Re: [Gnu-arch-users] Arch Versus CVS Versus Subversoin

[John A Meinel]

Andrew Suffield wrote:
> On Sun, Dec 05, 2004 at 08:09:30PM -0600, John A Meinel wrote:

[...]
> The difference in space consumed is not all that large for typical
> applications. It's not a big deal. As always, nobody cares about it
> enough to bother to implement it. Including you.

I don't know if I mentioned this earlier. *I* am fine with how arch does 
it. I agree that it isn't that big of a deal (I believe I've stated 
this). I think storing 2 copies of the file is a little expensive, but I 
perfectly understand why. It goes back to the idea that a single 
changeset is useful. If you don't store everything, then you need 
history, or something external.

I understand why, and I certainly don't care to change it.

*However* the original poster was asking if we do "binary diff like 
Subversion", and SVN's claim to fame is that they do delta compression 
on binary files.

So when you responded with "yes we do, it works perfectly the same as 
SVN." You were incorrect.

Actually, as far as stating that it should be called delta compression 
instead of binary diff goes against what I would consider the popular 
terminology. This is the first time I've heard it called delta 
compression, and though technically correct, it isn't commonly used. 
While I certainly will try to improve my usage, when someone asks about 
it, I would bet that 90% of the time they mean delta instead of diff. 
99.999% if they are also mentioning SVN at the same time.


To hopefully change topic slightly, does anyone know of a SCM that 
*doesn't* at least do something with binary files? I guess I read 
something that RCS had to uuencode and do a textual diff on the file. Do 
any of the modern ones fail to at least create full copies of the file?

John
=:->

signature.asc
Description: OpenPGP digital signature