Re: [gnugo-devel] Error in regression failure HTML view

[Trevor Morris]

Thanks for the bug report!  I always like to hear that these are useful.

At 08:49 PM 9/21/2002 +0200, you wrote:
>Just to let you know.  There seems to be a few problems with the
>otherwise excellent HTML view of the regression failures.  So far I
>have found:
>
>1. There are no move valuations in the traces.  Instead, there are two
>   copies of the move reason lists.
As Gunnar pointed out, the two copies probably aren't a problem.
When I run it here, the move reasons do exist.  Not sure why the
web site version doesn't have them.  I'll continue to investigate.


>2. If I try to view by category or unexpected results, I get a page
>   that says "Software error".
Fixed with the following patch.

Note that auto01.tst contains an invalid command:
auto01.tst:6 dragon_defend f3

Ironically, this test passed as "unknown command: 'dragon_defend'"
is in fact not equal to 0.

This is what was causing the problem.  I think owl_defend was
intended.  (Though should owl_defend be renamed to
dragon_defend? )  The corresponding command line option
is, after all, --decide-dragon, not --decide-owl.


Here's the patch, which I will bravely try via e-mail:

? regression/html
Index: regression/regress.plx
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/regress.plx,v
retrieving revision 1.18
diff -u -r1.18 regress.plx
--- regression/regress.plx      2 May 2002 21:09:32 -0000       1.18
+++ regression/regress.plx      21 Sep 2002 21:13:39 -0000
@@ -173,8 +173,8 @@
 
 sub insinglequote {
   my $s = shift;
-  $s =~ address@hidden@\\\\@;
-  $s =~ s@'@\\'@;
+  $s =~ address@hidden@address@hidden;
+  $s =~ s@'@\\'@g;
   return "'$s'";
 }