On Sat, Nov 5, 2011 at 12:11 AM, Stefan Küng <tortoisesvn_at_gmail.com> wrote:
> I didn't change anything regarding neon and/or SSPI between the 1.7.0RC and 1.7.1.
Thanks for the reply!
This is strange!
And from 1.6 to 1.7 was something changed?
Do you patch neon or use it as-is?
> But I noticed a problem in the thread you linked to here:
> http://permalink.gmane.org/gmane.comp.web.webdav.neon.general/797
>
> using http:// and SSPI with neon does not work at all and hasn't worked for years. Because neon has disabled this. Only when using https:// instead of http:// the SSPI authentication with neon actually works.
> Joe Orton (the developer of neon) once explained that SSPI over http is not secure and therefore disabled in neon.
> It's in the changelog for 0.26.0 (http://www.webdav.org/neon/history.html):
>
> Win32: SSPI support is now only enabled by default for SSL sessions.
Yes, I am aware of this.
However, this is not the problem.
As you can see from [1], I had two different servers, one CentOS and
one Gentoo, while with CentOS as server it works, with Gentoo as
server it does not work... Same client two different servers, same
kerberos REALM.
While with serf it works correctly in all configuration (as long as
SSPI is concerned), well except of the HTTPS long session issue.
Currently I have one configuration which is working on Windows:
TortoiseSVN-1.7.1+cmdline+Subclipse-1.8.2 with serf using HTTP and SSPI.
While I wish to return to the HTTPS+SSPI configuration, but subclipse
suffers from the neon issue I previously found in TortoiseSVN.
Either understand why neon behaves differently or find a way to debug serf...
Thanks!
Alon.
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2873716
To unsubscribe from this discussion, e-mail: [dev-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2011-11-04 23:52:07 CET