Re: dabbrev-expand behavior

[Dan Nicolaescu]

Chet Ramey <chet.ramey@case.edu> writes:

  > Dan Nicolaescu wrote:
  > > Hi,
  > > 
  > > Thanks for implementing dabbrev-expand in bash-4.0!
  > > 
  > > Unfortunately the behavior is not consistent with what dabbrev-expand
  > > does in Emacs (and tcsh), so it will be quite confusing for users to
  > > use.
  > 
  > Since the dabbrev-expand implementation combines existing mechanisms,
  > and fits into the bash completion framework, it will be consistent
  > with how bash completions behave.
  > 
  > That means that the possible completions will be generated the same
  > way as dynamic-complete-history; the possible completions will be
  > sorted and duplicates removed; and the completions will be displayed
  > as menu-complete does, which honors the setting of
  > show-all-if-ambiguous.
  > 
  > It might not be exactly the way emacs and tcsh do it, but it fits in
  > with how bash does things.

dabbrev-expand has been in emacs for probably 20 years, its behavior is
well defined, and expected by its users.  

Unfortunately what is now in bash is completely different, so it might
be a better idea to give it a different name, or even remove it
completely.  As it is now it is not very useful for dabbrev-expand
users.

If the current behavior makes it into a release it would be much harder
to do a proper implementation later.

Thanks