Message-ID: <41B3B589.8010309@arbash-meinel.com>
Date: Sun, 05 Dec 2004 19:27:37 -0600
From: John A Meinel <john@arbash-meinel.com>
User-Agent: Mozilla Thunderbird 0.9 (Windows/20041103)
MIME-Version: 1.0
To: Andrew Suffield <asuffield@debian.org>
Subject: Re: [Gnu-arch-users] Arch Versus CVS Versus Subversoin
References: <20041205000613.PNEI7152.lakermmtao09.cox.net@nonerjsnum1tkq>	<41B376B4.2060605@arbash-meinel.com>
	<20041206010853.GA2799@suffields.me.uk>
In-Reply-To: <20041206010853.GA2799@suffields.me.uk>
Cc: gnu-arch-users@gnu.org
Precedence: list
Content-Type: multipart/mixed; boundary="===============0972529019=="
Sender: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Errors-To: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Status: RO
Content-Length: 3947
Lines: 119
This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--===============0972529019==
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature";
	boundary="------------enig8CF3EE9C248370C1F5900CE5"

This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--------------enig8CF3EE9C248370C1F5900CE5
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Andrew Suffield wrote:
> On Sun, Dec 05, 2004 at 02:59:32PM -0600, John A Meinel wrote:
> 
>>Arch does not support binary diff in the subversion sense.
> 
> 
> Nonsense, it works just fine.
> 

No. In subversion, if you have a binary file, and you make changes, it 
only stores the changes to the file. (something like xdelta). So if you 
have a binary file that is 100k long, and you append 10k to it, the 
changeset that is stored in the archive is only ~10k in size. (I think 
the same is true if you modify only 10k of the file.)

However, with arch, if a file is declared binary and it is considered to 
be changed, then a completely new copy of the file is added to the 
repository. Actually, I just checked, and 2 new copies are added to the 
repository, the old version, and the new one.

Here are the steps to reproduce:

$ mkdir binary
$ cd binary
$ tla init-tree binary--dev--0
$ cp ~/image.png .
$ tla add image.png
$ tla import -S
$ dd if=/dev/zero bs=1000 count=10 >> image.png
$ tla changes
Mb image.png
$ tla commit -s "modified the binary."
$ tar xvzf 
$ARCHIVE/binary/binary--dev/binary--dev--0/patch-1/binary--dev--0--patch-1.patches.tar.gz
$ cd binary--dev--0--patch-1.patches/patches
$ ll
-rw-r--r--    1 jfmeinel users       29632 Dec  5 19:16 image.png.modified
-rw-r--r--    1 jfmeinel users       19632 Dec  5 19:16 image.png.original

This is most definitely NOT what the original poster was meaning by 
"binary diff".

Arch handles binary files, it does very well with them. But it *does 
not* just store the differences. It actually stores 2 whole copies for 
any change, just so you can compute the difference yourself if you so 
choose.

> 
>>In general, to do appropriate binary diffs, you must know the file 
>>format, and have an intelligent diff program.
> 
> 
> That's not a binary diff. That's some other format-specific diff.
> 
> 
>>I think the reason people want binary diff (it's why I wanted it) was to 
>>decrease the amount of space used in the archive.
> 
> 
> Compression is not related in any way.
> 

compression is not binary diff. True. But the reason people want binary 
diff is so that when you have your 10MB Word document, and you change 2 
lines in it, when you do the commit it doesn't have to upload 10MB of 
data. In arch's case, it actually will upload about 20MB of data. (It 
will compress first, but that is a different subject.)

However, since arch does compression, you should get pretty good 
compression. In fact, if your files are small enough, you should get 
great compression if one is mostly just a copy of the other one.

John
=:->

--------------enig8CF3EE9C248370C1F5900CE5
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Cygwin)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFBs7WNJdeBCYSNAAMRAlBKAKCV8efXw8NFFZhWh/lfrJVsMBtFUwCfSKpp
pVkIS0VAr70OI0gpxatqw/I=
=10uY
-----END PGP SIGNATURE-----

--------------enig8CF3EE9C248370C1F5900CE5--


--===============0972529019==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Gnu-arch-users mailing list
Gnu-arch-users@gnu.org
http://lists.gnu.org/mailman/listinfo/gnu-arch-users

GNU arch home page:
http://savannah.gnu.org/projects/gnu-arch/
--===============0972529019==--