| 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 | /* ----------------------------- regexp.h ----------------------------- */ |
| 129 | |
| 130 | PERL_STATIC_INLINE struct regexp * |
| 131 | S_ReANY(const REGEXP * const re) |
| 132 | { |
| 133 | assert(isREGEXP(re)); |
| 134 | return re->sv_u.svu_rx; |
| 135 | } |
| 136 | |
| 137 | /* ------------------------------- sv.h ------------------------------- */ |
| 138 | |
| 139 | PERL_STATIC_INLINE SV * |
| 140 | S_SvREFCNT_inc(SV *sv) |
| 141 | { |
| 142 | if (LIKELY(sv != NULL)) |
| 143 | SvREFCNT(sv)++; |
| 144 | return sv; |
| 145 | } |
| 146 | PERL_STATIC_INLINE SV * |
| 147 | S_SvREFCNT_inc_NN(SV *sv) |
| 148 | { |
| 149 | SvREFCNT(sv)++; |
| 150 | return sv; |
| 151 | } |
| 152 | PERL_STATIC_INLINE void |
| 153 | S_SvREFCNT_inc_void(SV *sv) |
| 154 | { |
| 155 | if (LIKELY(sv != NULL)) |
| 156 | SvREFCNT(sv)++; |
| 157 | } |
| 158 | PERL_STATIC_INLINE void |
| 159 | S_SvREFCNT_dec(pTHX_ SV *sv) |
| 160 | { |
| 161 | if (LIKELY(sv != NULL)) { |
| 162 | U32 rc = SvREFCNT(sv); |
| 163 | if (LIKELY(rc > 1)) |
| 164 | SvREFCNT(sv) = rc - 1; |
| 165 | else |
| 166 | Perl_sv_free2(aTHX_ sv, rc); |
| 167 | } |
| 168 | } |
| 169 | |
| 170 | PERL_STATIC_INLINE void |
| 171 | S_SvREFCNT_dec_NN(pTHX_ SV *sv) |
| 172 | { |
| 173 | U32 rc = SvREFCNT(sv); |
| 174 | if (LIKELY(rc > 1)) |
| 175 | SvREFCNT(sv) = rc - 1; |
| 176 | else |
| 177 | Perl_sv_free2(aTHX_ sv, rc); |
| 178 | } |
| 179 | |
| 180 | PERL_STATIC_INLINE void |
| 181 | SvAMAGIC_on(SV *sv) |
| 182 | { |
| 183 | assert(SvROK(sv)); |
| 184 | if (SvOBJECT(SvRV(sv))) HvAMAGIC_on(SvSTASH(SvRV(sv))); |
| 185 | } |
| 186 | PERL_STATIC_INLINE void |
| 187 | SvAMAGIC_off(SV *sv) |
| 188 | { |
| 189 | if (SvROK(sv) && SvOBJECT(SvRV(sv))) |
| 190 | HvAMAGIC_off(SvSTASH(SvRV(sv))); |
| 191 | } |
| 192 | |
| 193 | PERL_STATIC_INLINE U32 |
| 194 | S_SvPADSTALE_on(SV *sv) |
| 195 | { |
| 196 | assert(!(SvFLAGS(sv) & SVs_PADTMP)); |
| 197 | return SvFLAGS(sv) |= SVs_PADSTALE; |
| 198 | } |
| 199 | PERL_STATIC_INLINE U32 |
| 200 | S_SvPADSTALE_off(SV *sv) |
| 201 | { |
| 202 | assert(!(SvFLAGS(sv) & SVs_PADTMP)); |
| 203 | return SvFLAGS(sv) &= ~SVs_PADSTALE; |
| 204 | } |
| 205 | #if defined(PERL_CORE) || defined (PERL_EXT) |
| 206 | PERL_STATIC_INLINE STRLEN |
| 207 | S_sv_or_pv_pos_u2b(pTHX_ SV *sv, const char *pv, STRLEN pos, STRLEN *lenp) |
| 208 | { |
| 209 | PERL_ARGS_ASSERT_SV_OR_PV_POS_U2B; |
| 210 | if (SvGAMAGIC(sv)) { |
| 211 | U8 *hopped = utf8_hop((U8 *)pv, pos); |
| 212 | if (lenp) *lenp = (STRLEN)(utf8_hop(hopped, *lenp) - hopped); |
| 213 | return (STRLEN)(hopped - (U8 *)pv); |
| 214 | } |
| 215 | return sv_pos_u2b_flags(sv,pos,lenp,SV_CONST_RETURN); |
| 216 | } |
| 217 | #endif |
| 218 | |
| 219 | /* ------------------------------- handy.h ------------------------------- */ |
| 220 | |
| 221 | /* saves machine code for a common noreturn idiom typically used in Newx*() */ |
| 222 | #ifdef GCC_DIAG_PRAGMA |
| 223 | GCC_DIAG_IGNORE(-Wunused-function) /* Intentionally left semicolonless. */ |
| 224 | #endif |
| 225 | static void |
| 226 | S_croak_memory_wrap(void) |
| 227 | { |
| 228 | Perl_croak_nocontext("%s",PL_memory_wrap); |
| 229 | } |
| 230 | #ifdef GCC_DIAG_PRAGMA |
| 231 | GCC_DIAG_RESTORE /* Intentionally left semicolonless. */ |
| 232 | #endif |
| 233 | |
| 234 | /* ------------------------------- utf8.h ------------------------------- */ |
| 235 | |
| 236 | PERL_STATIC_INLINE void |
| 237 | S_append_utf8_from_native_byte(const U8 byte, U8** dest) |
| 238 | { |
| 239 | /* Takes an input 'byte' (Latin1 or EBCDIC) and appends it to the UTF-8 |
| 240 | * encoded string at '*dest', updating '*dest' to include it */ |
| 241 | |
| 242 | PERL_ARGS_ASSERT_APPEND_UTF8_FROM_NATIVE_BYTE; |
| 243 | |
| 244 | if (NATIVE_BYTE_IS_INVARIANT(byte)) |
| 245 | *(*dest)++ = byte; |
| 246 | else { |
| 247 | *(*dest)++ = UTF8_EIGHT_BIT_HI(byte); |
| 248 | *(*dest)++ = UTF8_EIGHT_BIT_LO(byte); |
| 249 | } |
| 250 | } |
| 251 | |
| 252 | /* |
| 253 | |
| 254 | A helper function for the macro isUTF8_CHAR(), which should be used instead of |
| 255 | this function. The macro will handle smaller code points directly saving time, |
| 256 | using this function as a fall-back for higher code points. |
| 257 | |
| 258 | Tests if the first bytes of string C<s> form a valid UTF-8 character. 0 is |
| 259 | returned if the bytes starting at C<s> up to but not including C<e> do not form a |
| 260 | complete well-formed UTF-8 character; otherwise the number of bytes in the |
| 261 | character is returned. |
| 262 | |
| 263 | Note that an INVARIANT (i.e. ASCII on non-EBCDIC) character is a valid UTF-8 |
| 264 | character. |
| 265 | |
| 266 | =cut */ |
| 267 | PERL_STATIC_INLINE STRLEN |
| 268 | S__is_utf8_char_slow(const U8 *s, const U8 *e) |
| 269 | { |
| 270 | dTHX; /* The function called below requires thread context */ |
| 271 | |
| 272 | STRLEN actual_len; |
| 273 | |
| 274 | PERL_ARGS_ASSERT__IS_UTF8_CHAR_SLOW; |
| 275 | |
| 276 | assert(e >= s); |
| 277 | utf8n_to_uvchr(s, e - s, &actual_len, UTF8_CHECK_ONLY); |
| 278 | |
| 279 | return (actual_len == (STRLEN) -1) ? 0 : actual_len; |
| 280 | } |
| 281 | |
| 282 | /* ------------------------------- perl.h ----------------------------- */ |
| 283 | |
| 284 | /* |
| 285 | =head1 Miscellaneous Functions |
| 286 | |
| 287 | =for apidoc AiR|bool|is_safe_syscall|const char *pv|STRLEN len|const char *what|const char *op_name |
| 288 | |
| 289 | Test that the given C<pv> doesn't contain any internal C<NUL> characters. |
| 290 | If it does, set C<errno> to ENOENT, optionally warn, and return FALSE. |
| 291 | |
| 292 | Return TRUE if the name is safe. |
| 293 | |
| 294 | Used by the IS_SAFE_SYSCALL() macro. |
| 295 | |
| 296 | =cut |
| 297 | */ |
| 298 | |
| 299 | PERL_STATIC_INLINE bool |
| 300 | S_is_safe_syscall(pTHX_ const char *pv, STRLEN len, const char *what, const char *op_name) { |
| 301 | /* While the Windows CE API provides only UCS-16 (or UTF-16) APIs |
| 302 | * perl itself uses xce*() functions which accept 8-bit strings. |
| 303 | */ |
| 304 | |
| 305 | PERL_ARGS_ASSERT_IS_SAFE_SYSCALL; |
| 306 | |
| 307 | if (len > 1) { |
| 308 | char *null_at; |
| 309 | if (UNLIKELY((null_at = (char *)memchr(pv, 0, len-1)) != NULL)) { |
| 310 | SETERRNO(ENOENT, LIB_INVARG); |
| 311 | Perl_ck_warner(aTHX_ packWARN(WARN_SYSCALLS), |
| 312 | "Invalid \\0 character in %s for %s: %s\\0%s", |
| 313 | what, op_name, pv, null_at+1); |
| 314 | return FALSE; |
| 315 | } |
| 316 | } |
| 317 | |
| 318 | return TRUE; |
| 319 | } |
| 320 | |
| 321 | /* |
| 322 | |
| 323 | Return true if the supplied filename has a newline character |
| 324 | immediately before the final NUL. |
| 325 | |
| 326 | My original look at this incorrectly used the len from SvPV(), but |
| 327 | that's incorrect, since we allow for a NUL in pv[len-1]. |
| 328 | |
| 329 | So instead, strlen() and work from there. |
| 330 | |
| 331 | This allow for the user reading a filename, forgetting to chomp it, |
| 332 | then calling: |
| 333 | |
| 334 | open my $foo, "$file\0"; |
| 335 | |
| 336 | */ |
| 337 | |
| 338 | #ifdef PERL_CORE |
| 339 | |
| 340 | PERL_STATIC_INLINE bool |
| 341 | S_should_warn_nl(const char *pv) { |
| 342 | STRLEN len; |
| 343 | |
| 344 | PERL_ARGS_ASSERT_SHOULD_WARN_NL; |
| 345 | |
| 346 | len = strlen(pv); |
| 347 | |
| 348 | return len > 0 && pv[len-1] == '\n'; |
| 349 | } |
| 350 | |
| 351 | #endif |
| 352 | |
| 353 | /* ------------------ pp.c, regcomp.c, toke.c, universal.c ------------ */ |
| 354 | |
| 355 | #define MAX_CHARSET_NAME_LENGTH 2 |
| 356 | |
| 357 | PERL_STATIC_INLINE const char * |
| 358 | get_regex_charset_name(const U32 flags, STRLEN* const lenp) |
| 359 | { |
| 360 | /* Returns a string that corresponds to the name of the regex character set |
| 361 | * given by 'flags', and *lenp is set the length of that string, which |
| 362 | * cannot exceed MAX_CHARSET_NAME_LENGTH characters */ |
| 363 | |
| 364 | *lenp = 1; |
| 365 | switch (get_regex_charset(flags)) { |
| 366 | case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS; |
| 367 | case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS; |
| 368 | case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS; |
| 369 | case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS; |
| 370 | case REGEX_ASCII_MORE_RESTRICTED_CHARSET: |
| 371 | *lenp = 2; |
| 372 | return ASCII_MORE_RESTRICT_PAT_MODS; |
| 373 | } |
| 374 | /* The NOT_REACHED; hides an assert() which has a rather complex |
| 375 | * definition in perl.h. */ |
| 376 | NOT_REACHED; /* NOTREACHED */ |
| 377 | return "?"; /* Unknown */ |
| 378 | } |
| 379 | |
| 380 | /* |
| 381 | * Local variables: |
| 382 | * c-indentation-style: bsd |
| 383 | * c-basic-offset: 4 |
| 384 | * indent-tabs-mode: nil |
| 385 | * End: |
| 386 | * |
| 387 | * ex: set ts=8 sts=4 sw=4 et: |
| 388 | */ |