Re: partial-completion-mode and deleting symlink problem

[Richard Stallman]

    Where exactly would you want the predicate to be specified? 
    In the partial-completion-mode code? (Like in the attached patch?)
    Or in read-file-name? 

It could be done in read-file-name.  Perhaps read-file-name,
when you specify "an existing file name", should always accept
names of symlinks to nonexistent files.

Another idea is to define another possible value for the MUSTMATCH
argument to read-file-name, which would stand for "existing file name
or symlink to nowhere".  Then this could be specified where
appropriate.

    If you look in one of my messages in this thread: 
    PC-do-completion calls  test-completions that calls
    read-file-name-internal which  returns Qnil because of this code:

A question occurs to me: is it really right for partial completion
to call test-completion here?  What operation did the user do?
Without partial completion mode, using the standard completion code,
would that operation call test-completion?  If not, then perhaps it
is wrong for partial completion mode to call it.  Or maybe it should
call it, but handle the return value of nil in a way more
consistent with ordinary completion.