Hi Mark,
> It isn't critical, but it works better when you have it. For example,
say
> you want to see the diff or log of a specific rev in the repository.
You
> can specify the complete URL to a specific path, but if it turns out
that
> path was not part of that revision, or does not exist at HEAD, you
will
> get an error. A CLI user would just type the root URL because that is
all
> you need if you are specifying a specific revision. We try to do the
> same. It just helps avoid problems that the GUI can get you into.
There
> are a few cases where TSVN suffers the same problem. I think that
they
> now have a low-level API they can use to get the repository root (I
know I
> have seen them ask for it.) The JavaHL API only makes available a
very
> small and specific subset of the Subversion API. If we want to use a
new
> feature we have to ask for it, wait for it to be added to JavaHL (or
add
> it ourselves) but then also wait for it to make its way to a release
of
> Subversion. In some cases, it is easier to solve the problem other
ways.
FWIW, what I did in svncopy.pl was to break down the path then build it
up
level by level, trying to access each level, until I stopped getting an
error. This allowed me to work out the root for myself.
For example, given
http://subversion/svn/repos/trunk/proj1/core/src/wibble.c
I run through the following steps:
svn log -r HEAD http://subversion/ <== FAILS
svn log -r HEAD http://subversion/svn/ <== FAILS
svn log -r HEAD http://subversion/svn/repos/ <== Succeeds.
>From this I conclude the repository root is http://subversion/svn/repos/
Cheers,
Ian Brockbank
Applications Software Team Leader
e: ian.brockbank@wolfsonmicro.com / apps@wolfsonmicro.com
scd: ian@scottishdance.net
t: +44 131 272 7145
f: +44 131 272 7001
Received on Tue Apr 19 19:25:16 2005