Julian Fitzell wrote:
But unfortunatly, I received no comments about it. Instead you core
developers waste your time and duplicate already existing
functionality.
It isn't exactly duplication. For people using ra_dav, a CGI
may be the easiest way to create repositories. But for people
using ra_svn or ra_local, setting up a web server to run the
CGI is likely annoying or impossible.
Oh, the script could be extended to handle ra_local and ra_svn, too.
It is _already_ extended beyond CGI: it acts as hook-scripts, too.
Huh? My point is that users using ra_local and ra_svn may not have a
web server running to serve the CGI and may not want to run
one.
You missed my point :-) I tried to say that they do _not_ need to run a
server. If you take a look at the script, you will find a dispatcher.
This dispatcher checks whether the script was called as a cgi or as a
hook-script. When called as a hook script, _no_ CGI-stuff is executed.
This dispatcher could be extended to check whether it was called
as a command-line-tool or as a Tk-GUI-tool and offer a different
interface to the same functionality.
Implementing all this stuff into svnadmin _could_ bring you further
than such a script if svnadmin would be extended to operate on
URLs instead of only on local paths. I have no clue where the decision
for svnadmin to operate on paths originated from. But as long as this
decision holds true, the benefits you are gaining when you implement
all this in svnadmin are limited.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 14 02:06:07 2006