X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexp.h;h=b1170f1e063df037ef90362fbe008aa17f8a46ce;hb=469bf43793110471bf2ceb5e99f0e9cc630d60fe;hp=33d9e3250e8a0c6fce8879e501ed16c8e8f23728;hpb=55204971972392ce5a252fbbd6d78b1c48ed70e3;p=p5sagit%2Fp5-mst-13.2.git

diff --git a/regexp.h b/regexp.h
index 33d9e32..b1170f1 100644
--- a/regexp.h
+++ b/regexp.h
@@ -1,3 +1,6 @@
+/*    regexp.h
+ */
+
 /*
  * Definitions etc. for regexp(3) routines.
  *
@@ -5,46 +8,102 @@
  * not the System V one.
  */
 
-/* $RCSfile: regexp.h,v $$Revision: 4.0.1.2 $$Date: 91/11/05 18:24:31 $
- *
- * $Log:	regexp.h,v $
- * Revision 4.0.1.2  91/11/05  18:24:31  lwall
- * patch11: minimum match length calculation in regexp is now cumulative
- * patch11: initial .* in pattern had dependency on value of $*
- * 
- * Revision 4.0.1.1  91/06/07  11:51:18  lwall
- * patch4: new copyright notice
- * patch4: // wouldn't use previous pattern if it started with a null character
- * patch4: $` was busted inside s///
- * 
- * 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? */
-	int minlen;		/* mininum possible length of $& */
-	int prelen;		/* length of precomp */
+	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;
 
-#define ROPT_ANCH 1
-#define ROPT_SKIP 2
-#define ROPT_IMPLICIT 4
+#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 REXEC_IGNOREPOS	8		/* \G matches at start. */
 
-regexp *regcomp();
-int regexec();
+#define ReREFCNT_inc(re) ((re && re->refcnt++), re)
+#define ReREFCNT_dec(re) pregfree(re)