[Samba] "net rpc vampire" case sensitivity problem

[John H Terpstra]

On Fri, 26 Mar 2004, Daniel Kasak wrote:

<snip>

> >Why might we want to impose the restriction needed because of a
> >brain-dead OS implementation on ALL OS platforms?
> >
> >- John T.
> >
> >
> ???
> To make the product easier to use maybe?

There are lots of things that can be done to make life easier, and not
enough resouces to make them all come true.

> It's all very well to say "Oh yeah. That doesn't work. You'll have to
> manually add that, and modify this script here to fit that, and wave a
> dead chicken over this thing here unless that over there has already
> done this, and ignore that error ... that's supposed to happen" and
> claim it's all Microsoft's fault in the process. Maybe it is. But here
> you have an end user ( and I'm relatively motivated compared to most end
> users, even though I probably fall short of a Samba developer ) saying
> that they want a little more integration of tools to make the process
> easier. Merely stating that this problem is Microsoft's fault does
> nothing to address the issue that the vampire command in it's default
> configuration doesn't work in many situations.

There is nothing wrong with asking. Every suggestion is worth making. That
is how priorities get set.

> What I'm suggesting is NOT that we all plague our OSs with Microsoft's
> mistakes to become compatible, but that we merely integrate the loose
> collection of scripts and how-tos and other assorted good advice into a
> coherent package that you can use to achieve the outcome that end users
> are after: to migrate from NT to Samba - not to research and debate the
> finer details of the differences between Microsoft's and Linux's
> implementation of feature X.
>
> I'm also not clear on what restrictions would be imposed by giving the
> user an option to run a script that converts NT user names into
> Linux-friendly user names.

Suggestion noted.

> Surely the amount of time it takes to add such scripts to the how-to
> collection and answer posts such as mine would be better spent in
> integrating the scripts into the actual product?

Wowa! Please understand that I was referring to the Linux implementation
being brain-dead - NOT MS Windows. I totally agree that Samba should
provide the ability to implement the same look and feel of MS Windows
servers.

Please understand that NOT all UNIX/Linux platforms have the limitation
that you can not have a space character, or upper case characters, in the
user name or group name.

In fact, Debian Linux does not have this restriction. It is my
understanding that the source tree for the applications that impose this
limitation in some Linux systems has already been fixed.

I took the trouble to research this and contacted the source developers of
the Linux utility package to make sure that this limitation would go away.
In future Linux releases you should NOT need a conversion script. It means
that you CAN have the same names for users and groups in Linux as you can
in Windows.

My question remains. Why would should we implement as part of the 'net rpc
vampire' process an imposition on ALL UNIX/Linux platforms just because
one platform wants to force you as the administrator to have only
contiguous lower case names for users and groups?

I agree that the ability to call a script would possibly be nice. My
experience is that most sites will use the opportunity to re-arrange the
deck chairs during the migration process. I am not convinced that the
ability to call a user/group name conversion script will mitigate the
overehad in doing that, so as far as I can see the value of this nice
feature is not at the "Must Have" level.

I did not state that this is Microsoft's fault. That is an invalid
conclusion you drew. Sorry, but I am not guilty as charged on that one.

We must question what new features really are important and what is not.
For the Samba-Team, it is important to know what features to spend time
on. We have limited resources. We are always happy to receive patches and
to integrate them into the code tree if they make sense.

At this time I am reluctant to ask one of our developers to take time away
from other high priority work to add what in essence is a "It would be
nice to have" feature. I totally agree that the features you describe
would be really neat, but does it sense for the project as a whole?

Not all of us cut code, I have spent the last year writing documentation.
My objective is to document Samba and it's deployment to help users like
you. I accept responsiblity for how I choose to use my time.

Feedback we obtain does make its way into the documentation. We also use
feedback to decide how best to improve Samba. We can not guarrantee to
adopt every suggestion for code change.

Cheers,
John T.
-- 
John H Terpstra
Email: jht at samba.org