On Thursday 03 of December 2009 23:04:19 Daniel Stenberg wrote:
> No, I too have a hard time to see how this happens... These cases tend to
> be very hard without actually being able to login and run and debug these
> things hands-on. They can in fact still be tricky even then!

I am able to get over the bug using a silly patch:

--- a/tests/libtest/lib513.c
+++ b/tests/libtest/lib513.c
@@ -12,12 +12,15 @@

 #include "memdebug.h"

+#include <unistd.h>
+
 static size_t read_callback(void *ptr, size_t size, size_t nmemb, void
*userp)
 {
   (void)ptr;
   (void)size;
   (void)nmemb;
   (void)userp;
+ sleep(2);
   return CURL_READFUNC_ABORT;
 }

The patch above makes it possible to run the same test in a loop without
a single failure. It might imply a synchronization problem, but no idea
where. Client? Server? Perl? Socket implementation on s390?

Any idea?

Kamil
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2009-12-04