| 1 | /* inline.h |
| 2 | * |
| 3 | * Copyright (C) 2012 by Larry Wall and others |
| 4 | * |
| 5 | * You may distribute under the terms of either the GNU General Public |
| 6 | * License or the Artistic License, as specified in the README file. |
| 7 | * |
| 8 | * This file is a home for static inline functions that cannot go in other |
| 9 | * headers files, because they depend on proto.h (included after most other |
| 10 | * headers) or struct definitions. |
| 11 | * |
| 12 | * Each section names the header file that the functions "belong" to. |
| 13 | */ |
| 14 | |
| 15 | /* ------------------------------- av.h ------------------------------- */ |
| 16 | |
| 17 | PERL_STATIC_INLINE SSize_t |
| 18 | S_av_top_index(pTHX_ AV *av) |
| 19 | { |
| 20 | PERL_ARGS_ASSERT_AV_TOP_INDEX; |
| 21 | assert(SvTYPE(av) == SVt_PVAV); |
| 22 | |
| 23 | return AvFILL(av); |
| 24 | } |
| 25 | |
| 26 | /* ------------------------------- cv.h ------------------------------- */ |
| 27 | |
| 28 | PERL_STATIC_INLINE GV * |
| 29 | S_CvGV(pTHX_ CV *sv) |
| 30 | { |
| 31 | return CvNAMED(sv) |
| 32 | ? Perl_cvgv_from_hek(aTHX_ sv) |
| 33 | : ((XPVCV*)MUTABLE_PTR(SvANY(sv)))->xcv_gv_u.xcv_gv; |
| 34 | } |
| 35 | |
| 36 | PERL_STATIC_INLINE I32 * |
| 37 | S_CvDEPTHp(const CV * const sv) |
| 38 | { |
| 39 | assert(SvTYPE(sv) == SVt_PVCV || SvTYPE(sv) == SVt_PVFM); |
| 40 | return &((XPVCV*)SvANY(sv))->xcv_depth; |
| 41 | } |
| 42 | |
| 43 | /* |
| 44 | CvPROTO returns the prototype as stored, which is not necessarily what |
| 45 | the interpreter should be using. Specifically, the interpreter assumes |
| 46 | that spaces have been stripped, which has been the case if the prototype |
| 47 | was added by toke.c, but is generally not the case if it was added elsewhere. |
| 48 | Since we can't enforce the spacelessness at assignment time, this routine |
| 49 | provides a temporary copy at parse time with spaces removed. |
| 50 | I<orig> is the start of the original buffer, I<len> is the length of the |
| 51 | prototype and will be updated when this returns. |
| 52 | */ |
| 53 | |
| 54 | #ifdef PERL_CORE |
| 55 | PERL_STATIC_INLINE char * |
| 56 | S_strip_spaces(pTHX_ const char * orig, STRLEN * const len) |
| 57 | { |
| 58 | SV * tmpsv; |
| 59 | char * tmps; |
| 60 | tmpsv = newSVpvn_flags(orig, *len, SVs_TEMP); |
| 61 | tmps = SvPVX(tmpsv); |
| 62 | while ((*len)--) { |
| 63 | if (!isSPACE(*orig)) |
| 64 | *tmps++ = *orig; |
| 65 | orig++; |
| 66 | } |
| 67 | *tmps = '\0'; |
| 68 | *len = tmps - SvPVX(tmpsv); |
| 69 | return SvPVX(tmpsv); |
| 70 | } |
| 71 | #endif |
| 72 | |
| 73 | /* ------------------------------- mg.h ------------------------------- */ |
| 74 | |
| 75 | #if defined(PERL_CORE) || defined(PERL_EXT) |
| 76 | /* assumes get-magic and stringification have already occurred */ |
| 77 | PERL_STATIC_INLINE STRLEN |
| 78 | S_MgBYTEPOS(pTHX_ MAGIC *mg, SV *sv, const char *s, STRLEN len) |
| 79 | { |
| 80 | assert(mg->mg_type == PERL_MAGIC_regex_global); |
| 81 | assert(mg->mg_len != -1); |
| 82 | if (mg->mg_flags & MGf_BYTES || !DO_UTF8(sv)) |
| 83 | return (STRLEN)mg->mg_len; |
| 84 | else { |
| 85 | const STRLEN pos = (STRLEN)mg->mg_len; |
| 86 | /* Without this check, we may read past the end of the buffer: */ |
| 87 | if (pos > sv_or_pv_len_utf8(sv, s, len)) return len+1; |
| 88 | return sv_or_pv_pos_u2b(sv, s, pos, NULL); |
| 89 | } |
| 90 | } |
| 91 | #endif |
| 92 | |
| 93 | /* ------------------------------- pad.h ------------------------------ */ |
| 94 | |
| 95 | #if defined(PERL_IN_PAD_C) || defined(PERL_IN_OP_C) |
| 96 | PERL_STATIC_INLINE bool |
| 97 | PadnameIN_SCOPE(const PADNAME * const pn, const U32 seq) |
| 98 | { |
| 99 | /* is seq within the range _LOW to _HIGH ? |
| 100 | * This is complicated by the fact that PL_cop_seqmax |
| 101 | * may have wrapped around at some point */ |
| 102 | if (COP_SEQ_RANGE_LOW(pn) == PERL_PADSEQ_INTRO) |
| 103 | return FALSE; /* not yet introduced */ |
| 104 | |
| 105 | if (COP_SEQ_RANGE_HIGH(pn) == PERL_PADSEQ_INTRO) { |
| 106 | /* in compiling scope */ |
| 107 | if ( |
| 108 | (seq > COP_SEQ_RANGE_LOW(pn)) |
| 109 | ? (seq - COP_SEQ_RANGE_LOW(pn) < (U32_MAX >> 1)) |
| 110 | : (COP_SEQ_RANGE_LOW(pn) - seq > (U32_MAX >> 1)) |
| 111 | ) |
| 112 | return TRUE; |
| 113 | } |
| 114 | else if ( |
| 115 | (COP_SEQ_RANGE_LOW(pn) > COP_SEQ_RANGE_HIGH(pn)) |
| 116 | ? |
| 117 | ( seq > COP_SEQ_RANGE_LOW(pn) |
| 118 | || seq <= COP_SEQ_RANGE_HIGH(pn)) |
| 119 | |
| 120 | : ( seq > COP_SEQ_RANGE_LOW(pn) |
| 121 | && seq <= COP_SEQ_RANGE_HIGH(pn)) |
| 122 | ) |
| 123 | return TRUE; |
| 124 | return FALSE; |
| 125 | } |
| 126 | #endif |
| 127 | |
| 128 | /* ------------------------------- pp.h ------------------------------- */ |
| 129 | |
| 130 | PERL_STATIC_INLINE I32 |
| 131 | S_TOPMARK(pTHX) |
| 132 | { |
| 133 | DEBUG_s(DEBUG_v(PerlIO_printf(Perl_debug_log, |
| 134 | "MARK top %p %"IVdf"\n", |
| 135 | PL_markstack_ptr, |
| 136 | (IV)*PL_markstack_ptr))); |
| 137 | return *PL_markstack_ptr; |
| 138 | } |
| 139 | |
| 140 | PERL_STATIC_INLINE I32 |
| 141 | S_POPMARK(pTHX) |
| 142 | { |
| 143 | DEBUG_s(DEBUG_v(PerlIO_printf(Perl_debug_log, |
| 144 | "MARK pop %p %"IVdf"\n", |
| 145 | (PL_markstack_ptr-1), |
| 146 | (IV)*(PL_markstack_ptr-1)))); |
| 147 | assert((PL_markstack_ptr > PL_markstack) || !"MARK underflow"); |
| 148 | return *PL_markstack_ptr--; |
| 149 | } |
| 150 | |
| 151 | /* ----------------------------- regexp.h ----------------------------- */ |
| 152 | |
| 153 | PERL_STATIC_INLINE struct regexp * |
| 154 | S_ReANY(const REGEXP * const re) |
| 155 | { |
| 156 | assert(isREGEXP(re)); |
| 157 | return re->sv_u.svu_rx; |
| 158 | } |
| 159 | |
| 160 | /* ------------------------------- sv.h ------------------------------- */ |
| 161 | |
| 162 | PERL_STATIC_INLINE SV * |
| 163 | S_SvREFCNT_inc(SV *sv) |
| 164 | { |
| 165 | if (LIKELY(sv != NULL)) |
| 166 | SvREFCNT(sv)++; |
| 167 | return sv; |
| 168 | } |
| 169 | PERL_STATIC_INLINE SV * |
| 170 | S_SvREFCNT_inc_NN(SV *sv) |
| 171 | { |
| 172 | SvREFCNT(sv)++; |
| 173 | return sv; |
| 174 | } |
| 175 | PERL_STATIC_INLINE void |
| 176 | S_SvREFCNT_inc_void(SV *sv) |
| 177 | { |
| 178 | if (LIKELY(sv != NULL)) |
| 179 | SvREFCNT(sv)++; |
| 180 | } |
| 181 | PERL_STATIC_INLINE void |
| 182 | S_SvREFCNT_dec(pTHX_ SV *sv) |
| 183 | { |
| 184 | if (LIKELY(sv != NULL)) { |
| 185 | U32 rc = SvREFCNT(sv); |
| 186 | if (LIKELY(rc > 1)) |
| 187 | SvREFCNT(sv) = rc - 1; |
| 188 | else |
| 189 | Perl_sv_free2(aTHX_ sv, rc); |
| 190 | } |
| 191 | } |
| 192 | |
| 193 | PERL_STATIC_INLINE void |
| 194 | S_SvREFCNT_dec_NN(pTHX_ SV *sv) |
| 195 | { |
| 196 | U32 rc = SvREFCNT(sv); |
| 197 | if (LIKELY(rc > 1)) |
| 198 | SvREFCNT(sv) = rc - 1; |
| 199 | else |
| 200 | Perl_sv_free2(aTHX_ sv, rc); |
| 201 | } |
| 202 | |
| 203 | PERL_STATIC_INLINE void |
| 204 | SvAMAGIC_on(SV *sv) |
| 205 | { |
| 206 | assert(SvROK(sv)); |
| 207 | if (SvOBJECT(SvRV(sv))) HvAMAGIC_on(SvSTASH(SvRV(sv))); |
| 208 | } |
| 209 | PERL_STATIC_INLINE void |
| 210 | SvAMAGIC_off(SV *sv) |
| 211 | { |
| 212 | if (SvROK(sv) && SvOBJECT(SvRV(sv))) |
| 213 | HvAMAGIC_off(SvSTASH(SvRV(sv))); |
| 214 | } |
| 215 | |
| 216 | PERL_STATIC_INLINE U32 |
| 217 | S_SvPADSTALE_on(SV *sv) |
| 218 | { |
| 219 | assert(!(SvFLAGS(sv) & SVs_PADTMP)); |
| 220 | return SvFLAGS(sv) |= SVs_PADSTALE; |
| 221 | } |
| 222 | PERL_STATIC_INLINE U32 |
| 223 | S_SvPADSTALE_off(SV *sv) |
| 224 | { |
| 225 | assert(!(SvFLAGS(sv) & SVs_PADTMP)); |
| 226 | return SvFLAGS(sv) &= ~SVs_PADSTALE; |
| 227 | } |
| 228 | #if defined(PERL_CORE) || defined (PERL_EXT) |
| 229 | PERL_STATIC_INLINE STRLEN |
| 230 | S_sv_or_pv_pos_u2b(pTHX_ SV *sv, const char *pv, STRLEN pos, STRLEN *lenp) |
| 231 | { |
| 232 | PERL_ARGS_ASSERT_SV_OR_PV_POS_U2B; |
| 233 | if (SvGAMAGIC(sv)) { |
| 234 | U8 *hopped = utf8_hop((U8 *)pv, pos); |
| 235 | if (lenp) *lenp = (STRLEN)(utf8_hop(hopped, *lenp) - hopped); |
| 236 | return (STRLEN)(hopped - (U8 *)pv); |
| 237 | } |
| 238 | return sv_pos_u2b_flags(sv,pos,lenp,SV_CONST_RETURN); |
| 239 | } |
| 240 | #endif |
| 241 | |
| 242 | /* ------------------------------- handy.h ------------------------------- */ |
| 243 | |
| 244 | /* saves machine code for a common noreturn idiom typically used in Newx*() */ |
| 245 | #ifdef GCC_DIAG_PRAGMA |
| 246 | GCC_DIAG_IGNORE(-Wunused-function) /* Intentionally left semicolonless. */ |
| 247 | #endif |
| 248 | static void |
| 249 | S_croak_memory_wrap(void) |
| 250 | { |
| 251 | Perl_croak_nocontext("%s",PL_memory_wrap); |
| 252 | } |
| 253 | #ifdef GCC_DIAG_PRAGMA |
| 254 | GCC_DIAG_RESTORE /* Intentionally left semicolonless. */ |
| 255 | #endif |
| 256 | |
| 257 | /* ------------------------------- utf8.h ------------------------------- */ |
| 258 | |
| 259 | /* |
| 260 | =head1 Unicode Support |
| 261 | */ |
| 262 | |
| 263 | PERL_STATIC_INLINE void |
| 264 | S_append_utf8_from_native_byte(const U8 byte, U8** dest) |
| 265 | { |
| 266 | /* Takes an input 'byte' (Latin1 or EBCDIC) and appends it to the UTF-8 |
| 267 | * encoded string at '*dest', updating '*dest' to include it */ |
| 268 | |
| 269 | PERL_ARGS_ASSERT_APPEND_UTF8_FROM_NATIVE_BYTE; |
| 270 | |
| 271 | if (NATIVE_BYTE_IS_INVARIANT(byte)) |
| 272 | *(*dest)++ = byte; |
| 273 | else { |
| 274 | *(*dest)++ = UTF8_EIGHT_BIT_HI(byte); |
| 275 | *(*dest)++ = UTF8_EIGHT_BIT_LO(byte); |
| 276 | } |
| 277 | } |
| 278 | |
| 279 | /* |
| 280 | =for apidoc valid_utf8_to_uvchr |
| 281 | Like L</utf8_to_uvchr_buf>(), but should only be called when it is known that |
| 282 | the next character in the input UTF-8 string C<s> is well-formed (I<e.g.>, |
| 283 | it passes C<L</isUTF8_CHAR>>. Surrogates, non-character code points, and |
| 284 | non-Unicode code points are allowed. |
| 285 | |
| 286 | =cut |
| 287 | |
| 288 | */ |
| 289 | |
| 290 | PERL_STATIC_INLINE UV |
| 291 | Perl_valid_utf8_to_uvchr(const U8 *s, STRLEN *retlen) |
| 292 | { |
| 293 | UV expectlen = UTF8SKIP(s); |
| 294 | const U8* send = s + expectlen; |
| 295 | UV uv = *s; |
| 296 | |
| 297 | PERL_ARGS_ASSERT_VALID_UTF8_TO_UVCHR; |
| 298 | |
| 299 | if (retlen) { |
| 300 | *retlen = expectlen; |
| 301 | } |
| 302 | |
| 303 | /* An invariant is trivially returned */ |
| 304 | if (expectlen == 1) { |
| 305 | return uv; |
| 306 | } |
| 307 | |
| 308 | /* Remove the leading bits that indicate the number of bytes, leaving just |
| 309 | * the bits that are part of the value */ |
| 310 | uv = NATIVE_UTF8_TO_I8(uv) & UTF_START_MASK(expectlen); |
| 311 | |
| 312 | /* Now, loop through the remaining bytes, accumulating each into the |
| 313 | * working total as we go. (I khw tried unrolling the loop for up to 4 |
| 314 | * bytes, but there was no performance improvement) */ |
| 315 | for (++s; s < send; s++) { |
| 316 | uv = UTF8_ACCUMULATE(uv, *s); |
| 317 | } |
| 318 | |
| 319 | return UNI_TO_NATIVE(uv); |
| 320 | |
| 321 | } |
| 322 | |
| 323 | /* |
| 324 | =for apidoc is_utf8_invariant_string |
| 325 | |
| 326 | Returns true iff the first C<len> bytes of the string C<s> are the same |
| 327 | regardless of the UTF-8 encoding of the string (or UTF-EBCDIC encoding on |
| 328 | EBCDIC machines). That is, if they are UTF-8 invariant. On ASCII-ish |
| 329 | machines, all the ASCII characters and only the ASCII characters fit this |
| 330 | definition. On EBCDIC machines, the ASCII-range characters are invariant, but |
| 331 | so also are the C1 controls and C<\c?> (which isn't in the ASCII range on |
| 332 | EBCDIC). |
| 333 | |
| 334 | If C<len> is 0, it will be calculated using C<strlen(s)>, (which means if you |
| 335 | use this option, that C<s> can't have embedded C<NUL> characters and has to |
| 336 | have a terminating C<NUL> byte). |
| 337 | |
| 338 | See also L</is_utf8_string>(), L</is_utf8_string_loclen>(), and |
| 339 | L</is_utf8_string_loc>(). |
| 340 | |
| 341 | =cut |
| 342 | */ |
| 343 | |
| 344 | PERL_STATIC_INLINE bool |
| 345 | S_is_utf8_invariant_string(const U8* const s, const STRLEN len) |
| 346 | { |
| 347 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
| 348 | const U8* x = s; |
| 349 | |
| 350 | PERL_ARGS_ASSERT_IS_UTF8_INVARIANT_STRING; |
| 351 | |
| 352 | for (; x < send; ++x) { |
| 353 | if (!UTF8_IS_INVARIANT(*x)) |
| 354 | return FALSE; |
| 355 | } |
| 356 | |
| 357 | return TRUE; |
| 358 | } |
| 359 | |
| 360 | /* |
| 361 | =for apidoc is_utf8_string |
| 362 | |
| 363 | Returns true if the first C<len> bytes of string C<s> form a valid |
| 364 | UTF-8 string, false otherwise. If C<len> is 0, it will be calculated |
| 365 | using C<strlen(s)> (which means if you use this option, that C<s> can't have |
| 366 | embedded C<NUL> characters and has to have a terminating C<NUL> byte). Note |
| 367 | that all characters being ASCII constitute 'a valid UTF-8 string'. |
| 368 | |
| 369 | See also L</is_utf8_invariant_string>(), L</is_utf8_string_loclen>(), and |
| 370 | L</is_utf8_string_loc>(). |
| 371 | |
| 372 | =cut |
| 373 | */ |
| 374 | |
| 375 | bool |
| 376 | Perl_is_utf8_string(const U8 *s, STRLEN len) |
| 377 | { |
| 378 | /* This is now marked pure in embed.fnc, because isUTF8_CHAR now is pure. |
| 379 | * Be aware of possible changes to that */ |
| 380 | |
| 381 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
| 382 | const U8* x = s; |
| 383 | |
| 384 | PERL_ARGS_ASSERT_IS_UTF8_STRING; |
| 385 | |
| 386 | while (x < send) { |
| 387 | STRLEN len = isUTF8_CHAR(x, send); |
| 388 | if (UNLIKELY(! len)) { |
| 389 | return FALSE; |
| 390 | } |
| 391 | x += len; |
| 392 | } |
| 393 | |
| 394 | return TRUE; |
| 395 | } |
| 396 | |
| 397 | /* |
| 398 | Implemented as a macro in utf8.h |
| 399 | |
| 400 | =for apidoc is_utf8_string_loc |
| 401 | |
| 402 | Like L</is_utf8_string> but stores the location of the failure (in the |
| 403 | case of "utf8ness failure") or the location C<s>+C<len> (in the case of |
| 404 | "utf8ness success") in the C<ep>. |
| 405 | |
| 406 | See also L</is_utf8_string_loclen>() and L</is_utf8_string>(). |
| 407 | |
| 408 | =for apidoc is_utf8_string_loclen |
| 409 | |
| 410 | Like L</is_utf8_string>() but stores the location of the failure (in the |
| 411 | case of "utf8ness failure") or the location C<s>+C<len> (in the case of |
| 412 | "utf8ness success") in the C<ep>, and the number of UTF-8 |
| 413 | encoded characters in the C<el>. |
| 414 | |
| 415 | See also L</is_utf8_string_loc>() and L</is_utf8_string>(). |
| 416 | |
| 417 | =cut |
| 418 | */ |
| 419 | |
| 420 | bool |
| 421 | Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el) |
| 422 | { |
| 423 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
| 424 | const U8* x = s; |
| 425 | STRLEN outlen = 0; |
| 426 | |
| 427 | PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN; |
| 428 | |
| 429 | while (x < send) { |
| 430 | STRLEN len = isUTF8_CHAR(x, send); |
| 431 | if (UNLIKELY(! len)) { |
| 432 | break; |
| 433 | } |
| 434 | x += len; |
| 435 | outlen++; |
| 436 | } |
| 437 | |
| 438 | if (el) |
| 439 | *el = outlen; |
| 440 | |
| 441 | if (ep) { |
| 442 | *ep = x; |
| 443 | } |
| 444 | |
| 445 | return (x == send); |
| 446 | } |
| 447 | |
| 448 | /* |
| 449 | =for apidoc utf8_distance |
| 450 | |
| 451 | Returns the number of UTF-8 characters between the UTF-8 pointers C<a> |
| 452 | and C<b>. |
| 453 | |
| 454 | WARNING: use only if you *know* that the pointers point inside the |
| 455 | same UTF-8 buffer. |
| 456 | |
| 457 | =cut |
| 458 | */ |
| 459 | |
| 460 | PERL_STATIC_INLINE IV |
| 461 | Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b) |
| 462 | { |
| 463 | PERL_ARGS_ASSERT_UTF8_DISTANCE; |
| 464 | |
| 465 | return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a); |
| 466 | } |
| 467 | |
| 468 | /* |
| 469 | =for apidoc utf8_hop |
| 470 | |
| 471 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
| 472 | forward or backward. |
| 473 | |
| 474 | WARNING: do not use the following unless you *know* C<off> is within |
| 475 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
| 476 | on the first byte of character or just after the last byte of a character. |
| 477 | |
| 478 | =cut |
| 479 | */ |
| 480 | |
| 481 | PERL_STATIC_INLINE U8 * |
| 482 | Perl_utf8_hop(const U8 *s, SSize_t off) |
| 483 | { |
| 484 | PERL_ARGS_ASSERT_UTF8_HOP; |
| 485 | |
| 486 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
| 487 | * the bitops (especially ~) can create illegal UTF-8. |
| 488 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
| 489 | |
| 490 | if (off >= 0) { |
| 491 | while (off--) |
| 492 | s += UTF8SKIP(s); |
| 493 | } |
| 494 | else { |
| 495 | while (off++) { |
| 496 | s--; |
| 497 | while (UTF8_IS_CONTINUATION(*s)) |
| 498 | s--; |
| 499 | } |
| 500 | } |
| 501 | return (U8 *)s; |
| 502 | } |
| 503 | |
| 504 | /* |
| 505 | |
| 506 | =for apidoc is_utf8_valid_partial_char |
| 507 | |
| 508 | Returns 1 if the first few bytes of the string starting at C<s> and |
| 509 | looking no further than S<C<e - 1>> are legal initial bytes of |
| 510 | well-formed UTF-8; otherwise it returns 0. |
| 511 | |
| 512 | This is useful when some fixed-length buffer is being tested for being |
| 513 | well-formed UTF-8, but the final few bytes in it are fewer than necessary to |
| 514 | fully determine what code point they might mean. (Presumably when the buffer |
| 515 | is refreshed with the next chunk of data, the new first bytes will complete the |
| 516 | partial code point.) This function is used to verify that the final bytes in |
| 517 | the current buffer are in fact the legal beginning of some code point, so that |
| 518 | if they aren't, the failure can be signalled without having to wait for the |
| 519 | next read. |
| 520 | |
| 521 | If the bytes terminated at S<C<e - 1>> are a full character (or more), 0 is |
| 522 | returned. |
| 523 | |
| 524 | =cut |
| 525 | */ |
| 526 | bool |
| 527 | S_is_utf8_valid_partial_char(const U8 * const s, const U8 * const e) |
| 528 | { |
| 529 | |
| 530 | PERL_ARGS_ASSERT_IS_UTF8_VALID_PARTIAL_CHAR; |
| 531 | |
| 532 | if (s >= e || s + UTF8SKIP(s) >= e) { |
| 533 | return FALSE; |
| 534 | } |
| 535 | |
| 536 | return cBOOL(_is_utf8_char_slow(s, e - s)); |
| 537 | } |
| 538 | |
| 539 | /* ------------------------------- perl.h ----------------------------- */ |
| 540 | |
| 541 | /* |
| 542 | =head1 Miscellaneous Functions |
| 543 | |
| 544 | =for apidoc AiR|bool|is_safe_syscall|const char *pv|STRLEN len|const char *what|const char *op_name |
| 545 | |
| 546 | Test that the given C<pv> doesn't contain any internal C<NUL> characters. |
| 547 | If it does, set C<errno> to C<ENOENT>, optionally warn, and return FALSE. |
| 548 | |
| 549 | Return TRUE if the name is safe. |
| 550 | |
| 551 | Used by the C<IS_SAFE_SYSCALL()> macro. |
| 552 | |
| 553 | =cut |
| 554 | */ |
| 555 | |
| 556 | PERL_STATIC_INLINE bool |
| 557 | S_is_safe_syscall(pTHX_ const char *pv, STRLEN len, const char *what, const char *op_name) { |
| 558 | /* While the Windows CE API provides only UCS-16 (or UTF-16) APIs |
| 559 | * perl itself uses xce*() functions which accept 8-bit strings. |
| 560 | */ |
| 561 | |
| 562 | PERL_ARGS_ASSERT_IS_SAFE_SYSCALL; |
| 563 | |
| 564 | if (len > 1) { |
| 565 | char *null_at; |
| 566 | if (UNLIKELY((null_at = (char *)memchr(pv, 0, len-1)) != NULL)) { |
| 567 | SETERRNO(ENOENT, LIB_INVARG); |
| 568 | Perl_ck_warner(aTHX_ packWARN(WARN_SYSCALLS), |
| 569 | "Invalid \\0 character in %s for %s: %s\\0%s", |
| 570 | what, op_name, pv, null_at+1); |
| 571 | return FALSE; |
| 572 | } |
| 573 | } |
| 574 | |
| 575 | return TRUE; |
| 576 | } |
| 577 | |
| 578 | /* |
| 579 | |
| 580 | Return true if the supplied filename has a newline character |
| 581 | immediately before the first (hopefully only) NUL. |
| 582 | |
| 583 | My original look at this incorrectly used the len from SvPV(), but |
| 584 | that's incorrect, since we allow for a NUL in pv[len-1]. |
| 585 | |
| 586 | So instead, strlen() and work from there. |
| 587 | |
| 588 | This allow for the user reading a filename, forgetting to chomp it, |
| 589 | then calling: |
| 590 | |
| 591 | open my $foo, "$file\0"; |
| 592 | |
| 593 | */ |
| 594 | |
| 595 | #ifdef PERL_CORE |
| 596 | |
| 597 | PERL_STATIC_INLINE bool |
| 598 | S_should_warn_nl(const char *pv) { |
| 599 | STRLEN len; |
| 600 | |
| 601 | PERL_ARGS_ASSERT_SHOULD_WARN_NL; |
| 602 | |
| 603 | len = strlen(pv); |
| 604 | |
| 605 | return len > 0 && pv[len-1] == '\n'; |
| 606 | } |
| 607 | |
| 608 | #endif |
| 609 | |
| 610 | /* ------------------ pp.c, regcomp.c, toke.c, universal.c ------------ */ |
| 611 | |
| 612 | #define MAX_CHARSET_NAME_LENGTH 2 |
| 613 | |
| 614 | PERL_STATIC_INLINE const char * |
| 615 | get_regex_charset_name(const U32 flags, STRLEN* const lenp) |
| 616 | { |
| 617 | /* Returns a string that corresponds to the name of the regex character set |
| 618 | * given by 'flags', and *lenp is set the length of that string, which |
| 619 | * cannot exceed MAX_CHARSET_NAME_LENGTH characters */ |
| 620 | |
| 621 | *lenp = 1; |
| 622 | switch (get_regex_charset(flags)) { |
| 623 | case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS; |
| 624 | case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS; |
| 625 | case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS; |
| 626 | case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS; |
| 627 | case REGEX_ASCII_MORE_RESTRICTED_CHARSET: |
| 628 | *lenp = 2; |
| 629 | return ASCII_MORE_RESTRICT_PAT_MODS; |
| 630 | } |
| 631 | /* The NOT_REACHED; hides an assert() which has a rather complex |
| 632 | * definition in perl.h. */ |
| 633 | NOT_REACHED; /* NOTREACHED */ |
| 634 | return "?"; /* Unknown */ |
| 635 | } |
| 636 | |
| 637 | /* |
| 638 | |
| 639 | Return false if any get magic is on the SV other than taint magic. |
| 640 | |
| 641 | */ |
| 642 | |
| 643 | PERL_STATIC_INLINE bool |
| 644 | S_sv_only_taint_gmagic(SV *sv) { |
| 645 | MAGIC *mg = SvMAGIC(sv); |
| 646 | |
| 647 | PERL_ARGS_ASSERT_SV_ONLY_TAINT_GMAGIC; |
| 648 | |
| 649 | while (mg) { |
| 650 | if (mg->mg_type != PERL_MAGIC_taint |
| 651 | && !(mg->mg_flags & MGf_GSKIP) |
| 652 | && mg->mg_virtual->svt_get) { |
| 653 | return FALSE; |
| 654 | } |
| 655 | mg = mg->mg_moremagic; |
| 656 | } |
| 657 | |
| 658 | return TRUE; |
| 659 | } |
| 660 | |
| 661 | /* ------------------ cop.h ------------------------------------------- */ |
| 662 | |
| 663 | |
| 664 | /* Enter a block. Push a new base context and return its address. */ |
| 665 | |
| 666 | PERL_STATIC_INLINE PERL_CONTEXT * |
| 667 | S_cx_pushblock(pTHX_ U8 type, U8 gimme, SV** sp, I32 saveix) |
| 668 | { |
| 669 | PERL_CONTEXT * cx; |
| 670 | |
| 671 | PERL_ARGS_ASSERT_CX_PUSHBLOCK; |
| 672 | |
| 673 | CXINC; |
| 674 | cx = CX_CUR(); |
| 675 | cx->cx_type = type; |
| 676 | cx->blk_gimme = gimme; |
| 677 | cx->blk_oldsaveix = saveix; |
| 678 | cx->blk_oldsp = (I32)(sp - PL_stack_base); |
| 679 | cx->blk_oldcop = PL_curcop; |
| 680 | cx->blk_oldmarksp = (I32)(PL_markstack_ptr - PL_markstack); |
| 681 | cx->blk_oldscopesp = PL_scopestack_ix; |
| 682 | cx->blk_oldpm = PL_curpm; |
| 683 | cx->blk_old_tmpsfloor = PL_tmps_floor; |
| 684 | |
| 685 | PL_tmps_floor = PL_tmps_ix; |
| 686 | CX_DEBUG(cx, "PUSH"); |
| 687 | return cx; |
| 688 | } |
| 689 | |
| 690 | |
| 691 | /* Exit a block (RETURN and LAST). */ |
| 692 | |
| 693 | PERL_STATIC_INLINE void |
| 694 | S_cx_popblock(pTHX_ PERL_CONTEXT *cx) |
| 695 | { |
| 696 | PERL_ARGS_ASSERT_CX_POPBLOCK; |
| 697 | |
| 698 | CX_DEBUG(cx, "POP"); |
| 699 | /* these 3 are common to cx_popblock and cx_topblock */ |
| 700 | PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp; |
| 701 | PL_scopestack_ix = cx->blk_oldscopesp; |
| 702 | PL_curpm = cx->blk_oldpm; |
| 703 | |
| 704 | /* LEAVE_SCOPE() should have made this true. /(?{})/ cheats |
| 705 | * and leaves a CX entry lying around for repeated use, so |
| 706 | * skip for multicall */ \ |
| 707 | assert( (CxTYPE(cx) == CXt_SUB && CxMULTICALL(cx)) |
| 708 | || PL_savestack_ix == cx->blk_oldsaveix); |
| 709 | PL_curcop = cx->blk_oldcop; |
| 710 | PL_tmps_floor = cx->blk_old_tmpsfloor; |
| 711 | } |
| 712 | |
| 713 | /* Continue a block elsewhere (e.g. NEXT, REDO, GOTO). |
| 714 | * Whereas cx_popblock() restores the state to the point just before |
| 715 | * cx_pushblock() was called, cx_topblock() restores it to the point just |
| 716 | * *after* cx_pushblock() was called. */ |
| 717 | |
| 718 | PERL_STATIC_INLINE void |
| 719 | S_cx_topblock(pTHX_ PERL_CONTEXT *cx) |
| 720 | { |
| 721 | PERL_ARGS_ASSERT_CX_TOPBLOCK; |
| 722 | |
| 723 | CX_DEBUG(cx, "TOP"); |
| 724 | /* these 3 are common to cx_popblock and cx_topblock */ |
| 725 | PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp; |
| 726 | PL_scopestack_ix = cx->blk_oldscopesp; |
| 727 | PL_curpm = cx->blk_oldpm; |
| 728 | |
| 729 | PL_stack_sp = PL_stack_base + cx->blk_oldsp; |
| 730 | } |
| 731 | |
| 732 | |
| 733 | PERL_STATIC_INLINE void |
| 734 | S_cx_pushsub(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, bool hasargs) |
| 735 | { |
| 736 | U8 phlags = CX_PUSHSUB_GET_LVALUE_MASK(Perl_was_lvalue_sub); |
| 737 | |
| 738 | PERL_ARGS_ASSERT_CX_PUSHSUB; |
| 739 | |
| 740 | PERL_DTRACE_PROBE_ENTRY(cv); |
| 741 | cx->blk_sub.cv = cv; |
| 742 | cx->blk_sub.olddepth = CvDEPTH(cv); |
| 743 | cx->blk_sub.prevcomppad = PL_comppad; |
| 744 | cx->cx_type |= (hasargs) ? CXp_HASARGS : 0; |
| 745 | cx->blk_sub.retop = retop; |
| 746 | SvREFCNT_inc_simple_void_NN(cv); |
| 747 | cx->blk_u16 = PL_op->op_private & (phlags|OPpDEREF); |
| 748 | } |
| 749 | |
| 750 | |
| 751 | /* subsets of cx_popsub() */ |
| 752 | |
| 753 | PERL_STATIC_INLINE void |
| 754 | S_cx_popsub_common(pTHX_ PERL_CONTEXT *cx) |
| 755 | { |
| 756 | CV *cv; |
| 757 | |
| 758 | PERL_ARGS_ASSERT_CX_POPSUB_COMMON; |
| 759 | assert(CxTYPE(cx) == CXt_SUB); |
| 760 | |
| 761 | PL_comppad = cx->blk_sub.prevcomppad; |
| 762 | PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL; |
| 763 | cv = cx->blk_sub.cv; |
| 764 | CvDEPTH(cv) = cx->blk_sub.olddepth; |
| 765 | cx->blk_sub.cv = NULL; |
| 766 | SvREFCNT_dec(cv); |
| 767 | } |
| 768 | |
| 769 | |
| 770 | /* handle the @_ part of leaving a sub */ |
| 771 | |
| 772 | PERL_STATIC_INLINE void |
| 773 | S_cx_popsub_args(pTHX_ PERL_CONTEXT *cx) |
| 774 | { |
| 775 | AV *av; |
| 776 | |
| 777 | PERL_ARGS_ASSERT_CX_POPSUB_ARGS; |
| 778 | assert(CxTYPE(cx) == CXt_SUB); |
| 779 | assert(AvARRAY(MUTABLE_AV( |
| 780 | PadlistARRAY(CvPADLIST(cx->blk_sub.cv))[ |
| 781 | CvDEPTH(cx->blk_sub.cv)])) == PL_curpad); |
| 782 | |
| 783 | CX_POP_SAVEARRAY(cx); |
| 784 | av = MUTABLE_AV(PAD_SVl(0)); |
| 785 | if (UNLIKELY(AvREAL(av))) |
| 786 | /* abandon @_ if it got reified */ |
| 787 | clear_defarray(av, 0); |
| 788 | else { |
| 789 | CLEAR_ARGARRAY(av); |
| 790 | } |
| 791 | } |
| 792 | |
| 793 | |
| 794 | PERL_STATIC_INLINE void |
| 795 | S_cx_popsub(pTHX_ PERL_CONTEXT *cx) |
| 796 | { |
| 797 | PERL_ARGS_ASSERT_CX_POPSUB; |
| 798 | assert(CxTYPE(cx) == CXt_SUB); |
| 799 | |
| 800 | PERL_DTRACE_PROBE_RETURN(cx->blk_sub.cv); |
| 801 | |
| 802 | if (CxHASARGS(cx)) |
| 803 | cx_popsub_args(cx); |
| 804 | cx_popsub_common(cx); |
| 805 | } |
| 806 | |
| 807 | |
| 808 | PERL_STATIC_INLINE void |
| 809 | S_cx_pushformat(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, GV *gv) |
| 810 | { |
| 811 | PERL_ARGS_ASSERT_CX_PUSHFORMAT; |
| 812 | |
| 813 | cx->blk_format.cv = cv; |
| 814 | cx->blk_format.retop = retop; |
| 815 | cx->blk_format.gv = gv; |
| 816 | cx->blk_format.dfoutgv = PL_defoutgv; |
| 817 | cx->blk_format.prevcomppad = PL_comppad; |
| 818 | cx->blk_u16 = 0; |
| 819 | |
| 820 | SvREFCNT_inc_simple_void_NN(cv); |
| 821 | CvDEPTH(cv)++; |
| 822 | SvREFCNT_inc_void(cx->blk_format.dfoutgv); |
| 823 | } |
| 824 | |
| 825 | |
| 826 | PERL_STATIC_INLINE void |
| 827 | S_cx_popformat(pTHX_ PERL_CONTEXT *cx) |
| 828 | { |
| 829 | CV *cv; |
| 830 | GV *dfout; |
| 831 | |
| 832 | PERL_ARGS_ASSERT_CX_POPFORMAT; |
| 833 | assert(CxTYPE(cx) == CXt_FORMAT); |
| 834 | |
| 835 | dfout = cx->blk_format.dfoutgv; |
| 836 | setdefout(dfout); |
| 837 | cx->blk_format.dfoutgv = NULL; |
| 838 | SvREFCNT_dec_NN(dfout); |
| 839 | |
| 840 | PL_comppad = cx->blk_format.prevcomppad; |
| 841 | PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL; |
| 842 | cv = cx->blk_format.cv; |
| 843 | cx->blk_format.cv = NULL; |
| 844 | --CvDEPTH(cv); |
| 845 | SvREFCNT_dec_NN(cv); |
| 846 | } |
| 847 | |
| 848 | |
| 849 | PERL_STATIC_INLINE void |
| 850 | S_cx_pusheval(pTHX_ PERL_CONTEXT *cx, OP *retop, SV *namesv) |
| 851 | { |
| 852 | PERL_ARGS_ASSERT_CX_PUSHEVAL; |
| 853 | |
| 854 | cx->blk_eval.retop = retop; |
| 855 | cx->blk_eval.old_namesv = namesv; |
| 856 | cx->blk_eval.old_eval_root = PL_eval_root; |
| 857 | cx->blk_eval.cur_text = PL_parser ? PL_parser->linestr : NULL; |
| 858 | cx->blk_eval.cv = NULL; /* later set by doeval_compile() */ |
| 859 | cx->blk_eval.cur_top_env = PL_top_env; |
| 860 | |
| 861 | assert(!(PL_in_eval & ~ 0x7F)); |
| 862 | assert(!(PL_op->op_type & ~0x1FF)); |
| 863 | cx->blk_u16 = (PL_in_eval & 0x7F) | ((U16)PL_op->op_type << 7); |
| 864 | } |
| 865 | |
| 866 | |
| 867 | PERL_STATIC_INLINE void |
| 868 | S_cx_popeval(pTHX_ PERL_CONTEXT *cx) |
| 869 | { |
| 870 | SV *sv; |
| 871 | |
| 872 | PERL_ARGS_ASSERT_CX_POPEVAL; |
| 873 | assert(CxTYPE(cx) == CXt_EVAL); |
| 874 | |
| 875 | PL_in_eval = CxOLD_IN_EVAL(cx); |
| 876 | PL_eval_root = cx->blk_eval.old_eval_root; |
| 877 | sv = cx->blk_eval.cur_text; |
| 878 | if (sv && SvSCREAM(sv)) { |
| 879 | cx->blk_eval.cur_text = NULL; |
| 880 | SvREFCNT_dec_NN(sv); |
| 881 | } |
| 882 | |
| 883 | sv = cx->blk_eval.old_namesv; |
| 884 | if (sv) { |
| 885 | cx->blk_eval.old_namesv = NULL; |
| 886 | SvREFCNT_dec_NN(sv); |
| 887 | } |
| 888 | } |
| 889 | |
| 890 | |
| 891 | /* push a plain loop, i.e. |
| 892 | * { block } |
| 893 | * while (cond) { block } |
| 894 | * for (init;cond;continue) { block } |
| 895 | * This loop can be last/redo'ed etc. |
| 896 | */ |
| 897 | |
| 898 | PERL_STATIC_INLINE void |
| 899 | S_cx_pushloop_plain(pTHX_ PERL_CONTEXT *cx) |
| 900 | { |
| 901 | PERL_ARGS_ASSERT_CX_PUSHLOOP_PLAIN; |
| 902 | cx->blk_loop.my_op = cLOOP; |
| 903 | } |
| 904 | |
| 905 | |
| 906 | /* push a true for loop, i.e. |
| 907 | * for var (list) { block } |
| 908 | */ |
| 909 | |
| 910 | PERL_STATIC_INLINE void |
| 911 | S_cx_pushloop_for(pTHX_ PERL_CONTEXT *cx, void *itervarp, SV* itersave) |
| 912 | { |
| 913 | PERL_ARGS_ASSERT_CX_PUSHLOOP_FOR; |
| 914 | |
| 915 | /* this one line is common with cx_pushloop_plain */ |
| 916 | cx->blk_loop.my_op = cLOOP; |
| 917 | |
| 918 | cx->blk_loop.itervar_u.svp = (SV**)itervarp; |
| 919 | cx->blk_loop.itersave = itersave; |
| 920 | #ifdef USE_ITHREADS |
| 921 | cx->blk_loop.oldcomppad = PL_comppad; |
| 922 | #endif |
| 923 | } |
| 924 | |
| 925 | |
| 926 | /* pop all loop types, including plain */ |
| 927 | |
| 928 | PERL_STATIC_INLINE void |
| 929 | S_cx_poploop(pTHX_ PERL_CONTEXT *cx) |
| 930 | { |
| 931 | PERL_ARGS_ASSERT_CX_POPLOOP; |
| 932 | |
| 933 | assert(CxTYPE_is_LOOP(cx)); |
| 934 | if ( CxTYPE(cx) == CXt_LOOP_ARY |
| 935 | || CxTYPE(cx) == CXt_LOOP_LAZYSV) |
| 936 | { |
| 937 | /* Free ary or cur. This assumes that state_u.ary.ary |
| 938 | * aligns with state_u.lazysv.cur. See cx_dup() */ |
| 939 | SV *sv = cx->blk_loop.state_u.lazysv.cur; |
| 940 | cx->blk_loop.state_u.lazysv.cur = NULL; |
| 941 | SvREFCNT_dec_NN(sv); |
| 942 | if (CxTYPE(cx) == CXt_LOOP_LAZYSV) { |
| 943 | sv = cx->blk_loop.state_u.lazysv.end; |
| 944 | cx->blk_loop.state_u.lazysv.end = NULL; |
| 945 | SvREFCNT_dec_NN(sv); |
| 946 | } |
| 947 | } |
| 948 | if (cx->cx_type & (CXp_FOR_PAD|CXp_FOR_GV)) { |
| 949 | SV *cursv; |
| 950 | SV **svp = (cx)->blk_loop.itervar_u.svp; |
| 951 | if ((cx->cx_type & CXp_FOR_GV)) |
| 952 | svp = &GvSV((GV*)svp); |
| 953 | cursv = *svp; |
| 954 | *svp = cx->blk_loop.itersave; |
| 955 | cx->blk_loop.itersave = NULL; |
| 956 | SvREFCNT_dec(cursv); |
| 957 | } |
| 958 | } |
| 959 | |
| 960 | |
| 961 | PERL_STATIC_INLINE void |
| 962 | S_cx_pushwhen(pTHX_ PERL_CONTEXT *cx) |
| 963 | { |
| 964 | PERL_ARGS_ASSERT_CX_PUSHWHEN; |
| 965 | |
| 966 | cx->blk_givwhen.leave_op = cLOGOP->op_other; |
| 967 | } |
| 968 | |
| 969 | |
| 970 | PERL_STATIC_INLINE void |
| 971 | S_cx_popwhen(pTHX_ PERL_CONTEXT *cx) |
| 972 | { |
| 973 | PERL_ARGS_ASSERT_CX_POPWHEN; |
| 974 | assert(CxTYPE(cx) == CXt_WHEN); |
| 975 | |
| 976 | PERL_UNUSED_ARG(cx); |
| 977 | PERL_UNUSED_CONTEXT; |
| 978 | /* currently NOOP */ |
| 979 | } |
| 980 | |
| 981 | |
| 982 | PERL_STATIC_INLINE void |
| 983 | S_cx_pushgiven(pTHX_ PERL_CONTEXT *cx, SV *orig_defsv) |
| 984 | { |
| 985 | PERL_ARGS_ASSERT_CX_PUSHGIVEN; |
| 986 | |
| 987 | cx->blk_givwhen.leave_op = cLOGOP->op_other; |
| 988 | cx->blk_givwhen.defsv_save = orig_defsv; |
| 989 | } |
| 990 | |
| 991 | |
| 992 | PERL_STATIC_INLINE void |
| 993 | S_cx_popgiven(pTHX_ PERL_CONTEXT *cx) |
| 994 | { |
| 995 | SV *sv; |
| 996 | |
| 997 | PERL_ARGS_ASSERT_CX_POPGIVEN; |
| 998 | assert(CxTYPE(cx) == CXt_GIVEN); |
| 999 | |
| 1000 | sv = GvSV(PL_defgv); |
| 1001 | GvSV(PL_defgv) = cx->blk_givwhen.defsv_save; |
| 1002 | cx->blk_givwhen.defsv_save = NULL; |
| 1003 | SvREFCNT_dec(sv); |
| 1004 | } |
| 1005 | |
| 1006 | /* |
| 1007 | * ex: set ts=8 sts=4 sw=4 et: |
| 1008 | */ |