Scott McDermott wrote:
> >
> > Handling + in auto_master would be nice. The + as an entry in a map
> > is utterly and fundamentally broken (a map is an associative lookup,
> > d**** it!)
>
> This may be so, and I've heard you argue this repeatedly. While I agree
> on a technical premise, I think you should think about how long NIS has
> been around, how pervasive and ubiquitous it is, and how many UNIX
> implementations out there support the maps the way Sun wrote them. I
> think it's nice for Linux to sit side-by-side with our UNIX servers
> instead of having to hack them to make them play nice.
>
The existing implementation works just fine and dandy with NIS. That's
what we use at Transmeta. It's just the Sun brain-damaged way of
*ASKING* for an NIS map in a local file that's the problem.
> > although the same (and better) functionality can be had with "multi"
> > maps. Unfortunately, I don't know of any sane way to specify multi
> > maps in the auto_master format; it's probably time to try to recommend
> > using something like /etc/autofstab with auto_master as a
> > compatibility fallback.
>
> How about a compile-time or run-time option to turn on "Sun dain
> bramage?" in the original auto?masters *and* in the maps themselves?
> Note that I don't know the code so I don't know if this is fundamentally
> difficult to implement. I do think it's essential to a heterogenous
> UNIX environment though.
>
See the above. It's the local files that are issues.
-hpa
