On 6/9/06, T. Wassermann <mail@tobias-wassermann.de> wrote:
> Hi,
>
> I'm writing a german book about subversion and I also use subversion
> very extensive for my project.
>
> Now I have a question without any answer: When should I use svnserve -d
> and when svnserve -d -T. The questions is: What's better on forking
> svnserve new connections or what's better on create a thread for a new
> connection? If there is no direct answer (or answers like "it's your
> choice..."): What is the matter of the -T option? Why does svnserve have
> it (i believe, this must make sense :-) )?
From a user's perspective there's very little difference between
forking and threaded svnserve. The threaded code is required for
systems where fork is not available (like windows), and thus it exists
on unix systems as well. From a developer's perspective it's often
easier to debug problems when you're running in threaded mode
(debuggers often have trouble with forking processes), but unless
you're the kind of person who likes to poke around at the internals of
the process in gdb it probably doesnt matter to you.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Fri Jun 9 22:09:06 2006