Re: [Gnu-arch-users] Ruminations on Arch Desiderata

[Paul Snively]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Stephen!

On Tuesday, September 16, 2003, at 04:05  AM, Stephen J. Turnbull wrote:

> Is that the one with all the colons?

Yes, the MFS and HFS families of filesystems do indeed use colons as a 
path separator. OTOH, Macintosh software almost never used paths for 
anything pre-Mac OS X.

>   IIRC it was basically Unix file
> system safe if you just left it alone.

That's incorrect, which is why we have MacBinary, MacBinary II, 
MacBinary III, AppleSingle, AppleDouble, and binhex encoding formats 
for these files.

>   Is there a reason why you need
> to be able to read this other than from a Mac?

It could be quite useful, e.g. when a file from an HFS filesystem has 
non-default metadata so the file is encoded with AppleSingle upon 
commiting to an archive, to be able to extract just the "data fork" on 
non-HFS filesystems. Obviously this should be optional behavior.

>   I mean, ISTR it was
> transparent to Mac users, so it should be transparent to tla, too, if
> you run tla from a Mac running AppleTalk or whatever.
Perhaps. I'm not interested in a Mac-only tla at all, however. And in 
any case it would be a lot of work: the diffing etc. tools would have 
to be rewritten to know about resource forks and HFS metadata. Ugh.

> You don't want to do this if you can possibly avoid it.  Let Apple do
> it for you.  They've been dealing with the corner cases for more than
> a decade, maybe two, and the system calls are the right interface to
> insinuate in any case.  (The idea is that either the string really is
> a file system object's name, or what are you doing handing it to the
> system call in the first place?  And the server knows it's an
> AppleTalk server!  But how does tla know if it's a native object or an
> AppleTalk server resource?)  If you try to emulate it within tla for
> non-Apple machines accessing AppleTalk, you probably aren't going to
> get it right.  If you absolutely must, then do it as close to the file
> system interface as you can.

I don't need either tla or Apple to do it for me; I have access to 
AppleSingle encoding/decoding libraries. As I mentioned, this format is 
being used every day in tools like netatalk, MacCVSPro, and CVL, all of 
which routinely talk to non-Mac servers (let me assure you that both of 
those CVS clients virtually never talk to CVS servers running on a 
Mac). :-) So my question remains: does anyone know what an appropriate 
bottleneck location in tla would be for integrating AppleSingle 
support, taking into consideration local filesystem access, access to 
network resources via tla's libraries, access to WebDAV via libneon, 
etc.?
> --
> Institute of Policy and Planning Sciences     
> http://turnbull.sk.tsukuba.ac.jp
> University of Tsukuba                    Tennodai 1-1-1 Tsukuba 
> 305-8573 JAPAN
>                Ask not how you can "do" free software business;
>               ask what your business can "do for" free software.

Best regards,
Paul

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (Darwin)

iEYEARECAAYFAj9nIDgACgkQugPBK9DeteoTiACgpFGcGAVygZIc0N9i3a27wXTj
/IwAoNNaVtbsFZF/W5tIMyLOcJBVwbi5
=/OKf
-----END PGP SIGNATURE-----