> How about 'svn cat -rX http://server/repo/foo'? I suppose that would
> output "foo". What about 'svn cat -rX+1 http://server/repo/foo'? I
> suppose that returns an error "foo not found". It seems that working
> copy paths and URLs produce different results.
yes, they would produce different results. When explicitely giving an URL
I would expect it to error if that "link" did not exist in the given
revision,
but giving a wc path should follow the moves/copies since its still the
same file, only at another location / with a different name.
Imagine a developer who wants an earlier version of a single file.
He does not know that the file was moved/renamed by another
developer before - he just wants to know how the file looked
at a specific revision. Then he'll use svn cat -rX wc/file > file_at_revX
and he's got the file.
Sure, it's possible to check the logs to find out at which revision(s)
the file was moved/renamed, but only if a) the log messages
are good (means the developer wrote about the move in the log)
and b) the number of revisions of the project are lower than about 100 -
or do you really think that a developer would search through more
than a 100 logs to find something? Maybe on *nix with grep and
good luck to give the correct search pattern, but never on windows.
Another way of getting the file's content would be the svn diff command,
but that does not produce a complete file but only the differences - and
many times that's enough. But this method is useless if you need the
complete file.
Stefan
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Mar 27 23:44:40 2003