}
#endif
+#if 0
STATIC I32
S_utf16_textfilter(pTHX_ int idx, SV *sv, int maxlen)
{
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);
-
}
return count;
}
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);
-
}
return count;
}
+#endif
/*
* Perl_lex_start
}
}
+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);
+ char *end = start + len;
+ while (start < end) {
+ I32 skip;
+ UV n;
+ if (utf)
+ n = utf8_to_uv((U8*)start, &skip);
+ 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.
S_force_version(pTHX_ char *s)
{
OP *version = Nullop;
- bool is_vstr = FALSE;
char *d;
s = skipspace(s);
d = s;
- if (*d == 'v') {
- is_vstr = TRUE;
+ 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);
- /* real VERSION number -- GBARR */
version = yylval.opval;
- if (is_vstr) {
- SV *ver = cSVOPx(version)->op_sv;
- SvUPGRADE(ver, SVt_PVIV);
- SvIOKp_on(ver); /* hint that it is a version */
+ 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 */
}
}
}
bool dorange = FALSE; /* are we in a translit range? */
bool has_utf = FALSE; /* embedded \x{} */
I32 len; /* ? */
+ 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;
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;
/* (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) {
- has_utf = TRUE;
- while (len--)
- *d++ = *s++;
- continue;
- }
- }
- else
- has_utf = TRUE; /* assume valid utf8 */
+ (void)utf8_to_uv((U8*)s, &len);
+ if (len == 1) {
+ /* illegal UTF8, 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_utf = TRUE;
+ continue;
}
/* backslashes */
default:
{
dTHR;
- if (ckWARN(WARN_MISC) && isALPHA(*s))
+ if (ckWARN(WARN_MISC) && isALNUM(*s))
Perl_warner(aTHX_ WARN_MISC,
"Unrecognized escape \\%c passed through",
*s);
/* \132 indicates an octal constant */
case '0': case '1': case '2': case '3':
case '4': case '5': case '6': case '7':
- *d++ = (char)scan_oct(s, 3, &len);
+ len = 0; /* disallow underscores */
+ uv = (UV)scan_oct(s, 3, &len);
s += len;
- continue;
+ goto NUM_ESCAPE_INSERT;
/* \x24 indicates a hex constant */
case 'x':
++s;
if (*s == '{') {
char* e = strchr(s, '}');
- UV uv;
-
if (!e) {
yyerror("Missing right brace on \\x{}");
e = s;
}
- /* note: utf always shorter than hex */
- uv = (UV)scan_hex(s + 1, e - s - 1, &len);
- if (uv > 127) {
- d = (char*)uv_to_utf8((U8*)d, uv);
- has_utf = TRUE;
- }
- else
- *d++ = (char)uv;
- s = e + 1;
+ len = 1; /* allow underscores */
+ uv = (UV)scan_hex(s + 1, e - s - 1, &len);
+ s = e + 1;
}
else {
- /* XXX collapse this branch into the one above */
- 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 */
+ len = 0; /* disallow underscores */
+ uv = (UV)scan_hex(s, 2, &len);
+ s += len;
+ }
+
+ 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) {
+ if (!thisutf && !has_utf && 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++;
+ }
+ 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 (thisutf || uv > 255) {
+ d = (char*)uv_to_utf8((U8*)d, uv);
has_utf = TRUE;
- }
+ }
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",
- (int)len,s,(int)len,s);
- }
- *d++ = (char)uv;
+ *d++ = (char)uv;
}
- s += len;
+ }
+ else {
+ *d++ = (char)uv;
}
continue;
++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{}");
res = new_constant( Nullch, 0, "charnames",
res, Nullsv, "\\N{...}" );
str = SvPV(res,len);
+ if (!has_utf && SvUTF8(res)) {
+ char *ostart = SvPVX(sv);
+ SvCUR_set(sv, d - ostart);
+ SvPOK_on(sv);
+ sv_utf8_upgrade(sv);
+ d = SvPVX(sv) + SvCUR(sv);
+ has_utf = TRUE;
+ }
if (len > e - s + 4) {
char *odest = SvPVX(sv);
*d = *s++;
if (isLOWER(*d))
*d = toUPPER(*d);
- *d++ = toCTRL(*d);
+ *d = toCTRL(*d);
+ d++;
#else
len = *s++;
*d++ = toCTRL(len);
}
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 */
{
#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");
+ "\t(Maybe you didn't strip carriage returns after a network transfer?)\n");
#endif
case ' ': case '\t': case '\f': case 013:
s++;
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++;
/* 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);
}
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:
char *start = s;
start++;
start++;
- while (isDIGIT(*start))
+ while (isDIGIT(*start) || *start == '_')
start++;
if (*start == '.' && isDIGIT(start[1])) {
s = scan_num(s);
case 'z': case 'Z':
keylookup: {
- STRLEN n_a;
gv = Nullgv;
gvp = 0;
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"))
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;
/* 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);
}
}
- /* 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)) {
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;
LOP(OP_BIND,XTERM);
case KEY_binmode:
- UNI(OP_BINMODE);
+ LOP(OP_BINMODE,XTERM);
case KEY_bless:
LOP(OP_BLESS,XTERM);
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) {
STATIC SV *
S_new_constant(pTHX_ char *s, STRLEN len, const char *key, SV *sv, SV *pv,
- const char *type)
+ const char *type)
{
dSP;
HV *table = GvHV(PL_hintgv); /* ^H */
SV *res;
SV **cvp;
SV *cv, *typesv;
- const char *why, *why1, *why2;
+ const char *why1, *why2, *why3;
- if (!(PL_hints & HINT_LOCALIZE_HH)) {
+ if (!table || !(PL_hints & HINT_LOCALIZE_HH)) {
SV *msg;
- why = "%^H is not localized";
- report_short:
- why1 = why2 = "";
+ why1 = "%^H is not consistent";
+ why2 = strEQ(key,"charnames")
+ ? " (missing \"use charnames ...\"?)"
+ : "";
+ why3 = "";
report:
msg = Perl_newSVpvf(aTHX_ "constant(%s): %s%s%s",
- (type ? type: "undef"), why1, why2, why);
+ (type ? type: "undef"), why1, why2, why3);
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 = "$^H{";
why2 = key;
+ why3 = "} is not defined";
goto report;
}
sv_2mortal(sv); /* Parent created it permanently */
SPAGAIN ;
/* Check the eval first */
- if (!PL_in_eval && SvTRUE(ERRSV))
- {
+ if (!PL_in_eval && SvTRUE(ERRSV)) {
STRLEN n_a;
sv_catpv(ERRSV, "Propagated");
yyerror(SvPV(ERRSV, n_a)); /* Duplicates the message inside eval */
POPSTACK;
if (!SvOK(res)) {
- why = "}} did not return a defined value";
why1 = "Call to &{$^H{";
why2 = key;
+ why3 = "}} did not return a defined value";
sv = res;
goto report;
- }
+ }
- return res;
+ return res;
}
STATIC char *
d++;
if (UTF) {
e = s;
- while (e < send && isALNUM_lazy_if(e,UTF) || *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);
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 = Nullsv; /* place to put the converted number */
bool floatit; /* boolean: int or float? */
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 */
/* make an sv from the string */
sv = NEWSV(92,0);
+ /* unfortunately this monster needs to be on one line or
+ makedepend will be confused. */
+#if (defined(USE_64_BIT_INT) && (!defined(HAS_STRTOLL)|| !defined(HAS_STRTOULL))) || (!defined(USE_64_BIT_INT) && (!defined(HAS_STRTOL) || !defined(HAS_STRTOUL)))
+
+ /*
+ No working strto[u]l[l]. Since atoi() doesn't do range checks,
+ we need to do this the hard way.
+ */
+
value = Atof(PL_tokenbuf);
/*
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
+ {
+ UV tryuv = U_V(value);
+ if (!floatit && (NV)tryuv == value) {
+ if (tryuv <= IV_MAX)
+ sv_setiv(sv, (IV)tryuv);
+ else
+ sv_setuv(sv, tryuv);
+ }
+ else
+ sv_setnv(sv, value);
+ }
+#else
+ /*
+ 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.
+ */
+
+ 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
+ sv_setuv(sv, uv);
+ }
+ if (floatit) {
+ value = Atof(PL_tokenbuf);
sv_setnv(sv, value);
+ }
+#endif
if ( floatit ? (PL_hints & HINT_NEW_FLOAT) :
(PL_hints & HINT_NEW_INTEGER) )
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 version number */
+
+ /* if it starts with a v, it could be a v-string */
case 'v':
+vstring:
{
char *pos = s;
pos++;
- while (isDIGIT(*pos))
+ while (isDIGIT(*pos) || *pos == '_')
pos++;
if (!isALPHA(*pos)) {
UV rev;
for (;;) {
if (*s == '0' && isDIGIT(s[1]))
yyerror("Octal number in vector unsupported");
- rev = atoi(s);
+ 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);
s = pos;
break;
}
- while (isDIGIT(*pos))
+ while (isDIGIT(*pos) || *pos == '_')
pos++;
}
Perl_warn(aTHX_ "%"SVf, msg);
else
qerror(msg);
- if (PL_error_count >= 10)
- Perl_croak(aTHX_ "%s has too many errors.\n", CopFILE(PL_curcop));
+ if (PL_error_count >= 10) {
+ if (PL_in_eval && SvCUR(ERRSV))
+ Perl_croak(aTHX_ "%_%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;
projects / p5sagit/p5-mst-13.2.git / blobdiff