X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=toke.c;h=cd6ed1d32cde2b8095cbcea0c59bca467a6e8335;hb=ef71a6f50e5f44625432580f1c0df3b478282b95;hp=49dc8b757c5b3fc760b82439bd96e3a99d043a9e;hpb=e1f15930cb6cb2f68ab17efdb7c081df813554a9;p=p5sagit%2Fp5-mst-13.2.git diff --git a/toke.c b/toke.c index 49dc8b7..cd6ed1d 100644 --- a/toke.c +++ b/toke.c @@ -1,6 +1,6 @@ /* toke.c * - * Copyright (c) 1991-1999, Larry Wall + * Copyright (c) 1991-2000, 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. @@ -13,7 +13,7 @@ /* * This file is the lexer for Perl. It's closely linked to the - * parser, perly.y. + * parser, perly.y. * * The main routine is yylex(), which returns the next token. */ @@ -28,27 +28,28 @@ static char ident_too_long[] = "Identifier too long"; static void restore_rsfp(pTHXo_ void *f); -static void restore_expect(pTHXo_ void *e); -static void restore_lex_expect(pTHXo_ void *e); +#ifndef PERL_NO_UTF16_FILTER +static I32 utf16_textfilter(pTHXo_ int idx, SV *sv, int maxlen); +static I32 utf16rev_textfilter(pTHXo_ int idx, SV *sv, int maxlen); +#endif + +#define XFAKEBRACK 128 +#define XENUMMASK 127 +/*#define UTF (SvUTF8(PL_linestr) && !(PL_hints & HINT_BYTE))*/ #define UTF (PL_hints & HINT_UTF8) -/* - * Note: we try to be careful never to call the isXXX_utf8() functions - * unless we're pretty sure we've seen the beginning of a UTF-8 character - * (that is, the two high bits are set). Otherwise we risk loading in the - * heavy-duty SWASHINIT and SWASHGET routines unnecessarily. - */ -#define isIDFIRST_lazy(p) ((!UTF || (*((U8*)p) < 0xc0)) \ - ? isIDFIRST(*(p)) \ - : isIDFIRST_utf8((U8*)p)) -#define isALNUM_lazy(p) ((!UTF || (*((U8*)p) < 0xc0)) \ - ? isALNUM(*(p)) \ - : isALNUM_utf8((U8*)p)) - -/* In variables name $^X, these are the legal values for X. + +/* In variables name $^X, these are the legal values for X. * 1999-02-27 mjd-perl-patch@plover.com */ #define isCONTROLVAR(x) (isUPPER(x) || strchr("[\\]^_?", (x))) +/* On MacOS, respect nonbreaking spaces */ +#ifdef MACOS_TRADITIONAL +#define SPACE_OR_TAB(c) ((c)==' '||(c)=='\312'||(c)=='\t') +#else +#define SPACE_OR_TAB(c) ((c)==' '||(c)=='\t') +#endif + /* LEX_* are values for PL_lex_state, the state of the lexer. * They are arranged oddly so that the guard on the switch statement * can get by with a single comparison (if the compiler is smart enough). @@ -68,33 +69,24 @@ static void restore_lex_expect(pTHXo_ void *e); #define LEX_FORMLINE 1 #define LEX_KNOWNEXT 0 -#ifdef I_FCNTL -#include -#endif -#ifdef I_SYS_FILE -#include -#endif - -/* XXX If this causes problems, set i_unistd=undef in the hint file. */ -#ifdef I_UNISTD -# include /* Needed for execv() */ -#endif - - #ifdef ff_next #undef ff_next #endif #ifdef USE_PURE_BISON -YYSTYPE* yylval_pointer = NULL; -int* yychar_pointer = NULL; +# ifndef YYMAXLEVEL +# define YYMAXLEVEL 100 +# endif +YYSTYPE* yylval_pointer[YYMAXLEVEL]; +int* yychar_pointer[YYMAXLEVEL]; +int yyactlevel = 0; # undef yylval # undef yychar -# define yylval (*yylval_pointer) -# define yychar (*yychar_pointer) -# define PERL_YYLEX_PARAM yylval_pointer,yychar_pointer +# define yylval (*yylval_pointer[yyactlevel]) +# define yychar (*yychar_pointer[yyactlevel]) +# define PERL_YYLEX_PARAM yylval_pointer[yyactlevel],yychar_pointer[yyactlevel] # undef yylex -# define yylex() Perl_yylex(aTHX_ yylval_pointer, yychar_pointer) +# define yylex() Perl_yylex_r(aTHX_ yylval_pointer[yyactlevel],yychar_pointer[yyactlevel]) #endif #include "keywords.h" @@ -104,7 +96,7 @@ int* yychar_pointer = NULL; #ifdef CLINE #undef CLINE #endif -#define CLINE (PL_copline = (PL_curcop->cop_line < PL_copline ? PL_curcop->cop_line : PL_copline)) +#define CLINE (PL_copline = (CopLINE(PL_curcop) < PL_copline ? CopLINE(PL_curcop) : PL_copline)) /* * Convenience functions to return different tokens and prime the @@ -129,7 +121,7 @@ int* yychar_pointer = NULL; * Aop : addition-level operator * Mop : multiplication-level operator * Eop : equality-testing operator - * Rop : relational operator <= != gt + * Rop : relational operator <= != gt * * Also see LOP and lop() below. */ @@ -215,22 +207,22 @@ S_no_op(pTHX_ char *what, char *s) if (!s) s = oldbp; - else { - assert(s >= oldbp); + else PL_bufptr = s; - } yywarn(Perl_form(aTHX_ "%s found where operator expected", what)); if (is_first) Perl_warn(aTHX_ "\t(Missing semicolon on previous line?)\n"); - else if (PL_oldoldbufptr && isIDFIRST_lazy(PL_oldoldbufptr)) { + else if (PL_oldoldbufptr && isIDFIRST_lazy_if(PL_oldoldbufptr,UTF)) { char *t; - for (t = PL_oldoldbufptr; *t && (isALNUM_lazy(t) || *t == ':'); t++) ; + for (t = PL_oldoldbufptr; *t && (isALNUM_lazy_if(t,UTF) || *t == ':'); t++) ; if (t < PL_bufptr && isSPACE(*t)) Perl_warn(aTHX_ "\t(Do you need to predeclare %.*s?)\n", t - PL_oldoldbufptr, PL_oldoldbufptr); } - else + else { + assert(s >= oldbp); Perl_warn(aTHX_ "\t(Missing operator before %.*s?)\n", s - oldbp, oldbp); + } PL_bufptr = oldbp; } @@ -282,7 +274,6 @@ S_missingterm(pTHX_ char *s) void Perl_deprecate(pTHX_ char *s) { - dTHR; if (ckWARN(WARN_DEPRECATED)) Perl_warner(aTHX_ WARN_DEPRECATED, "Use of %s is deprecated", s); } @@ -303,47 +294,38 @@ S_depcom(pTHX) * utf16-to-utf8-reversed. */ -#ifdef WIN32 - -STATIC I32 -S_win32_textfilter(pTHX_ int idx, SV *sv, int maxlen) -{ - I32 count = FILTER_READ(idx+1, sv, maxlen); - if (count > 0 && !maxlen) - win32_strip_return(sv); - return count; -} -#endif - -STATIC I32 -S_utf16_textfilter(pTHX_ int idx, SV *sv, int maxlen) +#ifdef PERL_CR_FILTER +static void +strip_return(SV *sv) { - I32 count = FILTER_READ(idx+1, sv, maxlen); - if (count) { - U8* tmps; - U8* tend; - New(898, tmps, SvCUR(sv) * 3 / 2 + 1, U8); - tend = utf16_to_utf8((U16*)SvPVX(sv), tmps, SvCUR(sv)); - sv_usepvn(sv, (char*)tmps, tend - tmps); - + register char *s = SvPVX(sv); + register char *e = s + SvCUR(sv); + /* outer loop optimized to do nothing if there are no CR-LFs */ + while (s < e) { + if (*s++ == '\r' && *s == '\n') { + /* hit a CR-LF, need to copy the rest */ + register char *d = s - 1; + *d++ = *s++; + while (s < e) { + if (*s == '\r' && s[1] == '\n') + s++; + *d++ = *s++; + } + SvCUR(sv) -= s - d; + return; + } } - return count; } STATIC I32 -S_utf16rev_textfilter(pTHX_ int idx, SV *sv, int maxlen) +S_cr_textfilter(pTHX_ int idx, SV *sv, int maxlen) { I32 count = FILTER_READ(idx+1, sv, maxlen); - if (count) { - U8* tmps; - U8* tend; - New(898, tmps, SvCUR(sv) * 3 / 2 + 1, U8); - tend = utf16_to_utf8_reversed((U16*)SvPVX(sv), tmps, SvCUR(sv)); - sv_usepvn(sv, (char*)tmps, tend - tmps); - - } + if (count > 0 && !maxlen) + strip_return(sv); return count; } +#endif /* * Perl_lex_start @@ -354,19 +336,25 @@ S_utf16rev_textfilter(pTHX_ int idx, SV *sv, int maxlen) void Perl_lex_start(pTHX_ SV *line) { - dTHR; char *s; STRLEN len; SAVEI32(PL_lex_dojoin); SAVEI32(PL_lex_brackets); - SAVEI32(PL_lex_fakebrack); SAVEI32(PL_lex_casemods); SAVEI32(PL_lex_starts); SAVEI32(PL_lex_state); - SAVESPTR(PL_lex_inpat); + SAVEVPTR(PL_lex_inpat); SAVEI32(PL_lex_inwhat); - SAVEI16(PL_curcop->cop_line); + if (PL_lex_state == LEX_KNOWNEXT) { + I32 toke = PL_nexttoke; + while (--toke >= 0) { + SAVEI32(PL_nexttype[toke]); + SAVEVPTR(PL_nextval[toke]); + } + SAVEI32(PL_nexttoke); + } + SAVECOPLINE(PL_curcop); SAVEPPTR(PL_bufptr); SAVEPPTR(PL_bufend); SAVEPPTR(PL_oldbufptr); @@ -380,14 +368,13 @@ Perl_lex_start(pTHX_ SV *line) SAVEI32(PL_lex_defer); SAVEI32(PL_sublex_info.sub_inwhat); SAVESPTR(PL_lex_repl); - SAVEDESTRUCTOR_X(restore_expect, PL_tokenbuf + PL_expect); /* encode as pointer */ - SAVEDESTRUCTOR_X(restore_lex_expect, PL_tokenbuf + PL_expect); + SAVEINT(PL_expect); + SAVEINT(PL_lex_expect); PL_lex_state = LEX_NORMAL; PL_lex_defer = 0; PL_expect = XSTATE; PL_lex_brackets = 0; - PL_lex_fakebrack = 0; New(899, PL_lex_brackstack, 120, char); New(899, PL_lex_casestack, 12, char); SAVEFREEPV(PL_lex_brackstack); @@ -399,6 +386,7 @@ Perl_lex_start(pTHX_ SV *line) PL_lex_stuff = Nullsv; PL_lex_repl = Nullsv; PL_lex_inpat = 0; + PL_nexttoke = 0; PL_lex_inwhat = 0; PL_sublex_info.sub_inwhat = 0; PL_linestr = line; @@ -434,7 +422,7 @@ Perl_lex_end(pTHX) * S_incline * This subroutine has nothing to do with tilting, whether at windmills * or pinball tables. Its name is short for "increment line". It - * increments the current line number in PL_curcop->cop_line and checks + * increments the current line number in CopLINE(PL_curcop) and checks * to see whether the line starts with a comment of the form * # line 500 "foo.pm" * If so, it sets the current line number and file to the values in the comment. @@ -443,42 +431,56 @@ Perl_lex_end(pTHX) STATIC void S_incline(pTHX_ char *s) { - dTHR; char *t; char *n; + char *e; char ch; - int sawline = 0; - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); if (*s++ != '#') return; - while (*s == ' ' || *s == '\t') s++; - if (strnEQ(s, "line ", 5)) { - s += 5; - sawline = 1; - } + while (SPACE_OR_TAB(*s)) s++; + if (strnEQ(s, "line", 4)) + s += 4; + else + return; + if (*s == ' ' || *s == '\t') + s++; + else + return; + while (SPACE_OR_TAB(*s)) s++; if (!isDIGIT(*s)) return; n = s; while (isDIGIT(*s)) s++; - while (*s == ' ' || *s == '\t') + while (SPACE_OR_TAB(*s)) s++; - if (*s == '"' && (t = strchr(s+1, '"'))) + if (*s == '"' && (t = strchr(s+1, '"'))) { s++; + e = t + 1; + } else { - if (!sawline) - return; /* false alarm */ for (t = s; !isSPACE(*t); t++) ; + e = t; } + while (SPACE_OR_TAB(*e) || *e == '\r' || *e == '\f') + e++; + if (*e != '\n' && *e != '\0') + return; /* false alarm */ + ch = *t; *t = '\0'; - if (t - s > 0) - PL_curcop->cop_filegv = gv_fetchfile(s); - else - PL_curcop->cop_filegv = gv_fetchfile(PL_origfilename); + if (t - s > 0) { +#ifdef USE_ITHREADS + Safefree(CopFILE(PL_curcop)); +#else + SvREFCNT_dec(CopFILEGV(PL_curcop)); +#endif + CopFILE_set(PL_curcop, s); + } *t = ch; - PL_curcop->cop_line = atoi(n)-1; + CopLINE_set(PL_curcop, atoi(n)-1); } /* @@ -490,9 +492,8 @@ S_incline(pTHX_ char *s) STATIC char * S_skipspace(pTHX_ register char *s) { - dTHR; if (PL_lex_formbrack && PL_lex_brackets <= PL_lex_formbrack) { - while (s < PL_bufend && (*s == ' ' || *s == '\t')) + while (s < PL_bufend && SPACE_OR_TAB(*s)) s++; return s; } @@ -590,7 +591,7 @@ S_skipspace(pTHX_ register char *s) sv_upgrade(sv, SVt_PVMG); sv_setpvn(sv,PL_bufptr,PL_bufend-PL_bufptr); - av_store(GvAV(PL_curcop->cop_filegv),(I32)PL_curcop->cop_line,sv); + av_store(CopFILEAV(PL_curcop),(I32)CopLINE(PL_curcop),sv); } } } @@ -609,20 +610,19 @@ S_check_uni(pTHX) { char *s; char *t; - dTHR; if (PL_oldoldbufptr != PL_last_uni) return; while (isSPACE(*PL_last_uni)) PL_last_uni++; - for (s = PL_last_uni; isALNUM_lazy(s) || *s == '-'; s++) ; + for (s = PL_last_uni; isALNUM_lazy_if(s,UTF) || *s == '-'; s++) ; if ((t = strchr(s, '(')) && t < PL_bufptr) return; if (ckWARN_d(WARN_AMBIGUOUS)){ char ch = *s; *s = '\0'; - Perl_warner(aTHX_ WARN_AMBIGUOUS, - "Warning: Use of \"%s\" without parens is ambiguous", + Perl_warner(aTHX_ WARN_AMBIGUOUS, + "Warning: Use of \"%s\" without parens is ambiguous", PL_last_uni); *s = ch; } @@ -673,9 +673,8 @@ S_uni(pTHX_ I32 f, char *s) */ STATIC I32 -S_lop(pTHX_ I32 f, expectation x, char *s) +S_lop(pTHX_ I32 f, int x, char *s) { - dTHR; yylval.ival = f; CLINE; PL_expect = x; @@ -702,7 +701,7 @@ S_lop(pTHX_ I32 f, expectation x, char *s) * handles the token correctly. */ -STATIC void +STATIC void S_force_next(pTHX_ I32 type) { PL_nexttype[PL_nexttoke] = type; @@ -735,10 +734,10 @@ S_force_word(pTHX_ register char *start, int token, int check_keyword, int allow { register char *s; STRLEN len; - + start = skipspace(start); s = start; - if (isIDFIRST_lazy(s) || + if (isIDFIRST_lazy_if(s,UTF) || (allow_pack && *s == ':') || (allow_initial_tick && *s == '\'') ) { @@ -777,7 +776,6 @@ S_force_ident(pTHX_ register char *s, int kind) PL_nextval[PL_nexttoke].opval = o; force_next(WORD); if (kind) { - dTHR; /* just for in_eval */ o->op_private = OPpCONST_ENTERED; /* XXX see note in pp_entereval() for why we forgo typo warnings if the symbol must be introduced in an eval. @@ -792,7 +790,32 @@ S_force_ident(pTHX_ register char *s, int kind) } } -/* +NV +Perl_str_to_version(pTHX_ SV *sv) +{ + NV retval = 0.0; + NV nshift = 1.0; + STRLEN len; + char *start = SvPVx(sv,len); + bool utf = SvUTF8(sv) ? TRUE : FALSE; + char *end = start + len; + while (start < end) { + STRLEN skip; + UV n; + if (utf) + n = utf8_to_uv((U8*)start, len, &skip, 0); + else { + n = *(U8*)start; + skip = 1; + } + retval += ((NV)n)/nshift; + start += skip; + nshift *= 1000; + } + return retval; +} + +/* * S_force_version * Forces the next token to be a version number. */ @@ -801,25 +824,31 @@ STATIC char * S_force_version(pTHX_ char *s) { OP *version = Nullop; + char *d; s = skipspace(s); - /* default VERSION number -- GBARR */ - - if(isDIGIT(*s)) { - char *d; - int c; - for( d=s, c = 1; isDIGIT(*d) || *d == '_' || (*d == '.' && c--); d++); - if((*d == ';' || isSPACE(*d)) && *(skipspace(d)) != ',') { - s = scan_num(s); - /* real VERSION number -- GBARR */ + d = s; + if (*d == 'v') + d++; + if (isDIGIT(*d)) { + for (; isDIGIT(*d) || *d == '_' || *d == '.'; d++); + if (*d == ';' || isSPACE(*d) || *d == '}' || !*d) { + SV *ver; + s = scan_num(s, &yylval); version = yylval.opval; + ver = cSVOPx(version)->op_sv; + if (SvPOK(ver) && !SvNIOK(ver)) { + (void)SvUPGRADE(ver, SVt_PVNV); + SvNVX(ver) = str_to_version(ver); + SvNOK_on(ver); /* hint that it is a version */ + } } } /* NOTE: The parser sees the package name and the VERSION swapped */ PL_nextval[PL_nexttoke].opval = version; - force_next(WORD); + force_next(WORD); return (s); } @@ -845,7 +874,7 @@ S_tokeq(pTHX_ SV *sv) goto finish; s = SvPV_force(sv, len); - if (SvIVX(sv) == -1) + if (SvTYPE(sv) >= SVt_PVIV && SvIVX(sv) == -1) goto finish; send = s + len; while (s < send && *s != '\\') @@ -923,9 +952,11 @@ S_sublex_start(pTHX) p = SvPV(sv, len); nsv = newSVpvn(p, len); + if (SvUTF8(sv)) + SvUTF8_on(nsv); SvREFCNT_dec(sv); sv = nsv; - } + } yylval.opval = (OP*)newSVOP(op_type, 0, sv); PL_lex_stuff = Nullsv; return THING; @@ -957,19 +988,17 @@ S_sublex_start(pTHX) STATIC I32 S_sublex_push(pTHX) { - dTHR; ENTER; PL_lex_state = PL_sublex_info.super_state; SAVEI32(PL_lex_dojoin); SAVEI32(PL_lex_brackets); - SAVEI32(PL_lex_fakebrack); SAVEI32(PL_lex_casemods); SAVEI32(PL_lex_starts); SAVEI32(PL_lex_state); - SAVESPTR(PL_lex_inpat); + SAVEVPTR(PL_lex_inpat); SAVEI32(PL_lex_inwhat); - SAVEI16(PL_curcop->cop_line); + SAVECOPLINE(PL_curcop); SAVEPPTR(PL_bufptr); SAVEPPTR(PL_oldbufptr); SAVEPPTR(PL_oldoldbufptr); @@ -988,7 +1017,6 @@ S_sublex_push(pTHX) PL_lex_dojoin = FALSE; PL_lex_brackets = 0; - PL_lex_fakebrack = 0; New(899, PL_lex_brackstack, 120, char); New(899, PL_lex_casestack, 12, char); SAVEFREEPV(PL_lex_brackstack); @@ -997,7 +1025,7 @@ S_sublex_push(pTHX) *PL_lex_casestack = '\0'; PL_lex_starts = 0; PL_lex_state = LEX_INTERPCONCAT; - PL_curcop->cop_line = PL_multi_start; + CopLINE_set(PL_curcop, PL_multi_start); PL_lex_inwhat = PL_sublex_info.sub_inwhat; if (PL_lex_inwhat == OP_MATCH || PL_lex_inwhat == OP_QR || PL_lex_inwhat == OP_SUBST) @@ -1036,7 +1064,6 @@ S_sublex_done(pTHX) SAVEFREESV(PL_linestr); PL_lex_dojoin = FALSE; PL_lex_brackets = 0; - PL_lex_fakebrack = 0; PL_lex_casemods = 0; *PL_lex_casestack = '\0'; PL_lex_starts = 0; @@ -1134,7 +1161,7 @@ S_sublex_done(pTHX) } (end switch) } (end if backslash) } (end while character to read) - + */ STATIC char * @@ -1145,15 +1172,18 @@ S_scan_const(pTHX_ char *start) register char *s = start; /* start of the constant */ register char *d = SvPVX(sv); /* destination for copies */ bool dorange = FALSE; /* are we in a translit range? */ - I32 len; /* ? */ + bool didrange = FALSE; /* did we just finish a range? */ + bool has_utf8 = FALSE; /* embedded \x{} */ + UV uv; + I32 utf = (PL_lex_inwhat == OP_TRANS && PL_sublex_info.sub_op) ? (PL_sublex_info.sub_op->op_private & (OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF)) : UTF; - I32 thisutf = (PL_lex_inwhat == OP_TRANS && PL_sublex_info.sub_op) + I32 this_utf8 = (PL_lex_inwhat == OP_TRANS && PL_sublex_info.sub_op) ? (PL_sublex_info.sub_op->op_private & (PL_lex_repl ? OPpTRANS_FROM_UTF : OPpTRANS_TO_UTF)) : UTF; - char *leaveit = /* set of acceptably-backslashed characters */ + const char *leaveit = /* set of acceptably-backslashed characters */ PL_lex_inpat ? "\\.^$@AGZdDwWsSbBpPXC+*?|()-nrtfeaxcz0123456789[{]} \t\n\r\f\v#" : ""; @@ -1175,6 +1205,12 @@ S_scan_const(pTHX_ char *start) min = (U8)*d; /* first char in range */ max = (U8)d[1]; /* last char in range */ + if (min > max) { + Perl_croak(aTHX_ + "Invalid [] range \"%c-%c\" in transliteration operator", + (char)min, (char)max); + } + #ifndef ASCIIish if ((isLOWER(min) && isLOWER(max)) || (isUPPER(min) && isUPPER(max))) { @@ -1195,11 +1231,15 @@ S_scan_const(pTHX_ char *start) /* mark the range as done, and continue */ dorange = FALSE; + didrange = TRUE; continue; } /* range begins (ignore - as first or last char) */ else if (*s == '-' && s+1 < send && s != start) { + if (didrange) { + Perl_croak(aTHX_ "Ambiguous range in transliteration operator"); + } if (utf) { *d++ = (char)0xff; /* use illegal utf8 byte--see pmtrans */ s++; @@ -1208,6 +1248,9 @@ S_scan_const(pTHX_ char *start) dorange = TRUE; s++; } + else { + didrange = FALSE; + } } /* if we get here, we're not doing a transliteration */ @@ -1218,8 +1261,10 @@ S_scan_const(pTHX_ char *start) if (s[2] == '#') { while (s < send && *s != ')') *d++ = *s++; - } else if (s[2] == '{' - || s[2] == 'p' && s[3] == '{') { /* This should march regcomp.c */ + } + else if (s[2] == '{' /* This should match regcomp.c */ + || ((s[2] == 'p' || s[2] == '?') && s[3] == '{')) + { I32 count = 1; char *regparse = s + (s[2] == '{' ? 3 : 4); char c; @@ -1227,9 +1272,9 @@ S_scan_const(pTHX_ char *start) while (count && (c = *regparse)) { if (c == '\\' && regparse[1]) regparse++; - else if (c == '{') + else if (c == '{') count++; - else if (c == '}') + else if (c == '}') count--; regparse++; } @@ -1249,8 +1294,11 @@ S_scan_const(pTHX_ char *start) *d++ = *s++; } - /* check for embedded arrays (@foo, @:foo, @'foo, @{foo}, @$foo) */ - else if (*s == '@' && s[1] && (isALNUM_lazy(s+1) || strchr(":'{$", s[1]))) + /* check for embedded arrays + (@foo, @:foo, @'foo, @{foo}, @$foo, @+, @-) + */ + else if (*s == '@' && s[1] + && (isALNUM_lazy_if(s+1,UTF) || strchr(":'{$+-", s[1]))) break; /* check for embedded scalars. only stop if we're sure it's a @@ -1265,16 +1313,25 @@ S_scan_const(pTHX_ char *start) /* (now in tr/// code again) */ - if (*s & 0x80 && thisutf) { - dTHR; /* only for ckWARN */ - if (ckWARN(WARN_UTF8)) { - (void)utf8_to_uv((U8*)s, &len); /* could cvt latin-1 to utf8 here... */ - if (len) { - while (len--) - *d++ = *s++; - continue; - } + if (*s & 0x80 && (this_utf8 || has_utf8)) { + STRLEN len = (STRLEN) -1; + UV uv; + if (this_utf8) { + uv = utf8_to_uv((U8*)s, send - s, &len, UTF8_CHECK_ONLY); + } + if (len == (STRLEN)-1) { + /* Illegal UTF8 (a high-bit byte), make it valid. */ + char *old_pvx = SvPVX(sv); + /* need space for one extra char (NOTE: SvCUR() not set here) */ + d = SvGROW(sv, SvLEN(sv) + 1) + (d - old_pvx); + d = (char*)uv_to_utf8((U8*)d, (U8)*s++); } + else { + while (len--) + *d++ = *s++; + } + has_utf8 = TRUE; + continue; } /* backslashes */ @@ -1292,7 +1349,6 @@ S_scan_const(pTHX_ char *start) if (PL_lex_inwhat == OP_SUBST && !PL_lex_inpat && isDIGIT(*s) && *s != '0' && !isDIGIT(s[1])) { - dTHR; /* only for ckWARN */ if (ckWARN(WARN_SYNTAX)) Perl_warner(aTHX_ WARN_SYNTAX, "\\%c better written as $%c", *s, *s); *--s = '$'; @@ -1317,9 +1373,8 @@ S_scan_const(pTHX_ char *start) /* FALL THROUGH */ default: { - dTHR; - if (ckWARN(WARN_UNSAFE) && isALPHA(*s)) - Perl_warner(aTHX_ WARN_UNSAFE, + if (ckWARN(WARN_MISC) && isALNUM(*s)) + Perl_warner(aTHX_ WARN_MISC, "Unrecognized escape \\%c passed through", *s); /* default action is to copy the quoted character */ @@ -1330,49 +1385,85 @@ S_scan_const(pTHX_ char *start) /* \132 indicates an octal constant */ case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': - *d++ = scan_oct(s, 3, &len); - s += len; - continue; + { + STRLEN len = 0; /* disallow underscores */ + uv = (UV)scan_oct(s, 3, &len); + s += len; + } + goto NUM_ESCAPE_INSERT; /* \x24 indicates a hex constant */ case 'x': ++s; if (*s == '{') { char* e = strchr(s, '}'); - if (!e) { yyerror("Missing right brace on \\x{}"); e = s; } - if (!utf) { - dTHR; - if (ckWARN(WARN_UTF8)) - Perl_warner(aTHX_ WARN_UTF8, - "Use of \\x{} without utf8 declaration"); + else { + STRLEN len = 1; /* allow underscores */ + uv = (UV)scan_hex(s + 1, e - s - 1, &len); + has_utf8 = TRUE; } - /* note: utf always shorter than hex */ - d = (char*)uv_to_utf8((U8*)d, - scan_hex(s + 1, e - s - 1, &len)); s = e + 1; } else { - UV uv = (UV)scan_hex(s, 2, &len); - if (utf && PL_lex_inwhat == OP_TRANS && - utf != (OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF)) { - d = (char*)uv_to_utf8((U8*)d, uv); /* doing a CU or UC */ + STRLEN len = 0; /* disallow underscores */ + uv = (UV)scan_hex(s, 2, &len); + s += len; } - else { - if (uv >= 127 && UTF) { - dTHR; - if (ckWARN(WARN_UTF8)) - Perl_warner(aTHX_ WARN_UTF8, - "\\x%.*s will produce malformed UTF-8 character; use \\x{%.*s} for that", - len,s,len,s); + } + + NUM_ESCAPE_INSERT: + /* Insert oct or hex escaped character. + * There will always enough room in sv since such escapes will + * be longer than any utf8 sequence they can end up as + */ + if (uv > 127 || has_utf8) { + if (!this_utf8 && !has_utf8 && uv > 255) { + /* might need to recode whatever we have accumulated so far + * if it contains any hibit chars + */ + int hicount = 0; + char *c; + for (c = SvPVX(sv); c < d; c++) { + if (*c & 0x80) + hicount++; } - *d++ = (char)uv; + if (hicount) { + char *old_pvx = SvPVX(sv); + char *src, *dst; + d = SvGROW(sv, SvCUR(sv) + hicount + 1) + (d - old_pvx); + + src = d - 1; + d += hicount; + dst = d - 1; + + while (src < dst) { + if (*src & 0x80) { + dst--; + uv_to_utf8((U8*)dst, (U8)*src--); + dst--; + } + else { + *dst-- = *src--; + } + } + } + } + + if (has_utf8 || uv > 255) { + d = (char*)uv_to_utf8((U8*)d, uv); + has_utf8 = TRUE; + } + else { + *d++ = (char)uv; } - s += len; + } + else { + *d++ = (char)uv; } continue; @@ -1381,22 +1472,30 @@ S_scan_const(pTHX_ char *start) ++s; if (*s == '{') { char* e = strchr(s, '}'); - HV *hv; - SV **svp; - SV *res, *cv; + SV *res; STRLEN len; char *str; - char *why = Nullch; - + if (!e) { yyerror("Missing right brace on \\N{}"); e = s - 1; goto cont_scan; } res = newSVpvn(s + 1, e - s - 1); - res = new_constant( Nullch, 0, "charnames", + res = new_constant( Nullch, 0, "charnames", res, Nullsv, "\\N{...}" ); str = SvPV(res,len); + if (!has_utf8 && SvUTF8(res)) { + char *ostart = SvPVX(sv); + SvCUR_set(sv, d - ostart); + SvPOK_on(sv); + *d = '\0'; + sv_utf8_upgrade(sv); + /* this just broke our allocation above... */ + SvGROW(sv, send - start); + d = SvPVX(sv) + SvCUR(sv); + has_utf8 = TRUE; + } if (len > e - s + 4) { char *odest = SvPVX(sv); @@ -1420,10 +1519,13 @@ S_scan_const(pTHX_ char *start) *d = *s++; if (isLOWER(*d)) *d = toUPPER(*d); - *d++ = toCTRL(*d); + *d = toCTRL(*d); + d++; #else - len = *s++; - *d++ = toCTRL(len); + { + U8 c = *s++; + *d++ = toCTRL(c); + } #endif continue; @@ -1471,6 +1573,8 @@ S_scan_const(pTHX_ char *start) *d = '\0'; SvCUR_set(sv, d - SvPVX(sv)); SvPOK_on(sv); + if (has_utf8) + SvUTF8_on(sv); /* shrink the sv if we allocated more than we used */ if (SvCUR(sv) + 5 < SvLEN(sv)) { @@ -1481,9 +1585,9 @@ S_scan_const(pTHX_ char *start) /* return the substring (via yylval) only if we parsed anything */ if (s > PL_bufptr) { if ( PL_hints & ( PL_lex_inpat ? HINT_NEW_RE : HINT_NEW_STRING ) ) - sv = new_constant(start, s - start, (PL_lex_inpat ? "qr" : "q"), + sv = new_constant(start, s - start, (PL_lex_inpat ? "qr" : "q"), sv, Nullsv, - ( PL_lex_inwhat == OP_TRANS + ( PL_lex_inwhat == OP_TRANS ? "tr" : ( (PL_lex_inwhat == OP_SUBST && !PL_lex_inpat) ? "s" @@ -1579,7 +1683,7 @@ S_intuit_more(pTHX_ register char *s) case '&': case '$': weight -= seen[un_char] * 10; - if (isALNUM_lazy(s+1)) { + if (isALNUM_lazy_if(s+1,UTF)) { scan_ident(s, send, tmpbuf, sizeof tmpbuf, FALSE); if ((int)strlen(tmpbuf) > 1 && gv_fetchpv(tmpbuf,FALSE, SVt_PV)) weight -= 100; @@ -1663,7 +1767,7 @@ S_intuit_more(pTHX_ register char *s) * Not a method if it's really "print foo $bar" * Method if it's really "foo package::" (interpreted as package->foo) * Not a method if bar is known to be a subroutne ("sub bar; foo bar") - * Not a method if bar is a filehandle or package, but is quotd with + * Not a method if bar is a filehandle or package, but is quoted with * => */ @@ -1754,7 +1858,7 @@ S_incl_perldb(pTHX) /* Encoded script support. filter_add() effectively inserts a - * 'pre-processing' function into the current source input stream. + * 'pre-processing' function into the current source input stream. * Note that the filter function only applies to the current source file * (e.g., it will not affect files 'require'd or 'use'd by this one). * @@ -1764,7 +1868,8 @@ S_incl_perldb(pTHX) * store private buffers and state information. * * The supplied datasv parameter is upgraded to a PVIO type - * and the IoDIRP field is used to store the function pointer. + * and the IoDIRP/IoANY field is used to store the function pointer, + * and IOf_FAKE_DIRP is enabled on datasv to mark this as such. * Note that IoTOP_NAME, IoFMT_NAME, IoBOTTOM_NAME, if set for * private use must be set using malloc'd pointers. */ @@ -1772,42 +1877,38 @@ S_incl_perldb(pTHX) SV * Perl_filter_add(pTHX_ filter_t funcp, SV *datasv) { - if (!funcp){ /* temporary handy debugging hack to be deleted */ - PL_filter_debug = atoi((char*)datasv); - return NULL; - } + if (!funcp) + return Nullsv; + if (!PL_rsfp_filters) PL_rsfp_filters = newAV(); if (!datasv) datasv = NEWSV(255,0); if (!SvUPGRADE(datasv, SVt_PVIO)) Perl_die(aTHX_ "Can't upgrade filter_add data to SVt_PVIO"); - IoDIRP(datasv) = (DIR*)funcp; /* stash funcp into spare field */ -#ifdef DEBUGGING - if (PL_filter_debug) { - STRLEN n_a; - Perl_warn(aTHX_ "filter_add func %p (%s)", funcp, SvPV(datasv, n_a)); - } -#endif /* DEBUGGING */ + IoANY(datasv) = (void *)funcp; /* stash funcp into spare field */ + IoFLAGS(datasv) |= IOf_FAKE_DIRP; + DEBUG_P(PerlIO_printf(Perl_debug_log, "filter_add func %p (%s)\n", + funcp, SvPV_nolen(datasv))); av_unshift(PL_rsfp_filters, 1); av_store(PL_rsfp_filters, 0, datasv) ; return(datasv); } - + /* Delete most recently added instance of this filter function. */ void Perl_filter_del(pTHX_ filter_t funcp) { -#ifdef DEBUGGING - if (PL_filter_debug) - Perl_warn(aTHX_ "filter_del func %p", funcp); -#endif /* DEBUGGING */ + SV *datasv; + DEBUG_P(PerlIO_printf(Perl_debug_log, "filter_del func %p", funcp)); if (!PL_rsfp_filters || AvFILLp(PL_rsfp_filters)<0) return; /* if filter is on top of stack (usual case) just pop it off */ - if (IoDIRP(FILTER_DATA(AvFILLp(PL_rsfp_filters))) == (DIR*)funcp){ - IoDIRP(FILTER_DATA(AvFILLp(PL_rsfp_filters))) = NULL; + datasv = FILTER_DATA(AvFILLp(PL_rsfp_filters)); + if (IoANY(datasv) == (void *)funcp) { + IoFLAGS(datasv) &= ~IOf_FAKE_DIRP; + IoANY(datasv) = (void *)NULL; sv_free(av_pop(PL_rsfp_filters)); return; @@ -1820,8 +1921,8 @@ Perl_filter_del(pTHX_ filter_t funcp) /* Invoke the n'th filter function for the current rsfp. */ I32 Perl_filter_read(pTHX_ int idx, SV *buf_sv, int maxlen) - - + + /* 0 = read one text line */ { filter_t funcp; @@ -1832,11 +1933,9 @@ Perl_filter_read(pTHX_ int idx, SV *buf_sv, int maxlen) if (idx > AvFILLp(PL_rsfp_filters)){ /* Any more filters? */ /* Provide a default input filter to make life easy. */ /* Note that we append to the line. This is handy. */ -#ifdef DEBUGGING - if (PL_filter_debug) - Perl_warn(aTHX_ "filter_read %d: from rsfp\n", idx); -#endif /* DEBUGGING */ - if (maxlen) { + DEBUG_P(PerlIO_printf(Perl_debug_log, + "filter_read %d: from rsfp\n", idx)); + if (maxlen) { /* Want a block */ int len ; int old_len = SvCUR(buf_sv) ; @@ -1863,21 +1962,16 @@ Perl_filter_read(pTHX_ int idx, SV *buf_sv, int maxlen) } /* Skip this filter slot if filter has been deleted */ if ( (datasv = FILTER_DATA(idx)) == &PL_sv_undef){ -#ifdef DEBUGGING - if (PL_filter_debug) - Perl_warn(aTHX_ "filter_read %d: skipped (filter deleted)\n", idx); -#endif /* DEBUGGING */ + DEBUG_P(PerlIO_printf(Perl_debug_log, + "filter_read %d: skipped (filter deleted)\n", + idx)); return FILTER_READ(idx+1, buf_sv, maxlen); /* recurse */ } /* Get function pointer hidden within datasv */ - funcp = (filter_t)IoDIRP(datasv); -#ifdef DEBUGGING - if (PL_filter_debug) { - STRLEN n_a; - Perl_warn(aTHX_ "filter_read %d: via function %p (%s)\n", - idx, funcp, SvPV(datasv,n_a)); - } -#endif /* DEBUGGING */ + funcp = (filter_t)IoANY(datasv); + DEBUG_P(PerlIO_printf(Perl_debug_log, + "filter_read %d: via function %p (%s)\n", + idx, funcp, SvPV_nolen(datasv))); /* Call function. The function is expected to */ /* call "FILTER_READ(idx+1, buf_sv)" first. */ /* Return: <0:error, =0:eof, >0:not eof */ @@ -1887,9 +1981,9 @@ Perl_filter_read(pTHX_ int idx, SV *buf_sv, int maxlen) STATIC char * S_filter_gets(pTHX_ register SV *sv, register PerlIO *fp, STRLEN append) { -#ifdef WIN32FILTER +#ifdef PERL_CR_FILTER if (!PL_rsfp_filters) { - filter_add(win32_textfilter,NULL); + filter_add(S_cr_textfilter,NULL); } #endif if (PL_rsfp_filters) { @@ -1905,6 +1999,31 @@ S_filter_gets(pTHX_ register SV *sv, register PerlIO *fp, STRLEN append) return (sv_gets(sv, fp, append)); } +STATIC HV * +S_find_in_my_stash(pTHX_ char *pkgname, I32 len) +{ + GV *gv; + + if (len == 11 && *pkgname == '_' && strEQ(pkgname, "__PACKAGE__")) + return PL_curstash; + + if (len > 2 && + (pkgname[len - 2] == ':' && pkgname[len - 1] == ':') && + (gv = gv_fetchpv(pkgname, FALSE, SVt_PVHV))) + { + return GvHV(gv); /* Foo:: */ + } + + /* use constant CLASS => 'MyClass' */ + if ((gv = gv_fetchpv(pkgname, FALSE, SVt_PVCV))) { + SV *sv; + if (GvCV(gv) && (sv = cv_const_sv(GvCV(gv)))) { + pkgname = SvPV_nolen(sv); + } + } + + return gv_stashpv(pkgname, FALSE); +} #ifdef DEBUGGING static char* exp_name[] = @@ -1938,6 +2057,33 @@ S_filter_gets(pTHX_ register SV *sv, register PerlIO *fp, STRLEN append) if we already built the token before, use it. */ +#ifdef USE_PURE_BISON +#ifdef __SC__ +#pragma segment Perl_yylex_r +#endif +int +Perl_yylex_r(pTHX_ YYSTYPE *lvalp, int *lcharp) +{ + int r; + + yylval_pointer[yyactlevel] = lvalp; + yychar_pointer[yyactlevel] = lcharp; + yyactlevel++; + if (yyactlevel >= YYMAXLEVEL) + Perl_croak(aTHX_ "panic: YYMAXLEVEL"); + + r = Perl_yylex(aTHX); + + yyactlevel--; + + return r; +} +#endif + +#ifdef __SC__ +#pragma segment Perl_yylex +#endif + int #ifdef USE_PURE_BISON Perl_yylex(pTHX_ YYSTYPE *lvalp, int *lcharp) @@ -1945,7 +2091,6 @@ Perl_yylex(pTHX_ YYSTYPE *lvalp, int *lcharp) Perl_yylex(pTHX) #endif { - dTHR; register char *s; register char *d; register I32 tmp; @@ -1953,17 +2098,15 @@ Perl_yylex(pTHX) GV *gv = Nullgv; GV **gvp = 0; -#ifdef USE_PURE_BISON - yylval_pointer = lvalp; - yychar_pointer = lcharp; -#endif - /* check if there's an identifier for us to look at */ if (PL_pending_ident) { /* pit holds the identifier we read and pending_ident is reset */ char pit = PL_pending_ident; PL_pending_ident = 0; + DEBUG_T({ PerlIO_printf(Perl_debug_log, + "### Tokener saw identifier '%s'\n", PL_tokenbuf); }) + /* if we're in a my(), we can't allow dynamics here. $foo'bar has already been turned into $foo::bar, so just check for colons. @@ -1972,6 +2115,10 @@ Perl_yylex(pTHX) */ if (PL_in_my) { if (PL_in_my == KEY_our) { /* "our" is merely analogous to "my" */ + if (strchr(PL_tokenbuf,':')) + yyerror(Perl_form(aTHX_ "No package name allowed for " + "variable %s in \"our\"", + PL_tokenbuf)); tmp = pad_allocmy(PL_tokenbuf); } else { @@ -1984,7 +2131,7 @@ Perl_yylex(pTHX) } } - /* + /* build the ops for accesses to a my() variable. Deny my($a) or my($b) in a sort block, *if* $a or $b is @@ -2009,15 +2156,19 @@ Perl_yylex(pTHX) } #endif /* USE_THREADS */ if ((tmp = pad_findmy(PL_tokenbuf)) != NOT_IN_PAD) { + SV *namesv = AvARRAY(PL_comppad_name)[tmp]; /* might be an "our" variable" */ - if (SvFLAGS(AvARRAY(PL_comppad_name)[tmp]) & SVpad_OUR) { + if (SvFLAGS(namesv) & SVpad_OUR) { /* build ops for a bareword */ - yylval.opval = (OP*)newSVOP(OP_CONST, 0, newSVpv(PL_tokenbuf+1, 0)); + SV *sym = newSVpv(HvNAME(GvSTASH(namesv)),0); + sv_catpvn(sym, "::", 2); + sv_catpv(sym, PL_tokenbuf+1); + yylval.opval = (OP*)newSVOP(OP_CONST, 0, sym); yylval.opval->op_private = OPpCONST_ENTERED; - gv_fetchpv(PL_tokenbuf+1, + gv_fetchpv(SvPVX(sym), (PL_in_eval - ? (GV_ADDMULTI | GV_ADDINEVAL | GV_ADDOUR) - : GV_ADDOUR + ? (GV_ADDMULTI | GV_ADDINEVAL) + : TRUE ), ((PL_tokenbuf[0] == '$') ? SVt_PV : (PL_tokenbuf[0] == '@') ? SVt_PVAV @@ -2055,9 +2206,14 @@ Perl_yylex(pTHX) */ if (pit == '@' && PL_lex_state != LEX_NORMAL && !PL_lex_brackets) { GV *gv = gv_fetchpv(PL_tokenbuf+1, FALSE, SVt_PVAV); - if (!gv || ((PL_tokenbuf[0] == '@') ? !GvAV(gv) : !GvHV(gv))) - yyerror(Perl_form(aTHX_ "In string, %s now must be written as \\%s", - PL_tokenbuf, PL_tokenbuf)); + if ((!gv || ((PL_tokenbuf[0] == '@') ? !GvAV(gv) : !GvHV(gv))) + && ckWARN(WARN_AMBIGUOUS)) + { + /* Downgraded from fatal to warning 20000522 mjd */ + Perl_warner(aTHX_ WARN_AMBIGUOUS, + "Possible unintended interpolation of %s in string", + PL_tokenbuf); + } } /* build ops for a bareword */ @@ -2088,6 +2244,10 @@ Perl_yylex(pTHX) PL_expect = PL_lex_expect; PL_lex_defer = LEX_NORMAL; } + DEBUG_T({ PerlIO_printf(Perl_debug_log, + "### Next token after '%s' was known, type %"IVdf"\n", PL_bufptr, + (IV)PL_nexttype[PL_nexttoke]); }) + return(PL_nexttype[PL_nexttoke]); /* interpolated case modifiers like \L \U, including \Q and \E. @@ -2119,6 +2279,8 @@ Perl_yylex(pTHX) return yylex(); } else { + DEBUG_T({ PerlIO_printf(Perl_debug_log, + "### Saw case modifier at '%s'\n", PL_bufptr); }) s = PL_bufptr + 1; if (strnEQ(s, "L\\u", 3) || strnEQ(s, "U\\l", 3)) tmp = *s, *s = s[2], s[2] = tmp; /* misordered... */ @@ -2169,6 +2331,8 @@ Perl_yylex(pTHX) case LEX_INTERPSTART: if (PL_bufptr == PL_bufend) return sublex_done(); + DEBUG_T({ PerlIO_printf(Perl_debug_log, + "### Interpolated variable at '%s'\n", PL_bufptr); }) PL_expect = XTERM; PL_lex_dojoin = (*PL_bufptr == '@'); PL_lex_state = LEX_INTERPNORMAL; @@ -2265,7 +2429,7 @@ Perl_yylex(pTHX) s = PL_bufptr; PL_oldoldbufptr = PL_oldbufptr; PL_oldbufptr = s; - DEBUG_p( { + DEBUG_T( { PerlIO_printf(Perl_debug_log, "### Tokener expecting %s at %s\n", exp_name[PL_expect], s); } ) @@ -2273,7 +2437,7 @@ Perl_yylex(pTHX) retry: switch (*s) { default: - if (isIDFIRST_lazy(s)) + if (isIDFIRST_lazy_if(s,UTF)) goto keylookup; Perl_croak(aTHX_ "Unrecognized character \\x%02X", *s & 255); case 4: @@ -2285,6 +2449,9 @@ Perl_yylex(pTHX) PL_last_lop = 0; if (PL_lex_brackets) yyerror("Missing right curly or square bracket"); + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Tokener got EOF\n"); + } ) TOKEN(0); } if (s++ < PL_bufend) @@ -2345,12 +2512,39 @@ Perl_yylex(pTHX) sv_upgrade(sv, SVt_PVMG); sv_setsv(sv,PL_linestr); - av_store(GvAV(PL_curcop->cop_filegv),(I32)PL_curcop->cop_line,sv); + av_store(CopFILEAV(PL_curcop),(I32)CopLINE(PL_curcop),sv); } goto retry; } do { - if ((s = filter_gets(PL_linestr, PL_rsfp, 0)) == Nullch) { + bool bof = PL_rsfp ? TRUE : FALSE; + if (bof) { +#ifdef PERLIO_IS_STDIO +# ifdef __GNU_LIBRARY__ +# if __GNU_LIBRARY__ == 1 /* Linux glibc5 */ +# define FTELL_FOR_PIPE_IS_BROKEN +# endif +# else +# ifdef __GLIBC__ +# if __GLIBC__ == 1 /* maybe some glibc5 release had it like this? */ +# define FTELL_FOR_PIPE_IS_BROKEN +# endif +# endif +# endif +#endif +#ifdef FTELL_FOR_PIPE_IS_BROKEN + /* This loses the possibility to detect the bof + * situation on perl -P when the libc5 is being used. + * Workaround? Maybe attach some extra state to PL_rsfp? + */ + if (!PL_preprocess) + bof = PerlIO_tell(PL_rsfp) == 0; +#else + bof = PerlIO_tell(PL_rsfp) == 0; +#endif + } + s = filter_gets(PL_linestr, PL_rsfp, 0); + if (s == Nullch) { fake_eof: if (PL_rsfp) { if (PL_preprocess && !PL_in_eval) @@ -2373,6 +2567,9 @@ Perl_yylex(pTHX) PL_oldoldbufptr = PL_oldbufptr = s = PL_linestart = SvPVX(PL_linestr); sv_setpv(PL_linestr,""); TOKEN(';'); /* not infinite loop because rsfp is NULL now */ + } else if (bof) { + PL_bufend = SvPVX(PL_linestr) + SvCUR(PL_linestr); + s = swallow_bom((U8*)s); } if (PL_doextract) { if (*s == '#' && s[1] == '!' && instr(s,"perl")) @@ -2394,10 +2591,10 @@ Perl_yylex(pTHX) sv_upgrade(sv, SVt_PVMG); sv_setsv(sv,PL_linestr); - av_store(GvAV(PL_curcop->cop_filegv),(I32)PL_curcop->cop_line,sv); + av_store(CopFILEAV(PL_curcop),(I32)CopLINE(PL_curcop),sv); } PL_bufend = SvPVX(PL_linestr) + SvCUR(PL_linestr); - if (PL_curcop->cop_line == 1) { + if (CopLINE(PL_curcop) == 1) { while (s < PL_bufend && isSPACE(*s)) s++; if (*s == ':' && s[1] != ':') /* for csh execing sh scripts */ @@ -2435,7 +2632,7 @@ Perl_yylex(pTHX) */ SV *x = GvSV(gv_fetchpv("\030", TRUE, SVt_PV)); assert(SvPOK(x) || SvGMAGICAL(x)); - if (sv_eq(x, GvSV(PL_curcop->cop_filegv))) { + if (sv_eq(x, CopFILESV(PL_curcop))) { sv_setpvn(x, ipath, ipathend - ipath); SvSETMAGIC(x); } @@ -2485,6 +2682,7 @@ Perl_yylex(pTHX) *s = '#'; /* Don't try to parse shebang line */ } #endif /* ALTERNATE_SHEBANG */ +#ifndef MACOS_TRADITIONAL if (!d && *s == '#' && ipathend > ipath && @@ -2509,16 +2707,17 @@ Perl_yylex(pTHX) else newargv = PL_origargv; newargv[0] = ipath; - PerlProc_execv(ipath, newargv); + PerlProc_execv(ipath, EXEC_ARGV_CAST(newargv)); Perl_croak(aTHX_ "Can't exec %s", ipath); } +#endif if (d) { U32 oldpdb = PL_perldb; bool oldn = PL_minus_n; bool oldp = PL_minus_p; while (*d && !isSPACE(*d)) d++; - while (*d == ' ' || *d == '\t') d++; + while (SPACE_OR_TAB(*d)) d++; if (*d++ == '-') { do { @@ -2530,8 +2729,8 @@ Perl_yylex(pTHX) } d = moreswitches(d); } while (d); - if (PERLDB_LINE && !oldpdb || - ( PL_minus_n || PL_minus_p ) && !(oldn || oldp) ) + if ((PERLDB_LINE && !oldpdb) || + ((PL_minus_n || PL_minus_p) && !(oldn || oldp))) /* if we have already added "LINE: while (<>) {", we must not do it again */ { @@ -2556,15 +2755,23 @@ Perl_yylex(pTHX) case '\r': #ifdef PERL_STRICT_CR Perl_warn(aTHX_ "Illegal character \\%03o (carriage return)", '\r'); - Perl_croak(aTHX_ - "(Maybe you didn't strip carriage returns after a network transfer?)\n"); + Perl_croak(aTHX_ + "\t(Maybe you didn't strip carriage returns after a network transfer?)\n"); #endif case ' ': case '\t': case '\f': case 013: +#ifdef MACOS_TRADITIONAL + case '\312': +#endif s++; goto retry; case '#': case '\n': if (PL_lex_state != LEX_NORMAL || (PL_in_eval && !PL_rsfp)) { + if (*s == '#' && s == PL_linestart && PL_in_eval && !PL_rsfp) { + /* handle eval qq[#line 1 "foo"\n ...] */ + CopLINE_dec(PL_curcop); + incline(s); + } d = PL_bufend; while (s < d && *s != '\n') s++; @@ -2584,51 +2791,75 @@ Perl_yylex(pTHX) goto retry; case '-': if (s[1] && isALPHA(s[1]) && !isALNUM(s[2])) { + I32 ftst = 0; + s++; PL_bufptr = s; tmp = *s++; - while (s < PL_bufend && (*s == ' ' || *s == '\t')) + while (s < PL_bufend && SPACE_OR_TAB(*s)) s++; if (strnEQ(s,"=>",2)) { s = force_word(PL_bufptr,WORD,FALSE,FALSE,FALSE); + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Saw unary minus before =>, forcing word '%s'\n", s); + } ) OPERATOR('-'); /* unary minus */ } PL_last_uni = PL_oldbufptr; - PL_last_lop_op = OP_FTEREAD; /* good enough */ switch (tmp) { - case 'r': FTST(OP_FTEREAD); - case 'w': FTST(OP_FTEWRITE); - case 'x': FTST(OP_FTEEXEC); - case 'o': FTST(OP_FTEOWNED); - case 'R': FTST(OP_FTRREAD); - case 'W': FTST(OP_FTRWRITE); - case 'X': FTST(OP_FTREXEC); - case 'O': FTST(OP_FTROWNED); - case 'e': FTST(OP_FTIS); - case 'z': FTST(OP_FTZERO); - case 's': FTST(OP_FTSIZE); - case 'f': FTST(OP_FTFILE); - case 'd': FTST(OP_FTDIR); - case 'l': FTST(OP_FTLINK); - case 'p': FTST(OP_FTPIPE); - case 'S': FTST(OP_FTSOCK); - case 'u': FTST(OP_FTSUID); - case 'g': FTST(OP_FTSGID); - case 'k': FTST(OP_FTSVTX); - case 'b': FTST(OP_FTBLK); - case 'c': FTST(OP_FTCHR); - case 't': FTST(OP_FTTTY); - case 'T': FTST(OP_FTTEXT); - case 'B': FTST(OP_FTBINARY); - case 'M': gv_fetchpv("\024",TRUE, SVt_PV); FTST(OP_FTMTIME); - case 'A': gv_fetchpv("\024",TRUE, SVt_PV); FTST(OP_FTATIME); - case 'C': gv_fetchpv("\024",TRUE, SVt_PV); FTST(OP_FTCTIME); + case 'r': ftst = OP_FTEREAD; break; + case 'w': ftst = OP_FTEWRITE; break; + case 'x': ftst = OP_FTEEXEC; break; + case 'o': ftst = OP_FTEOWNED; break; + case 'R': ftst = OP_FTRREAD; break; + case 'W': ftst = OP_FTRWRITE; break; + case 'X': ftst = OP_FTREXEC; break; + case 'O': ftst = OP_FTROWNED; break; + case 'e': ftst = OP_FTIS; break; + case 'z': ftst = OP_FTZERO; break; + case 's': ftst = OP_FTSIZE; break; + case 'f': ftst = OP_FTFILE; break; + case 'd': ftst = OP_FTDIR; break; + case 'l': ftst = OP_FTLINK; break; + case 'p': ftst = OP_FTPIPE; break; + case 'S': ftst = OP_FTSOCK; break; + case 'u': ftst = OP_FTSUID; break; + case 'g': ftst = OP_FTSGID; break; + case 'k': ftst = OP_FTSVTX; break; + case 'b': ftst = OP_FTBLK; break; + case 'c': ftst = OP_FTCHR; break; + case 't': ftst = OP_FTTTY; break; + case 'T': ftst = OP_FTTEXT; break; + case 'B': ftst = OP_FTBINARY; break; + case 'M': case 'A': case 'C': + gv_fetchpv("\024",TRUE, SVt_PV); + switch (tmp) { + case 'M': ftst = OP_FTMTIME; break; + case 'A': ftst = OP_FTATIME; break; + case 'C': ftst = OP_FTCTIME; break; + default: break; + } + break; default: - Perl_croak(aTHX_ "Unrecognized file test: -%c", (int)tmp); break; } + if (ftst) { + PL_last_lop_op = ftst; + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Saw file test %c\n", ftst); + } ) + FTST(ftst); + } + else { + /* Assume it was a minus followed by a one-letter named + * subroutine call (or a -bareword), then. */ + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### %c looked like a file test but was not\n", ftst); + } ) + s -= 2; + } } tmp = *s++; if (*s == tmp) { @@ -2641,7 +2872,7 @@ Perl_yylex(pTHX) else if (*s == '>') { s++; s = skipspace(s); - if (isIDFIRST_lazy(s)) { + if (isIDFIRST_lazy_if(s,UTF)) { s = force_word(s,METHOD,FALSE,TRUE,FALSE); TOKEN(ARROW); } @@ -2737,8 +2968,23 @@ Perl_yylex(pTHX) grabattrs: s = skipspace(s); attrs = Nullop; - while (isIDFIRST_lazy(s)) { + while (isIDFIRST_lazy_if(s,UTF)) { d = scan_word(s, PL_tokenbuf, sizeof PL_tokenbuf, FALSE, &len); + if (isLOWER(*s) && (tmp = keyword(PL_tokenbuf, len))) { + if (tmp < 0) tmp = -tmp; + switch (tmp) { + case KEY_or: + case KEY_and: + case KEY_for: + case KEY_unless: + case KEY_if: + case KEY_while: + case KEY_until: + goto got_attrs; + default: + break; + } + } if (*d == '(') { d = scan_str(d,TRUE,TRUE); if (!d) { @@ -2770,11 +3016,13 @@ Perl_yylex(pTHX) newSVpvn(s, len))); } s = skipspace(d); - while (*s == ',') + if (*s == ':' && s[1] != ':') s = skipspace(s+1); + else if (s == d) + break; /* require real whitespace or :'s */ } - tmp = (PL_expect == XOPERATOR ? '=' : '{'); /*'}' for vi */ - if (*s != ';' && *s != tmp) { + tmp = (PL_expect == XOPERATOR ? '=' : '{'); /*'}(' for vi */ + if (*s != ';' && *s != tmp && (tmp != '=' || *s != ')')) { char q = ((*s == '\'') ? '"' : '\''); /* If here for an expression, and parsed no attrs, back off. */ if (tmp == '=' && !attrs) { @@ -2794,6 +3042,7 @@ Perl_yylex(pTHX) op_free(attrs); OPERATOR(':'); } + got_attrs: if (attrs) { PL_nextval[PL_nexttoke].opval = attrs; force_next(THING); @@ -2809,8 +3058,7 @@ Perl_yylex(pTHX) PL_expect = XTERM; TOKEN('('); case ';': - if (PL_curcop->cop_line < PL_copline) - PL_copline = PL_curcop->cop_line; + CLINE; tmp = *s++; OPERATOR(tmp); case ')': @@ -2854,20 +3102,20 @@ Perl_yylex(pTHX) PL_lex_brackstack[PL_lex_brackets++] = XOPERATOR; OPERATOR(HASHBRACK); case XOPERATOR: - while (s < PL_bufend && (*s == ' ' || *s == '\t')) + while (s < PL_bufend && SPACE_OR_TAB(*s)) s++; d = s; PL_tokenbuf[0] = '\0'; if (d < PL_bufend && *d == '-') { PL_tokenbuf[0] = '-'; d++; - while (d < PL_bufend && (*d == ' ' || *d == '\t')) + while (d < PL_bufend && SPACE_OR_TAB(*d)) d++; } - if (d < PL_bufend && isIDFIRST_lazy(d)) { + if (d < PL_bufend && isIDFIRST_lazy_if(d,UTF)) { d = scan_word(d, PL_tokenbuf + 1, sizeof PL_tokenbuf - 1, FALSE, &len); - while (d < PL_bufend && (*d == ' ' || *d == '\t')) + while (d < PL_bufend && SPACE_OR_TAB(*d)) d++; if (*d == '}') { char minus = (PL_tokenbuf[0] == '-'); @@ -2923,7 +3171,8 @@ Perl_yylex(pTHX) if (++t < PL_bufend && (!isALNUM(*t) || ((*t == 'q' || *t == 'x') && ++t < PL_bufend - && !isALNUM(*t)))) { + && !isALNUM(*t)))) + { char *tmps; char open, close, term; I32 brackets = 1; @@ -2954,8 +3203,10 @@ Perl_yylex(pTHX) } t++; } - else if (isIDFIRST_lazy(s)) { - for (t++; t < PL_bufend && isALNUM_lazy(t); t++) ; + else if (isALNUM_lazy_if(t,UTF)) { + t += UTF8SKIP(t); + while (t < PL_bufend && isALNUM_lazy_if(t,UTF)) + t += UTF8SKIP(t); } while (t < PL_bufend && isSPACE(*t)) t++; @@ -2973,7 +3224,7 @@ Perl_yylex(pTHX) } break; } - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); if (isSPACE(*s) || *s == '#') PL_copline = NOLINE; /* invalidate current command line number */ TOKEN('{'); @@ -2984,11 +3235,12 @@ Perl_yylex(pTHX) yyerror("Unmatched right curly bracket"); else PL_expect = (expectation)PL_lex_brackstack[--PL_lex_brackets]; - if (PL_lex_brackets < PL_lex_formbrack) + if (PL_lex_brackets < PL_lex_formbrack && PL_lex_state != LEX_INTERPNORMAL) PL_lex_formbrack = 0; if (PL_lex_state == LEX_INTERPNORMAL) { if (PL_lex_brackets == 0) { - if (PL_lex_fakebrack) { + if (PL_expect & XFAKEBRACK) { + PL_expect &= XENUMMASK; PL_lex_state = LEX_INTERPEND; PL_bufptr = s; return yylex(); /* ignore fake brackets */ @@ -2999,9 +3251,9 @@ Perl_yylex(pTHX) PL_lex_state = LEX_INTERPEND; } } - if (PL_lex_brackets < PL_lex_fakebrack) { + if (PL_expect & XFAKEBRACK) { + PL_expect &= XENUMMASK; PL_bufptr = s; - PL_lex_fakebrack = 0; return yylex(); /* ignore fake brackets */ } force_next('}'); @@ -3013,10 +3265,12 @@ Perl_yylex(pTHX) AOPERATOR(ANDAND); s--; if (PL_expect == XOPERATOR) { - if (ckWARN(WARN_SEMICOLON) && isIDFIRST_lazy(s) && PL_bufptr == PL_linestart) { - PL_curcop->cop_line--; + if (ckWARN(WARN_SEMICOLON) + && isIDFIRST_lazy_if(s,UTF) && PL_bufptr == PL_linestart) + { + CopLINE_dec(PL_curcop); Perl_warner(aTHX_ WARN_SEMICOLON, PL_warn_nosemi); - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); } BAop(OP_BIT_AND); } @@ -3078,9 +3332,9 @@ Perl_yylex(pTHX) if (PL_lex_brackets < PL_lex_formbrack) { char *t; #ifdef PERL_STRICT_CR - for (t = s; *t == ' ' || *t == '\t'; t++) ; + for (t = s; SPACE_OR_TAB(*t); t++) ; #else - for (t = s; *t == ' ' || *t == '\t' || *t == '\r'; t++) ; + for (t = s; SPACE_OR_TAB(*t) || *t == '\r'; t++) ; #endif if (*t == '\n' || *t == '#') { s--; @@ -3143,7 +3397,7 @@ Perl_yylex(pTHX) } } - if (s[1] == '#' && (isIDFIRST_lazy(s+2) || strchr("{$:+-", s[2]))) { + if (s[1] == '#' && (isIDFIRST_lazy_if(s+2,UTF) || strchr("{$:+-", s[2]))) { PL_tokenbuf[0] = '@'; s = scan_ident(s + 1, PL_bufend, PL_tokenbuf + 1, sizeof PL_tokenbuf - 1, FALSE); @@ -3170,7 +3424,7 @@ Perl_yylex(pTHX) /* This kludge not intended to be bulletproof. */ if (PL_tokenbuf[1] == '[' && !PL_tokenbuf[2]) { yylval.opval = newSVOP(OP_CONST, 0, - newSViv((IV)PL_compiling.cop_arybase)); + newSViv(PL_compiling.cop_arybase)); yylval.opval->op_private = OPpCONST_ARYBASE; TERM(THING); } @@ -3186,7 +3440,7 @@ Perl_yylex(pTHX) PL_tokenbuf[0] = '@'; if (ckWARN(WARN_SYNTAX)) { for(t = s + 1; - isSPACE(*t) || isALNUM_lazy(t) || *t == '$'; + isSPACE(*t) || isALNUM_lazy_if(t,UTF) || *t == '$'; t++) ; if (*t++ == ',') { PL_bufptr = skipspace(PL_bufptr); @@ -3206,7 +3460,7 @@ Perl_yylex(pTHX) char tmpbuf[sizeof PL_tokenbuf]; STRLEN len; for (t++; isSPACE(*t); t++) ; - if (isIDFIRST_lazy(t)) { + if (isIDFIRST_lazy_if(t,UTF)) { t = scan_word(t, tmpbuf, sizeof tmpbuf, TRUE, &len); for (; isSPACE(*t); t++) ; if (*t == ';' && get_cv(tmpbuf, FALSE)) @@ -3224,12 +3478,12 @@ Perl_yylex(pTHX) PL_expect = XOPERATOR; else if (strchr("$@\"'`q", *s)) PL_expect = XTERM; /* e.g. print $fh "foo" */ - else if (strchr("&*<%", *s) && isIDFIRST_lazy(s+1)) + else if (strchr("&*<%", *s) && isIDFIRST_lazy_if(s+1,UTF)) PL_expect = XTERM; /* e.g. print $fh &sub */ - else if (isIDFIRST_lazy(s)) { + else if (isIDFIRST_lazy_if(s,UTF)) { char tmpbuf[sizeof PL_tokenbuf]; scan_word(s, tmpbuf, sizeof tmpbuf, TRUE, &len); - if (tmp = keyword(tmpbuf, len)) { + if ((tmp = keyword(tmpbuf, len))) { /* binary operators exclude handle interpretations */ switch (tmp) { case -KEY_x: @@ -3284,7 +3538,7 @@ Perl_yylex(pTHX) if (ckWARN(WARN_SYNTAX)) { if (*s == '[' || *s == '{') { char *t = s + 1; - while (*t && (isALNUM_lazy(t) || strchr(" \t$#+-'\"", *t))) + while (*t && (isALNUM_lazy_if(t,UTF) || strchr(" \t$#+-'\"", *t))) t++; if (*t == '}' || *t == ']') { t++; @@ -3303,9 +3557,10 @@ Perl_yylex(pTHX) case '?': /* may either be conditional or pattern */ if (PL_expect != XOPERATOR) { /* Disable warning on "study /blah/" */ - if (PL_oldoldbufptr == PL_last_uni - && (*PL_last_uni != 's' || s - PL_last_uni < 5 - || memNE(PL_last_uni, "study", 5) || isALNUM_lazy(PL_last_uni+5))) + if (PL_oldoldbufptr == PL_last_uni + && (*PL_last_uni != 's' || s - PL_last_uni < 5 + || memNE(PL_last_uni, "study", 5) + || isALNUM_lazy_if(PL_last_uni+5,UTF))) check_uni(); s = scan_pat(s,OP_MATCH); TERM(sublex_start()); @@ -3347,13 +3602,19 @@ Perl_yylex(pTHX) /* FALL THROUGH */ case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': - s = scan_num(s); + s = scan_num(s, &yylval); + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Saw number in '%s'\n", s); + } ) if (PL_expect == XOPERATOR) no_op("Number",s); TERM(THING); case '\'': s = scan_str(s,FALSE,FALSE); + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Saw string in '%s'\n", s); + } ) if (PL_expect == XOPERATOR) { if (PL_lex_formbrack && PL_lex_brackets == PL_lex_formbrack) { PL_expect = XTERM; @@ -3370,6 +3631,9 @@ Perl_yylex(pTHX) case '"': s = scan_str(s,FALSE,FALSE); + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Saw string in '%s'\n", s); + } ) if (PL_expect == XOPERATOR) { if (PL_lex_formbrack && PL_lex_brackets == PL_lex_formbrack) { PL_expect = XTERM; @@ -3392,6 +3656,9 @@ Perl_yylex(pTHX) case '`': s = scan_str(s,FALSE,FALSE); + DEBUG_T( { PerlIO_printf(Perl_debug_log, + "### Saw backtick string in '%s'\n", s); + } ) if (PL_expect == XOPERATOR) no_op("Backticks",s); if (!s) @@ -3409,6 +3676,31 @@ Perl_yylex(pTHX) no_op("Backslash",s); OPERATOR(REFGEN); + case 'v': + if (isDIGIT(s[1]) && PL_expect != XOPERATOR) { + char *start = s; + start++; + start++; + while (isDIGIT(*start) || *start == '_') + start++; + if (*start == '.' && isDIGIT(start[1])) { + s = scan_num(s, &yylval); + TERM(THING); + } + /* avoid v123abc() or $h{v1}, allow C */ + else if (!isALPHA(*start) && (PL_expect == XTERM || PL_expect == XREF)) { + char c = *start; + GV *gv; + *start = '\0'; + gv = gv_fetchpv(s, FALSE, SVt_PVCV); + *start = c; + if (!gv) { + s = scan_num(s, &yylval); + TERM(THING); + } + } + } + goto keylookup; case 'x': if (isDIGIT(s[1]) && PL_expect == XOPERATOR) { s++; @@ -3438,14 +3730,13 @@ Perl_yylex(pTHX) case 's': case 'S': case 't': case 'T': case 'u': case 'U': - case 'v': case 'V': + case 'V': case 'w': case 'W': case 'X': case 'y': case 'Y': case 'z': case 'Z': keylookup: { - STRLEN n_a; gv = Nullgv; gvp = 0; @@ -3453,10 +3744,10 @@ Perl_yylex(pTHX) s = scan_word(s, PL_tokenbuf, sizeof PL_tokenbuf, FALSE, &len); /* Some keywords can be followed by any delimiter, including ':' */ - tmp = (len == 1 && strchr("msyq", PL_tokenbuf[0]) || - len == 2 && ((PL_tokenbuf[0] == 't' && PL_tokenbuf[1] == 'r') || - (PL_tokenbuf[0] == 'q' && - strchr("qwxr", PL_tokenbuf[1])))); + tmp = ((len == 1 && strchr("msyq", PL_tokenbuf[0])) || + (len == 2 && ((PL_tokenbuf[0] == 't' && PL_tokenbuf[1] == 'r') || + (PL_tokenbuf[0] == 'q' && + strchr("qwxr", PL_tokenbuf[1]))))); /* x::* is just a word, unless x is "CORE" */ if (!tmp && *s == ':' && s[1] == ':' && strNE(PL_tokenbuf, "CORE")) @@ -3479,7 +3770,7 @@ Perl_yylex(pTHX) tmp = keyword(PL_tokenbuf, len); /* Is this a word before a => operator? */ - if (strnEQ(d,"=>",2)) { + if (*d == '=' && d[1] == '>') { CLINE; yylval.opval = (OP*)newSVOP(OP_CONST, 0, newSVpv(PL_tokenbuf,0)); yylval.opval->op_private = OPpCONST_BARE; @@ -3512,6 +3803,7 @@ Perl_yylex(pTHX) } else if (gv && !gvp && -tmp==KEY_lock /* XXX generalizable kludge */ + && GvCVu(gv) && !hv_fetch(GvHVn(PL_incgv), "Thread.pm", 9, FALSE)) { tmp = 0; /* any sub overrides "weak" keyword */ @@ -3538,7 +3830,7 @@ Perl_yylex(pTHX) /* Get the rest if it looks like a package qualifier */ - if (*s == '\'' || *s == ':' && s[1] == ':') { + if (*s == '\'' || (*s == ':' && s[1] == ':')) { STRLEN morelen; s = scan_word(s, PL_tokenbuf + len, sizeof PL_tokenbuf - len, TRUE, &morelen); @@ -3550,9 +3842,9 @@ Perl_yylex(pTHX) if (PL_expect == XOPERATOR) { if (PL_bufptr == PL_linestart) { - PL_curcop->cop_line--; + CopLINE_dec(PL_curcop); Perl_warner(aTHX_ WARN_SEMICOLON, PL_warn_nosemi); - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); } else no_op("Bareword",s); @@ -3565,8 +3857,8 @@ Perl_yylex(pTHX) if (len > 2 && PL_tokenbuf[len - 2] == ':' && PL_tokenbuf[len - 1] == ':') { - if (ckWARN(WARN_UNSAFE) && ! gv_fetchpv(PL_tokenbuf, FALSE, SVt_PVHV)) - Perl_warner(aTHX_ WARN_UNSAFE, + if (ckWARN(WARN_BAREWORD) && ! gv_fetchpv(PL_tokenbuf, FALSE, SVt_PVHV)) + Perl_warner(aTHX_ WARN_BAREWORD, "Bareword \"%s\" refers to nonexistent package", PL_tokenbuf); len -= 2; @@ -3617,7 +3909,7 @@ Perl_yylex(pTHX) /* Two barewords in a row may indicate method call. */ - if ((isIDFIRST_lazy(s) || *s == '$') && (tmp=intuit_method(s,gv))) + if ((isIDFIRST_lazy_if(s,UTF) || *s == '$') && (tmp=intuit_method(s,gv))) return tmp; /* If not a declared subroutine, it's an indirect object. */ @@ -3633,14 +3925,22 @@ Perl_yylex(pTHX) } } - /* If followed by a paren, it's certainly a subroutine. */ PL_expect = XOPERATOR; s = skipspace(s); + + /* Is this a word before a => operator? */ + if (*s == '=' && s[1] == '>') { + CLINE; + sv_setpv(((SVOP*)yylval.opval)->op_sv, PL_tokenbuf); + TERM(WORD); + } + + /* If followed by a paren, it's certainly a subroutine. */ if (*s == '(') { CLINE; if (gv && GvCVu(gv)) { - for (d = s + 1; *d == ' ' || *d == '\t'; d++) ; + for (d = s + 1; SPACE_OR_TAB(*d); d++) ; if (*d == ')' && (sv = cv_const_sv(GvCV(gv)))) { s = d + 1; goto its_constant; @@ -3663,7 +3963,7 @@ Perl_yylex(pTHX) /* If followed by a bareword, see if it looks like indir obj. */ - if ((isIDFIRST_lazy(s) || *s == '$') && (tmp = intuit_method(s,gv))) + if ((isIDFIRST_lazy_if(s,UTF) || *s == '$') && (tmp = intuit_method(s,gv))) return tmp; /* Not a method, so call it a subroutine (if defined) */ @@ -3739,12 +4039,12 @@ Perl_yylex(pTHX) case KEY___FILE__: yylval.opval = (OP*)newSVOP(OP_CONST, 0, - newSVsv(GvSV(PL_curcop->cop_filegv))); + newSVpv(CopFILE(PL_curcop),0)); TERM(THING); case KEY___LINE__: yylval.opval = (OP*)newSVOP(OP_CONST, 0, - Perl_newSVpvf(aTHX_ "%"IVdf, (IV)PL_curcop->cop_line)); + Perl_newSVpvf(aTHX_ "%"IVdf, (IV)CopLINE(PL_curcop))); TERM(THING); case KEY___PACKAGE__: @@ -3777,11 +4077,33 @@ Perl_yylex(pTHX) /* Mark this internal pseudo-handle as clean */ IoFLAGS(GvIOp(gv)) |= IOf_UNTAINT; if (PL_preprocess) - IoTYPE(GvIOp(gv)) = '|'; + IoTYPE(GvIOp(gv)) = IoTYPE_PIPE; else if ((PerlIO*)PL_rsfp == PerlIO_stdin()) - IoTYPE(GvIOp(gv)) = '-'; + IoTYPE(GvIOp(gv)) = IoTYPE_STD; else - IoTYPE(GvIOp(gv)) = '<'; + IoTYPE(GvIOp(gv)) = IoTYPE_RDONLY; +#if defined(WIN32) && !defined(PERL_TEXTMODE_SCRIPTS) + /* if the script was opened in binmode, we need to revert + * it to text mode for compatibility; but only iff it has CRs + * XXX this is a questionable hack at best. */ + if (PL_bufend-PL_bufptr > 2 + && PL_bufend[-1] == '\n' && PL_bufend[-2] == '\r') + { + Off_t loc = 0; + if (IoTYPE(GvIOp(gv)) == IoTYPE_RDONLY) { + loc = PerlIO_tell(PL_rsfp); + (void)PerlIO_seek(PL_rsfp, 0L, 0); + } + if (PerlLIO_setmode(PerlIO_fileno(PL_rsfp), O_TEXT) != -1) { +#if defined(__BORLANDC__) + /* XXX see note in do_binmode() */ + ((FILE*)PL_rsfp)->flags |= _F_BIN; +#endif + if (loc > 0) + PerlIO_seek(PL_rsfp, loc, 0); + } + } +#endif PL_rsfp = Nullfp; } goto fake_eof; @@ -3790,8 +4112,9 @@ Perl_yylex(pTHX) case KEY_AUTOLOAD: case KEY_DESTROY: case KEY_BEGIN: - case KEY_END: + case KEY_CHECK: case KEY_INIT: + case KEY_END: if (PL_expect == XSTATE) { s = PL_bufptr; goto really_sub; @@ -3803,7 +4126,8 @@ Perl_yylex(pTHX) s += 2; d = s; s = scan_word(s, PL_tokenbuf, sizeof PL_tokenbuf, FALSE, &len); - tmp = keyword(PL_tokenbuf, len); + if (!(tmp = keyword(PL_tokenbuf, len))) + Perl_croak(aTHX_ "CORE::%s is not a keyword", PL_tokenbuf); if (tmp < 0) tmp = -tmp; goto reserved_word; @@ -3829,7 +4153,7 @@ Perl_yylex(pTHX) LOP(OP_BIND,XTERM); case KEY_binmode: - UNI(OP_BINMODE); + LOP(OP_BINMODE,XTERM); case KEY_bless: LOP(OP_BLESS,XTERM); @@ -3858,17 +4182,19 @@ Perl_yylex(pTHX) case KEY_crypt: #ifdef FCRYPT - if (!PL_cryptseen++) + if (!PL_cryptseen) { + PL_cryptseen = TRUE; init_des(); + } #endif LOP(OP_CRYPT,XTERM); case KEY_chmod: - if (ckWARN(WARN_OCTAL)) { + if (ckWARN(WARN_CHMOD)) { for (d = s; d < PL_bufend && (isSPACE(*d) || *d == '('); d++) ; if (*d != '0' && isDIGIT(*d)) - Perl_warner(aTHX_ WARN_OCTAL, - "chmod: mode argument is missing initial 0"); + Perl_warner(aTHX_ WARN_CHMOD, + "chmod() mode argument is missing initial 0"); } LOP(OP_CHMOD,XTERM); @@ -3920,7 +4246,7 @@ Perl_yylex(pTHX) PREBLOCK(ELSE); case KEY_elsif: - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); OPERATOR(ELSIF); case KEY_eq: @@ -3928,7 +4254,7 @@ Perl_yylex(pTHX) case KEY_exists: UNI(OP_EXISTS); - + case KEY_exit: UNI(OP_EXIT); @@ -3970,9 +4296,9 @@ Perl_yylex(pTHX) case KEY_for: case KEY_foreach: - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); s = skipspace(s); - if (PL_expect == XSTATE && isIDFIRST_lazy(s)) { + if (PL_expect == XSTATE && isIDFIRST_lazy_if(s,UTF)) { char *p = s; if ((PL_bufend - p) >= 3 && strnEQ(p, "my", 2) && isSPACE(*(p + 2))) @@ -3981,7 +4307,7 @@ Perl_yylex(pTHX) strnEQ(p, "our", 3) && isSPACE(*(p + 3))) p += 3; p = skipspace(p); - if (isIDFIRST_lazy(p)) { + if (isIDFIRST_lazy_if(p,UTF)) { p = scan_ident(p, PL_bufend, PL_tokenbuf, sizeof PL_tokenbuf, TRUE); p = skipspace(p); @@ -4013,7 +4339,7 @@ Perl_yylex(pTHX) Rop(OP_SGE); case KEY_grep: - LOP(OP_GREPSTART, *s == '(' ? XTERM : XREF); + LOP(OP_GREPSTART, XREF); case KEY_goto: s = force_word(s,WORD,TRUE,FALSE,FALSE); @@ -4108,7 +4434,7 @@ Perl_yylex(pTHX) UNI(OP_HEX); case KEY_if: - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); OPERATOR(IF); case KEY_index: @@ -4132,7 +4458,7 @@ Perl_yylex(pTHX) case KEY_last: s = force_word(s,WORD,TRUE,FALSE,FALSE); LOOPX(OP_LAST); - + case KEY_lc: UNI(OP_LC); @@ -4175,7 +4501,7 @@ Perl_yylex(pTHX) TERM(sublex_start()); case KEY_map: - LOP(OP_MAPSTART, *s == '(' ? XTERM : XREF); + LOP(OP_MAPSTART, XREF); case KEY_mkdir: LOP(OP_MKDIR,XTERM); @@ -4196,11 +4522,11 @@ Perl_yylex(pTHX) case KEY_my: PL_in_my = tmp; s = skipspace(s); - if (isIDFIRST_lazy(s)) { + if (isIDFIRST_lazy_if(s,UTF)) { s = scan_word(s, PL_tokenbuf, sizeof PL_tokenbuf, TRUE, &len); if (len == 3 && strnEQ(PL_tokenbuf, "sub", 3)) goto really_sub; - PL_in_my_stash = gv_stashpv(PL_tokenbuf, FALSE); + PL_in_my_stash = find_in_my_stash(PL_tokenbuf, len); if (!PL_in_my_stash) { char tmpbuf[1024]; PL_bufptr = s; @@ -4234,12 +4560,12 @@ Perl_yylex(pTHX) case KEY_open: s = skipspace(s); - if (isIDFIRST_lazy(s)) { + if (isIDFIRST_lazy_if(s,UTF)) { char *t; - for (d = s; isALNUM_lazy(d); d++) ; + for (d = s; isALNUM_lazy_if(d,UTF); d++) ; t = skipspace(d); - if (strchr("|&*+-=!?:.", *t) && ckWARN_d(WARN_AMBIGUOUS)) - Perl_warner(aTHX_ WARN_AMBIGUOUS, + if (strchr("|&*+-=!?:.", *t) && ckWARN_d(WARN_PRECEDENCE)) + Perl_warner(aTHX_ WARN_PRECEDENCE, "Precedence problem: open %.*s should be open(%.*s)", d-s,s, d-s,s); } @@ -4277,7 +4603,7 @@ Perl_yylex(pTHX) case KEY_pos: UNI(OP_POS); - + case KEY_pack: LOP(OP_PACK,XTERM); @@ -4311,15 +4637,15 @@ Perl_yylex(pTHX) for (; isSPACE(*d) && len; --len, ++d) ; if (len) { char *b = d; - if (!warned && ckWARN(WARN_SYNTAX)) { + if (!warned && ckWARN(WARN_QW)) { for (; !isSPACE(*d) && len; --len, ++d) { if (*d == ',') { - Perl_warner(aTHX_ WARN_SYNTAX, + Perl_warner(aTHX_ WARN_QW, "Possible attempt to separate words with commas"); ++warned; } else if (*d == '#') { - Perl_warner(aTHX_ WARN_SYNTAX, + Perl_warner(aTHX_ WARN_QW, "Possible attempt to put comments in qw() list"); ++warned; } @@ -4329,7 +4655,7 @@ Perl_yylex(pTHX) for (; !isSPACE(*d) && len; --len, ++d) ; } words = append_elem(OP_LIST, words, - newSVOP(OP_CONST, 0, newSVpvn(b, d-b))); + newSVOP(OP_CONST, 0, tokeq(newSVpvn(b, d-b)))); } } if (words) { @@ -4368,12 +4694,18 @@ Perl_yylex(pTHX) OLDLOP(OP_RETURN); case KEY_require: - *PL_tokenbuf = '\0'; - s = force_word(s,WORD,TRUE,TRUE,FALSE); - if (isIDFIRST_lazy(PL_tokenbuf)) - gv_stashpvn(PL_tokenbuf, strlen(PL_tokenbuf), TRUE); - else if (*s == '<') - yyerror("<> should be quotes"); + s = skipspace(s); + if (isDIGIT(*s) || (*s == 'v' && isDIGIT(s[1]))) { + s = force_version(s); + } + else { + *PL_tokenbuf = '\0'; + s = force_word(s,WORD,TRUE,TRUE,FALSE); + if (isIDFIRST_lazy_if(PL_tokenbuf,UTF)) + gv_stashpvn(PL_tokenbuf, strlen(PL_tokenbuf), TRUE); + else if (*s == '<') + yyerror("<> should be quotes"); + } UNI(OP_REQUIRE); case KEY_reset: @@ -4433,7 +4765,7 @@ Perl_yylex(pTHX) case KEY_chomp: UNI(OP_CHOMP); - + case KEY_scalar: UNI(OP_SCALAR); @@ -4543,7 +4875,6 @@ Perl_yylex(pTHX) UNI(OP_STAT); case KEY_study: - PL_sawstudy++; UNI(OP_STUDY); case KEY_substr: @@ -4561,7 +4892,7 @@ Perl_yylex(pTHX) s = skipspace(s); - if (isIDFIRST_lazy(s) || *s == '\'' || + if (isIDFIRST_lazy_if(s,UTF) || *s == '\'' || (*s == ':' && s[1] == ':')) { PL_expect = XBLOCK; @@ -4701,11 +5032,11 @@ Perl_yylex(pTHX) UNI(OP_UNTIE); case KEY_until: - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); OPERATOR(UNTIL); case KEY_unless: - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); OPERATOR(UNLESS); case KEY_unlink: @@ -4721,10 +5052,10 @@ Perl_yylex(pTHX) LOP(OP_UTIME,XTERM); case KEY_umask: - if (ckWARN(WARN_OCTAL)) { + if (ckWARN(WARN_UMASK)) { for (d = s; d < PL_bufend && (isSPACE(*d) || *d == '('); d++) ; - if (*d != '0' && isDIGIT(*d)) - Perl_warner(aTHX_ WARN_OCTAL, + if (*d != '0' && isDIGIT(*d)) + Perl_warner(aTHX_ WARN_UMASK, "umask: argument is missing initial 0"); } UNI(OP_UMASK); @@ -4736,9 +5067,9 @@ Perl_yylex(pTHX) if (PL_expect != XSTATE) yyerror("\"use\" not allowed in expression"); s = skipspace(s); - if(isDIGIT(*s)) { + if (isDIGIT(*s) || (*s == 'v' && isDIGIT(s[1]))) { s = force_version(s); - if(*s == ';' || (s = skipspace(s), *s == ';')) { + if (*s == ';' || (s = skipspace(s), *s == ';')) { PL_nextval[PL_nexttoke].opval = Nullop; force_next(WORD); } @@ -4754,11 +5085,10 @@ Perl_yylex(pTHX) UNI(OP_VALUES); case KEY_vec: - PL_sawvec = TRUE; LOP(OP_VEC,XTERM); case KEY_while: - yylval.ival = PL_curcop->cop_line; + yylval.ival = CopLINE(PL_curcop); OPERATOR(WHILE); case KEY_warn: @@ -4779,7 +5109,7 @@ Perl_yylex(pTHX) { static char ctl_l[2]; - if (ctl_l[0] == '\0') + if (ctl_l[0] == '\0') ctl_l[0] = toCTRL('L'); gv_fetchpv(ctl_l,TRUE, SVt_PV); } @@ -4804,6 +5134,9 @@ Perl_yylex(pTHX) } }} } +#ifdef __SC__ +#pragma segment Main +#endif I32 Perl_keyword(pTHX_ register char *d, I32 len) @@ -4846,6 +5179,7 @@ Perl_keyword(pTHX_ register char *d, I32 len) break; case 'C': if (strEQ(d,"CORE")) return -KEY_CORE; + if (strEQ(d,"CHECK")) return KEY_CHECK; break; case 'c': switch (len) { @@ -4855,12 +5189,12 @@ Perl_keyword(pTHX_ register char *d, I32 len) if (strEQ(d,"cos")) return -KEY_cos; break; case 4: - if (strEQ(d,"chop")) return KEY_chop; + if (strEQ(d,"chop")) return -KEY_chop; break; case 5: if (strEQ(d,"close")) return -KEY_close; if (strEQ(d,"chdir")) return -KEY_chdir; - if (strEQ(d,"chomp")) return KEY_chomp; + if (strEQ(d,"chomp")) return -KEY_chomp; if (strEQ(d,"chmod")) return -KEY_chmod; if (strEQ(d,"chown")) return -KEY_chown; if (strEQ(d,"crypt")) return -KEY_crypt; @@ -4905,7 +5239,6 @@ Perl_keyword(pTHX_ register char *d, I32 len) } break; case 'E': - if (strEQ(d,"EQ")) { deprecate(d); return -KEY_eq;} if (strEQ(d,"END")) return KEY_END; break; case 'e': @@ -4922,7 +5255,7 @@ Perl_keyword(pTHX_ register char *d, I32 len) if (strEQ(d,"exit")) return -KEY_exit; if (strEQ(d,"eval")) return KEY_eval; if (strEQ(d,"exec")) return -KEY_exec; - if (strEQ(d,"each")) return KEY_each; + if (strEQ(d,"each")) return -KEY_each; break; case 5: if (strEQ(d,"elsif")) return KEY_elsif; @@ -4971,12 +5304,6 @@ Perl_keyword(pTHX_ register char *d, I32 len) break; } break; - case 'G': - if (len == 2) { - if (strEQ(d,"GT")) { deprecate(d); return -KEY_gt;} - if (strEQ(d,"GE")) { deprecate(d); return -KEY_ge;} - } - break; case 'g': if (strnEQ(d,"get",3)) { d += 3; @@ -5072,16 +5399,10 @@ Perl_keyword(pTHX_ register char *d, I32 len) break; case 'k': if (len == 4) { - if (strEQ(d,"keys")) return KEY_keys; + if (strEQ(d,"keys")) return -KEY_keys; if (strEQ(d,"kill")) return -KEY_kill; } break; - case 'L': - if (len == 2) { - if (strEQ(d,"LT")) { deprecate(d); return -KEY_lt;} - if (strEQ(d,"LE")) { deprecate(d); return -KEY_le;} - } - break; case 'l': switch (len) { case 2: @@ -5133,9 +5454,6 @@ Perl_keyword(pTHX_ register char *d, I32 len) break; } break; - case 'N': - if (strEQ(d,"NE")) { deprecate(d); return -KEY_ne;} - break; case 'n': if (strEQ(d,"next")) return KEY_next; if (strEQ(d,"ne")) return -KEY_ne; @@ -5163,11 +5481,11 @@ Perl_keyword(pTHX_ register char *d, I32 len) case 'p': switch (len) { case 3: - if (strEQ(d,"pop")) return KEY_pop; + if (strEQ(d,"pop")) return -KEY_pop; if (strEQ(d,"pos")) return KEY_pos; break; case 4: - if (strEQ(d,"push")) return KEY_push; + if (strEQ(d,"push")) return -KEY_push; if (strEQ(d,"pack")) return -KEY_pack; if (strEQ(d,"pipe")) return -KEY_pipe; break; @@ -5274,7 +5592,7 @@ Perl_keyword(pTHX_ register char *d, I32 len) case 'h': switch (len) { case 5: - if (strEQ(d,"shift")) return KEY_shift; + if (strEQ(d,"shift")) return -KEY_shift; break; case 6: if (strEQ(d,"shmctl")) return -KEY_shmctl; @@ -5303,7 +5621,7 @@ Perl_keyword(pTHX_ register char *d, I32 len) case 'p': if (strEQ(d,"split")) return KEY_split; if (strEQ(d,"sprintf")) return -KEY_sprintf; - if (strEQ(d,"splice")) return KEY_splice; + if (strEQ(d,"splice")) return -KEY_splice; break; case 'q': if (strEQ(d,"sqrt")) return -KEY_sqrt; @@ -5383,7 +5701,7 @@ Perl_keyword(pTHX_ register char *d, I32 len) if (strEQ(d,"unlink")) return -KEY_unlink; break; case 7: - if (strEQ(d,"unshift")) return KEY_unshift; + if (strEQ(d,"unshift")) return -KEY_unshift; if (strEQ(d,"ucfirst")) return -KEY_ucfirst; break; } @@ -5429,7 +5747,6 @@ S_checkcomma(pTHX_ register char *s, char *name, char *what) char *w; if (*s == ' ' && s[1] == '(') { /* XXX gotta be a better way */ - dTHR; /* only for ckWARN */ if (ckWARN(WARN_SYNTAX)) { int level = 1; for (w = s+2; *w && level; w++) { @@ -5451,9 +5768,9 @@ S_checkcomma(pTHX_ register char *s, char *name, char *what) s++; while (s < PL_bufend && isSPACE(*s)) s++; - if (isIDFIRST_lazy(s)) { + if (isIDFIRST_lazy_if(s,UTF)) { w = s++; - while (isALNUM_lazy(s)) + while (isALNUM_lazy_if(s,UTF)) s++; while (s < PL_bufend && isSPACE(*s)) s++; @@ -5475,37 +5792,45 @@ S_checkcomma(pTHX_ register char *s, char *name, char *what) and type is used with error messages only. */ STATIC SV * -S_new_constant(pTHX_ char *s, STRLEN len, char *key, SV *sv, SV *pv, char *type) +S_new_constant(pTHX_ char *s, STRLEN len, const char *key, SV *sv, SV *pv, + const char *type) { dSP; HV *table = GvHV(PL_hintgv); /* ^H */ SV *res; SV **cvp; SV *cv, *typesv; - char *why, *why1, *why2; - - if (!(PL_hints & HINT_LOCALIZE_HH)) { + const char *why1, *why2, *why3; + + if (!table || !(PL_hints & HINT_LOCALIZE_HH)) { SV *msg; - why = "%^H is not localized"; - report_short: - why1 = why2 = ""; + why2 = strEQ(key,"charnames") + ? "(possibly a missing \"use charnames ...\")" + : ""; + msg = Perl_newSVpvf(aTHX_ "Constant(%s) unknown: %s", + (type ? type: "undef"), why2); + + /* This is convoluted and evil ("goto considered harmful") + * but I do not understand the intricacies of all the different + * failure modes of %^H in here. The goal here is to make + * the most probable error message user-friendly. --jhi */ + + goto msgdone; + report: - msg = Perl_newSVpvf(aTHX_ "constant(%s): %s%s%s", - (type ? type: "undef"), why1, why2, why); + msg = Perl_newSVpvf(aTHX_ "Constant(%s): %s%s%s", + (type ? type: "undef"), why1, why2, why3); + msgdone: yyerror(SvPVX(msg)); SvREFCNT_dec(msg); return sv; } - if (!table) { - why = "%^H is not defined"; - goto report_short; - } cvp = hv_fetch(table, key, strlen(key), FALSE); if (!cvp || !SvOK(*cvp)) { - why = "} is not defined"; why1 = "$^H{"; why2 = key; + why3 = "} is not defined"; goto report; } sv_2mortal(sv); /* Parent created it permanently */ @@ -5516,27 +5841,25 @@ S_new_constant(pTHX_ char *s, STRLEN len, char *key, SV *sv, SV *pv, char *type) typesv = sv_2mortal(newSVpv(type, 0)); else typesv = &PL_sv_undef; - + PUSHSTACKi(PERLSI_OVERLOAD); ENTER ; SAVETMPS; - + PUSHMARK(SP) ; - EXTEND(sp, 4); + EXTEND(sp, 3); if (pv) PUSHs(pv); PUSHs(sv); if (pv) PUSHs(typesv); - PUSHs(cv); PUTBACK; call_sv(cv, G_SCALAR | ( PL_in_eval ? 0 : G_EVAL)); - + SPAGAIN ; - + /* Check the eval first */ - if (!PL_in_eval && SvTRUE(ERRSV)) - { + if (!PL_in_eval && SvTRUE(ERRSV)) { STRLEN n_a; sv_catpv(ERRSV, "Propagated"); yyerror(SvPV(ERRSV, n_a)); /* Duplicates the message inside eval */ @@ -5547,23 +5870,23 @@ S_new_constant(pTHX_ char *s, STRLEN len, char *key, SV *sv, SV *pv, char *type) res = POPs; (void)SvREFCNT_inc(res); } - + PUTBACK ; FREETMPS ; LEAVE ; POPSTACK; - + if (!SvOK(res)) { - why = "}} did not return a defined value"; why1 = "Call to &{$^H{"; why2 = key; + why3 = "}} did not return a defined value"; sv = res; goto report; - } + } - return res; + return res; } - + STATIC char * S_scan_word(pTHX_ register char *s, char *dest, STRLEN destlen, int allow_package, STRLEN *slp) { @@ -5574,7 +5897,7 @@ S_scan_word(pTHX_ register char *s, char *dest, STRLEN destlen, int allow_packag Perl_croak(aTHX_ ident_too_long); if (isALNUM(*s)) /* UTF handled below */ *d++ = *s++; - else if (*s == '\'' && allow_package && isIDFIRST_lazy(s+1)) { + else if (*s == '\'' && allow_package && isIDFIRST_lazy_if(s+1,UTF)) { *d++ = ':'; *d++ = ':'; s++; @@ -5609,8 +5932,6 @@ S_scan_ident(pTHX_ register char *s, register char *send, char *dest, STRLEN des char *bracket = 0; char funny = *s++; - if (PL_lex_brackets == 0) - PL_lex_fakebrack = 0; if (isSPACE(*s)) s = skipspace(s); d = dest; @@ -5628,7 +5949,7 @@ S_scan_ident(pTHX_ register char *s, register char *send, char *dest, STRLEN des Perl_croak(aTHX_ ident_too_long); if (isALNUM(*s)) /* UTF handled below */ *d++ = *s++; - else if (*s == '\'' && isIDFIRST_lazy(s+1)) { + else if (*s == '\'' && isIDFIRST_lazy_if(s+1,UTF)) { *d++ = ':'; *d++ = ':'; s++; @@ -5659,7 +5980,7 @@ S_scan_ident(pTHX_ register char *s, register char *send, char *dest, STRLEN des return s; } if (*s == '$' && s[1] && - (isALNUM_lazy(s+1) || strchr("${", s[1]) || strnEQ(s+1,"::",2)) ) + (isALNUM_lazy_if(s+1,UTF) || strchr("${", s[1]) || strnEQ(s+1,"::",2)) ) { return s; } @@ -5680,17 +6001,17 @@ S_scan_ident(pTHX_ register char *s, register char *send, char *dest, STRLEN des if (isSPACE(s[-1])) { while (s < send) { char ch = *s++; - if (ch != ' ' && ch != '\t') { + if (!SPACE_OR_TAB(ch)) { *d = ch; break; } } } - if (isIDFIRST_lazy(d)) { + if (isIDFIRST_lazy_if(d,UTF)) { d++; if (UTF) { e = s; - while (e < send && isALNUM_lazy(e) || *e == ':') { + while ((e < send && isALNUM_lazy_if(e,UTF)) || *e == ':') { e += UTF8SKIP(e); while (e < send && *e & 0x80 && is_utf8_mark((U8*)e)) e += UTF8SKIP(e); @@ -5706,22 +6027,20 @@ S_scan_ident(pTHX_ register char *s, register char *send, char *dest, STRLEN des Perl_croak(aTHX_ ident_too_long); } *d = '\0'; - while (s < send && (*s == ' ' || *s == '\t')) s++; + while (s < send && SPACE_OR_TAB(*s)) s++; if ((*s == '[' || (*s == '{' && strNE(dest, "sub")))) { - dTHR; /* only for ckWARN */ if (ckWARN(WARN_AMBIGUOUS) && keyword(dest, d - dest)) { - char *brack = *s == '[' ? "[...]" : "{...}"; + const char *brack = *s == '[' ? "[...]" : "{...}"; Perl_warner(aTHX_ WARN_AMBIGUOUS, "Ambiguous use of %c{%s%s} resolved to %c%s%s", funny, dest, brack, funny, dest, brack); } - PL_lex_fakebrack = PL_lex_brackets+1; bracket++; - PL_lex_brackstack[PL_lex_brackets++] = XOPERATOR; + PL_lex_brackstack[PL_lex_brackets++] = (char)(XOPERATOR | XFAKEBRACK); return s; } - } - /* Handle extended ${^Foo} variables + } + /* Handle extended ${^Foo} variables * 1999-02-27 mjd-perl-patch@plover.com */ else if (!isALNUM(*d) && !isPRINT(*d) /* isCTRL(d) */ && isALNUM(*s)) @@ -5741,7 +6060,6 @@ S_scan_ident(pTHX_ register char *s, register char *send, char *dest, STRLEN des if (funny == '#') funny = '@'; if (PL_lex_state == LEX_NORMAL) { - dTHR; /* only for ckWARN */ if (ckWARN(WARN_AMBIGUOUS) && (keyword(dest, d - dest) || get_cv(dest, FALSE))) { @@ -5917,45 +6235,20 @@ S_scan_trans(pTHX_ char *start) Perl_croak(aTHX_ "Transliteration replacement not terminated"); } - if (UTF) { - o = newSVOP(OP_TRANS, 0, 0); - utf8 = OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF; - } - else { - New(803,tbl,256,short); - o = newPVOP(OP_TRANS, 0, (char*)tbl); - utf8 = 0; - } + New(803,tbl,256,short); + o = newPVOP(OP_TRANS, 0, (char*)tbl); complement = del = squash = 0; - while (strchr("cdsCU", *s)) { + while (strchr("cds", *s)) { if (*s == 'c') complement = OPpTRANS_COMPLEMENT; else if (*s == 'd') del = OPpTRANS_DELETE; else if (*s == 's') squash = OPpTRANS_SQUASH; - else { - switch (count++) { - case 0: - if (*s == 'C') - utf8 &= ~OPpTRANS_FROM_UTF; - else - utf8 |= OPpTRANS_FROM_UTF; - break; - case 1: - if (*s == 'C') - utf8 &= ~OPpTRANS_TO_UTF; - else - utf8 |= OPpTRANS_TO_UTF; - break; - default: - Perl_croak(aTHX_ "Too many /C and /U options"); - } - } s++; } - o->op_private = del|squash|complement|utf8; + o->op_private = del|squash|complement; PL_lex_op = o; yylval.ival = OP_TRANS; @@ -5965,7 +6258,6 @@ S_scan_trans(pTHX_ char *start) STATIC char * S_scan_heredoc(pTHX_ register char *s) { - dTHR; SV *herewas; I32 op_type = OP_SCALAR; I32 len; @@ -5981,7 +6273,7 @@ S_scan_heredoc(pTHX_ register char *s) e = PL_tokenbuf + sizeof PL_tokenbuf - 1; if (!outer) *d++ = '\n'; - for (peek = s; *peek == ' ' || *peek == '\t'; peek++) ; + for (peek = s; SPACE_OR_TAB(*peek); peek++) ; if (*peek && strchr("`'\"",*peek)) { s = peek; term = *s++; @@ -5995,9 +6287,9 @@ S_scan_heredoc(pTHX_ register char *s) s++, term = '\''; else term = '"'; - if (!isALNUM_lazy(s)) + if (!isALNUM_lazy_if(s,UTF)) deprecate("bare << to mean <<\"\""); - for (; isALNUM_lazy(s); s++) { + for (; isALNUM_lazy_if(s,UTF); s++) { if (d < e) *d++ = *s; } @@ -6050,7 +6342,7 @@ S_scan_heredoc(pTHX_ register char *s) } CLINE; - PL_multi_start = PL_curcop->cop_line; + PL_multi_start = CopLINE(PL_curcop); PL_multi_open = PL_multi_close = '<'; term = *PL_tokenbuf; if (PL_lex_inwhat == OP_SUBST && PL_in_eval && !PL_rsfp) { @@ -6064,10 +6356,10 @@ S_scan_heredoc(pTHX_ register char *s) while (s < bufend && (*s != term || memNE(s,PL_tokenbuf,len)) ) { if (*s++ == '\n') - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); } if (s >= bufend) { - PL_curcop->cop_line = PL_multi_start; + CopLINE_set(PL_curcop, PL_multi_start); missingterm(PL_tokenbuf); } sv_setpvn(herewas,bufptr,d-bufptr+1); @@ -6084,15 +6376,15 @@ S_scan_heredoc(pTHX_ register char *s) while (s < PL_bufend && (*s != term || memNE(s,PL_tokenbuf,len)) ) { if (*s++ == '\n') - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); } if (s >= PL_bufend) { - PL_curcop->cop_line = PL_multi_start; + CopLINE_set(PL_curcop, PL_multi_start); missingterm(PL_tokenbuf); } sv_setpvn(tmpstr,d+1,s-d); s += len - 1; - PL_curcop->cop_line++; /* the preceding stmt passes a newline */ + CopLINE_inc(PL_curcop); /* the preceding stmt passes a newline */ sv_catpvn(herewas,s,PL_bufend-s); sv_setsv(PL_linestr,herewas); @@ -6104,10 +6396,10 @@ S_scan_heredoc(pTHX_ register char *s) while (s >= PL_bufend) { /* multiple line string? */ if (!outer || !(PL_oldoldbufptr = PL_oldbufptr = s = PL_linestart = filter_gets(PL_linestr, PL_rsfp, 0))) { - PL_curcop->cop_line = PL_multi_start; + CopLINE_set(PL_curcop, PL_multi_start); missingterm(PL_tokenbuf); } - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); PL_bufend = SvPVX(PL_linestr) + SvCUR(PL_linestr); #ifndef PERL_STRICT_CR if (PL_bufend - PL_linestart >= 2) { @@ -6129,8 +6421,7 @@ S_scan_heredoc(pTHX_ register char *s) sv_upgrade(sv, SVt_PVMG); sv_setsv(sv,PL_linestr); - av_store(GvAV(PL_curcop->cop_filegv), - (I32)PL_curcop->cop_line,sv); + av_store(CopFILEAV(PL_curcop), (I32)CopLINE(PL_curcop),sv); } if (*s == term && memEQ(s,PL_tokenbuf,len)) { s = PL_bufend - 1; @@ -6145,7 +6436,7 @@ S_scan_heredoc(pTHX_ register char *s) } s++; retval: - PL_multi_end = PL_curcop->cop_line; + PL_multi_end = CopLINE(PL_curcop); if (SvCUR(tmpstr) + 5 < SvLEN(tmpstr)) { SvLEN_set(tmpstr, SvCUR(tmpstr) + 1); Renew(SvPVX(tmpstr), SvLEN(tmpstr), char); @@ -6209,7 +6500,7 @@ S_scan_inputsymbol(pTHX_ char *start) if (*d == '$' && d[1]) d++; /* allow or */ - while (*d && (isALNUM_lazy(d) || *d == '\'' || *d == ':')) + while (*d && (isALNUM_lazy_if(d,UTF) || *d == '\'' || *d == ':')) d++; /* If we've tried to read what we allow filehandles to look like, and @@ -6301,7 +6592,7 @@ S_scan_inputsymbol(pTHX_ char *start) calls scan_str(). s/// makes yylex() call scan_subst() which calls scan_str(). tr/// and y/// make yylex() call scan_trans() which calls scan_str(). - + It skips whitespace before the string starts, and treats the first character as the delimiter. If the delimiter is one of ([{< then the corresponding "close" character )]}> is used as the closing @@ -6318,13 +6609,13 @@ S_scan_inputsymbol(pTHX_ char *start) STATIC char * S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) { - dTHR; SV *sv; /* scalar value: string */ char *tmps; /* temp string, used for delimiter matching */ register char *s = start; /* current position in the buffer */ register char term; /* terminating character */ register char *to; /* current position in the sv's data */ I32 brackets = 1; /* bracket nesting level */ + bool has_utf8 = FALSE; /* is there any utf8 content? */ /* skip space before the delimiter */ if (isSPACE(*s)) @@ -6335,8 +6626,11 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) /* after skipping whitespace, the next character is the terminator */ term = *s; + if ((term & 0x80) && UTF) + has_utf8 = TRUE; + /* mark where we are */ - PL_multi_start = PL_curcop->cop_line; + PL_multi_start = CopLINE(PL_curcop); PL_multi_open = term; /* find corresponding closing delimiter */ @@ -6366,7 +6660,7 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) for (; s < PL_bufend; s++,to++) { /* embedded newlines increment the current line number */ if (*s == '\n' && !PL_rsfp) - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); /* handle quoted delimiters */ if (*s == '\\' && s+1 < PL_bufend && term != '\\') { if (!keep_quoted && s[1] == term) @@ -6379,6 +6673,8 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) have found the terminator */ else if (*s == term) break; + else if (!has_utf8 && (*s & 0x80) && UTF) + has_utf8 = TRUE; *to = *s; } } @@ -6392,7 +6688,7 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) for (; s < PL_bufend; s++,to++) { /* embedded newlines increment the line count */ if (*s == '\n' && !PL_rsfp) - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); /* backslashes can escape the open or closing characters */ if (*s == '\\' && s+1 < PL_bufend) { if (!keep_quoted && @@ -6406,6 +6702,8 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) break; else if (*s == PL_multi_open) brackets++; + else if (!has_utf8 && (*s & 0x80) && UTF) + has_utf8 = TRUE; *to = *s; } } @@ -6417,7 +6715,8 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) * this next chunk reads more into the buffer if we're not done yet */ - if (s < PL_bufend) break; /* handle case where we are done yet :-) */ + if (s < PL_bufend) + break; /* handle case where we are done yet :-) */ #ifndef PERL_STRICT_CR if (to - SvPVX(sv) >= 2) { @@ -6441,11 +6740,11 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) if (!PL_rsfp || !(PL_oldoldbufptr = PL_oldbufptr = s = PL_linestart = filter_gets(PL_linestr, PL_rsfp, 0))) { sv_free(sv); - PL_curcop->cop_line = PL_multi_start; + CopLINE_set(PL_curcop, PL_multi_start); return Nullch; } /* we read a line, so increment our line counter */ - PL_curcop->cop_line++; + CopLINE_inc(PL_curcop); /* update debugger info */ if (PERLDB_LINE && PL_curstash != PL_debstash) { @@ -6453,19 +6752,20 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) sv_upgrade(sv, SVt_PVMG); sv_setsv(sv,PL_linestr); - av_store(GvAV(PL_curcop->cop_filegv), - (I32)PL_curcop->cop_line, sv); + av_store(CopFILEAV(PL_curcop), (I32)CopLINE(PL_curcop), sv); } /* having changed the buffer, we must update PL_bufend */ PL_bufend = SvPVX(PL_linestr) + SvCUR(PL_linestr); } - + /* at this point, we have successfully read the delimited string */ if (keep_delims) sv_catpvn(sv, s, 1); - PL_multi_end = PL_curcop->cop_line; + if (has_utf8) + SvUTF8_on(sv); + PL_multi_end = CopLINE(PL_curcop); s++; /* if we allocated too much space, give some back */ @@ -6477,7 +6777,7 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) /* decide whether this is the first or second quoted string we've read for this op */ - + if (PL_lex_stuff) PL_lex_repl = sv; else @@ -6506,16 +6806,15 @@ S_scan_str(pTHX_ char *start, int keep_quoted, int keep_delims) try converting the number to an integer and see if it can do so without loss of precision. */ - + char * -Perl_scan_num(pTHX_ char *start) +Perl_scan_num(pTHX_ char *start, YYSTYPE* lvalp) { register char *s = start; /* current position in buffer */ register char *d; /* destination in temp buffer */ register char *e; /* end of temp buffer */ - IV tryiv; /* used to see if it can be an IV */ - NV value; /* number read, as a double */ - SV *sv; /* place to put the converted number */ + NV nv; /* number read, as a double */ + SV *sv = Nullsv; /* place to put the converted number */ bool floatit; /* boolean: int or float? */ char *lastub = 0; /* position of last underbar */ static char number_too_long[] = "Number too long"; @@ -6525,10 +6824,9 @@ Perl_scan_num(pTHX_ char *start) switch (*s) { default: Perl_croak(aTHX_ "panic: scan_num"); - + /* if it starts with a 0, it could be an octal number, a decimal in - 0.13 disguise, or a hexadecimal number, or a binary number. - */ + 0.13 disguise, or a hexadecimal number, or a binary number. */ case '0': { /* variables: @@ -6541,7 +6839,6 @@ Perl_scan_num(pTHX_ char *start) we in octal/hex/binary?" indicator to disallow hex characters when in octal mode. */ - dTHR; NV n = 0.0; UV u = 0; I32 shift; @@ -6629,7 +6926,6 @@ Perl_scan_num(pTHX_ char *start) if ((x >> shift) != u && !(PL_hints & HINT_NEW_BINARY)) { - dTHR; overflowed = TRUE; n = (NV) u; if (ckWARN_d(WARN_OVERFLOW)) @@ -6661,7 +6957,6 @@ Perl_scan_num(pTHX_ char *start) out: sv = NEWSV(92,0); if (overflowed) { - dTHR; if (ckWARN(WARN_PORTABLE) && n > 4294967295.0) Perl_warner(aTHX_ WARN_PORTABLE, "%s number > %s non-portable", @@ -6670,7 +6965,6 @@ Perl_scan_num(pTHX_ char *start) } else { #if UVSIZE > 4 - dTHR; if (ckWARN(WARN_PORTABLE) && u > 0xffffffff) Perl_warner(aTHX_ WARN_PORTABLE, "%s number > %s non-portable", @@ -6696,11 +6990,10 @@ Perl_scan_num(pTHX_ char *start) /* read next group of digits and _ and copy into d */ while (isDIGIT(*s) || *s == '_') { - /* skip underscores, checking for misplaced ones + /* skip underscores, checking for misplaced ones if -w is on */ if (*s == '_') { - dTHR; /* only for ckWARN */ if (ckWARN(WARN_SYNTAX) && lastub && s - lastub != 3) Perl_warner(aTHX_ WARN_SYNTAX, "Misplaced _ in number"); lastub = ++s; @@ -6716,7 +7009,6 @@ Perl_scan_num(pTHX_ char *start) /* final misplaced underbar check */ if (lastub && s - lastub != 3) { - dTHR; if (ckWARN(WARN_SYNTAX)) Perl_warner(aTHX_ WARN_SYNTAX, "Misplaced _ in number"); } @@ -6739,6 +7031,11 @@ Perl_scan_num(pTHX_ char *start) if (*s != '_') *d++ = *s; } + if (*s == '.' && isDIGIT(s[1])) { + /* oops, it's really a v-string, but without the "v" */ + s = start - 1; + goto vstring; + } } /* read exponent part, if present */ @@ -6767,34 +7064,160 @@ Perl_scan_num(pTHX_ char *start) /* make an sv from the string */ sv = NEWSV(92,0); - value = Atof(PL_tokenbuf); +#if defined(Strtol) && defined(Strtoul) - /* - See if we can make do with an integer value without loss of - precision. We use I_V to cast to an int, because some - compilers have issues. Then we try casting it back and see - if it was the same. We only do this if we know we - specifically read an integer. + /* + strtol/strtoll sets errno to ERANGE if the number is too big + for an integer. We try to do an integer conversion first + if no characters indicating "float" have been found. + */ - Note: if floatit is true, then we don't need to do the - conversion at all. - */ - tryiv = I_V(value); - if (!floatit && (NV)tryiv == value) - sv_setiv(sv, tryiv); - else - sv_setnv(sv, value); + if (!floatit) { + IV iv; + UV uv; + errno = 0; + if (*PL_tokenbuf == '-') + iv = Strtol(PL_tokenbuf, (char**)NULL, 10); + else + uv = Strtoul(PL_tokenbuf, (char**)NULL, 10); + if (errno) + floatit = TRUE; /* Probably just too large. */ + else if (*PL_tokenbuf == '-') + sv_setiv(sv, iv); + else if (uv <= IV_MAX) + sv_setiv(sv, uv); /* Prefer IVs over UVs. */ + else + sv_setuv(sv, uv); + } + if (floatit) { + nv = Atof(PL_tokenbuf); + sv_setnv(sv, nv); + } +#else + /* + No working strtou?ll?. + + Unfortunately atol() doesn't do range checks (returning + LONG_MIN/LONG_MAX, and setting errno to ERANGE on overflows) + everywhere [1], so we cannot use use atol() (or atoll()). + If we could, they would be used, as Atol(), very much like + Strtol() and Strtoul() are used above. + + [1] XXX Configure test needed to check for atol() + (and atoll()) overflow behaviour XXX + + --jhi + + We need to do this the hard way. */ + + nv = Atof(PL_tokenbuf); + + /* See if we can make do with an integer value without loss of + precision. We use U_V to cast to a UV, because some + compilers have issues. Then we try casting it back and see + if it was the same [1]. We only do this if we know we + specifically read an integer. If floatit is true, then we + don't need to do the conversion at all. + + [1] Note that this is lossy if our NVs cannot preserve our + UVs. There are metaconfig defines NV_PRESERVES_UV (a boolean) + and NV_PRESERVES_UV_BITS (a number), but in general we really + do hope all such potentially lossy platforms have strtou?ll? + to do a lossless IV/UV conversion. + + Maybe could do some tricks with DBL_DIG, LDBL_DIG and + DBL_MANT_DIG and LDBL_MANT_DIG (these are already available + as NV_DIG and NV_MANT_DIG)? + + --jhi + */ + { + UV uv = U_V(nv); + if (!floatit && (NV)uv == nv) { + if (uv <= IV_MAX) + sv_setiv(sv, uv); /* Prefer IVs over UVs. */ + else + sv_setuv(sv, uv); + } + else + sv_setnv(sv, nv); + } +#endif if ( floatit ? (PL_hints & HINT_NEW_FLOAT) : (PL_hints & HINT_NEW_INTEGER) ) - sv = new_constant(PL_tokenbuf, d - PL_tokenbuf, + sv = new_constant(PL_tokenbuf, d - PL_tokenbuf, (floatit ? "float" : "integer"), sv, Nullsv, NULL); break; + + /* if it starts with a v, it could be a v-string */ + case 'v': +vstring: + { + char *pos = s; + pos++; + while (isDIGIT(*pos) || *pos == '_') + pos++; + if (!isALPHA(*pos)) { + UV rev; + U8 tmpbuf[UTF8_MAXLEN+1]; + U8 *tmpend; + bool utf8 = FALSE; + s++; /* get past 'v' */ + + sv = NEWSV(92,5); + sv_setpvn(sv, "", 0); + + for (;;) { + if (*s == '0' && isDIGIT(s[1])) + yyerror("Octal number in vector unsupported"); + rev = 0; + { + /* this is atoi() that tolerates underscores */ + char *end = pos; + UV mult = 1; + while (--end >= s) { + UV orev; + if (*end == '_') + continue; + orev = rev; + rev += (*end - '0') * mult; + mult *= 10; + if (orev > rev && ckWARN_d(WARN_OVERFLOW)) + Perl_warner(aTHX_ WARN_OVERFLOW, + "Integer overflow in decimal number"); + } + } + tmpend = uv_to_utf8(tmpbuf, rev); + utf8 = utf8 || rev > 127; + sv_catpvn(sv, (const char*)tmpbuf, tmpend - tmpbuf); + if (*pos == '.' && isDIGIT(pos[1])) + s = ++pos; + else { + s = pos; + break; + } + while (isDIGIT(*pos) || *pos == '_') + pos++; + } + + SvPOK_on(sv); + SvREADONLY_on(sv); + if (utf8) { + SvUTF8_on(sv); + sv_utf8_downgrade(sv, TRUE); + } + } + } + break; } /* make the op for the constant and return */ - yylval.opval = newSVOP(OP_CONST, 0, sv); + if (sv) + lvalp->opval = newSVOP(OP_CONST, 0, sv); + else + lvalp->opval = Nullop; return s; } @@ -6802,19 +7225,18 @@ Perl_scan_num(pTHX_ char *start) STATIC char * S_scan_formline(pTHX_ register char *s) { - dTHR; register char *eol; register char *t; SV *stuff = newSVpvn("",0); bool needargs = FALSE; while (!needargs) { - if (*s == '.' || *s == '}') { + if (*s == '.' || *s == /*{*/'}') { /*SUPPRESS 530*/ #ifdef PERL_STRICT_CR - for (t = s+1;*t == ' ' || *t == '\t'; t++) ; + for (t = s+1;SPACE_OR_TAB(*t); t++) ; #else - for (t = s+1;*t == ' ' || *t == '\t' || *t == '\r'; t++) ; + for (t = s+1;SPACE_OR_TAB(*t) || *t == '\r'; t++) ; #endif if (*t == '\n' || t == PL_bufend) break; @@ -6836,6 +7258,14 @@ S_scan_formline(pTHX_ register char *s) needargs = TRUE; } sv_catpvn(stuff, s, eol-s); +#ifndef PERL_STRICT_CR + if (eol-s > 1 && eol[-2] == '\r' && eol[-1] == '\n') { + char *end = SvPVX(stuff) + SvCUR(stuff); + end[-2] = '\n'; + end[-1] = '\0'; + SvCUR(stuff)--; + } +#endif } s = eol; if (PL_rsfp) { @@ -6885,7 +7315,6 @@ S_set_csh(pTHX) I32 Perl_start_subparse(pTHX_ I32 is_format, U32 flags) { - dTHR; I32 oldsavestack_ix = PL_savestack_ix; CV* outsidecv = PL_compcv; AV* comppadlist; @@ -6893,11 +7322,10 @@ Perl_start_subparse(pTHX_ I32 is_format, U32 flags) if (PL_compcv) { assert(SvTYPE(PL_compcv) == SVt_PVCV); } - save_I32(&PL_subline); + SAVEI32(PL_subline); save_item(PL_subname); SAVEI32(PL_padix); - SAVESPTR(PL_curpad); - SAVESPTR(PL_comppad); + SAVECOMPPAD(); SAVESPTR(PL_comppad_name); SAVESPTR(PL_compcv); SAVEI32(PL_comppad_name_fill); @@ -6916,7 +7344,7 @@ Perl_start_subparse(pTHX_ I32 is_format, U32 flags) PL_comppad_name_fill = 0; PL_min_intro_pending = 0; PL_padix = 0; - PL_subline = PL_curcop->cop_line; + PL_subline = CopLINE(PL_curcop); #ifdef USE_THREADS av_store(PL_comppad_name, 0, newSVpvn("@_", 2)); PL_curpad[0] = (SV*)newAV(); @@ -6942,7 +7370,6 @@ Perl_start_subparse(pTHX_ I32 is_format, U32 flags) int Perl_yywarn(pTHX_ char *s) { - dTHR; PL_in_eval |= EVAL_WARNONLY; yyerror(s); PL_in_eval &= ~EVAL_WARNONLY; @@ -6952,7 +7379,6 @@ Perl_yywarn(pTHX_ char *s) int Perl_yyerror(pTHX_ char *s) { - dTHR; char *where = NULL; char *context = NULL; int contlen = -1; @@ -6976,7 +7402,12 @@ Perl_yyerror(pTHX_ char *s) } else if (yychar > 255) where = "next token ???"; +#ifdef USE_PURE_BISON +/* GNU Bison sets the value -2 */ + else if (yychar == -2) { +#else else if ((yychar & 127) == 127) { +#endif if (PL_lex_state == LEX_NORMAL || (PL_lex_state == LEX_KNOWNEXT && PL_lex_defer == LEX_NORMAL)) where = "at end of line"; @@ -6996,32 +7427,111 @@ Perl_yyerror(pTHX_ char *s) where = SvPVX(where_sv); } msg = sv_2mortal(newSVpv(s, 0)); - Perl_sv_catpvf(aTHX_ msg, " at %_ line %"IVdf", ", - GvSV(PL_curcop->cop_filegv), (IV)PL_curcop->cop_line); + Perl_sv_catpvf(aTHX_ msg, " at %s line %"IVdf", ", + CopFILE(PL_curcop), (IV)CopLINE(PL_curcop)); if (context) Perl_sv_catpvf(aTHX_ msg, "near \"%.*s\"\n", contlen, context); else Perl_sv_catpvf(aTHX_ msg, "%s\n", where); - if (PL_multi_start < PL_multi_end && (U32)(PL_curcop->cop_line - PL_multi_end) <= 1) { + if (PL_multi_start < PL_multi_end && (U32)(CopLINE(PL_curcop) - PL_multi_end) <= 1) { Perl_sv_catpvf(aTHX_ msg, " (Might be a runaway multi-line %c%c string starting on line %"IVdf")\n", (int)PL_multi_open,(int)PL_multi_close,(IV)PL_multi_start); PL_multi_end = 0; } if (PL_in_eval & EVAL_WARNONLY) - Perl_warn(aTHX_ "%_", msg); + Perl_warn(aTHX_ "%"SVf, msg); else qerror(msg); - if (PL_error_count >= 10) - Perl_croak(aTHX_ "%_ has too many errors.\n", GvSV(PL_curcop->cop_filegv)); + if (PL_error_count >= 10) { + if (PL_in_eval && SvCUR(ERRSV)) + Perl_croak(aTHX_ "%"SVf"%s has too many errors.\n", + ERRSV, CopFILE(PL_curcop)); + else + Perl_croak(aTHX_ "%s has too many errors.\n", + CopFILE(PL_curcop)); + } PL_in_my = 0; PL_in_my_stash = Nullhv; return 0; } +STATIC char* +S_swallow_bom(pTHX_ U8 *s) +{ + STRLEN slen; + slen = SvCUR(PL_linestr); + switch (*s) { + case 0xFF: + if (s[1] == 0xFE) { + /* UTF-16 little-endian */ + if (s[2] == 0 && s[3] == 0) /* UTF-32 little-endian */ + Perl_croak(aTHX_ "Unsupported script encoding"); +#ifndef PERL_NO_UTF16_FILTER + DEBUG_p(PerlIO_printf(Perl_debug_log, "UTF-LE script encoding\n")); + s += 2; + if (PL_bufend > (char*)s) { + U8 *news; + I32 newlen; + + filter_add(utf16rev_textfilter, NULL); + New(898, news, (PL_bufend - (char*)s) * 3 / 2 + 1, U8); + PL_bufend = (char*)utf16_to_utf8_reversed(s, news, + PL_bufend - (char*)s - 1, + &newlen); + Copy(news, s, newlen, U8); + SvCUR_set(PL_linestr, newlen); + PL_bufend = SvPVX(PL_linestr) + newlen; + news[newlen++] = '\0'; + Safefree(news); + } +#else + Perl_croak(aTHX_ "Unsupported script encoding"); +#endif + } + break; + case 0xFE: + if (s[1] == 0xFF) { /* UTF-16 big-endian */ +#ifndef PERL_NO_UTF16_FILTER + DEBUG_p(PerlIO_printf(Perl_debug_log, "UTF-16BE script encoding\n")); + s += 2; + if (PL_bufend > (char *)s) { + U8 *news; + I32 newlen; + + filter_add(utf16_textfilter, NULL); + New(898, news, (PL_bufend - (char*)s) * 3 / 2 + 1, U8); + PL_bufend = (char*)utf16_to_utf8(s, news, + PL_bufend - (char*)s, + &newlen); + Copy(news, s, newlen, U8); + SvCUR_set(PL_linestr, newlen); + PL_bufend = SvPVX(PL_linestr) + newlen; + news[newlen++] = '\0'; + Safefree(news); + } +#else + Perl_croak(aTHX_ "Unsupported script encoding"); +#endif + } + break; + case 0xEF: + if (slen > 2 && s[1] == 0xBB && s[2] == 0xBF) { + DEBUG_p(PerlIO_printf(Perl_debug_log, "UTF-8 script encoding\n")); + s += 3; /* UTF-8 */ + } + break; + case 0: + if (slen > 3 && s[1] == 0 && /* UTF-32 big-endian */ + s[2] == 0xFE && s[3] == 0xFF) + { + Perl_croak(aTHX_ "Unsupported script encoding"); + } + } + return (char*)s; +} #ifdef PERL_OBJECT -#define NO_XSLOCKS #include "XSUB.h" #endif @@ -7042,28 +7552,42 @@ restore_rsfp(pTHXo_ void *f) PL_rsfp = fp; } -/* - * restore_expect - * Restores the state of PL_expect when the lexing that begun with a - * start_lex() call has ended. - */ - -static void -restore_expect(pTHXo_ void *e) +#ifndef PERL_NO_UTF16_FILTER +static I32 +utf16_textfilter(pTHXo_ int idx, SV *sv, int maxlen) { - /* a safe way to store a small integer in a pointer */ - PL_expect = (expectation)((char *)e - PL_tokenbuf); -} + I32 count = FILTER_READ(idx+1, sv, maxlen); + if (count) { + U8* tmps; + U8* tend; + I32 newlen; + New(898, tmps, SvCUR(sv) * 3 / 2 + 1, U8); + if (!*SvPV_nolen(sv)) + /* Game over, but don't feed an odd-length string to utf16_to_utf8 */ + return count; -/* - * restore_lex_expect - * Restores the state of PL_lex_expect when the lexing that begun with a - * start_lex() call has ended. - */ + tend = utf16_to_utf8((U8*)SvPVX(sv), tmps, SvCUR(sv), &newlen); + sv_usepvn(sv, (char*)tmps, tend - tmps); + } + return count; +} -static void -restore_lex_expect(pTHXo_ void *e) +static I32 +utf16rev_textfilter(pTHXo_ int idx, SV *sv, int maxlen) { - /* a safe way to store a small integer in a pointer */ - PL_lex_expect = (expectation)((char *)e - PL_tokenbuf); + I32 count = FILTER_READ(idx+1, sv, maxlen); + if (count) { + U8* tmps; + U8* tend; + I32 newlen; + if (!*SvPV_nolen(sv)) + /* Game over, but don't feed an odd-length string to utf16_to_utf8 */ + return count; + + New(898, tmps, SvCUR(sv) * 3 / 2 + 1, U8); + tend = utf16_to_utf8_reversed((U8*)SvPVX(sv), tmps, SvCUR(sv), &newlen); + sv_usepvn(sv, (char*)tmps, tend - tmps); + } + return count; } +#endif