Jens Peters wrote:
> Actually at the time when I wrote this email I was still playing with
> compiling bdb. The libs had compilid well, but I got some error in the
> example programs (excxx...exe) because they did not include db_config.h I
Ah, I see. I only compiled the dll - that's all I really need. The
example programs aren't used in TSVN.
> My own test shows that accessing an old bdb repos went fine and the standard
> subversion test suite succeeded as well (expect two tests with the
> property/wildcard things, but this is expected and already fixed in trunk as
> far as I know). So I still feel a little bit weird about changing the crt
> includes, but it seems to work :)
As long as you don't get any warnings or errors, you should be fine.
> By the way, in a first attempt I compiled subversion with the new
> neon-0.25.5 without the sspi support (#undef have_sspi). When I started the
> svn test suite the first time with mod_dav_svn the tests were failing
> instantly, because commands like:
>
> svn --username rjandom --password ...
>
> arrived at the apache with my domainname in front of it: domain\\rjandom. Of
> course apache returned a authorization failed. I took the 0.25.4 version,
> disabled sspi support, and got the same result. Did you got something like
> this in your builds?
I don't run the test suite. But I remember some mails on the Subversion
dev list about such test failures.
Also, I've enabled SSPI support again on the TSVN trunk (with the 0.25.5
neon version). That way, people in need of SSPI can at least use a
nightly, while we don't break anything for users using the official
versions.
I think Joe will implement some API changes to neon so
enabling/disabling SSPI support can be done at runtime - but it will
also need changes in Subversion to make this work. So I hope all this
will be done by the time Subversion 1.4.0 gets out.
> Now I'm running the suite with 0.25.5 sspi enabled and the tests are running
> quite well. But in this case --username and --password is simply completely
> ignored and neon always passes my own credentials. I don't know if I like
> that, I would prefer the old situation over that. But I already read that
> the neon guys have plans to implement changing the authenticate method at
> runtime. So lets wait for svn 1.4.
Yes, with SSPI enabled, you can't pass username/password anymore. If you
look at the neon source, you'll see that it first tries to authenticate
with SSPI, and only if that fails it tries the basic authentications.
But depending on the server setup, there might not be a basic
authentication or SSPI will succeed (e.g. with user "Guest"), then it
simply won't get to the basic authentication at all.
That's why we need neon to implement enabling/disabling SSPI at runtime.
Stefan
--
___
oo // \\ "De Chelonian Mobile"
(_,\/ \_/ \ TortoiseSVN
\ \_/_\_/> The coolest Interface to (Sub)Version Control
/_/ \_\ http://tortoisesvn.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tortoisesvn.tigris.org
For additional commands, e-mail: dev-help@tortoisesvn.tigris.org
Received on Sat Jan 21 18:09:35 2006