Nicholas Clark's tweaks to Encode compile process

 1. Use decimal constants to unit string arrays rather than
    very long string litterals (latter having possible "\0"
    on end issues, and needing ANSI C line splicing).
 2. Replace tail-recursion during insert process with itteration.
Nick C sees 12% speedup from these.

p4raw-id: //depot/perlio@14618

diff --git a/ext/Encode/compile b/ext/Encode/compile
index ee6d778..8fe6c0f 100755 (executable)
--- a/ext/Encode/compile
+++ b/ext/Encode/compile
@@ -403,23 +403,22 @@ sub enter
 {
  my ($a,$s,$d,$t,$fb) = @_;
  $t = $a if @_ < 4;
- my $b = substr($s,0,1);
- my $e = $a->{$b};
- unless ($e)
-  {     # 0  1  2  3         4  5
-   $e = [$b,$b,'',{},length($s),0,$fb];
-   $a->{$b} = $e;
-  }
- if (length($s) > 1)
-  {
-   enter($e->[3],substr($s,1),$d,$t,$fb);
-  }
- else
-  {
+
+ while (1) {
+  $s =~ s/(.)//s;
+  my $b = $1;
+  my $e = $a->{$b};
+  #                 0  1  2  3           4  5
+  $a->{$b} = $e = [$b,$b,'',{},1+length($s),0,$fb] unless $e;
+  unless (length($s)) {
    $e->[2] = $d;
    $e->[3] = $t;
    $e->[5] = length($d);
+   return;
   }
+  # Tail recursion.
+  $a = $e->[3];
+ }
 }
 
 
@@ -450,16 +449,12 @@ sub outstring
    }
    $strings{$s} = $sym = $name;
    $strings += length($s);
-   printf $fh "\nstatic const U8 %s[%d] =\n",$name,length($s);
-   # Do in chunks of 16 chars to constrain line length
-   # Assumes ANSI C adjacent string litteral concatenation
-   while (length($s))
-    {
-     my $c = substr($s,0,16,'');
-     print  $fh '"',join('',map(sprintf('\x%02x',ord($_)),split(//,$c))),'"';
-     print  $fh "\n" if length($s);
-    }
-   printf $fh ";\n";
+   my $definition = sprintf "static const U8 %s[%d] = { ",$name,length($s);
+   # Maybe we should assert that these are all <256.
+   $definition .= join(',',unpack "C*",$s);
+   # We have a single long line. Split it at convenient commas.
+   $definition =~ s/(.{74,77},)/$1\n/g;
+   print $fh "$definition };\n\n";
   }
  return $sym;
 }