Subclipse performance (Was: JavaSVN)

Mark Phippard wrote:

>[...] I suppose if the JavaSVN option were
>significantly faster then I could change my mind pretty easily.
>
>
If we're talking performance in the native Subversion client I think
that we could gain quite a bit from hitting the WC status at a lower
level (svn_wc instead of svn_client).
That way, we could control locking of the admin area in a finely grained
manner, and thas enumerate the statuses a lot cheaper.

Another similar is to make the repository browser reuse the RA session,
if it had access. As it works now, it is quite sluggish across WAN and
even LAN boundaries, of course especially through SSL.

Only problem is, those interfaces are only exposed in the C API (and
Python and Perl, through SWIG), but not the JavaHL API.

-Jesper
Received on Wed Dec 15 06:59:53 2004