BasementJack wrote:
> Stefan,
>
> Virus scanner sounds like a good avenue to look into, I'll need to
> try that! I can totally see it mingling with the http traffic.
> I am using eset nod32 AV - are you aware of any issues with this
> one?
>
> As for the comment about a missing log attachment and line 1&2, I
> was talking about my response table:
> (I can see now that it's not as clear in some newsreaders as it was
> when I wrote it due to line wrapping)
>
> Test results: (checking out a small 10k project with a handful of
> revisions)
>
> Line 1:
> F: Tortoise client build 14361 -> VisualSVN server 1.6.1 over http
> Slow, double sets of PROPFINDS found in trace. 5-6 seconds
> -----------------------------------------------------
> Line 2:
> F: Tortoise client build 14361 -> VisualSVN server 1.6.1 over https
> Slow - did not trace (did not expect to be able to decypher encrypted
> https traffic) 5-6 seconds
> ----------------------------------------------------
> Line 3:
> A: Subversion command line client 1.5.3 -> VisualSVN server 1.6.1
> over
> https: seemed reasonable (<1 second)
> ------------------------------------------------------
> Line 4:
> A+++: Tortoise client build 14361 -> SubversionTigrisServer 1.5.3
> over
> svn protocol FAST FAST FAST*** (<1 second) ***
Best guess: the VisualSVN server has SSPI authentication enabled, and
that's why it's slow for you: for every connection, it first tries SSPI
auth and then (if that doesn't authenticate) falls back to basic auth.
both the server from tigris and the command line client don't have SSPI
compiled in, so that's why those are fast in this situation (you can
enable SSPI for the tigris server too, but you have to do that manually
and install mod_auth_sspi yourself).
Stefan
--
___
oo // \\ "De Chelonian Mobile"
(_,\/ \_/ \ TortoiseSVN
\ \_/_\_/> The coolest Interface to (Sub)Version Control
/_/ \_\ http://tortoisesvn.net
Received on 2008-11-03 18:52:51 CET