Main Page | Class Hierarchy | Class List | File List | Class Members | File Members

regex.h

Go to the documentation of this file.
00001 /* Definitions for data structures and routines for the regular
00002    expression library, version 0.12.
00003 
00004    Copyright (C) 1985, 89, 90, 91, 92, 93, 95 Free Software Foundation, Inc.
00005 
00006    This program is free software; you can redistribute it and/or modify
00007    it under the terms of the GNU General Public License as published by
00008    the Free Software Foundation; either version 2, or (at your option)
00009    any later version.
00010 
00011    This program is distributed in the hope that it will be useful,
00012    but WITHOUT ANY WARRANTY; without even the implied warranty of
00013    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00014    GNU General Public License for more details.
00015 
00016    You should have received a copy of the GNU General Public License
00017    along with this program; if not, write to the Free Software
00018    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  */
00019 
00020 #ifndef __REGEXP_LIBRARY_H__
00021 #define __REGEXP_LIBRARY_H__
00022 
00023 /* POSIX says that <sys/types.h> must be included (by the caller) before
00024    <regex.h>.  */
00025 
00026 #if !defined (_POSIX_C_SOURCE) && !defined (_POSIX_SOURCE) && defined (VMS)
00027 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
00028    should be there.  */
00029 #include <stddef.h>
00030 #endif
00031 
00032 
00033 /* The following bits are used to determine the regexp syntax we
00034    recognize.  The set/not-set meanings are chosen so that Emacs syntax
00035    remains the value 0.  The bits are given in alphabetical order, and
00036    the definitions shifted by one from the previous bit; thus, when we
00037    add or remove a bit, only one other definition need change.  */
00038 typedef unsigned reg_syntax_t;
00039 
00040 /* If this bit is not set, then \ inside a bracket expression is literal.
00041    If set, then such a \ quotes the following character.  */
00042 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1)
00043 
00044 /* If this bit is not set, then + and ? are operators, and \+ and \? are
00045         literals. 
00046    If set, then \+ and \? are operators and + and ? are literals.  */
00047 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
00048 
00049 /* If this bit is set, then character classes are supported.  They are:
00050         [:alpha:], [:upper:], [:lower:],  [:digit:], [:alnum:], [:xdigit:],
00051         [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
00052    If not set, then character classes are not supported.  */
00053 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
00054 
00055 /* If this bit is set, then ^ and $ are always anchors (outside bracket
00056         expressions, of course).
00057    If this bit is not set, then it depends:
00058            ^  is an anchor if it is at the beginning of a regular
00059                  expression or after an open-group or an alternation operator;
00060            $  is an anchor if it is at the end of a regular expression, or
00061                  before a close-group or an alternation operator.  
00062 
00063    This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
00064    POSIX draft 11.2 says that * etc. in leading positions is undefined.
00065    We already implemented a previous draft which made those constructs
00066    invalid, though, so we haven't changed the code back.  */
00067 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
00068 
00069 /* If this bit is set, then special characters are always special
00070         regardless of where they are in the pattern.
00071    If this bit is not set, then special characters are special only in
00072         some contexts; otherwise they are ordinary.  Specifically, 
00073         * + ? and intervals are only special when not after the beginning,
00074         open-group, or alternation operator.  */
00075 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
00076 
00077 /* If this bit is set, then *, +, ?, and { cannot be first in an re or
00078         immediately after an alternation or begin-group operator.  */
00079 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
00080 
00081 /* If this bit is set, then . matches newline.
00082    If not set, then it doesn't.  */
00083 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
00084 
00085 /* If this bit is set, then . doesn't match NUL.
00086    If not set, then it does.  */
00087 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
00088 
00089 /* If this bit is set, nonmatching lists [^...] do not match newline.
00090    If not set, they do.  */
00091 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
00092 
00093 /* If this bit is set, either \{...\} or {...} defines an
00094         interval, depending on RE_NO_BK_BRACES. 
00095    If not set, \{, \}, {, and } are literals.  */
00096 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
00097 
00098 /* If this bit is set, +, ? and | aren't recognized as operators.
00099    If not set, they are.  */
00100 #define RE_LIMITED_OPS (RE_INTERVALS << 1)
00101 
00102 /* If this bit is set, newline is an alternation operator.
00103    If not set, newline is literal.  */
00104 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
00105 
00106 /* If this bit is set, then `{...}' defines an interval, and \{ and \}
00107         are literals.
00108   If not set, then `\{...\}' defines an interval.  */
00109 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
00110 
00111 /* If this bit is set, (...) defines a group, and \( and \) are literals.
00112    If not set, \(...\) defines a group, and ( and ) are literals.  */
00113 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
00114 
00115 /* If this bit is set, then <digit> matches <digit>.
00116    If not set, then <digit> is a back-reference.  */
00117 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
00118 
00119 /* If this bit is set, then | is an alternation operator, and \| is literal. 
00120    If not set, then \| is an alternation operator, and | is literal.  */
00121 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
00122 
00123 /* If this bit is set, then an ending range point collating higher
00124         than the starting range point, as in [z-a], is invalid.
00125    If not set, then when ending range point collates higher than the
00126         starting range point, the range is ignored.  */
00127 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
00128 
00129 /* If this bit is set, then an unmatched ) is ordinary.
00130    If not set, then an unmatched ) is invalid.  */
00131 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
00132 
00133 /* If this bit is set, succeed as soon as we match the whole pattern,
00134    without further backtracking.  */
00135 #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
00136 
00137 /* This global variable defines the particular regexp syntax to use (for
00138    some interfaces).  When a regexp is compiled, the syntax used is
00139    stored in the pattern buffer, so changing this does not affect
00140    already-compiled regexps.  */
00141 extern reg_syntax_t re_syntax_options;
00142 
00143 /* Define combinations of the above bits for the standard possibilities.
00144    (The [[[ comments delimit what gets put into the Texinfo file, so
00145    don't delete them!)  */ 
00146 /* [[[begin syntaxes]]] */
00147 #define RE_SYNTAX_EMACS 0
00148 
00149 #define RE_SYNTAX_AWK                                                   \
00150   (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL                       \
00151    | RE_NO_BK_PARENS            | RE_NO_BK_REFS                         \
00152    | RE_NO_BK_VBAR               | RE_NO_EMPTY_RANGES                   \
00153    | RE_UNMATCHED_RIGHT_PAREN_ORD)
00154 
00155 #define RE_SYNTAX_POSIX_AWK                                             \
00156   (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS)
00157 
00158 #define RE_SYNTAX_GREP                                                  \
00159   (RE_BK_PLUS_QM              | RE_CHAR_CLASSES                         \
00160    | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS                            \
00161    | RE_NEWLINE_ALT)
00162 
00163 #define RE_SYNTAX_EGREP                                                 \
00164   (RE_CHAR_CLASSES        | RE_CONTEXT_INDEP_ANCHORS                    \
00165    | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE                    \
00166    | RE_NEWLINE_ALT       | RE_NO_BK_PARENS                             \
00167    | RE_NO_BK_VBAR)
00168 
00169 #define RE_SYNTAX_POSIX_EGREP                                           \
00170   (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES)
00171 
00172 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff.  */
00173 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
00174 
00175 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
00176 
00177 /* Syntax bits common to both basic and extended POSIX regex syntax.  */
00178 #define _RE_SYNTAX_POSIX_COMMON                                         \
00179   (RE_CHAR_CLASSES | RE_DOT_NEWLINE      | RE_DOT_NOT_NULL              \
00180    | RE_INTERVALS  | RE_NO_EMPTY_RANGES)
00181 
00182 #define RE_SYNTAX_POSIX_BASIC                                           \
00183   (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
00184 
00185 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
00186    RE_LIMITED_OPS, i.e., \? \+ \| are not recognized.  Actually, this
00187    isn't minimal, since other operators, such as \`, aren't disabled.  */
00188 #define RE_SYNTAX_POSIX_MINIMAL_BASIC                                   \
00189   (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
00190 
00191 #define RE_SYNTAX_POSIX_EXTENDED                                        \
00192   (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS                   \
00193    | RE_CONTEXT_INDEP_OPS  | RE_NO_BK_BRACES                            \
00194    | RE_NO_BK_PARENS       | RE_NO_BK_VBAR                              \
00195    | RE_UNMATCHED_RIGHT_PAREN_ORD)
00196 
00197 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS
00198    replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added.  */
00199 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED                                \
00200   (_RE_SYNTAX_POSIX_COMMON  | RE_CONTEXT_INDEP_ANCHORS                  \
00201    | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES                           \
00202    | RE_NO_BK_PARENS        | RE_NO_BK_REFS                             \
00203    | RE_NO_BK_VBAR          | RE_UNMATCHED_RIGHT_PAREN_ORD)
00204 /* [[[end syntaxes]]] */
00205 
00206 /* Maximum number of duplicates an interval can allow.  Some systems
00207    (erroneously) define this in other header files, but we want our
00208    value, so remove any previous define.  */
00209 #ifdef RE_DUP_MAX
00210 #undef RE_DUP_MAX
00211 #endif
00212 #define RE_DUP_MAX ((1 << 15) - 1) 
00213 
00214 
00215 /* POSIX `cflags' bits (i.e., information for `regcomp').  */
00216 
00217 /* If this bit is set, then use extended regular expression syntax.
00218    If not set, then use basic regular expression syntax.  */
00219 #define REG_EXTENDED 1
00220 
00221 /* If this bit is set, then ignore case when matching.
00222    If not set, then case is significant.  */
00223 #define REG_ICASE (REG_EXTENDED << 1)
00224  
00225 /* If this bit is set, then anchors do not match at newline
00226         characters in the string.
00227    If not set, then anchors do match at newlines.  */
00228 #define REG_NEWLINE (REG_ICASE << 1)
00229 
00230 /* If this bit is set, then report only success or fail in regexec.
00231    If not set, then returns differ between not matching and errors.  */
00232 #define REG_NOSUB (REG_NEWLINE << 1)
00233 
00234 
00235 /* POSIX `eflags' bits (i.e., information for regexec).  */
00236 
00237 /* If this bit is set, then the beginning-of-line operator doesn't match
00238         the beginning of the string (presumably because it's not the
00239         beginning of a line).
00240    If not set, then the beginning-of-line operator does match the
00241         beginning of the string.  */
00242 #define REG_NOTBOL 1
00243 
00244 /* Like REG_NOTBOL, except for the end-of-line.  */
00245 #define REG_NOTEOL (1 << 1)
00246 
00247 
00248 /* If any error codes are removed, changed, or added, update the
00249    `re_error_msg' table in regex.c.  */
00250 typedef enum
00251 {
00252   REG_NOERROR = 0,      /* Success.  */
00253   REG_NOMATCH,          /* Didn't find a match (for regexec).  */
00254 
00255   /* POSIX regcomp return error codes.  (In the order listed in the
00256         standard.)  */
00257   REG_BADPAT,           /* Invalid pattern.  */
00258   REG_ECOLLATE,         /* Not implemented.  */
00259   REG_ECTYPE,           /* Invalid character class name.  */
00260   REG_EESCAPE,          /* Trailing backslash.  */
00261   REG_ESUBREG,          /* Invalid back reference.  */
00262   REG_EBRACK,           /* Unmatched left bracket.  */
00263   REG_EPAREN,           /* Parenthesis imbalance.  */ 
00264   REG_EBRACE,           /* Unmatched \{.  */
00265   REG_BADBR,            /* Invalid contents of \{\}.  */
00266   REG_ERANGE,           /* Invalid range end.  */
00267   REG_ESPACE,           /* Ran out of memory.  */
00268   REG_BADRPT,           /* No preceding re for repetition op.  */
00269 
00270   /* Error codes we've added.  */
00271   REG_EEND,             /* Premature end.  */
00272   REG_ESIZE,            /* Compiled pattern bigger than 2^16 bytes.  */
00273   REG_ERPAREN           /* Unmatched ) or \); not returned from regcomp.  */
00274 } reg_errcode_t;
00275 
00276 /* This data structure represents a compiled pattern.  Before calling
00277    the pattern compiler, the fields `buffer', `allocated', `fastmap',
00278    `translate', and `no_sub' can be set.  After the pattern has been
00279    compiled, the `re_nsub' field is available.  All other fields are
00280    private to the regex routines.  */
00281 
00282 struct re_pattern_buffer
00283 {
00284 /* [[[begin pattern_buffer]]] */
00285         /* Space that holds the compiled pattern.  It is declared as
00286                 `unsigned char *' because its elements are
00287                  sometimes used as array indexes.  */
00288   unsigned char *buffer;
00289 
00290         /* Number of bytes to which `buffer' points.  */
00291   unsigned long allocated;
00292 
00293         /* Number of bytes actually used in `buffer'.  */
00294   unsigned long used;   
00295 
00296            /* Syntax setting with which the pattern was compiled.  */
00297   reg_syntax_t syntax;
00298 
00299            /* Pointer to a fastmap, if any, otherwise zero.  re_search uses
00300                  the fastmap, if there is one, to skip over impossible
00301                  starting points for matches.  */
00302   char *fastmap;
00303 
00304            /* Either a translate table to apply to all characters before
00305                  comparing them, or zero for no translation.  The translation
00306                  is applied to a pattern when it is compiled and to a string
00307                  when it is matched.  */
00308   char *translate;
00309 
00310         /* Number of subexpressions found by the compiler.  */
00311   size_t re_nsub;
00312 
00313            /* Zero if this pattern cannot match the empty string, one else.
00314                  Well, in truth it's used only in `re_search_2', to see
00315                  whether or not we should use the fastmap, so we don't set
00316                  this absolutely perfectly; see `re_compile_fastmap' (the
00317                  `duplicate' case).  */
00318   unsigned can_be_null : 1;
00319 
00320            /* If REGS_UNALLOCATED, allocate space in the `regs' structure
00321                    for `max (RE_NREGS, re_nsub + 1)' groups.
00322                  If REGS_REALLOCATE, reallocate space if necessary.
00323                  If REGS_FIXED, use what's there.  */
00324 #define REGS_UNALLOCATED 0
00325 #define REGS_REALLOCATE 1
00326 #define REGS_FIXED 2
00327   unsigned regs_allocated : 2;
00328 
00329            /* Set to zero when `regex_compile' compiles a pattern; set to one
00330                  by `re_compile_fastmap' if it updates the fastmap.  */
00331   unsigned fastmap_accurate : 1;
00332 
00333            /* If set, `re_match_2' does not return information about
00334                  subexpressions.  */
00335   unsigned no_sub : 1;
00336 
00337            /* If set, a beginning-of-line anchor doesn't match at the
00338                  beginning of the string.  */ 
00339   unsigned not_bol : 1;
00340 
00341            /* Similarly for an end-of-line anchor.  */
00342   unsigned not_eol : 1;
00343 
00344            /* If true, an anchor at a newline matches.  */
00345   unsigned newline_anchor : 1;
00346 
00347 /* [[[end pattern_buffer]]] */
00348 };
00349 
00350 typedef struct re_pattern_buffer regex_t;
00351 
00352 /* Type for byte offsets within the string.  POSIX mandates this.  */
00353 typedef int regoff_t;
00354 
00355 
00356 /* This is the structure we store register match data in.  See
00357    regex.texinfo for a full description of what registers match.  */
00358 struct re_registers
00359 {
00360   unsigned num_regs;
00361   regoff_t *start;
00362   regoff_t *end;
00363 };
00364 
00365 
00366 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
00367    `re_match_2' returns information about at least this many registers
00368    the first time a `regs' structure is passed.  */
00369 #ifndef RE_NREGS
00370 #define RE_NREGS 30
00371 #endif
00372 
00373 
00374 /* POSIX specification for registers.  Aside from the different names than
00375    `re_registers', POSIX uses an array of structures, instead of a
00376    structure of arrays.  */
00377 typedef struct
00378 {
00379   regoff_t rm_so;  /* Byte offset from string's start to substring's start.  */
00380   regoff_t rm_eo;  /* Byte offset from string's start to substring's end.  */
00381 } regmatch_t;
00382 
00383 /* Declarations for routines.  */
00384 
00385 /* To avoid duplicating every routine declaration -- once with a
00386    prototype (if we are ANSI), and once without (if we aren't) -- we
00387    use the following macro to declare argument types.  This
00388    unfortunately clutters up the declarations a bit, but I think it's
00389    worth it.  */
00390 
00391 #if __STDC__
00392 
00393 #define _RE_ARGS(args) args
00394 
00395 #else /* not __STDC__ */
00396 
00397 #define _RE_ARGS(args) ()
00398 
00399 #endif /* not __STDC__ */
00400 
00401 /* Sets the current default syntax to SYNTAX, and return the old syntax.
00402    You can also simply assign to the `re_syntax_options' variable.  */
00403 extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax));
00404 
00405 /* Compile the regular expression PATTERN, with length LENGTH
00406    and syntax given by the global `re_syntax_options', into the buffer
00407    BUFFER.  Return NULL if successful, and an error string if not.  */
00408 extern const char *re_compile_pattern
00409   _RE_ARGS ((const char *pattern, int length,
00410                    struct re_pattern_buffer *buffer));
00411 
00412 
00413 /* Compile a fastmap for the compiled pattern in BUFFER; used to
00414    accelerate searches.  Return 0 if successful and -2 if was an
00415    internal error.  */
00416 extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer));
00417 
00418 
00419 /* Search in the string STRING (with length LENGTH) for the pattern
00420    compiled into BUFFER.  Start searching at position START, for RANGE
00421    characters.  Return the starting position of the match, -1 for no
00422    match, or -2 for an internal error.  Also return register
00423    information in REGS (if REGS and BUFFER->no_sub are nonzero).  */
00424 extern int re_search
00425   _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
00426                   int length, int start, int range, struct re_registers *regs));
00427 
00428 
00429 /* Like `re_search', but search in the concatenation of STRING1 and
00430    STRING2.  Also, stop searching at index START + STOP.  */
00431 extern int re_search_2
00432   _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
00433                    int length1, const char *string2, int length2,
00434                    int start, int range, struct re_registers *regs, int stop));
00435 
00436 
00437 /* Like `re_search', but return how many characters in STRING the regexp
00438    in BUFFER matched, starting at position START.  */
00439 extern int re_match
00440   _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string,
00441                    int length, int start, struct re_registers *regs));
00442 
00443 
00444 /* Relates to `re_match' as `re_search_2' relates to `re_search'.  */
00445 extern int re_match_2 
00446   _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1,
00447                    int length1, const char *string2, int length2,
00448                    int start, struct re_registers *regs, int stop));
00449 
00450 
00451 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and
00452    ENDS.  Subsequent matches using BUFFER and REGS will use this memory
00453    for recording register information.  STARTS and ENDS must be
00454    allocated with malloc, and must each be at least `NUM_REGS * sizeof
00455    (regoff_t)' bytes long.
00456 
00457    If NUM_REGS == 0, then subsequent matches should allocate their own
00458    register data.
00459 
00460    Unless this function is called, the first search or match using
00461    PATTERN_BUFFER will allocate its own register data, without
00462    freeing the old data.  */
00463 extern void re_set_registers
00464   _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs,
00465                    unsigned num_regs, regoff_t *starts, regoff_t *ends));
00466 
00467 #ifdef _REGEX_RE_COMP
00468 /* 4.2 bsd compatibility.  */
00469 extern char *re_comp _RE_ARGS ((const char *));
00470 extern int re_exec _RE_ARGS ((const char *));
00471 #endif
00472 
00473 /* POSIX compatibility.  */
00474 extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags));
00475 extern int regexec
00476   _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch,
00477                    regmatch_t pmatch[], int eflags));
00478 extern size_t regerror
00479   _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf,
00480                    size_t errbuf_size));
00481 extern void regfree _RE_ARGS ((regex_t *preg));
00482 
00483 #endif /* not __REGEXP_LIBRARY_H__ */
00484 
00485 /*
00486 Local variables:
00487 make-backup-files: t
00488 version-control: t
00489 trim-versions-without-asking: nil
00490 End:
00491 */

Generated on Fri May 7 21:15:43 2004 by doxygen 1.3.5