Re: rfn-eshadow.el and remote filenames

[Michael Albinus <michael.albinus@gmx.de>]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> Current rfn-eshadow.el does not assume anything about the behavior of
> substitute-in-file-name and I think we should keep it that way.  Obviously,
> it does have some preconceived notion about how substitute-in-file-name
> *might* work, but it works safely in "all" cases.
>
> Currently, rfn-eshadow.el only recognizes when substitute-in-file-name drops
> a prefix of the current file name, so we need to extend it to recognize
> cases where substitute-in-file-name drops some internal part instead.
> How to do that efficiently?  I'm not sure.
>
> I'd prefer we find a more general way to do it, but if not, maybe your code
> can be kept mostly, but then it should check its result against the return
> value of substitute-in-file-name.

A more general approach could be when Tramp hooks into
`insert-behind-hooks' of `rfn-eshadow-overlay'. rfn-eshadow.el would
stay unchanged then, likely.

I'll play a little bit with, checking also the performance.

>         Stefan

Best regards, Michael.