Subversion? (Re: HEADS UP: Importing csup into base)

[Stijn Hoop]

On Sun, Mar 05, 2006 at 08:42:18AM +0900, Hiroharu Tamaru wrote:
> Yes, you can find out where your banches came from, but
> IIRC, you cannot find out what happened to all the branches,
> when you are looking at it from the trunk, or what's
> happening on trunk by looking at a branch.

That's correct.

> I find it invaluable to see which feature is backported to
> which branches when I look into the FreeBSD cvsweb pages.

I think it's more of an artifact of the way CVS stores it's
information, but it's not possible in this way with Subversion as far
as I know.

> Is there any way, when you specify
> https://svn.sandcat.nl/repos/sws/trunk/include/main.php
>                                 ~~~~~~~
> that you'd know if there are any branches that have this
> file forked off from its trunk version,

I don't think there is a way to do that in Subversion right now.

> or know if and when the change made at revision foo (on trunk) was
> merged to a branch by looking at the log on the *trunk side* (you need
> it especially when you are not aware that such a branch
> exists).

I understand. It's a good point, and one that I didn't think of.

> I assume the latter would be solved when they support
> 'real-merging' (as they plan to), but what about the former?

Only time will tell I guess. Maybe it's possible to work around it by
setting a property on the copied files whenever they are copied and
then using that property in CVSWeb. I do hope that they will support
'real merges' as you put it soon.

> I'd loved to prooved wrong, since this is about the only
> thing I miss with subversion.
> 
> Or, well, vendor-branch may be another one..
> Has the 'import' function changed to allow an update to a
> vendor branch now, or do you need 'svn remove' on each
> deleted file and 'svn add' on each added file?

Not that I know of, but using svn_load_dirs.pl makes things a lot
easier. My personal wish is that they would make this a standard
command for 'svn' instead of a script in the tarball that isn't
even installed by default.

Getting back to the point, I'd like to stress again that I'm not
advocating switching _to Subversion_ at this point, simply because I
haven't been able to set up a test repository to test things like
this. I personally like Subversion but there might be better choices
for the project; I read Robert Ollivier's paper on distributed VCSs /
Mercurial yesterday and I like the thought of using a distributed VCS.
But it all depends on the gain, as written earlier.

--Stijn

-- 
Ubuntu is a Linux for human beings. Actually, the word "Ubuntu"
really is an ancient African word for "I am sick of compiling Gentoo".
		-- Jeff Waugh
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-arch/attachments/20060305/babf8848/attachment.bin