On Wed, 16 Jul 2003, Paul Johnson wrote:

> This has gotten a bit more worrisome now. I only seem to get the 52 response
> if I trace through the execution with the debugger. If I let it run, it goes
> away forever. If I break into the execution I'm always staring at an
> assembler ret instrucion in high memory that can't be traced.
>
> ugh :( Not had one of these for a long time....

I believe that particular effect is an artifact of your debugging and not a
bug in libcurl.

> btw, I'm pretty sure the person doing the server side of my app is writing
> http compliant code - he practically invented it! :)

Then I suggest you check what kind of data the server actually returns. If not
with libcurl's HEADER or DEBUGFUNCTION, use a network tool to capture the HTTP
traffic.

And check the server end. What does the server believe it has received and
what has it sent?

AFAIK, there's only one bug in 7.10.5 that can cause error 52 to occur
wrongly, and that only happens if the server returns non-compliant HTTP (which
is fixed in CVS). There might of course be others still unknown to me.

-- 
 Daniel Stenberg -- curl: been grokking URLs since 1998
-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0