/* doop.c
*
- * Copyright (c) 1991-2000, Larry Wall
+ * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
+ * 2000, 2001, 2002, 2004, by Larry Wall and others
*
* You may distribute under the terms of either the GNU General Public
* License or the Artistic License, as specified in the README file.
* "'So that was the job I felt I had to do when I started,' thought Sam."
*/
+/* This file contains some common functions needed to carry out certain
+ * ops. For example both pp_schomp() and pp_chomp() - scalar and array
+ * chomp operations - call the function do_chomp() found in this file.
+ */
+
#include "EXTERN.h"
#define PERL_IN_DOOP_C
#include "perl.h"
#ifndef PERL_MICRO
-#if !defined(NSIG) || defined(M_UNIX) || defined(M_XENIX)
#include <signal.h>
#endif
-#endif
-
-#define HALF_UTF8_UPGRADE(start,end) \
- STMT_START { \
- if ((start)<(end)) { \
- U8* NeWsTr; \
- STRLEN LeN = (end) - (start); \
- NeWsTr = bytes_to_utf8(start, &LeN); \
- Safefree(start); \
- (start) = NeWsTr; \
- (end) = (start) + LeN; \
- } \
- } STMT_END
STATIC I32
S_do_trans_simple(pTHX_ SV *sv)
U8 *send;
U8 *dstart;
I32 matches = 0;
- I32 sutf = SvUTF8(sv);
+ I32 grows = PL_op->op_private & OPpTRANS_GROWS;
STRLEN len;
short *tbl;
I32 ch;
tbl = (short*)cPVOP->op_pv;
if (!tbl)
- Perl_croak(aTHX_ "panic: do_trans");
+ Perl_croak(aTHX_ "panic: do_trans_simple line %d",__LINE__);
s = (U8*)SvPV(sv, len);
send = s + len;
- /* First, take care of non-UTF8 input strings, because they're easy */
- if (!sutf) {
+ /* First, take care of non-UTF-8 input strings, because they're easy */
+ if (!SvUTF8(sv)) {
while (s < send) {
if ((ch = tbl[*s]) >= 0) {
matches++;
- *s++ = ch;
+ *s++ = (U8)ch;
}
else
s++;
}
/* Allow for expansion: $_="a".chr(400); tr/a/\xFE/, FE needs encoding */
- Newz(0, d, len*2+1, U8);
+ if (grows)
+ New(0, d, len*2+1, U8);
+ else
+ d = s;
dstart = d;
while (s < send) {
STRLEN ulen;
- short c;
+ UV c;
- ulen = 1;
/* Need to check this, otherwise 128..255 won't match */
- c = utf8_to_uv(s, send - s, &ulen, 0);
- if (c < 0x100 && (ch = tbl[(short)c]) >= 0) {
+ c = utf8n_to_uvchr(s, send - s, &ulen, 0);
+ if (c < 0x100 && (ch = tbl[c]) >= 0) {
matches++;
- if (ch < 0x80)
- *d++ = ch;
- else
- d = uv_to_utf8(d,ch);
+ d = uvchr_to_utf8(d, ch);
s += ulen;
}
else { /* No match -> copy */
- while (ulen--)
- *d++ = *s++;
+ Move(s, d, ulen, U8);
+ d += ulen;
+ s += ulen;
}
}
- *d = '\0';
- sv_setpvn(sv, (const char*)dstart, d - dstart);
- Safefree(dstart);
+ if (grows) {
+ sv_setpvn(sv, (char*)dstart, d - dstart);
+ Safefree(dstart);
+ }
+ else {
+ *d = '\0';
+ SvCUR_set(sv, d - dstart);
+ }
SvUTF8_on(sv);
SvSETMAGIC(sv);
return matches;
}
STATIC I32
-S_do_trans_count(pTHX_ SV *sv)/* SPC - OK */
+S_do_trans_count(pTHX_ SV *sv)
{
U8 *s;
U8 *send;
I32 matches = 0;
- I32 hasutf = SvUTF8(sv);
STRLEN len;
short *tbl;
+ I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT;
tbl = (short*)cPVOP->op_pv;
if (!tbl)
- Perl_croak(aTHX_ "panic: do_trans");
+ Perl_croak(aTHX_ "panic: do_trans_count line %d",__LINE__);
s = (U8*)SvPV(sv, len);
send = s + len;
- while (s < send) {
- if (hasutf && *s & 0x80)
- s += UTF8SKIP(s);
- else {
- UV c;
- STRLEN ulen;
- ulen = 1;
- if (hasutf)
- c = utf8_to_uv(s, send - s, &ulen, 0);
- else
- c = *s;
- if (c < 0x100 && tbl[c] >= 0)
+ if (!SvUTF8(sv))
+ while (s < send) {
+ if (tbl[*s++] >= 0)
matches++;
- s += ulen;
- }
- }
+ }
+ else
+ while (s < send) {
+ UV c;
+ STRLEN ulen;
+ c = utf8n_to_uvchr(s, send - s, &ulen, 0);
+ if (c < 0x100) {
+ if (tbl[c] >= 0)
+ matches++;
+ } else if (complement)
+ matches++;
+ s += ulen;
+ }
return matches;
}
STATIC I32
-S_do_trans_complex(pTHX_ SV *sv)/* SPC - NOT OK */
+S_do_trans_complex(pTHX_ SV *sv)
{
U8 *s;
U8 *send;
U8 *d;
U8 *dstart;
- I32 hasutf = SvUTF8(sv);
+ I32 isutf8;
I32 matches = 0;
- STRLEN len;
+ I32 grows = PL_op->op_private & OPpTRANS_GROWS;
+ I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT;
+ I32 del = PL_op->op_private & OPpTRANS_DELETE;
+ STRLEN len, rlen = 0;
short *tbl;
I32 ch;
tbl = (short*)cPVOP->op_pv;
if (!tbl)
- Perl_croak(aTHX_ "panic: do_trans");
+ Perl_croak(aTHX_ "panic: do_trans_complex line %d",__LINE__);
s = (U8*)SvPV(sv, len);
+ isutf8 = SvUTF8(sv);
send = s + len;
- Newz(0, d, len*2+1, U8);
- dstart = d;
-
- if (PL_op->op_private & OPpTRANS_SQUASH) {
- U8* p = send;
-
- while (s < send) {
- if (hasutf && *s & 0x80)
- s += UTF8SKIP(s);
- else {
- if ((ch = tbl[*s]) >= 0) {
- *d = ch;
+ if (!isutf8) {
+ dstart = d = s;
+ if (PL_op->op_private & OPpTRANS_SQUASH) {
+ U8* p = send;
+ while (s < send) {
+ if ((ch = tbl[*s]) >= 0) {
+ *d = (U8)ch;
+ matches++;
+ if (p != d - 1 || *p != *d)
+ p = d++;
+ }
+ else if (ch == -1) /* -1 is unmapped character */
+ *d++ = *s;
+ else if (ch == -2) /* -2 is delete character */
matches++;
- if (p != d - 1 || *p != *d)
- p = d++;
- }
- else if (ch == -1) /* -1 is unmapped character */
- *d++ = *s; /* -2 is delete character */
- s++;
- }
+ s++;
+ }
}
- }
- else {
- while (s < send) {
- UV comp;
- if (hasutf && *s & 0x80)
- comp = utf8_to_uv_simple(s, NULL);
- else
- comp = *s;
-
- ch = tbl[comp];
-
- if (ch == -1) { /* -1 is unmapped character */
- ch = comp;
- matches--;
- }
-
- if (ch >= 0) {
- if (hasutf)
- d = uv_to_utf8(d, ch);
- else
- *d++ = ch;
- }
- matches++;
-
- s += hasutf && *s & 0x80 ? UNISKIP(*s) : 1;
-
+ else {
+ while (s < send) {
+ if ((ch = tbl[*s]) >= 0) {
+ matches++;
+ *d++ = (U8)ch;
+ }
+ else if (ch == -1) /* -1 is unmapped character */
+ *d++ = *s;
+ else if (ch == -2) /* -2 is delete character */
+ matches++;
+ s++;
+ }
}
+ *d = '\0';
+ SvCUR_set(sv, d - dstart);
}
+ else { /* isutf8 */
+ if (grows)
+ New(0, d, len*2+1, U8);
+ else
+ d = s;
+ dstart = d;
+ if (complement && !del)
+ rlen = tbl[0x100];
- *d = '\0';
+#ifdef MACOS_TRADITIONAL
+#define comp CoMP /* "comp" is a keyword in some compilers ... */
+#endif
- sv_setpvn(sv, (const char*)dstart, d - dstart);
- Safefree(dstart);
- if (hasutf)
- SvUTF8_on(sv);
+ if (PL_op->op_private & OPpTRANS_SQUASH) {
+ UV pch = 0xfeedface;
+ while (s < send) {
+ STRLEN len;
+ UV comp = utf8_to_uvchr(s, &len);
+
+ if (comp > 0xff) {
+ if (!complement) {
+ Copy(s, d, len, U8);
+ d += len;
+ }
+ else {
+ matches++;
+ if (!del) {
+ ch = (rlen == 0) ? comp :
+ (comp - 0x100 < rlen) ?
+ tbl[comp+1] : tbl[0x100+rlen];
+ if ((UV)ch != pch) {
+ d = uvchr_to_utf8(d, ch);
+ pch = (UV)ch;
+ }
+ s += len;
+ continue;
+ }
+ }
+ }
+ else if ((ch = tbl[comp]) >= 0) {
+ matches++;
+ if ((UV)ch != pch) {
+ d = uvchr_to_utf8(d, ch);
+ pch = (UV)ch;
+ }
+ s += len;
+ continue;
+ }
+ else if (ch == -1) { /* -1 is unmapped character */
+ Copy(s, d, len, U8);
+ d += len;
+ }
+ else if (ch == -2) /* -2 is delete character */
+ matches++;
+ s += len;
+ pch = 0xfeedface;
+ }
+ }
+ else {
+ while (s < send) {
+ STRLEN len;
+ UV comp = utf8_to_uvchr(s, &len);
+ if (comp > 0xff) {
+ if (!complement) {
+ Move(s, d, len, U8);
+ d += len;
+ }
+ else {
+ matches++;
+ if (!del) {
+ if (comp - 0x100 < rlen)
+ d = uvchr_to_utf8(d, tbl[comp+1]);
+ else
+ d = uvchr_to_utf8(d, tbl[0x100+rlen]);
+ }
+ }
+ }
+ else if ((ch = tbl[comp]) >= 0) {
+ d = uvchr_to_utf8(d, ch);
+ matches++;
+ }
+ else if (ch == -1) { /* -1 is unmapped character */
+ Copy(s, d, len, U8);
+ d += len;
+ }
+ else if (ch == -2) /* -2 is delete character */
+ matches++;
+ s += len;
+ }
+ }
+ if (grows) {
+ sv_setpvn(sv, (char*)dstart, d - dstart);
+ Safefree(dstart);
+ }
+ else {
+ *d = '\0';
+ SvCUR_set(sv, d - dstart);
+ }
+ SvUTF8_on(sv);
+ }
SvSETMAGIC(sv);
return matches;
-
}
STATIC I32
-S_do_trans_simple_utf8(pTHX_ SV *sv)/* SPC - OK */
+S_do_trans_simple_utf8(pTHX_ SV *sv)
{
U8 *s;
U8 *send;
U8 *d;
U8 *start;
- U8 *dstart;
+ U8 *dstart, *dend;
I32 matches = 0;
+ I32 grows = PL_op->op_private & OPpTRANS_GROWS;
STRLEN len;
SV* rv = (SV*)cSVOP->op_sv;
SV** svp = hv_fetch(hv, "NONE", 4, FALSE);
UV none = svp ? SvUV(*svp) : 0x7fffffff;
UV extra = none + 1;
- UV final;
+ UV final = 0;
UV uv;
- I32 isutf;
- I32 howmany;
+ I32 isutf8;
+ U8 hibit = 0;
- isutf = SvUTF8(sv);
s = (U8*)SvPV(sv, len);
+ isutf8 = SvUTF8(sv);
+ if (!isutf8) {
+ U8 *t = s, *e = s + len;
+ while (t < e) {
+ U8 ch = *t++;
+ if ((hibit = !NATIVE_IS_INVARIANT(ch)))
+ break;
+ }
+ if (hibit)
+ s = bytes_to_utf8(s, &len);
+ }
send = s + len;
start = s;
if (svp)
final = SvUV(*svp);
- /* d needs to be bigger than s, in case e.g. upgrading is required */
- Newz(0, d, len*2+1, U8);
- dstart = d;
+ if (grows) {
+ /* d needs to be bigger than s, in case e.g. upgrading is required */
+ New(0, d, len*3+UTF8_MAXLEN, U8);
+ dend = d + len * 3;
+ dstart = d;
+ }
+ else {
+ dstart = d = s;
+ dend = d + len;
+ }
+
while (s < send) {
- if ((uv = swash_fetch(rv, s)) < none) {
+ if ((uv = swash_fetch(rv, s, TRUE)) < none) {
s += UTF8SKIP(s);
matches++;
- if ((uv & 0x80) && !isutf++)
- HALF_UTF8_UPGRADE(dstart,d);
- d = uv_to_utf8(d, uv);
+ d = uvuni_to_utf8(d, uv);
}
else if (uv == none) {
- int i;
- i = UTF8SKIP(s);
- if (i > 1 && !isutf++)
- HALF_UTF8_UPGRADE(dstart,d);
- while(i--)
- *d++ = *s++;
+ int i = UTF8SKIP(s);
+ Move(s, d, i, U8);
+ d += i;
+ s += i;
}
else if (uv == extra) {
- int i;
- i = UTF8SKIP(s);
+ int i = UTF8SKIP(s);
s += i;
matches++;
- if (i > 1 && !isutf++)
- HALF_UTF8_UPGRADE(dstart,d);
- d = uv_to_utf8(d, final);
+ d = uvuni_to_utf8(d, final);
}
else
s += UTF8SKIP(s);
+
+ if (d > dend) {
+ STRLEN clen = d - dstart;
+ STRLEN nlen = dend - dstart + len + UTF8_MAXLEN;
+ if (!grows)
+ Perl_croak(aTHX_ "panic: do_trans_simple_utf8 line %d",__LINE__);
+ Renew(dstart, nlen+UTF8_MAXLEN, U8);
+ d = dstart + clen;
+ dend = dstart + nlen;
+ }
+ }
+ if (grows || hibit) {
+ sv_setpvn(sv, (char*)dstart, d - dstart);
+ Safefree(dstart);
+ if (grows && hibit)
+ Safefree(start);
+ }
+ else {
+ *d = '\0';
+ SvCUR_set(sv, d - dstart);
}
- *d = '\0';
- sv_setpvn(sv, (const char*)dstart, d - dstart);
SvSETMAGIC(sv);
- if (isutf)
- SvUTF8_on(sv);
+ SvUTF8_on(sv);
return matches;
}
STATIC I32
-S_do_trans_count_utf8(pTHX_ SV *sv)/* SPC - OK */
+S_do_trans_count_utf8(pTHX_ SV *sv)
{
U8 *s;
- U8 *send;
+ U8 *start = 0, *send;
I32 matches = 0;
STRLEN len;
HV* hv = (HV*)SvRV(rv);
SV** svp = hv_fetch(hv, "NONE", 4, FALSE);
UV none = svp ? SvUV(*svp) : 0x7fffffff;
+ UV extra = none + 1;
UV uv;
+ U8 hibit = 0;
s = (U8*)SvPV(sv, len);
- if (!SvUTF8(sv))
- s = bytes_to_utf8(s, &len);
+ if (!SvUTF8(sv)) {
+ U8 *t = s, *e = s + len;
+ while (t < e) {
+ U8 ch = *t++;
+ if ((hibit = !NATIVE_IS_INVARIANT(ch)))
+ break;
+ }
+ if (hibit)
+ start = s = bytes_to_utf8(s, &len);
+ }
send = s + len;
while (s < send) {
- if ((uv = swash_fetch(rv, s)) < none)
+ if ((uv = swash_fetch(rv, s, TRUE)) < none || uv == extra)
matches++;
s += UTF8SKIP(s);
}
+ if (hibit)
+ Safefree(start);
return matches;
}
STATIC I32
-S_do_trans_complex_utf8(pTHX_ SV *sv) /* SPC - NOT OK */
+S_do_trans_complex_utf8(pTHX_ SV *sv)
{
U8 *s;
- U8 *send;
+ U8 *start, *send;
U8 *d;
I32 matches = 0;
I32 squash = PL_op->op_private & OPpTRANS_SQUASH;
I32 del = PL_op->op_private & OPpTRANS_DELETE;
+ I32 grows = PL_op->op_private & OPpTRANS_GROWS;
SV* rv = (SV*)cSVOP->op_sv;
HV* hv = (HV*)SvRV(rv);
SV** svp = hv_fetch(hv, "NONE", 4, FALSE);
UV none = svp ? SvUV(*svp) : 0x7fffffff;
UV extra = none + 1;
- UV final;
+ UV final = 0;
+ bool havefinal = FALSE;
UV uv;
STRLEN len;
- U8 *dst;
- I32 isutf = SvUTF8(sv);
+ U8 *dstart, *dend;
+ I32 isutf8;
+ U8 hibit = 0;
s = (U8*)SvPV(sv, len);
+ isutf8 = SvUTF8(sv);
+ if (!isutf8) {
+ U8 *t = s, *e = s + len;
+ while (t < e) {
+ U8 ch = *t++;
+ if ((hibit = !NATIVE_IS_INVARIANT(ch)))
+ break;
+ }
+ if (hibit)
+ s = bytes_to_utf8(s, &len);
+ }
send = s + len;
+ start = s;
svp = hv_fetch(hv, "FINAL", 5, FALSE);
- if (svp)
+ if (svp) {
final = SvUV(*svp);
+ havefinal = TRUE;
+ }
- Newz(0, d, len*2+1, U8);
- dst = d;
+ if (grows) {
+ /* d needs to be bigger than s, in case e.g. upgrading is required */
+ New(0, d, len*3+UTF8_MAXLEN, U8);
+ dend = d + len * 3;
+ dstart = d;
+ }
+ else {
+ dstart = d = s;
+ dend = d + len;
+ }
if (squash) {
UV puv = 0xfeedface;
while (s < send) {
- if (SvUTF8(sv))
- uv = swash_fetch(rv, s);
- else {
- U8 tmpbuf[2];
- uv = *s++;
- if (uv < 0x80)
- tmpbuf[0] = uv;
- else {
- tmpbuf[0] = (( uv >> 6) | 0xc0);
- tmpbuf[1] = (( uv & 0x3f) | 0x80);
- }
- uv = swash_fetch(rv, tmpbuf);
+ uv = swash_fetch(rv, s, TRUE);
+
+ if (d > dend) {
+ STRLEN clen = d - dstart;
+ STRLEN nlen = dend - dstart + len + UTF8_MAXLEN;
+ if (!grows)
+ Perl_croak(aTHX_ "panic: do_trans_complex_utf8 line %d",__LINE__);
+ Renew(dstart, nlen+UTF8_MAXLEN, U8);
+ d = dstart + clen;
+ dend = dstart + nlen;
}
-
if (uv < none) {
matches++;
+ s += UTF8SKIP(s);
if (uv != puv) {
- if ((uv & 0x80) && !isutf++)
- HALF_UTF8_UPGRADE(dst,d);
- d = uv_to_utf8(d, uv);
+ d = uvuni_to_utf8(d, uv);
puv = uv;
}
- s += UTF8SKIP(s);
continue;
}
else if (uv == none) { /* "none" is unmapped character */
- STRLEN ulen;
- *d++ = (U8)utf8_to_uv(s, send - s, &ulen, 0);
- s += ulen;
+ int i = UTF8SKIP(s);
+ Move(s, d, i, U8);
+ d += i;
+ s += i;
puv = 0xfeedface;
continue;
}
else if (uv == extra && !del) {
matches++;
- if (uv != puv) {
- d = uv_to_utf8(d, final);
- puv = final;
+ if (havefinal) {
+ s += UTF8SKIP(s);
+ if (puv != final) {
+ d = uvuni_to_utf8(d, final);
+ puv = final;
+ }
+ }
+ else {
+ STRLEN len;
+ uv = utf8_to_uvuni(s, &len);
+ if (uv != puv) {
+ Move(s, d, len, U8);
+ d += len;
+ puv = uv;
+ }
+ s += len;
}
- s += UTF8SKIP(s);
continue;
}
matches++; /* "none+1" is delete character */
}
else {
while (s < send) {
- if (SvUTF8(sv))
- uv = swash_fetch(rv, s);
- else {
- U8 tmpbuf[2];
- uv = *s++;
- if (uv < 0x80)
- tmpbuf[0] = uv;
- else {
- tmpbuf[0] = (( uv >> 6) | 0xc0);
- tmpbuf[1] = (( uv & 0x3f) | 0x80);
- }
- uv = swash_fetch(rv, tmpbuf);
+ uv = swash_fetch(rv, s, TRUE);
+ if (d > dend) {
+ STRLEN clen = d - dstart;
+ STRLEN nlen = dend - dstart + len + UTF8_MAXLEN;
+ if (!grows)
+ Perl_croak(aTHX_ "panic: do_trans_complex_utf8 line %d",__LINE__);
+ Renew(dstart, nlen+UTF8_MAXLEN, U8);
+ d = dstart + clen;
+ dend = dstart + nlen;
}
if (uv < none) {
matches++;
- d = uv_to_utf8(d, uv);
s += UTF8SKIP(s);
+ d = uvuni_to_utf8(d, uv);
continue;
}
else if (uv == none) { /* "none" is unmapped character */
- STRLEN ulen;
- *d++ = (U8)utf8_to_uv(s, send - s, &ulen, 0);
- s += ulen;
+ int i = UTF8SKIP(s);
+ Move(s, d, i, U8);
+ d += i;
+ s += i;
continue;
}
else if (uv == extra && !del) {
matches++;
- d = uv_to_utf8(d, final);
s += UTF8SKIP(s);
+ d = uvuni_to_utf8(d, final);
continue;
}
matches++; /* "none+1" is delete character */
s += UTF8SKIP(s);
}
}
- if (dst)
- sv_usepvn(sv, (char*)dst, d - dst);
+ if (grows || hibit) {
+ sv_setpvn(sv, (char*)dstart, d - dstart);
+ Safefree(dstart);
+ if (grows && hibit)
+ Safefree(start);
+ }
else {
*d = '\0';
- SvCUR_set(sv, d - (U8*)SvPVX(sv));
+ SvCUR_set(sv, d - dstart);
}
+ SvUTF8_on(sv);
SvSETMAGIC(sv);
return matches;
I32 hasutf = (PL_op->op_private &
(OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF));
- if (SvREADONLY(sv) && !(PL_op->op_private & OPpTRANS_IDENTICAL))
- Perl_croak(aTHX_ PL_no_modify);
-
+ if (SvREADONLY(sv)) {
+ if (SvIsCOW(sv))
+ sv_force_normal_flags(sv, 0);
+ if (SvREADONLY(sv) && !(PL_op->op_private & OPpTRANS_IDENTICAL))
+ Perl_croak(aTHX_ PL_no_modify);
+ }
(void)SvPV(sv, len);
if (!len)
return 0;
- if (!SvPOKp(sv))
- (void)SvPV_force(sv, len);
- if (!(PL_op->op_private & OPpTRANS_IDENTICAL))
+ if (!(PL_op->op_private & OPpTRANS_IDENTICAL)) {
+ if (!SvPOKp(sv))
+ (void)SvPV_force(sv, len);
(void)SvPOK_only_UTF8(sv);
+ }
DEBUG_t( Perl_deb(aTHX_ "2.TBL\n"));
- switch (PL_op->op_private & ~hasutf & 63) {
+ switch (PL_op->op_private & ~hasutf & (
+ OPpTRANS_FROM_UTF|OPpTRANS_TO_UTF|OPpTRANS_IDENTICAL|
+ OPpTRANS_SQUASH|OPpTRANS_DELETE|OPpTRANS_COMPLEMENT)) {
case 0:
if (hasutf)
return do_trans_simple_utf8(sv);
return do_trans_simple(sv);
case OPpTRANS_IDENTICAL:
+ case OPpTRANS_IDENTICAL|OPpTRANS_COMPLEMENT:
if (hasutf)
return do_trans_count_utf8(sv);
else
register I32 items = sp - mark;
register STRLEN len;
STRLEN delimlen;
- register char *delim = SvPV(del, delimlen);
STRLEN tmplen;
+ (void) SvPV(del, delimlen); /* stringify and get the delimlen */
+ /* SvCUR assumes it's SvPOK() and woe betide you if it's not. */
+
mark++;
len = (items > 0 ? (delimlen * (items - 1) ) : 0);
(void)SvUPGRADE(sv, SVt_PV);
++mark;
}
- if (items-- > 0) {
- char *s;
+ sv_setpvn(sv, "", 0);
+ /* sv_setpv retains old UTF8ness [perl #24846] */
+ SvUTF8_off(sv);
- sv_setpv(sv, "");
+ if (PL_tainting && SvMAGICAL(sv))
+ SvTAINTED_off(sv);
+
+ if (items-- > 0) {
if (*mark)
sv_catsv(sv, *mark);
mark++;
}
- else
- sv_setpv(sv,"");
- len = delimlen;
- if (len) {
+
+ if (delimlen) {
for (; items > 0; items--,mark++) {
- sv_catpvn(sv,delim,len);
+ sv_catsv(sv,del);
sv_catsv(sv,*mark);
}
}
char *pat = SvPV(*sarg, patlen);
bool do_taint = FALSE;
+ SvUTF8_off(sv);
+ if (DO_UTF8(*sarg))
+ SvUTF8_on(sv);
sv_vsetpvfn(sv, pat, patlen, Null(va_list*), sarg + 1, len - 1, &do_taint);
SvSETMAGIC(sv);
if (do_taint)
else {
offset >>= 3; /* turn into byte offset */
if (size == 16) {
- if (offset >= srclen)
+ if ((STRLEN)offset >= srclen)
retnum = 0;
else
retnum = (UV) s[offset] << 8;
}
else if (size == 32) {
- if (offset >= srclen)
+ if ((STRLEN)offset >= srclen)
retnum = 0;
- else if (offset + 1 >= srclen)
+ else if ((STRLEN)(offset + 1) >= srclen)
retnum =
((UV) s[offset ] << 24);
- else if (offset + 2 >= srclen)
+ else if ((STRLEN)(offset + 2) >= srclen)
retnum =
((UV) s[offset ] << 24) +
((UV) s[offset + 1] << 16);
#ifdef UV_IS_QUAD
else if (size == 64) {
if (ckWARN(WARN_PORTABLE))
- Perl_warner(aTHX_ WARN_PORTABLE,
+ Perl_warner(aTHX_ packWARN(WARN_PORTABLE),
"Bit vector size > 32 non-portable");
if (offset >= srclen)
retnum = 0;
#ifdef UV_IS_QUAD
else if (size == 64) {
if (ckWARN(WARN_PORTABLE))
- Perl_warner(aTHX_ WARN_PORTABLE,
+ Perl_warner(aTHX_ packWARN(WARN_PORTABLE),
"Bit vector size > 32 non-portable");
retnum =
((UV) s[offset ] << 56) +
lval = SvUV(sv);
offset = LvTARGOFF(sv);
if (offset < 0)
- Perl_croak(aTHX_ "Assigning to negative offset in vec");
+ Perl_croak(aTHX_ "Negative offset to vec in lvalue context");
size = LvTARGLEN(sv);
if (size < 1 || (size & (size-1))) /* size < 1 or not a power of two */
Perl_croak(aTHX_ "Illegal number of bits in vec");
else {
offset >>= 3; /* turn into byte offset */
if (size == 8)
- s[offset ] = lval & 0xff;
+ s[offset ] = (U8)( lval & 0xff);
else if (size == 16) {
- s[offset ] = (lval >> 8) & 0xff;
- s[offset+1] = lval & 0xff;
+ s[offset ] = (U8)((lval >> 8) & 0xff);
+ s[offset+1] = (U8)( lval & 0xff);
}
else if (size == 32) {
- s[offset ] = (lval >> 24) & 0xff;
- s[offset+1] = (lval >> 16) & 0xff;
- s[offset+2] = (lval >> 8) & 0xff;
- s[offset+3] = lval & 0xff;
+ s[offset ] = (U8)((lval >> 24) & 0xff);
+ s[offset+1] = (U8)((lval >> 16) & 0xff);
+ s[offset+2] = (U8)((lval >> 8) & 0xff);
+ s[offset+3] = (U8)( lval & 0xff);
}
#ifdef UV_IS_QUAD
else if (size == 64) {
if (ckWARN(WARN_PORTABLE))
- Perl_warner(aTHX_ WARN_PORTABLE,
+ Perl_warner(aTHX_ packWARN(WARN_PORTABLE),
"Bit vector size > 32 non-portable");
- s[offset ] = (lval >> 56) & 0xff;
- s[offset+1] = (lval >> 48) & 0xff;
- s[offset+2] = (lval >> 40) & 0xff;
- s[offset+3] = (lval >> 32) & 0xff;
- s[offset+4] = (lval >> 24) & 0xff;
- s[offset+5] = (lval >> 16) & 0xff;
- s[offset+6] = (lval >> 8) & 0xff;
- s[offset+7] = lval & 0xff;
+ s[offset ] = (U8)((lval >> 56) & 0xff);
+ s[offset+1] = (U8)((lval >> 48) & 0xff);
+ s[offset+2] = (U8)((lval >> 40) & 0xff);
+ s[offset+3] = (U8)((lval >> 32) & 0xff);
+ s[offset+4] = (U8)((lval >> 24) & 0xff);
+ s[offset+5] = (U8)((lval >> 16) & 0xff);
+ s[offset+6] = (U8)((lval >> 8) & 0xff);
+ s[offset+7] = (U8)( lval & 0xff);
}
#endif
}
do_chop(astr,hv_iterval(hv,entry));
return;
}
- else if (SvREADONLY(sv))
- Perl_croak(aTHX_ PL_no_modify);
+ else if (SvREADONLY(sv)) {
+ if (SvFAKE(sv)) {
+ /* SV is copy-on-write */
+ sv_force_normal_flags(sv, 0);
+ }
+ if (SvREADONLY(sv))
+ Perl_croak(aTHX_ PL_no_modify);
+ }
s = SvPV(sv, len);
if (len && !SvPOK(sv))
s = SvPV_force(sv, len);
char *send = s + len;
char *start = s;
s = send - 1;
- while ((*s & 0xc0) == 0x80)
- --s;
- if (UTF8SKIP(s) != send - s && ckWARN_d(WARN_UTF8))
- Perl_warner(aTHX_ WARN_UTF8, "Malformed UTF-8 character");
- sv_setpvn(astr, s, send - s);
- *s = '\0';
- SvCUR_set(sv, s - start);
- SvNIOK_off(sv);
- SvUTF8_on(astr);
+ while (s > start && UTF8_IS_CONTINUATION(*s))
+ s--;
+ if (utf8_to_uvchr((U8*)s, 0)) {
+ sv_setpvn(astr, s, send - s);
+ *s = '\0';
+ SvCUR_set(sv, s - start);
+ SvNIOK_off(sv);
+ SvUTF8_on(astr);
+ }
}
else
sv_setpvn(astr, "", 0);
{
register I32 count;
STRLEN len;
+ STRLEN n_a;
char *s;
+ char *temp_buffer = NULL;
+ SV* svrecode = Nullsv;
if (RsSNARF(PL_rs))
return 0;
count += do_chomp(hv_iterval(hv,entry));
return count;
}
- else if (SvREADONLY(sv))
- Perl_croak(aTHX_ PL_no_modify);
+ else if (SvREADONLY(sv)) {
+ if (SvFAKE(sv)) {
+ /* SV is copy-on-write */
+ sv_force_normal_flags(sv, 0);
+ }
+ if (SvREADONLY(sv))
+ Perl_croak(aTHX_ PL_no_modify);
+ }
+
+ if (PL_encoding) {
+ if (!SvUTF8(sv)) {
+ /* XXX, here sv is utf8-ized as a side-effect!
+ If encoding.pm is used properly, almost string-generating
+ operations, including literal strings, chr(), input data, etc.
+ should have been utf8-ized already, right?
+ */
+ sv_recode_to_utf8(sv, PL_encoding);
+ }
+ }
+
s = SvPV(sv, len);
- if (len && !SvPOKp(sv))
- s = SvPV_force(sv, len);
if (s && len) {
s += --len;
if (RsPARA(PL_rs)) {
}
}
else {
- STRLEN rslen;
+ STRLEN rslen, rs_charlen;
char *rsptr = SvPV(PL_rs, rslen);
+
+ rs_charlen = SvUTF8(PL_rs)
+ ? sv_len_utf8(PL_rs)
+ : rslen;
+
+ if (SvUTF8(PL_rs) != SvUTF8(sv)) {
+ /* Assumption is that rs is shorter than the scalar. */
+ if (SvUTF8(PL_rs)) {
+ /* RS is utf8, scalar is 8 bit. */
+ bool is_utf8 = TRUE;
+ temp_buffer = (char*)bytes_from_utf8((U8*)rsptr,
+ &rslen, &is_utf8);
+ if (is_utf8) {
+ /* Cannot downgrade, therefore cannot possibly match
+ */
+ assert (temp_buffer == rsptr);
+ temp_buffer = NULL;
+ goto nope;
+ }
+ rsptr = temp_buffer;
+ }
+ else if (PL_encoding) {
+ /* RS is 8 bit, encoding.pm is used.
+ * Do not recode PL_rs as a side-effect. */
+ svrecode = newSVpvn(rsptr, rslen);
+ sv_recode_to_utf8(svrecode, PL_encoding);
+ rsptr = SvPV(svrecode, rslen);
+ rs_charlen = sv_len_utf8(svrecode);
+ }
+ else {
+ /* RS is 8 bit, scalar is utf8. */
+ temp_buffer = (char*)bytes_to_utf8((U8*)rsptr, &rslen);
+ rsptr = temp_buffer;
+ }
+ }
if (rslen == 1) {
if (*s != *rsptr)
goto nope;
s -= rslen - 1;
if (memNE(s, rsptr, rslen))
goto nope;
- count += rslen;
+ count += rs_charlen;
}
}
- *s = '\0';
+ s = SvPV_force(sv, n_a);
SvCUR_set(sv, len);
+ *SvEND(sv) = '\0';
SvNIOK_off(sv);
+ SvSETMAGIC(sv);
}
nope:
- SvSETMAGIC(sv);
+
+ if (svrecode)
+ SvREFCNT_dec(svrecode);
+
+ Safefree(temp_buffer);
return count;
}
char *rsave;
bool left_utf = DO_UTF8(left);
bool right_utf = DO_UTF8(right);
- I32 needlen;
+ I32 needlen = 0;
if (left_utf && !right_utf)
sv_utf8_upgrade(right);
if (sv != left || (optype != OP_BIT_AND && !SvOK(sv) && !SvGMAGICAL(sv)))
sv_setpvn(sv, "", 0); /* avoid undef warning on |= and ^= */
- lsave = lc = SvPV(left, leftlen);
- rsave = rc = SvPV(right, rightlen);
+ lsave = lc = SvPV_nomg(left, leftlen);
+ rsave = rc = SvPV_nomg(right, rightlen);
len = leftlen < rightlen ? leftlen : rightlen;
lensave = len;
if ((left_utf || right_utf) && (sv == left || sv == right)) {
}
else if (SvOK(sv) || SvTYPE(sv) > SVt_PVMG) {
STRLEN n_a;
- dc = SvPV_force(sv, n_a);
- if (SvCUR(sv) < len) {
- dc = SvGROW(sv, len + 1);
+ dc = SvPV_force_nomg(sv, n_a);
+ if (SvCUR(sv) < (STRLEN)len) {
+ dc = SvGROW(sv, (STRLEN)(len + 1));
(void)memzero(dc + SvCUR(sv), len - SvCUR(sv) + 1);
}
if (optype != OP_BIT_AND && (left_utf || right_utf))
switch (optype) {
case OP_BIT_AND:
while (lulen && rulen) {
- luc = utf8_to_uv((U8*)lc, lulen, &ulen, UTF8_ALLOW_ANYUV);
+ luc = utf8n_to_uvchr((U8*)lc, lulen, &ulen, UTF8_ALLOW_ANYUV);
lc += ulen;
lulen -= ulen;
- ruc = utf8_to_uv((U8*)rc, rulen, &ulen, UTF8_ALLOW_ANYUV);
+ ruc = utf8n_to_uvchr((U8*)rc, rulen, &ulen, UTF8_ALLOW_ANYUV);
rc += ulen;
rulen -= ulen;
duc = luc & ruc;
- dc = (char*)uv_to_utf8((U8*)dc, duc);
+ dc = (char*)uvchr_to_utf8((U8*)dc, duc);
}
if (sv == left || sv == right)
(void)sv_usepvn(sv, dcsave, needlen);
break;
case OP_BIT_XOR:
while (lulen && rulen) {
- luc = utf8_to_uv((U8*)lc, lulen, &ulen, UTF8_ALLOW_ANYUV);
+ luc = utf8n_to_uvchr((U8*)lc, lulen, &ulen, UTF8_ALLOW_ANYUV);
lc += ulen;
lulen -= ulen;
- ruc = utf8_to_uv((U8*)rc, rulen, &ulen, UTF8_ALLOW_ANYUV);
+ ruc = utf8n_to_uvchr((U8*)rc, rulen, &ulen, UTF8_ALLOW_ANYUV);
rc += ulen;
rulen -= ulen;
duc = luc ^ ruc;
- dc = (char*)uv_to_utf8((U8*)dc, duc);
+ dc = (char*)uvchr_to_utf8((U8*)dc, duc);
}
goto mop_up_utf;
case OP_BIT_OR:
while (lulen && rulen) {
- luc = utf8_to_uv((U8*)lc, lulen, &ulen, UTF8_ALLOW_ANYUV);
+ luc = utf8n_to_uvchr((U8*)lc, lulen, &ulen, UTF8_ALLOW_ANYUV);
lc += ulen;
lulen -= ulen;
- ruc = utf8_to_uv((U8*)rc, rulen, &ulen, UTF8_ALLOW_ANYUV);
+ ruc = utf8n_to_uvchr((U8*)rc, rulen, &ulen, UTF8_ALLOW_ANYUV);
rc += ulen;
rulen -= ulen;
duc = luc | ruc;
- dc = (char*)uv_to_utf8((U8*)dc, duc);
+ dc = (char*)uvchr_to_utf8((U8*)dc, duc);
}
mop_up_utf:
if (sv == left || sv == right)
*dc++ = *lc++ | *rc++;
mop_up:
len = lensave;
- if (rightlen > len)
+ if (rightlen > (STRLEN)len)
sv_catpvn(sv, rsave + len, rightlen - len);
- else if (leftlen > len)
+ else if (leftlen > (STRLEN)len)
sv_catpvn(sv, lsave + len, leftlen - len);
else
*SvEND(sv) = '\0';
OP *
Perl_do_kv(pTHX)
{
- djSP;
+ dSP;
HV *hv = (HV*)POPs;
HV *keys;
register HE *entry;
I32 gimme = GIMME_V;
I32 dokeys = (PL_op->op_type == OP_KEYS);
I32 dovalues = (PL_op->op_type == OP_VALUES);
- I32 realhv = (SvTYPE(hv) == SVt_PVHV);
if (PL_op->op_type == OP_RV2HV || PL_op->op_type == OP_PADHV)
dokeys = dovalues = TRUE;
if (!hv) {
- if (PL_op->op_flags & OPf_MOD) { /* lvalue */
+ if (PL_op->op_flags & OPf_MOD || LVRET) { /* lvalue */
dTARGET; /* make sure to clear its target here */
if (SvTYPE(TARG) == SVt_PVLV)
LvTARG(TARG) = Nullsv;
RETURN;
}
- keys = realhv ? hv : avhv_keys((AV*)hv);
+ keys = hv;
(void)hv_iterinit(keys); /* always reset iterator regardless */
if (gimme == G_VOID)
IV i;
dTARGET;
- if (PL_op->op_flags & OPf_MOD) { /* lvalue */
+ if (PL_op->op_flags & OPf_MOD || LVRET) { /* lvalue */
if (SvTYPE(TARG) < SVt_PVLV) {
sv_upgrade(TARG, SVt_PVLV);
- sv_magic(TARG, Nullsv, 'k', Nullch, 0);
+ sv_magic(TARG, Nullsv, PERL_MAGIC_nkeys, Nullch, 0);
}
LvTYPE(TARG) = 'k';
if (LvTARG(TARG) != (SV*)keys) {
RETURN;
}
- if (! SvTIED_mg((SV*)keys, 'P'))
+ if (! SvTIED_mg((SV*)keys, PERL_MAGIC_tied))
i = HvKEYS(keys);
else {
i = 0;
PUTBACK; /* hv_iternext and hv_iterval might clobber stack_sp */
while ((entry = hv_iternext(keys))) {
SPAGAIN;
- if (dokeys)
- XPUSHs(hv_iterkeysv(entry)); /* won't clobber stack_sp */
+ if (dokeys) {
+ SV* sv = hv_iterkeysv(entry);
+ XPUSHs(sv); /* won't clobber stack_sp */
+ }
if (dovalues) {
PUTBACK;
- tmpstr = realhv ?
- hv_iterval(hv,entry) : avhv_iterval((AV*)hv,entry);
+ tmpstr = hv_iterval(hv,entry);
DEBUG_H(Perl_sv_setpvf(aTHX_ tmpstr, "%lu%%%d=%lu",
(unsigned long)HeHASH(entry),
HvMAX(keys)+1,
https://perl5.git.perl.org / perl5.git / blobdiff