n_to_uvchr(s, send - s, &len, + UTF8_ALLOW_DEFAULT); I32 ch; if (comp > 0xff) { if (!complement) { @@ -273,7 +276,7 @@ S_do_trans_complex(pTHX_ SV *sv) matches++; } else if (ch == -1) { /* -1 is unmapped character */ - Copy(s, d, len, U8); + Move(s, d, len, U8); d += len; } else if (ch == -2) /* -2 is delete character */ @@ -296,8 +299,9 @@ S_do_trans_complex(pTHX_ SV *sv) } STATIC I32 -S_do_trans_simple_utf8(pTHX_ SV *sv) +S_do_trans_simple_utf8(pTHX_ SV * const sv) { + dVAR; U8 *s; U8 *send; U8 *d; @@ -306,19 +310,23 @@ S_do_trans_simple_utf8(pTHX_ SV *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 * svp = hv_fetch(hv, "NONE", 4, FALSE); + 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; - I32 isutf8; U8 hibit = 0; - s = (U8*)SvPV(sv, len); - isutf8 = SvUTF8(sv); - if (!isutf8) { + 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; while (t < e) { @@ -333,7 +341,7 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) send = s + len; start = s; - svp = hv_fetch(hv, "FINAL", 5, FALSE); + svp = hv_fetchs(hv, "FINAL", FALSE); if (svp) final = SvUV(*svp); @@ -396,22 +404,29 @@ S_do_trans_simple_utf8(pTHX_ SV *sv) } STATIC I32 -S_do_trans_count_utf8(pTHX_ SV *sv) +S_do_trans_count_utf8(pTHX_ SV * const sv) { + dVAR; const U8 *s; const U8 *start = NULL; const U8 *send; I32 matches = 0; STRLEN len; - - SV* const rv = (SV*)cSVOP->op_sv; - HV* const hv = (HV*)SvRV(rv); - SV* const * const svp = hv_fetch(hv, "NONE", 4, FALSE); + 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; @@ -439,17 +454,23 @@ S_do_trans_count_utf8(pTHX_ SV *sv) } STATIC I32 -S_do_trans_complex_utf8(pTHX_ SV *sv) +S_do_trans_complex_utf8(pTHX_ SV * const sv) { + dVAR; U8 *start, *send; U8 *d; I32 matches = 0; 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 *svp = hv_fetch(hv, "NONE", 4, FALSE); + 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; @@ -457,10 +478,11 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) STRLEN len; U8 *dstart, *dend; U8 hibit = 0; + U8 *s = (U8*)SvPV_nomg(sv, len); - U8 *s = (U8*)SvPV(sv, len); - const I32 isutf8 = SvUTF8(sv); - if (!isutf8) { + PERL_ARGS_ASSERT_DO_TRANS_COMPLEX_UTF8; + + if (!SvUTF8(sv)) { const U8 *t = s; const U8 * const e = s + len; while (t < e) { @@ -475,7 +497,7 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) send = s + len; start = s; - svp = hv_fetch(hv, "FINAL", 5, FALSE); + svp = hv_fetchs(hv, "FINAL", FALSE); if (svp) { final = SvUV(*svp); havefinal = TRUE; @@ -534,7 +556,7 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) } else { STRLEN len; - uv = utf8_to_uvuni(s, &len); + uv = utf8n_to_uvuni(s, send - s, &len, UTF8_ALLOW_DEFAULT); if (uv != puv) { Move(s, d, len, U8); d += len; @@ -602,15 +624,18 @@ S_do_trans_complex_utf8(pTHX_ SV *sv) I32 Perl_do_trans(pTHX_ SV *sv) { + dVAR; STRLEN len; 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) @@ -648,14 +673,17 @@ Perl_do_trans(pTHX_ SV *sv) } void -Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **sp) +Perl_do_join(pTHX_ register SV *sv, SV *delim, register SV **mark, register SV **sp) { + dVAR; SV ** const oldmark = mark; register I32 items = sp - mark; register STRLEN len; STRLEN delimlen; - (void) SvPV_const(del, delimlen); /* stringify and get the 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. */ mark++; @@ -677,7 +705,7 @@ Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **s ++mark; } - sv_setpvn(sv, "", 0); + sv_setpvs(sv, ""); /* sv_setpv retains old UTF8ness [perl #24846] */ SvUTF8_off(sv); @@ -692,7 +720,7 @@ Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **s if (delimlen) { for (; items > 0; items--,mark++) { - sv_catsv(sv,del); + sv_catsv(sv,delim); sv_catsv(sv,*mark); } } @@ -706,14 +734,25 @@ Perl_do_join(pTHX_ register SV *sv, SV *del, register SV **mark, register SV **s void Perl_do_sprintf(pTHX_ SV *sv, I32 len, SV **sarg) { + dVAR; STRLEN patlen; 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); - sv_vsetpvfn(sv, pat, patlen, Null(va_list*), sarg + 1, len - 1, &do_taint); + sv_vsetpvfn(sv, pat, patlen, NULL, sarg + 1, len - 1, &do_taint); SvSETMAGIC(sv); if (do_taint) SvTAINTED_on(sv); @@ -723,130 +762,137 @@ Perl_do_sprintf(pTHX_ SV *sv, I32 len, SV **sarg) UV Perl_do_vecget(pTHX_ SV *sv, I32 offset, I32 size) { - STRLEN srclen, len; + dVAR; + 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); - offset *= size; /* turn into bit offset */ - len = (offset + 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 { - offset >>= 3; /* turn into byte offset */ if (size == 16) { - if ((STRLEN)offset >= srclen) + if (uoffset >= srclen) retnum = 0; else - retnum = (UV) s[offset] << 8; + retnum = (UV) s[uoffset] << 8; } else if (size == 32) { - if ((STRLEN)offset >= srclen) + if (uoffset >= srclen) retnum = 0; - else if ((STRLEN)(offset + 1) >= srclen) + else if (uoffset + 1 >= srclen) retnum = - ((UV) s[offset ] << 24); - else if ((STRLEN)(offset + 2) >= srclen) + ((UV) s[uoffset ] << 24); + else if (uoffset + 2 >= srclen) retnum = - ((UV) s[offset ] << 24) + - ((UV) s[offset + 1] << 16); + ((UV) s[uoffset ] << 24) + + ((UV) s[uoffset + 1] << 16); else retnum = - ((UV) s[offset ] << 24) + - ((UV) s[offset + 1] << 16) + - ( s[offset + 2] << 8); + ((UV) s[uoffset ] << 24) + + ((UV) s[uoffset + 1] << 16) + + ( s[uoffset + 2] << 8); } #ifdef UV_IS_QUAD else if (size == 64) { - if (ckWARN(WARN_PORTABLE)) - Perl_warner(aTHX_ packWARN(WARN_PORTABLE), - "Bit vector size > 32 non-portable"); - if (offset >= srclen) + Perl_ck_warner(aTHX_ packWARN(WARN_PORTABLE), + "Bit vector size > 32 non-portable"); + if (uoffset >= srclen) retnum = 0; - else if (offset + 1 >= srclen) + else if (uoffset + 1 >= srclen) retnum = - (UV) s[offset ] << 56; - else if (offset + 2 >= srclen) + (UV) s[uoffset ] << 56; + else if (uoffset + 2 >= srclen) retnum = - ((UV) s[offset ] << 56) + - ((UV) s[offset + 1] << 48); - else if (offset + 3 >= srclen) + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48); + else if (uoffset + 3 >= srclen) retnum = - ((UV) s[offset ] << 56) + - ((UV) s[offset + 1] << 48) + - ((UV) s[offset + 2] << 40); - else if (offset + 4 >= srclen) + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48) + + ((UV) s[uoffset + 2] << 40); + else if (uoffset + 4 >= srclen) retnum = - ((UV) s[offset ] << 56) + - ((UV) s[offset + 1] << 48) + - ((UV) s[offset + 2] << 40) + - ((UV) s[offset + 3] << 32); - else if (offset + 5 >= srclen) + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48) + + ((UV) s[uoffset + 2] << 40) + + ((UV) s[uoffset + 3] << 32); + else if (uoffset + 5 >= srclen) retnum = - ((UV) s[offset ] << 56) + - ((UV) s[offset + 1] << 48) + - ((UV) s[offset + 2] << 40) + - ((UV) s[offset + 3] << 32) + - ( s[offset + 4] << 24); - else if (offset + 6 >= srclen) + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48) + + ((UV) s[uoffset + 2] << 40) + + ((UV) s[uoffset + 3] << 32) + + ( s[uoffset + 4] << 24); + else if (uoffset + 6 >= srclen) retnum = - ((UV) s[offset ] << 56) + - ((UV) s[offset + 1] << 48) + - ((UV) s[offset + 2] << 40) + - ((UV) s[offset + 3] << 32) + - ((UV) s[offset + 4] << 24) + - ((UV) s[offset + 5] << 16); + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48) + + ((UV) s[uoffset + 2] << 40) + + ((UV) s[uoffset + 3] << 32) + + ((UV) s[uoffset + 4] << 24) + + ((UV) s[uoffset + 5] << 16); else retnum = - ((UV) s[offset ] << 56) + - ((UV) s[offset + 1] << 48) + - ((UV) s[offset + 2] << 40) + - ((UV) s[offset + 3] << 32) + - ((UV) s[offset + 4] << 24) + - ((UV) s[offset + 5] << 16) + - ( s[offset + 6] << 8); + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48) + + ((UV) s[uoffset + 2] << 40) + + ((UV) s[uoffset + 3] << 32) + + ((UV) s[uoffset + 4] << 24) + + ((UV) s[uoffset + 5] << 16) + + ( s[uoffset + 6] << 8); } #endif } } else if (size < 8) - retnum = (s[offset >> 3] >> (offset & 7)) & ((1 << size) - 1); + retnum = (s[uoffset] >> bitoffs) & ((1 << size) - 1); else { - offset >>= 3; /* turn into byte offset */ if (size == 8) - retnum = s[offset]; + retnum = s[uoffset]; else if (size == 16) retnum = - ((UV) s[offset] << 8) + - s[offset + 1]; + ((UV) s[uoffset] << 8) + + s[uoffset + 1]; else if (size == 32) retnum = - ((UV) s[offset ] << 24) + - ((UV) s[offset + 1] << 16) + - ( s[offset + 2] << 8) + - s[offset + 3]; + ((UV) s[uoffset ] << 24) + + ((UV) s[uoffset + 1] << 16) + + ( s[uoffset + 2] << 8) + + 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[offset ] << 56) + - ((UV) s[offset + 1] << 48) + - ((UV) s[offset + 2] << 40) + - ((UV) s[offset + 3] << 32) + - ((UV) s[offset + 4] << 24) + - ((UV) s[offset + 5] << 16) + - ( s[offset + 6] << 8) + - s[offset + 7]; + ((UV) s[uoffset ] << 56) + + ((UV) s[uoffset + 1] << 48) + + ((UV) s[uoffset + 2] << 40) + + ((UV) s[uoffset + 3] << 32) + + ((UV) s[uoffset + 4] << 24) + + ((UV) s[uoffset + 5] << 16) + + ( s[uoffset + 6] << 8) + + s[uoffset + 7]; } #endif } @@ -861,7 +907,8 @@ Perl_do_vecget(pTHX_ SV *sv, I32 offset, I32 size) void Perl_do_vecset(pTHX_ SV *sv) { - register I32 offset; + dVAR; + register I32 offset, bitoffs = 0; register I32 size; register unsigned char *s; register UV lval; @@ -870,6 +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); @@ -890,8 +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); @@ -900,14 +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) { @@ -922,9 +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); @@ -940,218 +991,9 @@ Perl_do_vecset(pTHX_ SV *sv) } void -Perl_do_chop(pTHX_ register SV *astr, register SV *sv) -{ - 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) -{ - register I32 count; - STRLEN len; - char *s; - char *temp_buffer = NULL; - SV* svrecode = Nullsv; - - 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; #ifdef LIBERAL register long *dl; register long *ll; @@ -1162,50 +1004,77 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) STRLEN rightlen; register const char *lc; register const char *rc; - register I32 len; - I32 lensave; + register STRLEN len; + STRLEN lensave; const char *lsave; const char *rsave; - const bool left_utf = DO_UTF8(left); - const bool right_utf = DO_UTF8(right); - I32 needlen = 0; + bool left_utf; + bool right_utf; + STRLEN needlen = 0; - if (left_utf && !right_utf) - sv_utf8_upgrade(right); - else if (!left_utf && right_utf) - sv_utf8_upgrade(left); + 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 + fired off. */ + + left_utf = DO_UTF8(left); + right_utf = DO_UTF8(right); + + if (left_utf && !right_utf) { + /* Avoid triggering overloading again by using temporaries. + Maybe there should be a variant of sv_utf8_upgrade that takes pvn + */ + 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 = newSVpvn_flags(lsave, leftlen, SVs_TEMP); + sv_utf8_upgrade(left); + lsave = lc = SvPV_nomg_const(left, leftlen); + left_utf = TRUE; + } + len = leftlen < rightlen ? leftlen : rightlen; lensave = len; + SvCUR_set(sv, len); + (void)SvPOK_only(sv); if ((left_utf || right_utf) && (sv == left || sv == right)) { needlen = optype == OP_BIT_AND ? len : leftlen + rightlen; Newxz(dc, needlen + 1, char); } else if (SvOK(sv) || SvTYPE(sv) > SVt_PVMG) { dc = SvPV_force_nomg_nolen(sv); - if (SvLEN(sv) < (STRLEN)(len + 1)) { - dc = SvGROW(sv, (STRLEN)(len + 1)); + if (SvLEN(sv) < len + 1) { + dc = SvGROW(sv, len + 1); (void)memzero(dc + SvCUR(sv), len - SvCUR(sv) + 1); } if (optype != OP_BIT_AND && (left_utf || right_utf)) dc = SvGROW(sv, leftlen + rightlen + 1); } else { - needlen = ((optype == OP_BIT_AND) - ? len : (leftlen > rightlen ? leftlen : rightlen)); + needlen = optype == OP_BIT_AND + ? len : (leftlen > rightlen ? leftlen : rightlen); Newxz(dc, needlen + 1, char); - (void)sv_usepvn(sv, dc, needlen); + sv_usepvn_flags(sv, dc, needlen, SV_HAS_TRAILING_NUL); dc = SvPVX(sv); /* sv_usepvn() calls Renew() */ } - SvCUR_set(sv, len); - (void)SvPOK_only(sv); if (left_utf || right_utf) { UV duc, luc, ruc; - char * const dcsave = dc; + char *dcorig = dc; + char *dcsave = NULL; STRLEN lulen = leftlen; STRLEN rulen = rightlen; STRLEN ulen; @@ -1223,8 +1092,8 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) dc = (char*)uvchr_to_utf8((U8*)dc, duc); } if (sv == left || sv == right) - (void)sv_usepvn(sv, dcsave, needlen); - SvCUR_set(sv, dc - dcsave); + (void)sv_usepvn(sv, dcorig, needlen); + SvCUR_set(sv, dc - dcorig); break; case OP_BIT_XOR: while (lulen && rulen) { @@ -1250,16 +1119,26 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) dc = (char*)uvchr_to_utf8((U8*)dc, duc); } mop_up_utf: + if (rulen) + dcsave = savepvn(rc, rulen); + else if (lulen) + dcsave = savepvn(lc, lulen); if (sv == left || sv == right) - (void)sv_usepvn(sv, dcsave, needlen); - SvCUR_set(sv, dc - dcsave); + (void)sv_usepvn(sv, dcorig, needlen); /* Uses Renew(). */ + SvCUR_set(sv, dc - dcorig); if (rulen) - sv_catpvn(sv, rc, rulen); + sv_catpvn(sv, dcsave, rulen); else if (lulen) - sv_catpvn(sv, lc, lulen); + sv_catpvn(sv, dcsave, lulen); else *SvEND(sv) = '\0'; + Safefree(dcsave); break; + default: + if (sv == left || sv == right) + Safefree(dcorig); + Perl_croak(aTHX_ "panic: do_vop called for op %u (%s)", + (unsigned)optype, PL_op_name[optype]); } SvUTF8_on(sv); goto finish; @@ -1267,11 +1146,11 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) else #ifdef LIBERAL if (len >= sizeof(long)*4 && - !((long)dc % sizeof(long)) && - !((long)lc % sizeof(long)) && - !((long)rc % sizeof(long))) /* It's almost always aligned... */ + !((unsigned long)dc % sizeof(long)) && + !((unsigned long)lc % sizeof(long)) && + !((unsigned long)rc % sizeof(long))) /* It's almost always aligned... */ { - const I32 remainder = len % (sizeof(long)*4); + const STRLEN remainder = len % (sizeof(long)*4); len /= (sizeof(long)*4); dl = (long*)dc; @@ -1327,7 +1206,7 @@ Perl_do_vop(pTHX_ I32 optype, SV *sv, SV *left, SV *right) *dc++ = *lc++ | *rc++; mop_up: len = lensave; - if (rightlen > (STRLEN)len) + if (rightlen > len) sv_catpvn(sv, rsave + len, rightlen - len); else if (leftlen > (STRLEN)len) sv_catpvn(sv, lsave + len, leftlen - len); @@ -1343,61 +1222,46 @@ finish: OP * Perl_do_kv(pTHX) { + dVAR; dSP; - HV * const hv = (HV*)POPs; - HV *keys; + HV * const keys = MUTABLE_HV(POPs); 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 */ - dTARGET; /* make sure to clear its target here */ - if (SvTYPE(TARG) == SVt_PVLV) - LvTARG(TARG) = Nullsv; - PUSHs(TARG); - } - RETURN; - } - - keys = hv; (void)hv_iterinit(keys); /* always reset iterator regardless */ if (gimme == G_VOID) 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, Nullsv, PERL_MAGIC_nkeys, Nullch, 0); - } - LvTYPE(TARG) = 'k'; - if (LvTARG(TARG) != (SV*)keys) { - if (LvTARG(TARG)) - SvREFCNT_dec(LvTARG(TARG)); - LvTARG(TARG) = SvREFCNT_inc(keys); - } - PUSHs(TARG); - RETURN; + 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); } - - if (! SvTIED_mg((SV*)keys, PERL_MAGIC_tied)) - i = HvKEYS(keys); 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))) { @@ -1409,7 +1273,7 @@ Perl_do_kv(pTHX) if (dovalues) { SV *tmpstr; PUTBACK; - tmpstr = hv_iterval(hv,entry); + tmpstr = hv_iterval(keys,entry); DEBUG_H(Perl_sv_setpvf(aTHX_ tmpstr, "%lu%%%d=%lu", (unsigned long)HeHASH(entry), (int)HvMAX(keys)+1,