Allow anchored patterns to use "first code unit" optimization.

git-svn-id: svn://vcs.exim.org/pcre2/code/trunk@837 6239d852-aaf2-0410-a92c-79f79f948069

1 files changed, 8 insertions, 26 deletions

diff --git a/testdata/testoutput5 b/testdata/testoutput5
index 619942c..ff438e6 100644
--- a/testdata/testoutput5
+++ b/testdata/testoutput5
@@ -194,6 +194,7 @@ Subject length lower bound = 3
 Capturing subpattern count = 0
 Compile options: utf
 Overall options: anchored utf
+Starting code units: a b 
 Subject length lower bound = 1
     bar
  0: b
@@ -205,28 +206,6 @@ No match
     \x{100}
 No match
 
-/^[^ab]/IB,utf
-------------------------------------------------------------------
-        Bra
-        ^
-        [\x00-`c-\xff] (neg)
-        Ket
-        End
-------------------------------------------------------------------
-Capturing subpattern count = 0
-Compile options: utf
-Overall options: anchored utf
-Subject length lower bound = 1
-    c
- 0: c
-    \x{ff}
- 0: \x{ff}
-    \x{100}
- 0: \x{100}
-\= Expect no match
-    aaa
-No match
-
 /\x{100}*(\d+|"(?1)")/utf
     1234
  0: 1234
@@ -479,7 +458,10 @@ Subject length lower bound = 0
     \x{100}X
  0: X
 
-/^\ሴ/IB,utf
+# Use no_start_optimize because the first code unit is different in 8-bit from
+# the wider modes.
+
+/^\ሴ/IB,utf,no_start_optimize
 ------------------------------------------------------------------
         Bra
         ^
@@ -488,9 +470,9 @@ Subject length lower bound = 0
         End
 ------------------------------------------------------------------
 Capturing subpattern count = 0
-Compile options: utf
-Overall options: anchored utf
-Subject length lower bound = 1
+Compile options: no_start_optimize utf
+Overall options: anchored no_start_optimize utf
+Subject length lower bound = 0
 
 /()()()()()()()()()()
  ()()()()()()()()()()