From e357fc671749d483a102a060a774d41a9b256b46 Mon Sep 17 00:00:00 2001
From: Yves Orton <demerphq@gmail.com>
Date: Tue, 9 Jan 2007 01:46:31 +0100
Subject: [PATCH] Let the regex parser decide if we have a special pattern or
 not. Message-ID:
 <9b18b3110701081546n2c3f19acy29adc0d233bb848c@mail.gmail.com>

p4raw-id: //depot/perl@29730
---
 op.c      |  4 +++-
 op.h      |  8 +++++---
 pp.c      |  2 +-
 pp_ctl.c  |  2 +-
 regcomp.c |  5 +++++
 regexp.h  | 30 ++++++++++++++++++------------
 6 files changed, 33 insertions(+), 18 deletions(-)

diff --git a/op.c b/op.c
index a495db8..df2e90b 100644
--- a/op.c
+++ b/op.c
@@ -3256,8 +3256,10 @@ Perl_pmruntime(pTHX_ OP *o, OP *expr, bool isreg)
 	    pm->op_pmdynflags |= PMdf_UTF8;
 	/* FIXME - can we make this function take const char * args?  */
 	PM_SETRE(pm, CALLREGCOMP((char*)p, (char*)p + plen, pm));
-	if (strEQ("\\s+", PM_GETRE(pm)->precomp))
+	if (PM_GETRE(pm)->extflags & RXf_WHITE)
 	    pm->op_pmflags |= PMf_WHITE;
+	else
+	    pm->op_pmflags &= ~PMf_WHITE;
 #ifdef PERL_MAD
 	op_getmad(expr,(OP*)pm,'e');
 #else
diff --git a/op.h b/op.h
index f84f123..1ac4aa0 100644
--- a/op.h
+++ b/op.h
@@ -367,15 +367,17 @@ struct pmop {
 #define PMf_GLOBAL	0x0100		/* pattern had a g modifier */
 #define PMf_CONTINUE	0x0200		/* don't reset pos() if //g fails */
 #define PMf_EVAL	0x0400		/* evaluating replacement as expr */
+
+/* The following flags have exact equivalents in regcomp.h with the prefix RXf_
+ * which are stored in the regexp->extflags member.
+ */
 #define PMf_LOCALE	0x0800		/* use locale for character types */
 #define PMf_MULTILINE	0x1000		/* assume multiple lines */
 #define PMf_SINGLELINE	0x2000		/* assume single line */
 #define PMf_FOLD	0x4000		/* case insensitivity */
 #define PMf_EXTENDED	0x8000		/* chuck embedded whitespace */
 
-/* mask of bits stored in regexp->extflags
-   these all are also called RXf_PMf_xyz
- */
+/* mask of bits that need to be transfered to re->extflags */
 #define PMf_COMPILETIME	(PMf_MULTILINE|PMf_SINGLELINE|PMf_LOCALE|PMf_FOLD|PMf_EXTENDED)
 
 #ifdef USE_ITHREADS
diff --git a/pp.c b/pp.c
index 4505890..4523584 100644
--- a/pp.c
+++ b/pp.c
@@ -4629,7 +4629,7 @@ PP(pp_split)
 		++s;
 	}
     }
-    else if (rx->precomp[0] == '^' && rx->precomp[1] == '\0') {
+    else if (rx->extflags & RXf_START_ONLY) {
 	while (--limit) {
 	    for (m = s; m < strend && *m != '\n'; m++)
 		;
diff --git a/pp_ctl.c b/pp_ctl.c
index 8506daa..b439c3e 100644
--- a/pp_ctl.c
+++ b/pp_ctl.c
@@ -177,7 +177,7 @@ PP(pp_regcomp)
 
     if (!PM_GETRE(pm)->prelen && PL_curpm)
 	pm = PL_curpm;
-    else if (strEQ("\\s+", PM_GETRE(pm)->precomp))
+    else if (PM_GETRE(pm)->extflags & RXf_WHITE)
 	pm->op_pmflags |= PMf_WHITE;
     else
 	pm->op_pmflags &= ~PMf_WHITE;
diff --git a/regcomp.c b/regcomp.c
index 8772237..c1c141a 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -4582,6 +4582,11 @@ reStudy:
         r->paren_names = (HV*)SvREFCNT_inc(RExC_paren_names);
     else
         r->paren_names = NULL;
+    if (r->prelen == 3 && strEQ("\\s+", r->precomp))
+	r->extflags |= RXf_WHITE;
+    else if (r->prelen == 1 && r->precomp[0] == '^')
+        r->extflags |= RXf_START_ONLY;
+
 #ifdef DEBUGGING
     if (RExC_paren_names) {
         ri->name_list_idx = add_data( pRExC_state, 1, "p" );
diff --git a/regexp.h b/regexp.h
index f7c9833..d02b321 100644
--- a/regexp.h
+++ b/regexp.h
@@ -116,6 +116,10 @@ typedef struct regexp_engine {
 
 /* Flags stored in regexp->extflags 
  * These are used by code external to the regexp engine
+ *
+ * Note that flags starting with RXf_PMf_ have exact equivalents
+ * stored in op_pmflags and which are defined in op.h, they are defined
+ * numerically here only for clarity.
  */
 
 /* Anchor and GPOS related stuff */
@@ -125,20 +129,22 @@ typedef struct regexp_engine {
 #define RXf_ANCH_GPOS   	0x00000008
 #define RXf_GPOS_SEEN   	0x00000010
 #define RXf_GPOS_FLOAT  	0x00000020
-/* five bits here */
+/* two bits here */
 #define RXf_ANCH        	(RXf_ANCH_BOL|RXf_ANCH_MBOL|RXf_ANCH_GPOS|RXf_ANCH_SBOL)
 #define RXf_GPOS_CHECK          (RXf_GPOS_SEEN|RXf_ANCH_GPOS)
-#define RXf_ANCH_SINGLE         (RXf_ANCH_SBOL|RXf_ANCH_GPOS)        
-/* 
- * 0xF800 of extflags is used by PMf_COMPILETIME 
- * These are the regex equivelent of the PMf_xyz stuff defined 
- * in op.h
- */
-#define RXf_PMf_LOCALE  	0x00000800
-#define RXf_PMf_MULTILINE	0x00001000
-#define RXf_PMf_SINGLELINE	0x00002000
-#define RXf_PMf_FOLD    	0x00004000
-#define RXf_PMf_EXTENDED	0x00008000
+#define RXf_ANCH_SINGLE         (RXf_ANCH_SBOL|RXf_ANCH_GPOS)
+
+/* Flags indicating special patterns */
+#define RXf_START_ONLY		0x00000200 /* Pattern is /^/ */
+#define RXf_WHITE		0x00000400 /* Pattern is /\s+/ */
+
+/* 0xF800 of extflags is used by (RXf_)PMf_COMPILETIME */
+#define RXf_PMf_LOCALE  	0x00000800 /* use locale */
+#define RXf_PMf_MULTILINE	0x00001000 /* /m         */
+#define RXf_PMf_SINGLELINE	0x00002000 /* /s         */
+#define RXf_PMf_FOLD    	0x00004000 /* /i         */
+#define RXf_PMf_EXTENDED	0x00008000 /* /x         */
+/* these flags are transfered from the PMOP->op_pmflags member during compilation */
 #define RXf_PMf_COMPILETIME	(RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_LOCALE|RXf_PMf_FOLD|RXf_PMf_EXTENDED)
 
 /* What we have seen */
-- 
2.7.4