Subject: Re: [netatalk-admins] End the Beta!
From: Mike (dugan@libwais.sonoma.edu)
Date: Thu Nov 11 1999 - 20:23:27 EST
On Fri, 12 Nov 1999 webmaster@euronetwork.org wrote:
> Subject: Re: [netatalk-admins] End the Beta!
> Maybe I send you an small extration of this log file, so you can see
> what a lot of problems accured?
>
> After loosing a lot of date, nearly 350MB, an no help form nowhere, what
> would you do. This company has to work with this machine, not trying
> this and that. It's not fo fun what they are doing. And I had spend a lo
> of hour an money to solve these problems, but they still remain.
>
> There are three other companies who have get servers form our company,
> working very well with Linux, but they are using PC's as client with
> SAMBA. There was never ever any problem or mistery information in the
> message logfile or any lost of data. This only happens with netatalk!
The source code is there. If you really want to examine a problem, you are
free to examine the source code, and figure out where the problems are
occuring. I am sure the error messages are documented in the source, and
you can search for what conditions must be true in source for a specific
error to be logged. You want better verbage for error messages? Suggest
some.
If you find the problem, then you can notify the maintainer of the error,
and it is very likely the patch will be incorporated in the next release.
I use netatalk with samba for file sharing of files, and have not had any
major problems. Most of the problems have to do with training users for
proper exchange of files, and naming conventions. (For things like
extentions for proper resource mapping.) Samba has a ./configure option
"--with-netatalk" which is supposed to deal with renaming resource fork
info held in the .AppleDouble folder whenever the data fork is renamed
from a samba session, but I have not finished testing this. (Read the
samba docs for info on this.)
The point is this:
If you have time to to complain about a project, then why not use that
time to contribute instead? If you have time to invest in examining
netatalk, and complaining about a lack of certain features, maybe you can
spend time examining code; it will make your netatalk service/daemons more
stable, and help other admins to have a more stable product. You want
something new added? You want a new feature? Add it, send a diff patch
file to the maintainer and see if they will add it.
When Apple offered the NetBooting mac system with Mac OS X, I used a
sniffer and reverse engineered the bootp/dhcp,tftp,AFS/IP process and
documented it. I patch dhcp source to allow for a Linux box to offer net
booting to the newer Mac machines. Certainly, this was an examination to
see what was going on top try an solve problems I found in Mac OS X that I
could not fix, but it was work I started so that others could choose to
use it.(It is not a simple procedure, but you can see what I found in my
analysis at http://libweb.sonoma.edu/mike/macnc/ )
I found a problem, read other people complaints about it, searched the
internet asked questions about it on USENET, and even contacted Apple to
see if they had a solution. I could not find one, so I implemented my own
for testing. In the end a modification to DHCP source was the biggest
issue. If I was interrested, I did get feedback from the maintainers of
DHCP source that I might be able to get a certain kind of patch added to
the new dhcp source tree if I implemented it. I chose to not implement the
full feature addition to dhcp source to give it back to the original
maintainers of dhcp because I had no present use for it.
You have a complaint about netatalk, and problems? Try examining the
problems and see what is wrong and then provide a patch. You have free
time? Contribute to A. Sun with time and code. You want a closed system
where you buy something and pay for support calls, and then cannot do
anything or hire anyone else to fix a problem you find in a product? Talk
to Microsoft, or another closed system product.
Beta or not? Netatalk is more stable than many commercial closed product
in "final release."
This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:17:24 EST