Further tweaks to make it easier to create regexp engine plug ins.

Message-ID: <9b18b3110612050713g77cac516x46fb5baac99b47c9@mail.gmail.com>

(with tweaks)

p4raw-id: //depot/perl@29468

diff --git a/regcomp.c b/regcomp.c
index b7510a9..1611eb4 100644 (file)
--- a/regcomp.c
+++ b/regcomp.c
@@ -8707,15 +8707,17 @@ Perl_re_dup(pTHX_ const regexp *r, CLONE_PARAMS *param)
     Newx(ret->endp, npar, I32);
     Copy(r->endp, ret->endp, npar, I32);
 
-    Newx(ret->substrs, 1, struct reg_substr_data);
-    for (s = ret->substrs->data, i = 0; i < 3; i++, s++) {
-       s->min_offset = r->substrs->data[i].min_offset;
-       s->max_offset = r->substrs->data[i].max_offset;
-       s->end_shift  = r->substrs->data[i].end_shift;
-       s->substr     = sv_dup_inc(r->substrs->data[i].substr, param);
-       s->utf8_substr = sv_dup_inc(r->substrs->data[i].utf8_substr, param);
-    }
-    
+    if (ret->substrs) {
+        Newx(ret->substrs, 1, struct reg_substr_data);
+        for (s = ret->substrs->data, i = 0; i < 3; i++, s++) {
+            s->min_offset = r->substrs->data[i].min_offset;
+            s->max_offset = r->substrs->data[i].max_offset;
+            s->end_shift  = r->substrs->data[i].end_shift;
+            s->substr     = sv_dup_inc(r->substrs->data[i].substr, param);
+            s->utf8_substr = sv_dup_inc(r->substrs->data[i].utf8_substr, param);
+        }
+    } else 
+        ret->substrs = NULL;    
 
     ret->precomp        = SAVEPVN(r->precomp, r->prelen);
     ret->refcnt         = r->refcnt;
@@ -8743,7 +8745,8 @@ Perl_re_dup(pTHX_ const regexp *r, CLONE_PARAMS *param)
 #endif
     
     ret->pprivate = r->pprivate;
-    RXi_SET(ret,CALLREGDUPE_PVT(ret,param));
+    if (ret->pprivate) 
+        RXi_SET(ret,CALLREGDUPE_PVT(ret,param));
     
     ptr_table_store(PL_ptr_table, r, ret);
     return ret;

diff --git a/regcomp.h b/regcomp.h
index 1075080..ce96da1 100644 (file)
--- a/regcomp.h
+++ b/regcomp.h
@@ -92,7 +92,17 @@ typedef OP OP_4tree;                 /* Will be redefined later. */
  * special test to reverse the sign of BACK pointers since the offset is
  * stored negative.]
  */
-typedef struct regexp_internal {
+
+/* This is the stuff that used to live in regexp.h that was truly
+   private to the engine itself. It now lives here. */
+
+/* swap buffer for paren structs */
+typedef struct regexp_paren_ofs {
+    I32 *startp;
+    I32 *endp;
+} regexp_paren_ofs;
+
+ typedef struct regexp_internal {
         regexp_paren_ofs *swap; /* Swap copy of *startp / *endp */
        U32 *offsets;           /* offset annotations 20001228 MJD 
                                    data about mapping the program to the 
@@ -109,6 +119,20 @@ typedef struct regexp_internal {
 #define RXi_SET(x,y) (x)->pprivate = (void*)(y)   
 #define RXi_GET(x)   ((regexp_internal *)((x)->pprivate))
 #define RXi_GET_DECL(r,ri) regexp_internal *ri = RXi_GET(r)
+/*
+ * Flags stored in regexp->intflags
+ * These are used only internally to the regexp engine
+ *
+ * See regexp.h for flags used externally to the regexp engine
+ */
+#define PREGf_SKIP             0x00000001
+#define PREGf_IMPLICIT         0x00000002 /* Converted .* to ^.* */
+#define PREGf_NAUGHTY          0x00000004 /* how exponential is this pattern? */
+#define PREGf_VERBARG_SEEN     0x00000008
+#define PREGf_CUTGROUP_SEEN    0x00000010
+
+
+/* this is where the old regcomp.h started */
 
 struct regnode_string {
     U8 str_len;
@@ -460,18 +484,6 @@ struct reg_data {
     void* data[1];
 };
 
-struct reg_substr_datum {
-    I32 min_offset;
-    I32 max_offset;
-    SV *substr;                /* non-utf8 variant */
-    SV *utf8_substr;   /* utf8 variant */
-    I32 end_shift;
-};
-
-struct reg_substr_data {
-    struct reg_substr_datum data[3];   /* Actual array */
-};
-
 #define anchored_substr substrs->data[0].substr
 #define anchored_utf8 substrs->data[0].utf8_substr
 #define anchored_offset substrs->data[0].min_offset

diff --git a/regexp.h b/regexp.h
index 85af91e..248090a 100644 (file)
--- a/regexp.h
+++ b/regexp.h
@@ -15,8 +15,8 @@
  * not the System V one.
  */
 #ifndef PLUGGABLE_RE_EXTENSION
-/* we don't want to include this stuff if we are inside Nicholas'
- * pluggable regex engine code */
+/* we don't want to include this stuff if we are inside of
+   an external regex engine based on the core one - like re 'debug'*/
 
 struct regnode {
     U8 flags;
@@ -32,10 +32,16 @@ struct reg_data;
 
 struct regexp_engine;
 
-typedef struct regexp_paren_ofs {
-    I32 *startp;
-    I32 *endp;
-} regexp_paren_ofs;
+struct reg_substr_datum {
+    I32 min_offset;
+    I32 max_offset;
+    SV *substr;                /* non-utf8 variant */
+    SV *utf8_substr;   /* utf8 variant */
+    I32 end_shift;
+};
+struct reg_substr_data {
+    struct reg_substr_datum data[3];   /* Actual array */
+};
 
 #ifdef PERL_OLD_COPY_ON_WRITE
 #define SV_SAVED_COPY   SV *saved_copy; /* If non-NULL, SV which is COW from original */
@@ -83,13 +89,16 @@ typedef struct regexp {
        I32 refcnt;             /* Refcount of this regexp */
 } regexp;
 
-
+/* used for high speed searches */
 typedef struct re_scream_pos_data_s
 {
     char **scream_olds;                /* match pos */
     I32 *scream_pos;           /* Internal iterator of scream. */
 } re_scream_pos_data;
 
+/* regexp_engine structure. This is the dispatch table for regexes.
+ * Any regex engine implementation must be able to build one of these.
+ */
 typedef struct regexp_engine {
     regexp* (*comp) (pTHX_ char* exp, char* xend, PMOP* pm);
     I32            (*exec) (pTHX_ regexp* prog, char* stringarg, char* strend,
@@ -105,17 +114,6 @@ typedef struct regexp_engine {
 #endif    
 } regexp_engine;
 
-/* 
- * Flags stored in regexp->intflags 
- * These are used only internally to the regexp engine
- */
-#define PREGf_SKIP             0x00000001
-#define PREGf_IMPLICIT         0x00000002 /* Converted .* to ^.* */
-#define PREGf_NAUGHTY          0x00000004 /* how exponential is this pattern? */
-#define PREGf_VERBARG_SEEN      0x00000008
-#define PREGf_CUTGROUP_SEEN    0x00000010
-
-
 /* Flags stored in regexp->extflags 
  * These are used by code external to the regexp engine
  */