bug#2028: 23.0.60; (1) dired-guess-default should remove duplicates (2) doc string

bug#2028: 23.0.60; (1) dired-guess-default should remove duplicates (2) doc string

[Drew Adams]

`dired-guess-default' returns a list of default shell commands
appropriate for its input, FILES.  It should remove duplicates from
this list before returning it.
 
Also, the doc string is:
 
"Guess a shell commands for FILES.  Return command or list of commands.
See `dired-guess-shell-alist-user'."
 
This is not correct grammatically. Suggestion:
 
"Return a shell command, or a list of such, appropriate for FILES.
See `dired-guess-shell-alist-user'."
 

BTW (I didn't check the code - just a naive question): Why do we
return sometimes a single command instead of a singleton list of that
command? What is gained by this added complexity? Is there some
calling code that *must* receive a command and not a list of commands?
It seems like any calling code would need to be able to handle the
list case, including a singleton list, anyway.
 
 
 
In GNU Emacs 23.0.60.1 (i386-mingw-nt5.1.2600)
 of 2009-01-04 on LENNART-69DE564
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (3.4) --no-opt --cflags -Ic:/g/include
-fno-crossjumping'
 

bug#2028: 23.0.60; (1) dired-guess-default should remove duplicates (2) doc string

[Lars Magne Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

> `dired-guess-default' returns a list of default shell commands
> appropriate for its input, FILES.  It should remove duplicates from
> this list before returning it.

Ok; done.

> Also, the doc string is:
>
> "Guess a shell commands for FILES.  Return command or list of commands.
> See `dired-guess-shell-alist-user'."
>
> This is not correct grammatically. Suggestion:
>
> "Return a shell command, or a list of such, appropriate for FILES.
> See `dired-guess-shell-alist-user'."

I've changed it to something like that.

> BTW (I didn't check the code - just a naive question): Why do we
> return sometimes a single command instead of a singleton list of that
> command? What is gained by this added complexity? Is there some
> calling code that *must* receive a command and not a list of commands?

I'm guessing somebody extended the function at some point.

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/