On Dec 19, 2007, at 10:32 AM, Erik Huelsmann wrote:
> Issue 2305 is about 'svn info' failing when the current working
> directory isn't a working copy but the target is a symlink to a
> working copy.
>
> I posit that running 'svn info' on a symlink target like that is
> ambiguous, but that we should in effect be consistent accros symlinks
> to directories and files.
>
> If the symlink to a versioned file would fail, then also the symlink
> to a versioned directory should fail.
>
> Since we currently exhibit the above behaviour, I think the bugreport
> is INVALID.
>
> Comments?
FWIW, I run into this because I've got a structure similar to this:
..../vol1/wc1/
..../vol1/wc3/
..../vol2/wc2/
..../wc/wc1 -> ../vol1/wc1
..../wc/wc2 -> ../vol2/wc2
..../wc/wc3 -> ../vol1/wc3
And I've a env var wc=..../wc
Because some of the wc's are large working copies of debug builds and
I can only have volumes of up to 2 Gb on AFS. I sometimes have to
shuffle things about, but it gets frustrating remembering which vol#
a given wc is in. Hence the symlinked structure.
I do find it a bit frustrating that "svn info $wc/wc1" fails.
Apparently I'm not the only one (hence the issue).
If a symlink is not inside a versioned directory (which is why svn
reports an error) and points to a wc dir, is it better
1) to assume the user is shooting himself in the foot by calling
'info' on a file that he intended to have versioned inside a wc but
somehow messed up that whole intention wildly by putting the symlink
in the wrong place and then calling 'info' on it in that wrong place, or
2) to assume that since the symlink is not inside a wc and points to
a wc dir, he probably wants 'info' on the target of the symlink ?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Dec 19 21:45:01 2007