Re: remote vs local symlinking

[Eli Zaretskii]

On Tue, 20 Mar 2001 Tom_Roche@ncsu.edu wrote:

> Perhaps you can agree that it's a bit
> "simple-minded" to seek to test whether a remote OS can perform an
> operation by testing whether the local client (or its OS) can perform
> that operation? or to constrain a remote session based on local
> capabilities?

Of course it's simple-minded!  But that's how the code is written, and
rewriting it should not cause any serious breakage.

> Is Emacs so brittle that throwing an error would break it?

Perhaps it is, perhaps it isn't.  All I'm saying is that someone
should analyze this, not assume that it will magically work.  The
entire issue of handling symlinks should be analyzed as a whole,
because you need all the operations that deal with symlinks to be 
consistent in what they say and do.  For example, if some 
symlink-related primitive doesn't have a file handler, it should be 
added as part of this change.