tramp-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Strange error when doing M-g on a group in Gnus


From: Kai Großjohann
Subject: Re: Strange error when doing M-g on a group in Gnus
Date: Fri, 06 Sep 2002 17:19:06 +0200
User-agent: Gnus/5.090008 (Oort Gnus v0.08) Emacs/21.3.50 (i686-pc-linux-gnu)

Michael Albinus <address@hidden> writes:

> Hmm, it's my code, indeed. Expanding "/home-local" there's a check
> whether there's something to do with file name completion (because it
> matches the regexp in file-name-handler-alist). For this I check,
> beside other indications, `last-input-char'. Looks like M-g is
> expanded to -134217625 ... strange.

Why do you check last-input-char at all?  My thinking is this:

  - If the filename has no colon, then you should complete.

  - If the filename has one colon, then it might look like
    "/method:bla".  Then you should do user/hostname completion.  Or
    it might look like "/foo:bar" where foo is not a method.  Then foo
    must be of the form host or address@hidden and thus the filename is
    complete, and the non-completion part of Tramp should do it.

    If the filename looks like "/method:bla" but "bla" contains a
    slash, then the filename is bogus.  Or maybe the "method" part
    should be interpreted as a hostname.  Better stay away from such
    strange filenames.

  - If the filename has more than one colon, then it could be a
    multi-hop filename.  But you don't deal with those, yet.  So
    tramp-completion-mode should return nil.

Does this make sense?  Do these rules do the right thing?

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)




reply via email to

[Prev in Thread] Current Thread [Next in Thread]