From 61a36c016d02518f6bc19ca7f7d1b5a51cc47d49 Mon Sep 17 00:00:00 2001
From: Jarkko Hietaniemi <jhi@iki.fi>
Date: Sat, 12 Jan 2002 18:06:49 +0000
Subject: [PATCH] Special treatment for U+03B0 and U+0390 in //i, this means
 that we can remove the minlen pessimisations introduced by the #14096.

p4raw-id: //depot/perl@14219
---
 pp_hot.c  |  4 +---
 regcomp.c | 44 ++++++++++++++++++++++++++++++++++++++++++++
 regexec.c |  4 +---
 3 files changed, 46 insertions(+), 6 deletions(-)

diff --git a/pp_hot.c b/pp_hot.c
index df52bb1..b759153 100644
--- a/pp_hot.c
+++ b/pp_hot.c
@@ -1235,9 +1235,7 @@ PP(pp_match)
 	pm = PL_curpm;
 	rx = PM_GETRE(pm);
     }
-    if (rx->minlen > len &&
-	!PL_reg_match_utf8 /* ANYOFs can balloon to EXACTFs */
-	)
+    if (rx->minlen > len)
       goto failure;
 
     truebase = t = s;
diff --git a/regcomp.c b/regcomp.c
index 65b0a5f..bf1b42f 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -736,6 +736,50 @@ S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp, I32 *deltap, reg
 		    n = nnext;
 		}
 	    }
+
+	    if (UTF && OP(scan) == EXACTF) {
+/*
+  Two problematic code points in Unicode casefolding of EXACT nodes:
+
+   U+0390 - GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS
+   U+03B0 - GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS
+
+   which casefold to
+
+   Unicode			UTF-8
+
+   U+03B9 U+0308 U+0301		0xCE 0xB9 0xCC 0x88 0xCC 0x81
+   U+03C5 U+0308 U+0301		0xCF 0x85 0xCC 0x88 0xCC 0x81
+
+   This means that in case-insensitive matching (or "loose matching",
+   as Unicode calls it), an EXACTF of length six (the UTF-8 encoded byte
+   length of the above casefolded versions) can match a target string
+   of length two (the byte length of UTF-8 encoded U+0390 or U+03B0).
+   This would rather mess up the minimum length computation.
+
+   What we'll do is to look for the tail four bytes, and then peek
+   at the preceding two bytes to see whether we need to decrease
+   the minimum length by four (six minus two).
+
+   Thanks to the design of UTF-8, there cannot be false matches:
+   A sequence of valid UTF-8 bytes cannot be a subsequence of
+   another valid sequence of UTF-8 bytes.
+
+*/
+		 char *s0 = STRING(scan), *s, *t;
+		 char *s1 = s0 + STR_LEN(scan) - 1, *s2 = s1 - 4;
+		 char *t0 = "\xcc\x88\xcc\x81";
+		 char *t1 = t0 + 3;
+		 
+		 for (s = s0 + 2;
+		      s < s2 && (t = ninstr(s, s1, t0, t1));
+		      s = t + 4) {
+		      if (((U8)t[-1] == 0xB9 && (U8)t[-2] == 0xCE) ||
+			  ((U8)t[-1] == 0x85 && (U8)t[-2] == 0xCF))
+			   min -= 4;
+		 }
+	    }
+
 #ifdef DEBUGGING
 	    /* Allow dumping */
 	    n = scan + NODE_SZ_STR(scan);
diff --git a/regexec.c b/regexec.c
index 0ceff78..78c4e24 100644
--- a/regexec.c
+++ b/regexec.c
@@ -1554,9 +1554,7 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *
     }
 
     minlen = prog->minlen;
-    if (strend - startpos < minlen &&
-	!PL_reg_match_utf8 /* ANYOFs can balloon to EXACTFs */
-	) {
+    if (strend - startpos < minlen) {
         DEBUG_r(PerlIO_printf(Perl_debug_log,
 			      "String too short [regexec_flags]...\n"));
 	goto phooey;
-- 
2.7.4