Mercurial > hg > xemacs-beta
diff src/regex.c @ 446:1ccc32a20af4 r21-2-38
Import from CVS: tag r21-2-38
| author | cvs |
|---|---|
| date | Mon, 13 Aug 2007 11:37:21 +0200 |
| parents | 576fb035e263 |
| children | 98528da0b7fc |
line wrap: on
line diff
--- a/src/regex.c Mon Aug 13 11:36:20 2007 +0200 +++ b/src/regex.c Mon Aug 13 11:37:21 2007 +0200 @@ -104,7 +104,10 @@ { } -#endif /* not MULE */ +#endif /* MULE */ + +#define RE_TRANSLATE(ch) TRT_TABLE_OF (translate, (Emchar) ch) +#define TRANSLATE_P(tr) (!NILP (tr)) #else /* not emacs */ @@ -173,11 +176,14 @@ } } -#endif /* not SYNTAX_TABLE */ +#endif /* SYNTAX_TABLE */ #define SYNTAX_UNSAFE(ignored, c) re_syntax_table[c] -#endif /* not emacs */ +#define RE_TRANSLATE(c) translate[(unsigned char) (c)] +#define TRANSLATE_P(tr) tr + +#endif /* emacs */ /* Under XEmacs, this is needed because we don't define it elsewhere. */ #ifdef SWITCH_ENUM_BUG @@ -288,8 +294,8 @@ #ifndef _AIX /* Already did AIX, up at the top. */ void *alloca (); #endif /* not _AIX */ -#endif /* not HAVE_ALLOCA_H */ -#endif /* not __GNUC__ */ +#endif /* HAVE_ALLOCA_H */ +#endif /* __GNUC__ */ #endif /* not alloca */ @@ -304,7 +310,7 @@ /* No need to do anything to free, after alloca. */ #define REGEX_FREE(arg) ((void)0) /* Do nothing! But inhibit gcc warning. */ -#endif /* not REGEX_MALLOC */ +#endif /* REGEX_MALLOC */ /* Define how to allocate the failure stack. */ @@ -333,8 +339,8 @@ /* No need to explicitly free anything. */ #define REGEX_FREE_STACK(arg) -#endif /* not REGEX_MALLOC */ -#endif /* not REL_ALLOC */ +#endif /* REGEX_MALLOC */ +#endif /* REL_ALLOC */ /* True if `size1' is non-NULL and PTR is pointing anywhere inside @@ -359,6 +365,9 @@ #define MAX(a, b) ((a) > (b) ? (a) : (b)) #define MIN(a, b) ((a) < (b) ? (a) : (b)) +/* Type of source-pattern and string chars. */ +typedef const unsigned char re_char; + typedef char boolean; #define false 0 #define true 1 @@ -554,7 +563,7 @@ #ifdef DEBUG static void -extract_number (int *dest, unsigned char *source) +extract_number (int *dest, re_char *source) { int temp = SIGN_EXTEND_CHAR (*(source + 1)); *dest = *source & 0377; @@ -657,11 +666,11 @@ the START pointer into it and ending just before the pointer END. */ static void -print_partial_compiled_pattern (unsigned char *start, unsigned char *end) +print_partial_compiled_pattern (re_char *start, re_char *end) { int mcnt, mcnt2; - unsigned char *p = start; - unsigned char *pend = end; + unsigned char *p = (unsigned char *) start; + re_char *pend = end; if (start == NULL) { @@ -939,7 +948,7 @@ static void print_compiled_pattern (struct re_pattern_buffer *bufp) { - unsigned char *buffer = bufp->buffer; + re_char *buffer = bufp->buffer; print_partial_compiled_pattern (buffer, buffer + bufp->used); printf ("%ld bytes used/%ld bytes allocated.\n", bufp->used, @@ -965,8 +974,8 @@ static void -print_double_string (const char *where, const char *string1, int size1, - const char *string2, int size2) +print_double_string (re_char *where, re_char *string1, int size1, + re_char *string2, int size2) { if (where == NULL) printf ("(null)"); @@ -1000,7 +1009,7 @@ #define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) #define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) -#endif /* not DEBUG */ +#endif /* DEBUG */ /* Set by `re_set_syntax' to the current regexp syntax to recognize. Can also be assigned to arbitrarily: each pattern buffer stores its own @@ -1123,7 +1132,7 @@ union fail_stack_elt { - unsigned char *pointer; + re_char *pointer; int integer; }; @@ -1132,8 +1141,8 @@ typedef struct { fail_stack_elt_t *stack; - unsigned size; - unsigned avail; /* Offset of next open position. */ + size_t size; + size_t avail; /* Offset of next open position. */ } fail_stack_type; #define FAIL_STACK_EMPTY() (fail_stack.avail == 0) @@ -1309,7 +1318,7 @@ DEBUG_PRINT2 (" Pushing high active reg: %d\n", highest_active_reg);\ PUSH_FAILURE_INT (highest_active_reg); \ \ - DEBUG_PRINT2 (" Pushing pattern 0x%lx: ", (long) pattern_place); \ + DEBUG_PRINT2 (" Pushing pattern 0x%lx: \n", (long) pattern_place); \ DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \ PUSH_FAILURE_POINTER (pattern_place); \ \ @@ -1385,7 +1394,7 @@ saved NULL, thus retaining our current position in the string. */ \ string_temp = POP_FAILURE_POINTER (); \ if (string_temp != NULL) \ - str = (const char *) string_temp; \ + str = string_temp; \ \ DEBUG_PRINT2 (" Popping string 0x%lx: `", (long) str); \ DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \ @@ -1410,10 +1419,10 @@ DEBUG_PRINT2 (" info: 0x%lx\n", \ * (long *) ®_info[this_reg]); \ \ - regend[this_reg] = (const char *) POP_FAILURE_POINTER (); \ + regend[this_reg] = POP_FAILURE_POINTER (); \ DEBUG_PRINT2 (" end: 0x%lx\n", (long) regend[this_reg]); \ \ - regstart[this_reg] = (const char *) POP_FAILURE_POINTER (); \ + regstart[this_reg] = POP_FAILURE_POINTER (); \ DEBUG_PRINT2 (" start: 0x%lx\n", (long) regstart[this_reg]); \ } \ \ @@ -1476,7 +1485,7 @@ while (0) /* Registers are set to a sentinel when they haven't yet matched. */ -static char reg_unset_dummy; +static unsigned char reg_unset_dummy; #define REG_UNSET_VALUE (®_unset_dummy) #define REG_UNSET(e) ((e) == REG_UNSET_VALUE) @@ -1487,10 +1496,9 @@ string passed to us by the user to an unsigned char that we can use as an array index (in, e.g., `translate'). */ #define PATFETCH(c) \ - do {if (p == pend) return REG_EEND; \ - assert (p < pend); \ - c = (unsigned char) *p++; \ - if (translate) c = (unsigned char) translate[c]; \ + do { \ + PATFETCH_RAW (c); \ + c = TRANSLATE (c); \ } while (0) /* Fetch the next character in the uncompiled pattern, with no @@ -1498,11 +1506,12 @@ #define PATFETCH_RAW(c) \ do {if (p == pend) return REG_EEND; \ assert (p < pend); \ - c = (unsigned char) *p++; \ + c = charptr_emchar (p); \ + INC_CHARPTR (p); \ } while (0) /* Go backwards one character in the pattern. */ -#define PATUNFETCH p-- +#define PATUNFETCH DEC_CHARPTR (p) #ifdef MULE @@ -1511,8 +1520,8 @@ assert (p < pend); \ emch = charptr_emchar ((const Bufbyte *) p); \ INC_CHARPTR (p); \ - if (translate && emch < 0x80) \ - emch = (Emchar) (unsigned char) translate[emch]; \ + if (TRANSLATE_P (translate) && emch < 0x80) \ + emch = (Emchar) (unsigned char) RE_TRANSLATE (emch); \ } while (0) #define PATFETCH_RAW_EXTENDED(emch) \ @@ -1554,18 +1563,18 @@ #define PATFETCH_RAW_EITHER(emch) PATFETCH_RAW (emch) #define PATUNFETCH_EITHER PATUNFETCH -#endif /* not MULE */ +#endif /* MULE */ /* If `translate' is non-null, return translate[D], else just D. We cast the subscript to translate because some data is declared as `char *', to avoid warnings when a string constant is passed. But when we use a character as a subscript we must make it unsigned. */ -#define TRANSLATE(d) (translate ? translate[(unsigned char) (d)] : (d)) +#define TRANSLATE(d) (TRANSLATE_P (translate) ? RE_TRANSLATE (d) : (d)) #ifdef MULE #define TRANSLATE_EXTENDED_UNSAFE(emch) \ - (translate && emch < 0x80 ? translate[emch] : (emch)) + (TRANSLATE_P (translate) && emch < 0x80 ? RE_TRANSLATE (emch) : (emch)) #endif @@ -1576,14 +1585,14 @@ /* Make sure we have at least N more bytes of space in buffer. */ #define GET_BUFFER_SPACE(n) \ - while (b - bufp->buffer + (n) > bufp->allocated) \ + while (buf_end - bufp->buffer + (n) > bufp->allocated) \ EXTEND_BUFFER () /* Make sure we have one more byte of buffer space and then add C to it. */ #define BUF_PUSH(c) \ do { \ GET_BUFFER_SPACE (1); \ - *b++ = (unsigned char) (c); \ + *buf_end++ = (unsigned char) (c); \ } while (0) @@ -1591,8 +1600,8 @@ #define BUF_PUSH_2(c1, c2) \ do { \ GET_BUFFER_SPACE (2); \ - *b++ = (unsigned char) (c1); \ - *b++ = (unsigned char) (c2); \ + *buf_end++ = (unsigned char) (c1); \ + *buf_end++ = (unsigned char) (c2); \ } while (0) @@ -1600,9 +1609,9 @@ #define BUF_PUSH_3(c1, c2, c3) \ do { \ GET_BUFFER_SPACE (3); \ - *b++ = (unsigned char) (c1); \ - *b++ = (unsigned char) (c2); \ - *b++ = (unsigned char) (c3); \ + *buf_end++ = (unsigned char) (c1); \ + *buf_end++ = (unsigned char) (c2); \ + *buf_end++ = (unsigned char) (c3); \ } while (0) @@ -1615,13 +1624,15 @@ #define STORE_JUMP2(op, loc, to, arg) \ store_op2 (op, loc, (to) - (loc) - 3, arg) -/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */ +/* Like `STORE_JUMP', but for inserting. Assume `buf_end' is the + buffer end. */ #define INSERT_JUMP(op, loc, to) \ - insert_op1 (op, loc, (to) - (loc) - 3, b) - -/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */ + insert_op1 (op, loc, (to) - (loc) - 3, buf_end) + +/* Like `STORE_JUMP2', but for inserting. Assume `buf_end' is the + buffer end. */ #define INSERT_JUMP2(op, loc, to, arg) \ - insert_op2 (op, loc, (to) - (loc) - 3, arg, b) + insert_op2 (op, loc, (to) - (loc) - 3, arg, buf_end) /* This is not an arbitrary limit: the arguments which represent offsets @@ -1636,7 +1647,7 @@ being larger than MAX_BUF_SIZE, then flag memory exhausted. */ #define EXTEND_BUFFER() \ do { \ - unsigned char *old_buffer = bufp->buffer; \ + re_char *old_buffer = bufp->buffer; \ if (bufp->allocated == MAX_BUF_SIZE) \ return REG_ESIZE; \ bufp->allocated <<= 1; \ @@ -1648,7 +1659,7 @@ /* If the buffer moved, move all the pointers into it. */ \ if (old_buffer != bufp->buffer) \ { \ - b = (b - old_buffer) + bufp->buffer; \ + buf_end = (buf_end - old_buffer) + bufp->buffer; \ begalt = (begalt - old_buffer) + bufp->buffer; \ if (fixup_alt_jump) \ fixup_alt_jump = (fixup_alt_jump - old_buffer) + bufp->buffer;\ @@ -1705,7 +1716,7 @@ /* Set the bit for character C in a bit vector. */ #define SET_LIST_BIT(c) \ - (b[((unsigned char) (c)) / BYTEWIDTH] \ + (buf_end[((unsigned char) (c)) / BYTEWIDTH] \ |= 1 << (((unsigned char) c) % BYTEWIDTH)) #ifdef MULE @@ -1762,18 +1773,19 @@ unsigned char *end); static void insert_op2 (re_opcode_t op, unsigned char *loc, int arg1, int arg2, unsigned char *end); -static boolean at_begline_loc_p (const char *pattern, const char *p, +static boolean at_begline_loc_p (re_char *pattern, re_char *p, reg_syntax_t syntax); -static boolean at_endline_loc_p (const char *p, const char *pend, int syntax); +static boolean at_endline_loc_p (re_char *p, re_char *pend, int syntax); static boolean group_in_compile_stack (compile_stack_type compile_stack, regnum_t regnum); -static reg_errcode_t compile_range (const char **p_ptr, const char *pend, - char *translate, reg_syntax_t syntax, +static reg_errcode_t compile_range (re_char **p_ptr, re_char *pend, + RE_TRANSLATE_TYPE translate, + reg_syntax_t syntax, unsigned char *b); #ifdef MULE -static reg_errcode_t compile_extended_range (const char **p_ptr, - const char *pend, - char *translate, +static reg_errcode_t compile_extended_range (re_char **p_ptr, + re_char *pend, + RE_TRANSLATE_TYPE translate, reg_syntax_t syntax, Lisp_Object rtab); #endif /* MULE */ @@ -1786,10 +1798,10 @@ unsigned char *end, register_info_type *reg_info); static int bcmp_translate (const unsigned char *s1, const unsigned char *s2, - REGISTER int len, char *translate); + REGISTER int len, RE_TRANSLATE_TYPE translate); static int re_match_2_internal (struct re_pattern_buffer *bufp, - const char *string1, int size1, - const char *string2, int size2, int pos, + re_char *string1, int size1, + re_char *string2, int size2, int pos, struct re_registers *regs, int stop); #ifndef MATCH_MAY_ALLOCATE @@ -1808,11 +1820,11 @@ but never make them smaller. */ static int regs_allocated_size; -static const char ** regstart, ** regend; -static const char ** old_regstart, ** old_regend; -static const char **best_regstart, **best_regend; +static re_char ** regstart, ** regend; +static re_char ** old_regstart, ** old_regend; +static re_char **best_regstart, **best_regend; static register_info_type *reg_info; -static const char **reg_dummy; +static re_char **reg_dummy; static register_info_type *reg_info_dummy; /* Make the register vectors big enough for NUM_REGS registers, @@ -1823,14 +1835,14 @@ { if (num_regs > regs_allocated_size) { - RETALLOC_IF (regstart, num_regs, const char *); - RETALLOC_IF (regend, num_regs, const char *); - RETALLOC_IF (old_regstart, num_regs, const char *); - RETALLOC_IF (old_regend, num_regs, const char *); - RETALLOC_IF (best_regstart, num_regs, const char *); - RETALLOC_IF (best_regend, num_regs, const char *); + RETALLOC_IF (regstart, num_regs, re_char *); + RETALLOC_IF (regend, num_regs, re_char *); + RETALLOC_IF (old_regstart, num_regs, re_char *); + RETALLOC_IF (old_regend, num_regs, re_char *); + RETALLOC_IF (best_regstart, num_regs, re_char *); + RETALLOC_IF (best_regend, num_regs, re_char *); RETALLOC_IF (reg_info, num_regs, register_info_type); - RETALLOC_IF (reg_dummy, num_regs, const char *); + RETALLOC_IF (reg_dummy, num_regs, re_char *); RETALLOC_IF (reg_info_dummy, num_regs, register_info_type); regs_allocated_size = num_regs; @@ -1862,7 +1874,7 @@ return (free (compile_stack.stack), value) static reg_errcode_t -regex_compile (const char *pattern, int size, reg_syntax_t syntax, +regex_compile (re_char *pattern, int size, reg_syntax_t syntax, struct re_pattern_buffer *bufp) { /* We fetch characters from PATTERN here. We declare these as int @@ -1874,20 +1886,20 @@ REGISTER EMACS_INT c, c1; /* A random temporary spot in PATTERN. */ - const char *p1; + re_char *p1; /* Points to the end of the buffer, where we should append. */ - REGISTER unsigned char *b; + REGISTER unsigned char *buf_end; /* Keeps track of unclosed groups. */ compile_stack_type compile_stack; /* Points to the current (ending) position in the pattern. */ - const char *p = pattern; - const char *pend = pattern + size; + re_char *p = pattern; + re_char *pend = pattern + size; /* How to translate the characters in the pattern. */ - char *translate = bufp->translate; + RE_TRANSLATE_TYPE translate = bufp->translate; /* Address of the count-byte of the most recently inserted `exactn' command. This makes it possible to tell if a new exact-match @@ -1905,7 +1917,7 @@ /* Place in the uncompiled pattern (i.e., the {) to which to go back if the interval is invalid. */ - const char *beg_interval; + re_char *beg_interval; /* Address of the place where a forward jump should go to the end of the containing expression. Each alternative of an `or' -- except the @@ -1972,7 +1984,7 @@ bufp->allocated = INIT_BUF_SIZE; } - begalt = b = bufp->buffer; + begalt = buf_end = bufp->buffer; /* Loop through the uncompiled pattern until we're at the end. */ while (p != pend) @@ -2103,10 +2115,10 @@ 9: end of pattern. */ GET_BUFFER_SPACE (6); - INSERT_JUMP (jump, laststart, b + 3); - b += 3; + INSERT_JUMP (jump, laststart, buf_end + 3); + buf_end += 3; INSERT_JUMP (on_failure_jump, laststart, laststart + 6); - b += 3; + buf_end += 3; } else if (zero_times_ok) { @@ -2117,10 +2129,10 @@ 9: end of pattern. */ GET_BUFFER_SPACE (6); - INSERT_JUMP (jump, laststart, b + 3); - b += 3; - STORE_JUMP (on_failure_jump, b, laststart + 3); - b += 3; + INSERT_JUMP (jump, laststart, buf_end + 3); + buf_end += 3; + STORE_JUMP (on_failure_jump, buf_end, laststart + 3); + buf_end += 3; } else { @@ -2130,8 +2142,8 @@ 6: end of pattern. */ GET_BUFFER_SPACE (3); - STORE_JUMP (on_failure_jump, b, laststart); - b += 3; + STORE_JUMP (on_failure_jump, buf_end, laststart); + buf_end += 3; } } else @@ -2140,10 +2152,11 @@ boolean keep_string_p = false; if (many_times_ok) - { /* More than one repetition is allowed, so put in at the - end a backward relative jump from `b' to before the next - jump we're going to put in below (which jumps from - laststart to after this jump). + { /* More than one repetition is allowed, so put in + at the end a backward relative jump from + `buf_end' to before the next jump we're going + to put in below (which jumps from laststart to + after this jump). But if we are at the `*' in the exact sequence `.*\n', insert an unconditional jump backwards to the ., @@ -2161,29 +2174,30 @@ character after the `*'. Do we have to do something analogous here for null bytes, because of RE_DOT_NOT_NULL? */ - if (TRANSLATE (*(p - 2)) == TRANSLATE ('.') + if (*(p - 2) == '.' && zero_times_ok - && p < pend && TRANSLATE (*p) == TRANSLATE ('\n') + && p < pend && *p == '\n' && !(syntax & RE_DOT_NEWLINE)) { /* We have .*\n. */ - STORE_JUMP (jump, b, laststart); + STORE_JUMP (jump, buf_end, laststart); keep_string_p = true; } else /* Anything else. */ - STORE_JUMP (maybe_pop_jump, b, laststart - 3); + STORE_JUMP (maybe_pop_jump, buf_end, laststart - 3); /* We've added more stuff to the buffer. */ - b += 3; + buf_end += 3; } - /* On failure, jump from laststart to b + 3, which will be the - end of the buffer after this jump is inserted. */ + /* On failure, jump from laststart to buf_end + 3, + which will be the end of the buffer after this jump + is inserted. */ GET_BUFFER_SPACE (3); INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump : on_failure_jump, - laststart, b + 3); - b += 3; + laststart, buf_end + 3); + buf_end += 3; if (!zero_times_ok) { @@ -2194,7 +2208,7 @@ we hit that loop. */ GET_BUFFER_SPACE (3); INSERT_JUMP (dummy_failure_jump, laststart, laststart + 6); - b += 3; + buf_end += 3; } } pending_exact = 0; @@ -2203,7 +2217,7 @@ case '.': - laststart = b; + laststart = buf_end; BUF_PUSH (anychar); break; @@ -2223,7 +2237,7 @@ opcode, the length count, and the bitset; 34 bytes in all. */ GET_BUFFER_SPACE (34); - laststart = b; + laststart = buf_end; /* We test `*p == '^' twice, instead of using an if statement, so we only need one BUF_PUSH. */ @@ -2238,10 +2252,10 @@ BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH); /* Clear the whole map. */ - memset (b, 0, (1 << BYTEWIDTH) / BYTEWIDTH); + memset (buf_end, 0, (1 << BYTEWIDTH) / BYTEWIDTH); /* charset_not matches newline according to a syntax bit. */ - if ((re_opcode_t) b[-2] == charset_not + if ((re_opcode_t) buf_end[-2] == charset_not && (syntax & RE_HAT_LISTS_NOT_NEWLINE)) SET_LIST_BIT ('\n'); @@ -2251,18 +2265,18 @@ { /* There are extended chars here, which means we need to start over and shift to unified range-table format. */ - if (b[-2] == charset) - b[-2] = charset_mule; + if (buf_end[-2] == charset) + buf_end[-2] = charset_mule; else - b[-2] = charset_mule_not; - b--; + buf_end[-2] = charset_mule_not; + buf_end--; p = p1; /* go back to the beginning of the charset, after a possible ^. */ rtab = Vthe_lisp_rangetab; Fclear_range_table (rtab); /* charset_not matches newline according to a syntax bit. */ - if ((re_opcode_t) b[-1] == charset_mule_not + if ((re_opcode_t) buf_end[-1] == charset_mule_not && (syntax & RE_HAT_LISTS_NOT_NEWLINE)) SET_EITHER_BIT ('\n'); } @@ -2273,7 +2287,7 @@ { if (p == pend) FREE_STACK_RETURN (REG_EBRACK); - PATFETCH_EITHER (c); + PATFETCH (c); #ifdef MULE if (c >= 0x80 && !has_extended_chars) @@ -2292,7 +2306,7 @@ { if (p == pend) FREE_STACK_RETURN (REG_EESCAPE); - PATFETCH_EITHER (c1); + PATFETCH (c1); #ifdef MULE if (c1 >= 0x80 && !has_extended_chars) { @@ -2321,7 +2335,7 @@ operator. */ if (c == '-' && !(p - 2 >= pattern && p[-2] == '[') - && !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^') + && !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^') && *p != ']') { reg_errcode_t ret; @@ -2337,7 +2351,7 @@ syntax, rtab); else #endif /* MULE */ - ret = compile_range (&p, pend, translate, syntax, b); + ret = compile_range (&p, pend, translate, syntax, buf_end); if (ret != REG_NOERROR) FREE_STACK_RETURN (ret); } @@ -2359,7 +2373,7 @@ syntax, rtab); else #endif /* MULE */ - ret = compile_range (&p, pend, translate, syntax, b); + ret = compile_range (&p, pend, translate, syntax, buf_end); if (ret != REG_NOERROR) FREE_STACK_RETURN (ret); } @@ -2378,12 +2392,9 @@ for (;;) { - /* Do not do PATFETCH_EITHER() here. We want - to just see if the bytes match particular - strings, and we put them all back if not. - - #### May need to be changed once trt tables - are working. */ + /* #### This code is unused. + Correctness is not checked after TRT + table change. */ PATFETCH (c); if (c == ':' || c == ']' || p == pend || c1 == CHAR_CLASS_MAX_LENGTH) @@ -2392,7 +2403,7 @@ } str[c1] = '\0'; - /* If isn't a word bracketed by `[:' and:`]': + /* If isn't a word bracketed by `[:' and `:]': undo the ending character, the letters, and leave the leading `:' and `[' (but set bits for them). */ if (c == ':' && *p == ']') @@ -2466,16 +2477,16 @@ int bytes_needed = unified_range_table_bytes_needed (rtab); GET_BUFFER_SPACE (bytes_needed); - unified_range_table_copy_data (rtab, b); - b += unified_range_table_bytes_used (b); + unified_range_table_copy_data (rtab, buf_end); + buf_end += unified_range_table_bytes_used (buf_end); break; } #endif /* MULE */ /* Discard any (non)matching list bytes that are all 0 at the end of the map. Decrease the map-length byte too. */ - while ((int) b[-1] > 0 && b[b[-1] - 1] == 0) - b[-1]--; - b += b[-1]; + while ((int) buf_end[-1] > 0 && buf_end[buf_end[-1] - 1] == 0) + buf_end[-1]--; + buf_end += buf_end[-1]; } break; @@ -2535,10 +2546,10 @@ if (!(syntax & RE_NO_SHY_GROUPS) && p != pend - && TRANSLATE(*p) == TRANSLATE('?')) + && *p == '?') { p++; - PATFETCH(c); + PATFETCH (c); switch (c) { case ':': /* shy groups */ @@ -2572,7 +2583,7 @@ COMPILE_STACK_TOP.begalt_offset = begalt - bufp->buffer; COMPILE_STACK_TOP.fixup_alt_jump = fixup_alt_jump ? fixup_alt_jump - bufp->buffer + 1 : 0; - COMPILE_STACK_TOP.laststart_offset = b - bufp->buffer; + COMPILE_STACK_TOP.laststart_offset = buf_end - bufp->buffer; COMPILE_STACK_TOP.regnum = r; /* We will eventually replace the 0 with the number of @@ -2582,7 +2593,7 @@ if (r <= MAX_REGNUM) { COMPILE_STACK_TOP.inner_group_offset - = b - bufp->buffer + 2; + = buf_end - bufp->buffer + 2; BUF_PUSH_3 (start_memory, r, 0); } @@ -2590,7 +2601,7 @@ fixup_alt_jump = 0; laststart = 0; - begalt = b; + begalt = buf_end; /* If we've reached MAX_REGNUM groups, then this open won't actually generate any code, so we'll have to clear pending_exact explicitly. */ @@ -2619,7 +2630,7 @@ /* We allocated space for this jump when we assigned to `fixup_alt_jump', in the `handle_alt' case below. */ - STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1); + STORE_JUMP (jump_past_alt, fixup_alt_jump, buf_end - 1); } /* See similar code for backslashed left paren above. */ @@ -2677,9 +2688,9 @@ /* Insert before the previous alternative a jump which jumps to this alternative if the former fails. */ GET_BUFFER_SPACE (3); - INSERT_JUMP (on_failure_jump, begalt, b + 6); + INSERT_JUMP (on_failure_jump, begalt, buf_end + 6); pending_exact = 0; - b += 3; + buf_end += 3; /* The alternative before this one has a jump after it which gets executed if it gets matched. Adjust that @@ -2698,17 +2709,17 @@ bytes which we'll fill in when we get to after `c'. */ if (fixup_alt_jump) - STORE_JUMP (jump_past_alt, fixup_alt_jump, b); + STORE_JUMP (jump_past_alt, fixup_alt_jump, buf_end); /* Mark and leave space for a jump after this alternative, to be filled in later either by next alternative or when know we're at the end of a series of alternatives. */ - fixup_alt_jump = b; + fixup_alt_jump = buf_end; GET_BUFFER_SPACE (3); - b += 3; + buf_end += 3; laststart = 0; - begalt = b; + begalt = buf_end; break; @@ -2781,7 +2792,7 @@ if (syntax & RE_CONTEXT_INVALID_OPS) FREE_STACK_RETURN (REG_BADRPT); else if (syntax & RE_CONTEXT_INDEP_OPS) - laststart = b; + laststart = buf_end; else goto unfetch_interval; } @@ -2792,8 +2803,8 @@ if (upper_bound == 0) { GET_BUFFER_SPACE (3); - INSERT_JUMP (jump, laststart, b + 3); - b += 3; + INSERT_JUMP (jump, laststart, buf_end + 3); + buf_end += 3; } /* Otherwise, we have a nontrivial interval. When @@ -2818,16 +2829,16 @@ because `re_compile_fastmap' needs to know. Jump to the `jump_n' we might insert below. */ INSERT_JUMP2 (succeed_n, laststart, - b + 5 + (upper_bound > 1) * 5, + buf_end + 5 + (upper_bound > 1) * 5, lower_bound); - b += 5; + buf_end += 5; /* Code to initialize the lower bound. Insert before the `succeed_n'. The `5' is the last two bytes of this `set_number_at', plus 3 bytes of the following `succeed_n'. */ - insert_op2 (set_number_at, laststart, 5, lower_bound, b); - b += 5; + insert_op2 (set_number_at, laststart, 5, lower_bound, buf_end); + buf_end += 5; if (upper_bound > 1) { /* More than one repetition is allowed, so @@ -2837,9 +2848,9 @@ When we've reached this during matching, we'll have matched the interval once, so jump back only `upper_bound - 1' times. */ - STORE_JUMP2 (jump_n, b, laststart + 5, + STORE_JUMP2 (jump_n, buf_end, laststart + 5, upper_bound - 1); - b += 5; + buf_end += 5; /* The location we want to set is the second parameter of the `jump_n'; that is `b-2' as @@ -2855,9 +2866,10 @@ We insert this at the beginning of the loop so that if we fail during matching, we'll reinitialize the bounds. */ - insert_op2 (set_number_at, laststart, b - laststart, - upper_bound - 1, b); - b += 5; + insert_op2 (set_number_at, laststart, + buf_end - laststart, + upper_bound - 1, buf_end); + buf_end += 5; } } pending_exact = 0; @@ -2889,7 +2901,7 @@ break; case 's': - laststart = b; + laststart = buf_end; PATFETCH (c); /* XEmacs addition */ if (c >= 0x80 || syntax_spec_code[c] == 0377) @@ -2898,7 +2910,7 @@ break; case 'S': - laststart = b; + laststart = buf_end; PATFETCH (c); /* XEmacs addition */ if (c >= 0x80 || syntax_spec_code[c] == 0377) @@ -2909,7 +2921,7 @@ #ifdef MULE /* 97.2.17 jhod merged in to XEmacs from mule-2.3 */ case 'c': - laststart = b; + laststart = buf_end; PATFETCH_RAW (c); if (c < 32 || c > 127) FREE_STACK_RETURN (REG_ECATEGORY); @@ -2917,7 +2929,7 @@ break; case 'C': - laststart = b; + laststart = buf_end; PATFETCH_RAW (c); if (c < 32 || c > 127) FREE_STACK_RETURN (REG_ECATEGORY); @@ -2929,13 +2941,13 @@ case 'w': - laststart = b; + laststart = buf_end; BUF_PUSH (wordchar); break; case 'W': - laststart = b; + laststart = buf_end; BUF_PUSH (notwordchar); break; @@ -2966,20 +2978,23 @@ case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': - if (syntax & RE_NO_BK_REFS) - goto normal_char; - - c1 = c - '0'; - - if (c1 > regnum) - FREE_STACK_RETURN (REG_ESUBREG); - - /* Can't back reference to a subexpression if inside of it. */ - if (group_in_compile_stack (compile_stack, c1)) - goto normal_char; - - laststart = b; - BUF_PUSH_2 (duplicate, c1); + { + regnum_t reg; + if (syntax & RE_NO_BK_REFS) + goto normal_char; + + reg = c - '0'; + + if (reg > regnum) + FREE_STACK_RETURN (REG_ESUBREG); + + /* Can't back reference to a subexpression if inside of it. */ + if (group_in_compile_stack (compile_stack, reg)) + goto normal_char; + + laststart = buf_end; + BUF_PUSH_2 (duplicate, reg); + } break; @@ -3008,14 +3023,13 @@ { /* XEmacs: modifications here for Mule. */ /* `q' points to the beginning of the next char. */ - const char *q = p - 1; - INC_CHARPTR (q); + re_char *q = p; /* If no exactn currently being built. */ if (!pending_exact /* If last exactn not at current position. */ - || pending_exact + *pending_exact + 1 != b + || pending_exact + *pending_exact + 1 != buf_end /* We have only one byte following the exactn for the count. */ || ((unsigned int) (*pending_exact + (q - p)) >= @@ -3033,21 +3047,30 @@ { /* Start building a new exactn. */ - laststart = b; + laststart = buf_end; BUF_PUSH_2 (exactn, 0); - pending_exact = b - 1; + pending_exact = buf_end - 1; } +#ifndef MULE BUF_PUSH (c); (*pending_exact)++; - - while (p < q) - { - PATFETCH (c); - BUF_PUSH (c); - (*pending_exact)++; - } +#else + { + Bytecount bt_count; + Bufbyte tmp_buf[MAX_EMCHAR_LEN]; + int i; + + bt_count = set_charptr_emchar (tmp_buf, c); + + for (i = 0; i < bt_count; i++) + { + BUF_PUSH (tmp_buf[i]); + (*pending_exact)++; + } + } +#endif break; } } /* switch (c) */ @@ -3057,7 +3080,7 @@ /* Through the pattern now. */ if (fixup_alt_jump) - STORE_JUMP (jump_past_alt, fixup_alt_jump, b); + STORE_JUMP (jump_past_alt, fixup_alt_jump, buf_end); if (!COMPILE_STACK_EMPTY) FREE_STACK_RETURN (REG_EPAREN); @@ -3070,7 +3093,7 @@ free (compile_stack.stack); /* We have succeeded; set the length of the buffer. */ - bufp->used = b - bufp->buffer; + bufp->used = buf_end - bufp->buffer; #ifdef DEBUG if (debug) @@ -3114,7 +3137,7 @@ = (fail_stack_elt_t *) realloc (fail_stack.stack, (fail_stack.size * sizeof (fail_stack_elt_t))); -#endif /* not emacs */ +#endif /* emacs */ } regex_grow_registers (num_regs); @@ -3184,9 +3207,9 @@ least one character before the ^. */ static boolean -at_begline_loc_p (const char *pattern, const char *p, reg_syntax_t syntax) +at_begline_loc_p (re_char *pattern, re_char *p, reg_syntax_t syntax) { - const char *prev = p - 2; + re_char *prev = p - 2; boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\'; return @@ -3201,11 +3224,11 @@ at least one character after the $, i.e., `P < PEND'. */ static boolean -at_endline_loc_p (const char *p, const char *pend, int syntax) +at_endline_loc_p (re_char *p, re_char *pend, int syntax) { - const char *next = p; + re_char *next = p; boolean next_backslash = *next == '\\'; - const char *next_next = p + 1 < pend ? p + 1 : 0; + re_char *next_next = p + 1 < pend ? p + 1 : 0; return /* Before a subexpression? */ @@ -3247,12 +3270,12 @@ `regex_compile' itself. */ static reg_errcode_t -compile_range (const char **p_ptr, const char *pend, char *translate, - reg_syntax_t syntax, unsigned char *b) +compile_range (re_char **p_ptr, re_char *pend, RE_TRANSLATE_TYPE translate, + reg_syntax_t syntax, unsigned char *buf_end) { unsigned this_char; - const char *p = *p_ptr; + re_char *p = *p_ptr; int range_start, range_end; if (p == pend) @@ -3292,7 +3315,8 @@ #ifdef MULE static reg_errcode_t -compile_extended_range (const char **p_ptr, const char *pend, char *translate, +compile_extended_range (re_char **p_ptr, re_char *pend, + RE_TRANSLATE_TYPE translate, reg_syntax_t syntax, Lisp_Object rtab) { Emchar this_char, range_start, range_end; @@ -3414,7 +3438,7 @@ /* Reset for next path. */ path_can_be_null = true; - p = fail_stack.stack[--fail_stack.avail].pointer; + p = (unsigned char *) fail_stack.stack[--fail_stack.avail].pointer; continue; } @@ -3463,10 +3487,10 @@ /* And all extended characters must be allowed, too. */ for (j = 0x80; j < 0xA0; j++) fastmap[j] = 1; -#else /* ! MULE */ +#else /* not MULE */ for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++) fastmap[j] = 1; -#endif /* ! MULE */ +#endif /* MULE */ for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))) @@ -3623,14 +3647,14 @@ } } } -#else /* ! MULE */ +#else /* not MULE */ for (j = 0; j < (1 << BYTEWIDTH); j++) if (SYNTAX_UNSAFE (XCHAR_TABLE (regex_emacs_buffer->mirror_syntax_table), j) == (enum syntaxcode) k) fastmap[j] = 1; -#endif /* ! MULE */ +#endif /* MULE */ break; @@ -3664,14 +3688,14 @@ } } } -#else /* ! MULE */ +#else /* not MULE */ for (j = 0; j < (1 << BYTEWIDTH); j++) if (SYNTAX_UNSAFE (XCHAR_TABLE (regex_emacs_buffer->mirror_syntax_table), j) != (enum syntaxcode) k) fastmap[j] = 1; -#endif /* ! MULE */ +#endif /* MULE */ break; #ifdef MULE @@ -3901,19 +3925,21 @@ stack overflow). */ int -re_search_2 (struct re_pattern_buffer *bufp, const char *string1, - int size1, const char *string2, int size2, int startpos, +re_search_2 (struct re_pattern_buffer *bufp, const char *str1, + int size1, const char *str2, int size2, int startpos, int range, struct re_registers *regs, int stop) { int val; + re_char *string1 = (re_char *) str1; + re_char *string2 = (re_char *) str2; REGISTER char *fastmap = bufp->fastmap; - REGISTER char *translate = bufp->translate; + REGISTER RE_TRANSLATE_TYPE translate = bufp->translate; int total_size = size1 + size2; int endpos = startpos + range; #ifdef REGEX_BEGLINE_CHECK int anchored_at_begline = 0; #endif - const unsigned char *d; + re_char *d; Charcount d_size; /* Check for out-of-range STARTPOS. */ @@ -3985,12 +4011,8 @@ DEC_CHARPTR(d); /* Ok, since startpos != size1. */ d_size = charcount_to_bytecount (d, 1); - if (translate) -#ifdef MULE - while (range > lim && (*d >= 0x80 || translate[*d] != '\n')) -#else - while (range > lim && translate[*d] != '\n') -#endif + if (TRANSLATE_P (translate)) + while (range > lim && *d != '\n') { d += d_size; /* Speedier INC_CHARPTR(d) */ d_size = charcount_to_bytecount (d, 1); @@ -4027,13 +4049,20 @@ /* Written out as an if-else to avoid testing `translate' inside the loop. */ - if (translate) - while (range > lim && + if (TRANSLATE_P (translate)) + while (range > lim) + { #ifdef MULE - *d < 0x80 && -#endif - !fastmap[(unsigned char)translate[*d]]) - { + Emchar buf_ch; + + buf_ch = charptr_emchar (d); + buf_ch = RE_TRANSLATE (buf_ch); + if (buf_ch >= 0200 || fastmap[(unsigned char) buf_ch]) + break; +#else + if (fastmap[(unsigned char)RE_TRANSLATE (*d)]) + break; +#endif /* MULE */ d_size = charcount_to_bytecount (d, 1); range -= d_size; d += d_size; /* Speedier INC_CHARPTR(d) */ @@ -4050,15 +4079,17 @@ } else /* Searching backwards. */ { - unsigned char c = (size1 == 0 || startpos >= size1 - ? string2[startpos - size1] - : string1[startpos]); + Emchar c = (size1 == 0 || startpos >= size1 + ? charptr_emchar (string2 + startpos - size1) + : charptr_emchar (string1 + startpos)); + c = TRANSLATE (c); #ifdef MULE - if (c < 0x80 && !fastmap[(unsigned char) TRANSLATE (c)]) + if (!(c >= 0200 || fastmap[(unsigned char) c])) + goto advance; #else - if (!fastmap[(unsigned char) TRANSLATE (c)]) + if (!fastmap[(unsigned char) c]) + goto advance; #endif - goto advance; } } @@ -4171,9 +4202,9 @@ FREE_VAR (reg_dummy); \ FREE_VAR (reg_info_dummy); \ } while (0) -#else +#else /* not MATCH_MAY_ALLOCATE */ #define FREE_VARIABLES() ((void)0) /* Do nothing! But inhibit gcc warning. */ -#endif /* not MATCH_MAY_ALLOCATE */ +#endif /* MATCH_MAY_ALLOCATE */ /* These values must meet several constraints. They must not be valid register values; since we have a limit of 255 registers (because @@ -4194,7 +4225,7 @@ re_match (struct re_pattern_buffer *bufp, const char *string, int size, int pos, struct re_registers *regs) { - int result = re_match_2_internal (bufp, NULL, 0, string, size, + int result = re_match_2_internal (bufp, NULL, 0, (re_char *) string, size, pos, regs, size); alloca (0); return result; @@ -4220,7 +4251,8 @@ int size1, const char *string2, int size2, int pos, struct re_registers *regs, int stop) { - int result = re_match_2_internal (bufp, string1, size1, string2, size2, + int result = re_match_2_internal (bufp, (re_char *) string1, size1, + (re_char *) string2, size2, pos, regs, stop); alloca (0); return result; @@ -4229,8 +4261,8 @@ /* This is a separate function so that we can force an alloca cleanup afterwards. */ static int -re_match_2_internal (struct re_pattern_buffer *bufp, const char *string1, - int size1, const char *string2, int size2, int pos, +re_match_2_internal (struct re_pattern_buffer *bufp, re_char *string1, + int size1, re_char *string2, int size2, int pos, struct re_registers *regs, int stop) { /* General temporaries. */ @@ -4239,14 +4271,14 @@ int should_succeed; /* XEmacs change */ /* Just past the end of the corresponding string. */ - const char *end1, *end2; + re_char *end1, *end2; /* Pointers into string1 and string2, just past the last characters in each to consider matching. */ - const char *end_match_1, *end_match_2; + re_char *end_match_1, *end_match_2; /* Where we are in the data, and the end of the current string. */ - const char *d, *dend; + re_char *d, *dend; /* Where we are in the pattern, and the end of the pattern. */ unsigned char *p = bufp->buffer; @@ -4254,10 +4286,10 @@ /* Mark the opcode just after a start_memory, so we can test for an empty subpattern when we get to the stop_memory. */ - unsigned char *just_past_start_mem = 0; + re_char *just_past_start_mem = 0; /* We use this to map every character in the string. */ - char *translate = bufp->translate; + RE_TRANSLATE_TYPE translate = bufp->translate; /* Failure point stack. Each place that can handle a failure further down the line pushes a failure point on this stack. It consists of @@ -4299,7 +4331,7 @@ stopped matching the regnum-th subexpression. (The zeroth register keeps track of what the whole pattern matches.) */ #ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ - const char **regstart, **regend; + re_char **regstart, **regend; #endif /* If a group that's operated upon by a repetition operator fails to @@ -4308,7 +4340,7 @@ are when we last see its open-group operator. Similarly for a register's end. */ #ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ - const char **old_regstart, **old_regend; + re_char **old_regstart, **old_regend; #endif /* The is_active field of reg_info helps us keep track of which (possibly @@ -4327,7 +4359,7 @@ turn happens only if we have not yet matched the entire string. */ unsigned best_regs_set = false; #ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ - const char **best_regstart, **best_regend; + re_char **best_regstart, **best_regend; #endif /* Logically, this is `best_regend[0]'. But we don't want to have to @@ -4338,14 +4370,14 @@ the end of the best match so far in a separate variable. We initialize this to NULL so that when we backtrack the first time and need to test it, it's not garbage. */ - const char *match_end = NULL; + re_char *match_end = NULL; /* This helps SET_REGS_MATCHED avoid doing redundant work. */ int set_regs_matched_done = 0; /* Used when we pop values we don't care about. */ #ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */ - const char **reg_dummy; + re_char **reg_dummy; register_info_type *reg_info_dummy; #endif @@ -4373,14 +4405,14 @@ array indexing. We should fix this. */ if (bufp->re_nsub) { - regstart = REGEX_TALLOC (num_regs, const char *); - regend = REGEX_TALLOC (num_regs, const char *); - old_regstart = REGEX_TALLOC (num_regs, const char *); - old_regend = REGEX_TALLOC (num_regs, const char *); - best_regstart = REGEX_TALLOC (num_regs, const char *); - best_regend = REGEX_TALLOC (num_regs, const char *); + regstart = REGEX_TALLOC (num_regs, re_char *); + regend = REGEX_TALLOC (num_regs, re_char *); + old_regstart = REGEX_TALLOC (num_regs, re_char *); + old_regend = REGEX_TALLOC (num_regs, re_char *); + best_regstart = REGEX_TALLOC (num_regs, re_char *); + best_regend = REGEX_TALLOC (num_regs, re_char *); reg_info = REGEX_TALLOC (num_regs, register_info_type); - reg_dummy = REGEX_TALLOC (num_regs, const char *); + reg_dummy = REGEX_TALLOC (num_regs, re_char *); reg_info_dummy = REGEX_TALLOC (num_regs, register_info_type); if (!(regstart && regend && old_regstart && old_regend && reg_info @@ -4420,7 +4452,6 @@ MATCHED_SOMETHING (reg_info[mcnt]) = 0; EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0; } - /* We move `string1' into `string2' if the latter's empty -- but not if `string1' is null. */ if (size2 == 0 && string1 != NULL) @@ -4462,7 +4493,7 @@ dend = end_match_2; } - DEBUG_PRINT1 ("The compiled pattern is: "); + DEBUG_PRINT1 ("The compiled pattern is: \n"); DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend); DEBUG_PRINT1 ("The string to match is: `"); DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2); @@ -4660,22 +4691,40 @@ /* This is written out as an if-else so we don't waste time testing `translate' inside the loop. */ - if (translate) + if (TRANSLATE_P (translate)) { do { +#ifdef MULE + Emchar pat_ch, buf_ch; + Bytecount pat_len; + PREFETCH (); - if (translate[(unsigned char) *d++] != (char) *p++) + pat_ch = charptr_emchar (p); + buf_ch = charptr_emchar (d); + if (RE_TRANSLATE (buf_ch) != pat_ch) goto fail; + + pat_len = charcount_to_bytecount (p, 1); + p += pat_len; + INC_CHARPTR (d); + + mcnt -= pat_len; +#else /* not MULE */ + PREFETCH (); + if ((unsigned char) RE_TRANSLATE (*d++) != *p++) + goto fail; + mcnt--; +#endif } - while (--mcnt); + while (mcnt > 0); } else { do { PREFETCH (); - if (*d++ != (char) *p++) goto fail; + if (*d++ != *p++) goto fail; } while (--mcnt); } @@ -4950,7 +4999,7 @@ followed by the numeric value of <digit> as the register number. */ case duplicate: { - REGISTER const char *d2, *dend2; + REGISTER re_char *d2, *dend2; int regno = *p++; /* Get which register to match against. */ DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno); @@ -4998,7 +5047,7 @@ /* Compare that many; failure if mismatch, else move past them. */ - if (translate + if (TRANSLATE_P (translate) ? bcmp_translate ((unsigned char *) d, (unsigned char *) d2, mcnt, translate) : memcmp (d, d2, mcnt)) @@ -5086,7 +5135,7 @@ EXTRACT_NUMBER_AND_INCR (mcnt, p); DEBUG_PRINT3 (" %d (to 0x%lx):\n", mcnt, (long) (p + mcnt)); - PUSH_FAILURE_POINT (p + mcnt, (char *) 0, -2); + PUSH_FAILURE_POINT (p + mcnt, (unsigned char *) 0, -2); break; @@ -5306,7 +5355,7 @@ `pop_failure_point'. */ unsigned dummy_low_reg, dummy_high_reg; unsigned char *pdummy; - const char *sdummy = NULL; + re_char *sdummy = NULL; DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n"); POP_FAILURE_POINT (sdummy, pdummy, @@ -5342,7 +5391,7 @@ DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n"); /* It doesn't matter what we push for the string here. What the code at `fail' tests is the value for the pattern. */ - PUSH_FAILURE_POINT ((unsigned char *) 0, (char *) 0, -2); + PUSH_FAILURE_POINT ((unsigned char *) 0, (unsigned char *) 0, -2); goto unconditional_jump; @@ -5355,7 +5404,7 @@ DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n"); /* See comments just above at `dummy_failure_jump' about the two zeroes. */ - PUSH_FAILURE_POINT ((unsigned char *) 0, (char *) 0, -2); + PUSH_FAILURE_POINT ((unsigned char *) 0, (unsigned char *) 0, -2); break; /* Have to succeed matching what follows at least n times. @@ -5611,7 +5660,7 @@ SET_REGS_MATCHED (); d++; break; -#endif /* not emacs */ +#endif /* emacs */ default: abort (); @@ -5921,15 +5970,34 @@ bytes; nonzero otherwise. */ static int -bcmp_translate (const unsigned char *s1, const unsigned char *s2, - REGISTER int len, char *translate) +bcmp_translate (re_char *s1, re_char *s2, + REGISTER int len, RE_TRANSLATE_TYPE translate) { REGISTER const unsigned char *p1 = s1, *p2 = s2; +#ifdef MULE + const unsigned char *p1_end = s1 + len; + const unsigned char *p2_end = s2 + len; + + while (p1 != p1_end && p2 != p2_end) + { + Emchar p1_ch, p2_ch; + + p1_ch = charptr_emchar (p1); + p2_ch = charptr_emchar (p2); + + if (RE_TRANSLATE (p1_ch) + != RE_TRANSLATE (p2_ch)) + return 1; + INC_CHARPTR (p1); + INC_CHARPTR (p2); + } +#else /* not MULE */ while (len) { - if (translate[*p1++] != translate[*p2++]) return 1; + if (RE_TRANSLATE (*p1++) != RE_TRANSLATE (*p2++)) return 1; len--; } +#endif /* MULE */ return 0; } @@ -5962,7 +6030,7 @@ /* Match anchors at newline. */ bufp->newline_anchor = 1; - ret = regex_compile (pattern, length, re_syntax_options, bufp); + ret = regex_compile ((unsigned char *) pattern, length, re_syntax_options, bufp); if (!ret) return NULL; @@ -6007,7 +6075,7 @@ /* Match anchors at newlines. */ re_comp_buf.newline_anchor = 1; - ret = regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf); + ret = regex_compile ((unsigned char *)s, strlen (s), re_syntax_options, &re_comp_buf); if (!ret) return NULL; @@ -6113,7 +6181,7 @@ /* POSIX says a null character in the pattern terminates it, so we can use strlen here in compiling the pattern. */ - ret = regex_compile (pattern, strlen (pattern), syntax, preg); + ret = regex_compile ((unsigned char *) pattern, strlen (pattern), syntax, preg); /* POSIX doesn't distinguish between an unmatched open-group and an unmatched close-group: both are REG_EPAREN. */