Hi, I run a web server for a university and am working on an online
portfolio project and I have been looking at webdav and subversion as
possibilities in the stuff I am working on and wanted to ask some questions
and see if I could get some feedback on the sorts of things I am
considering doing as to whether they are possible and a good idea.
On the webserving, I have had the entire web repository in cvs for about
a year now. There are a lots of different pages managed by lots of
different people and it was really useful to have the traceability and fine
grained revision control that cvs offered. (The down side was a complete
duplication of the web hierarchy and a variety of geek oriented
interfaces (as well as cvs's usual problems.)) I've been watching
subversion for a while and I am glad to see you reach alpha. It is looking
good.
I have been trying to come up with a way to have requests be transparently
mapped to the subversion repository so that I don't have to maintain a
separate checked out version for the webservers use. Not sure if I can
manage it (going to play with mod_rewrite), but if I could, will this have
any adverse affects on the repository? (being read accessed with the
frequency of a medium traffic webserver)
Oh, while I am on it. The webserver process has to have write access to
the repository regardless of whether it is writing or not? Well, that's
not really a question since I have verified that... Is there any way to
avoid this? With cvs I had a separate locks directory that the webserver
user could write to, but the main repository was restricted. It meant that
an attacker could screw with the locks and blow the webserver's quota, but
the repository as a whole was safe. Is this sort of workaround possible
with subversion?
The other questions I have surround a project I am working on for the
education department here. They have to be able that they are teaching
well, so they are getting the students work online and setting up an
online portfolio where it can be discussed. I thought that it would be
really handy if I made the file system stuff wedav based. That way they
could make changes and what not and I wouldn't have to worry about issues
that arise when there is only one copy. (Changes made after a review.
Changes made as part of a later assignment that change the nature of the
project. Etc.) It also makes it more widely accessible as deltaV compliant
clients become available.
The thing is that they will all be creating a couple hundred megs worth of
stuff (education majors make movies and power point slides and all
sorts of interesting things nowadays) and over time that will have to be
archived and moved away.
If I have a single repository then there is no way for me to remove a tree
completely, is there? From what I was reading of how the trees and deltas
and what-not are stored it seems like it would be hard. The other option I
have is lots of individual repositories, but there is no way to
dynamically add them to the server is there? I would have to create really
big apache config files and reload the server (there are 5-600 users at a
time here). It would make creating dynamic adding of users impossible it
seems. (Well, cumbersome at least.)
Any suggestions would be much appreciated. Congratulations again on the
good work.
Will Holcomb
P.S. I'm not on the list, so please cc in any replies.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Aug 1 05:06:33 2002