--On July 25, 2005 1:06:40 AM -0400 Greg Hudson <ghudson@MIT.EDU> wrote:
> On Sun, 2005-07-24 at 23:32 +0200, Justin Erenkrantz wrote:
>> What would be wrong with the ra_layer (server-side) sending down a baton
>> function that could either write to syslog() (in the case of ra_svn) or
>> ap_log_rerror() (in the case of ra_svn/mod_dav_svn).
>
> For starters, we have no server-side ra layer. We have two servers (an
> Apache module, mod_dav_svn, and a standalone daemon, svnserve), which
> make calls to libsvn_repos and libsvn_fs.
>
> So I don't really understand this proposal, since it refers to a layer
> we don't have. It also tries to define a mechanism without first
> defining what logging we will be doing, which is exactly the reason why
> we needed to back off and start this thread instead of continuing the
> old ones.
Of course we have a server-side implementation. I meant placing the logging
calls inside svnserve and mod_dav_svn - which are absolutely the equivalent of
a server-side RA layer. They may not share that much code; but they share the
same place in the protocol stack. -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jul 25 09:34:22 2005