"Gerrit P. Haase" <[EMAIL PROTECTED]> writes:

> uf_uristr: ==> mailto:[EMAIL PROTECTED]
> uf_uristok 13
> ok 14
> r: resolving [EMAIL PROTECTED]
> uf_uristr: ==> mailto:[EMAIL PROTECTED]

Ok.  I understand now.  After open(STDERR, ">&STDOUT") both file
handles run with their own buffers and they might flush at unfortunate
times.  And cygwin appears to have smaller buffers than most others.
Perhaps I could get away with *STDERR = *STDOUT?  Another fix could be
to turn on $| for both file handles, but I think I will just go with
something similar to the first patch we had.

Thanks for looking into this.


Index: t/heuristic.t
RCS file: /cvsroot/libwww-perl/uri/t/heuristic.t,v
retrieving revision 1.7
retrieving revision 1.8
diff -u -p -u -r1.7 -r1.8
--- t/heuristic.t       2001/02/27 17:19:33     1.7
+++ t/heuristic.t       2001/09/14 17:04:20     1.8
@@ -6,8 +6,10 @@ if (-f "OFFLINE") {
 print "1..14\n";
 use URI::Heuristic qw(uf_urlstr uf_url);
-open(STDERR, ">&STDOUT");  # redirect STDERR
+if (shift) {
+    $URI::Heuristic::DEBUG++;
+    open(STDERR, ">&STDOUT");  # redirect STDERR
 print "not " unless uf_urlstr("http://www.sn.no/";) eq "http://www.sn.no/";;
 print "ok 1\n";

Reply via email to