Separated the [:foo:] parsing to its own function.

[perl5.git] / regexp.h

diff --git a/regexp.h b/regexp.h
index 600e630..67410a5 100644 (file)
--- a/regexp.h
+++ b/regexp.h
@@ -1,3 +1,6 @@
+/*    regexp.h
+ */
+
 /*
  * Definitions etc. for regexp(3) routines.
  *
@@ -5,30 +8,101 @@
  * not the System V one.
  */
 
-/* $Header: regexp.h,v 4.0 91/03/20 01:39:23 lwall Locked $
- *
- * $Log:       regexp.h,v $
- * Revision 4.0  91/03/20  01:39:23  lwall
- * 4.0 baseline.
- * 
- */
+
+struct regnode {
+    U8 flags;
+    U8  type;
+    U16 next_off;
+};
+
+typedef struct regnode regnode;
+
+struct reg_data {
+    U32 count;
+    U8 *what;
+    void* data[1];
+};
+
+struct reg_substr_datum {
+    I32 min_offset;
+    I32 max_offset;
+    SV *substr;
+};
+
+struct reg_substr_data {
+    struct reg_substr_datum data[3];   /* Actual array */
+};
 
 typedef struct regexp {
+       I32 refcnt;
        char **startp;
        char **endp;
-       STR *regstart;          /* Internal use only. */
-       char *regstclass;
-       STR *regmust;           /* Internal use only. */
-       int regback;            /* Can regmust locate first try? */
+       regnode *regstclass;
+       I32 minlen;             /* mininum possible length of $& */
+       I32 prelen;             /* length of precomp */
+       U32 nparens;            /* number of parentheses */
+       U32 lastparen;          /* last paren matched */
        char *precomp;          /* pre-compilation regular expression */
        char *subbase;          /* saved string so \digit works forever */
+       char *subbeg;           /* same, but not responsible for allocation */
        char *subend;           /* end of subbase */
-       char reganch;           /* Internal use only. */
-       char do_folding;        /* do case-insensitive match? */
-       char lastparen;         /* last paren matched */
-       char nparens;           /* number of parentheses */
-       char program[1];        /* Unwarranted chumminess with compiler. */
+       U32 reganch;            /* Internal use only +
+                                  Tainted information used by regexec? */
+#if 0
+        SV *anchored_substr;   /* Substring at fixed position wrt start. */
+       I32 anchored_offset;    /* Position of it. */
+        SV *float_substr;      /* Substring at variable position wrt start. */
+       I32 float_min_offset;   /* Minimal position of it. */
+       I32 float_max_offset;   /* Maximal position of it. */
+        SV *check_substr;      /* Substring to check before matching. */
+        I32 check_offset_min;  /* Offset of the above. */
+        I32 check_offset_max;  /* Offset of the above. */
+#else
+        struct reg_substr_data *substrs;
+#endif
+        struct reg_data *data; /* Additional data. */
+       regnode program[1];     /* Unwarranted chumminess with compiler. */
 } regexp;
 
-regexp *regcomp();
-int regexec();
+#define anchored_substr substrs->data[0].substr
+#define anchored_offset substrs->data[0].min_offset
+#define float_substr substrs->data[1].substr
+#define float_min_offset substrs->data[1].min_offset
+#define float_max_offset substrs->data[1].max_offset
+#define check_substr substrs->data[2].substr
+#define check_offset_min substrs->data[2].min_offset
+#define check_offset_max substrs->data[2].max_offset
+
+#define ROPT_ANCH              (ROPT_ANCH_BOL|ROPT_ANCH_MBOL|ROPT_ANCH_GPOS)
+#define ROPT_ANCH_SINGLE       (ROPT_ANCH_BOL|ROPT_ANCH_GPOS)
+#define ROPT_ANCH_BOL          0x00001
+#define ROPT_ANCH_MBOL         0x00002
+#define ROPT_ANCH_GPOS         0x00004
+#define ROPT_SKIP              0x00008
+#define ROPT_IMPLICIT          0x00010 /* Converted .* to ^.* */
+#define ROPT_NOSCAN            0x00020 /* Check-string always at start. */
+#define ROPT_GPOS_SEEN         0x00040
+#define ROPT_CHECK_ALL         0x00080
+#define ROPT_LOOKBEHIND_SEEN   0x00100
+#define ROPT_EVAL_SEEN         0x00200
+#define ROPT_TAINTED_SEEN      0x00400
+#define ROPT_ANCH_SBOL         0x00800
+
+/* 0xf800 of reganch is used by PMf_COMPILETIME */
+
+#define ROPT_UTF8              0x10000
+#define ROPT_NAUGHTY           0x20000 /* how exponential is this pattern? */
+
+#define RX_MATCH_TAINTED(prog) ((prog)->reganch & ROPT_TAINTED_SEEN)
+#define RX_MATCH_TAINTED_on(prog) ((prog)->reganch |= ROPT_TAINTED_SEEN)
+#define RX_MATCH_TAINTED_off(prog) ((prog)->reganch &= ~ROPT_TAINTED_SEEN)
+#define RX_MATCH_TAINTED_set(prog, t) ((t) \
+                                      ? RX_MATCH_TAINTED_on(prog) \
+                                      : RX_MATCH_TAINTED_off(prog))
+
+#define REXEC_COPY_STR 1               /* Need to copy the string. */
+#define REXEC_CHECKED  2               /* check_substr already checked. */
+#define REXEC_SCREAM   4               /* use scream table. */
+
+#define ReREFCNT_inc(re) ((re && re->refcnt++), re)
+#define ReREFCNT_dec(re) pregfree(re)