WWW::HTTP again

Andreas Koenig (andreas@filomena)
Thu, 16 Mar 1995 14:25:00 +0000 (BST)


Apologies for my silence during the past days, I was tracking down a bug in MArtijn's code. I had no place in mind to follow the hierarchy problems :)

I'll be back in the discussion soon. Let me show you first the patch:


--- HTTP.pm~	Tue Mar 14 18:41:00 1995
+++ HTTP.pm	Thu Mar 16 13:52:28 1995
@@ -14,7 +14,8 @@
 
     my $this = bless &WWW::Response::new($class, $rc);
     $this->{'SOCKET'} = "WWW::HTTP::$socket";
-    %{$this->{'OUTHEADER'}} = %$headers; # hash copy
+#    %{$this->{'OUTHEADER'}} = %$headers; # hash copy
+    $this->{'OUTHEADER'} = $headers; # hash copy
     $this->{'TIMEOUT'} = $timeout if defined $timeout;
     $this;
 }
@@ -318,6 +319,8 @@
 	else {
 	    $headersref->{$key} = $val;
 	}
+    } else {
+	$headersref->{$key} = $val;
     }
 }
 

Not much, eh? Kept me crazy for a day :) Of course, I realize, that this chunk of code doesn't necessarily do the best ever possible, but it does what you intended, when I read your comments right.
-----

What made me grin for a while was Martijn's statement:

> Just because there is a name space doesn't mean we own it :-)

In fact, I don't agree. I'd hope, that all our efforts bring together different implementations for the mutual benefit. I'd suggest we make the best approach possible, establish the namespace, and take it. LWP as a name for the working group is certainly better than the long name, but let's keep it outside the code area.

BTW, your packages, Martijn, are already in a good shape. I like to use them. Thanks for the good work.

Regards,
andreas