* or may not be utf8.
*/
-STATIC I32
+STATIC Size_t
S_do_trans_simple(pTHX_ SV * const sv)
{
- I32 matches = 0;
+ Size_t matches = 0;
STRLEN len;
U8 *s = (U8*)SvPV_nomg(sv,len);
U8 * const send = s+len;
/* First, take care of non-UTF-8 input strings, because they're easy */
if (!SvUTF8(sv)) {
while (s < send) {
- const I32 ch = tbl->map[*s];
+ const short ch = tbl->map[*s];
if (ch >= 0) {
matches++;
*s = (U8)ch;
SvSETMAGIC(sv);
}
else {
- const I32 grows = PL_op->op_private & OPpTRANS_GROWS;
+ const bool grows = cBOOL(PL_op->op_private & OPpTRANS_GROWS);
U8 *d;
U8 *dstart;
dstart = d;
while (s < send) {
STRLEN ulen;
- I32 ch;
+ short ch;
/* Need to check this, otherwise 128..255 won't match */
const UV c = utf8n_to_uvchr(s, send - s, &ulen, UTF8_ALLOW_DEFAULT);
if (c < 0x100 && (ch = tbl->map[c]) >= 0) {
matches++;
- d = uvchr_to_utf8(d, ch);
+ d = uvchr_to_utf8(d, (UV)ch);
s += ulen;
}
else { /* No match -> copy */
* or may not be utf8.
*/
-STATIC I32
+STATIC Size_t
S_do_trans_count(pTHX_ SV * const sv)
{
STRLEN len;
const U8 *s = (const U8*)SvPV_nomg_const(sv, len);
const U8 * const send = s + len;
- I32 matches = 0;
+ Size_t matches = 0;
const OPtrans_map * const tbl = (OPtrans_map*)cPVOP->op_pv;
PERL_ARGS_ASSERT_DO_TRANS_COUNT;
}
}
else {
- const I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT;
+ const bool complement = cBOOL(PL_op->op_private & OPpTRANS_COMPLEMENT);
while (s < send) {
STRLEN ulen;
const UV c = utf8n_to_uvchr(s, send - s, &ulen, UTF8_ALLOW_DEFAULT);
* or may not be utf8.
*/
-STATIC I32
+STATIC Size_t
S_do_trans_complex(pTHX_ SV * const sv)
{
STRLEN len;
U8 *s = (U8*)SvPV_nomg(sv, len);
U8 * const send = s+len;
- I32 matches = 0;
- const OPtrans_map_ex * const extbl = (OPtrans_map_ex*)cPVOP->op_pv;
+ Size_t matches = 0;
+ const OPtrans_map * const tbl = (OPtrans_map*)cPVOP->op_pv;
PERL_ARGS_ASSERT_DO_TRANS_COMPLEX;
- if (!extbl)
+ if (!tbl)
Perl_croak(aTHX_ "panic: do_trans_complex line %d",__LINE__);
if (!SvUTF8(sv)) {
if (PL_op->op_private & OPpTRANS_SQUASH) {
const U8* p = send;
while (s < send) {
- const I32 ch = extbl->map[*s];
+ const short ch = tbl->map[*s];
if (ch >= 0) {
*d = (U8)ch;
matches++;
}
else {
while (s < send) {
- const I32 ch = extbl->map[*s];
+ const short ch = tbl->map[*s];
if (ch >= 0) {
matches++;
*d++ = (U8)ch;
SvCUR_set(sv, d - dstart);
}
else { /* is utf8 */
- const I32 complement = PL_op->op_private & OPpTRANS_COMPLEMENT;
- const I32 grows = PL_op->op_private & OPpTRANS_GROWS;
- const I32 del = PL_op->op_private & OPpTRANS_DELETE;
+ const bool squash = cBOOL(PL_op->op_private & OPpTRANS_SQUASH);
+ const bool grows = cBOOL(PL_op->op_private & OPpTRANS_GROWS);
U8 *d;
U8 *dstart;
- SSize_t excess = 0;
+ Size_t size = tbl->size;
+ UV pch = 0xfeedface;
if (grows)
Newx(d, len*2+1, U8);
else
d = s;
dstart = d;
- if (complement)
- /* number of replacement chars in excess of any 0x00..0xff
- * search characters */
- excess = extbl->excess_len;
- if (PL_op->op_private & OPpTRANS_SQUASH) {
- UV pch = 0xfeedface;
- while (s < send) {
- STRLEN len;
- const UV comp = utf8n_to_uvchr(s, send - s, &len,
- UTF8_ALLOW_DEFAULT);
- UV ch;
- short sch;
-
- if (comp > 0xff) {
- if (!complement) {
- Move(s, d, len, U8);
- d += len;
- }
- else {
- /* use the implicit 0x100..0x7fffffff search range */
- UV comp100 = comp - 0x100;
- matches++;
- ch = del
- /* setting ch to pch forces char to be deleted */
- ? ((excess > (IV)comp100)
- ? (UV)extbl->map_ex[comp100]
- : pch )
-
- : ( (excess == -1) ? comp :
- (UV)(( excess == 0
- || excess <= (IV)comp100)
- ? extbl->repeat_char
- : extbl->map_ex[comp100]
- )
- );
- if (ch != pch) {
- d = uvchr_to_utf8(d, ch);
- pch = ch;
- }
- s += len;
- continue;
- }
- }
- else if ((sch = extbl->map[comp]) >= 0) {
- ch = (UV)sch;
- matches++;
- if (ch != pch) {
- d = uvchr_to_utf8(d, ch);
- pch = ch;
- }
- s += len;
- continue;
- }
- else if (sch == -1) { /* -1 is unmapped character */
- Move(s, d, len, U8);
- d += len;
- }
- else if (sch == -2) /* -2 is delete character */
- matches++;
- s += len;
- pch = 0xfeedface;
- }
- }
- else {
- while (s < send) {
- STRLEN len;
- const UV comp = utf8n_to_uvchr(s, send - s, &len,
- UTF8_ALLOW_DEFAULT);
- UV ch;
- short sch;
- if (comp > 0xff) {
- if (!complement) {
- Move(s, d, len, U8);
- d += len;
- }
- else {
- /* use the implicit 0x100..0x7fffffff search range */
- UV comp100 = comp - 0x100;
- matches++;
- if (del) {
- if (excess > (IV)comp100) {
- ch = (UV)extbl->map_ex[comp100];
- d = uvchr_to_utf8(d, ch);
- }
- }
- else {
- /* tr/...//c should call S_do_trans_count
- * instead */
- assert(excess != -1);
- ch = ( excess == 0
- || excess <= (IV)comp100)
- ? (UV)extbl->repeat_char
- : (UV)extbl->map_ex[comp100];
- d = uvchr_to_utf8(d, ch);
- }
- }
- }
- else if ((sch = extbl->map[comp]) >= 0) {
- d = uvchr_to_utf8(d, (UV)sch);
- matches++;
- }
- else if (sch == -1) { /* -1 is unmapped character */
- Move(s, d, len, U8);
- d += len;
- }
- else if (sch == -2) /* -2 is delete character */
- matches++;
- s += len;
- }
- }
+ while (s < send) {
+ STRLEN len;
+ const UV comp = utf8n_to_uvchr(s, send - s, &len,
+ UTF8_ALLOW_DEFAULT);
+ UV ch;
+ short sch;
+
+ sch = tbl->map[comp >= size ? size : comp];
+
+ if (sch >= 0) {
+ ch = (UV)sch;
+ replace:
+ matches++;
+ if (LIKELY(!squash || ch != pch)) {
+ d = uvchr_to_utf8(d, ch);
+ pch = ch;
+ }
+ s += len;
+ continue;
+ }
+ else if (sch == -1) { /* -1 is unmapped character */
+ Move(s, d, len, U8);
+ d += len;
+ }
+ else if (sch == -2) /* -2 is delete character */
+ matches++;
+ else {
+ assert(sch == -3); /* -3 is empty replacement */
+ ch = comp;
+ goto replace;
+ }
+
+ s += len;
+ pch = 0xfeedface;
+ }
+
if (grows) {
sv_setpvn(sv, (char*)dstart, d - dstart);
Safefree(dstart);
* or may not be utf8.
*/
-STATIC I32
+STATIC Size_t
S_do_trans_simple_utf8(pTHX_ SV * const sv)
{
U8 *s;
U8 *d;
U8 *start;
U8 *dstart, *dend;
- I32 matches = 0;
- const I32 grows = PL_op->op_private & OPpTRANS_GROWS;
+ Size_t matches = 0;
+ const bool grows = cBOOL(PL_op->op_private & OPpTRANS_GROWS);
STRLEN len;
SV* const rv =
#ifdef USE_ITHREADS
* or may not be utf8.
*/
-STATIC I32
+STATIC Size_t
S_do_trans_count_utf8(pTHX_ SV * const sv)
{
const U8 *s;
const U8 *start = NULL;
const U8 *send;
- I32 matches = 0;
+ Size_t matches = 0;
STRLEN len;
SV* const rv =
#ifdef USE_ITHREADS
* or may not be utf8.
*/
-STATIC I32
+STATIC Size_t
S_do_trans_complex_utf8(pTHX_ SV * const sv)
{
U8 *start, *send;
U8 *d;
- I32 matches = 0;
- const I32 squash = PL_op->op_private & OPpTRANS_SQUASH;
- const I32 del = PL_op->op_private & OPpTRANS_DELETE;
- const I32 grows = PL_op->op_private & OPpTRANS_GROWS;
+ Size_t matches = 0;
+ const bool squash = cBOOL(PL_op->op_private & OPpTRANS_SQUASH);
+ const bool del = cBOOL(PL_op->op_private & OPpTRANS_DELETE);
+ const bool grows = cBOOL(PL_op->op_private & OPpTRANS_GROWS);
SV* const rv =
#ifdef USE_ITHREADS
PAD_SVl(cPADOP->op_padix);
* Returns a count of number of characters translated
*/
-I32
+Size_t
Perl_do_trans(pTHX_ SV *sv)
{
STRLEN len;
- const I32 flags = PL_op->op_private;
- const I32 hasutf = flags & (OPpTRANS_FROM_UTF | OPpTRANS_TO_UTF);
+ const U8 flags = PL_op->op_private;
+ const U8 hasutf = flags & (OPpTRANS_FROM_UTF | OPpTRANS_TO_UTF);
PERL_ARGS_ASSERT_DO_TRANS;
STRLEN rightlen;
const char *lc;
const char *rc;
- STRLEN len;
+ STRLEN len = 0;
STRLEN lensave;
const char *lsave;
const char *rsave;
* portion. That means that at least one of the operands has to be
* entirely non-UTF-8, and the length of that operand has to be before the
* first above-FF in the other */
- if (left_utf8) {
- if (right_utf8 || rightlen > leftlen) {
- Perl_croak(aTHX_ fatal_above_ff_msg, PL_op_desc[optype]);
+ if (left_utf8 || right_utf8) {
+ if (left_utf8) {
+ if (right_utf8 || rightlen > leftlen) {
+ Perl_croak(aTHX_ FATAL_ABOVE_FF_MSG, PL_op_desc[optype]);
+ }
+ len = rightlen;
}
- len = rightlen;
- }
- else if (right_utf8) {
- if (leftlen > rightlen) {
- Perl_croak(aTHX_ fatal_above_ff_msg, PL_op_desc[optype]);
+ else if (right_utf8) {
+ if (leftlen > rightlen) {
+ Perl_croak(aTHX_ FATAL_ABOVE_FF_MSG, PL_op_desc[optype]);
+ }
+ len = leftlen;
}
- len = leftlen;
+
+ Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED),
+ DEPRECATED_ABOVE_FF_MSG, PL_op_desc[optype]);
}
else { /* Neither is UTF-8 */
- len = leftlen < rightlen ? leftlen : rightlen;
+ len = MIN(leftlen, rightlen);
}
lensave = len;