Subject: Re: [Netatalk-devel] Re: Netatalk, Macos X and other issues
From: roger.day@globalgraphics.com
Date: Fri Jul 13 2001 - 03:45:04 EDT
I'm quite happy with CVS. It's just that elusive confidence factor. Why can't
they commit? was the comment I got when my boss saw the list pre-releases on the
website.
In objective quality terms there might be no difference btn a pre and a beta,
but we're talking irrational feel-goodness here:-)
At 13/07/2001 03:33:07, Joe Clarke <marcus@marcuscom.com> wrote:
# The instructions for getting CVS to work are quite clear on the CVS page.
#
# http://sourceforge.net/cvs/?group_id=8642
#
# Once you download the source, you run ./autogen.sh from the source
# directory, then configure as normal. Perhaps a nightly CVS snapshot build
# wouldn't be a bad idea. I'd be happy to set this up unless there are any
# objections.
#
# Joe Clarke
#
# On Thu, 12 Jul 2001, Troy R. Baldwin wrote:
#
# > What if one does not know how to use CVS? It sure would be nice if there
were a
# regular
# > beta release or something we can easily compile and use.
# >
# > Thanks,
# > --TRBo
# >
# > Joe Clarke wrote:
# >
# > > I submitted a patch to netatalk that "fixes" the AFP problem with opening
# > > and copying files in OS X. The reason I say "fixes" is that it relies on
# > > the old DID algorithm. I have not tried the mtab stuff yet, so I don't
# > > know if that will work.
# > >
# > > Anyway, if you want OS X support with netatalk, download the CVS version
# > > from http://sourceforge.net/cvs/?group_id=8642, and configure with the
# > > added option --with-did=last. You should be able to use netatalk with OS
# > > X after that.
# > >
# > > The big goal I think is to find the right way to do DID calculations so
# > > that collisions do not exist and DIDs are persistent. If Apple wants to
# > > contribute, that would be great. I'm not sure if anyone else is
currently
# > > working on the DID "Rossetta Stone."
# > >
# > > I'm currently working on porting AFP 3.0 to netatalk. I have most of the
# > > volume and directory code done. I'm not sure how well this will work,
but
# > > it should provide for longer filename support.
# > >
# > > Joe Clarke
# > >
# > > On Wed, 11 Jul 2001 roger.day@globalgraphics.com wrote:
# > >
# > > > We tried the 1.5pre6, which is why I was trying that wacky version
there.
# > > >
# > > > The problem with pre6 was that it caused NFS to stop working -
authentication
# > > > stopped working. Reverting to a previous release fixed this.
# > > >
# > > > I have a proper description of the problem:-
# > > >
# > > > <GMT17-May-2001 23:49:56GMT> B.S.:
# > > > Problem is that netatlk is not making their fileIDs unique and the
AppleShare
# > > > client uses those ids for its hash entries. Thats why we are hitting
the
# assert.
# > > > I will file a bug against netatalk and also change the AppleShare
client so
# that
# > > > it will not assert any more. Although anytime someone tries to do a
ResolveID
# > > > (not too common, I think), it will get back some indeterminate file and
that
# > > > will be bad.
# > > >
# > > > <GMT23-May-2001 21:31:24GMT> B.S.:
# > > > Fix is too catch this case and not assert, but to fail the create call
with
# some
# > > > error. Real fix is to get an updated netatalk installed on the server.
I am
# > > > working on getting a fixed netatalk submitted.
# > > >
# > > > >from here
# > > >
# > > >
#
http://www.publicsource.apple.com/bugs/X%20Server/AppleShare%20X%20Client/2666595.h
# tml
# > > >
# > > > I can't see any bug filed anywhere on the netatalk site - that's the
first
# place
# > > > I looked.
# > > >
# > > > Would this be fixed in pre7?
# > > >
# > > > Another question: why so many pre's?
# > > >
# > > >
# > > >
# > > > At 11/07/2001 16:31:48, Jeff <jeff@univrel.pr.uconn.edu> wrote:
# > > > # roger.day@globalgraphics.com wrote:
# > > > # > # Which version of netatalk are you using?
# > > > # >
# > > > # > The tar file is called pre-asun2.1.4-37b.tar
# > > > # >
# > > > # > it decompresses into a directory called
# > > > # >
# > > > # > netatalk-1.4b2+asun2.1.4
# > > > # >
# > > > # > Is there a deterministic way of finding the version?
# > > > #
# > > > # That's version 1.4b2+asun2.1.4. The series on sourceforge is
currently
# > > > # at 1.5pre6.
# > > > #
# > > > # > # > This is quite important to us as we're relying on this to try
and
# get
# > > > a
# > > > # > product
# > > > # > # > out shortly.
# > > > # > # >
# > > > # > # > We would also like to know -who- is supporting netatalk these
days.
# > > > I've
# > > > # > heard a
# > > > # > # > rumour that Apple are looking for a netatalk developer.
# > > > # > #
# > > > # > # Interesting. Currently most support seems to be via the admin
and
# > > > # > # developer lists.
# > > > # >
# > > > # > This was from someone on the darwin-developers list.
# > > > # >
# > > > # > Also the lack of archive here
# > > > http://netatalk.sourceforge.net/mailing_lists.php
# > > > # > for 2000/2001 does give an appearance of cobwebness to the page.
# > > > #
# > > > # Sorry about that. The netatalk development is in full swing. Check
out
# > > > # http://sourceforge.net/projects/netatalk/ for a little more
information
# > > > # on that.
# > > > #
# > > > # I'd suggest trying 1.5pre6 or the upcoming 1.5pre7, as it probably
# > > > # resolves that issue.
# > > > #
# > > > # Jeff
# > > > #
# > > > Roger
# > > >
# > > >
# > > >
# > > > _______________________________________________
# > > > Netatalk-devel mailing list
# > > > Netatalk-devel@lists.sourceforge.net
# > > > http://lists.sourceforge.net/lists/listinfo/netatalk-devel
# > > >
# > > >
# >
# >
# >
#
Roger
This archive was generated by hypermail 2b28 : Sun Oct 14 2001 - 03:04:44 EDT