more shell-pcomplete UI regressions

more shell-pcomplete UI regressions

[Noah Friedman]

I cannot complete file names that have spaces in them, even if I put
backslashes in front of the spaces.

For instance this used to work:

    ~ $ ls /com/mp3/albums/King\ Cr<TAB>
    -| ~ $ ls /com/mp3/albums/King\ Crimson/

But now I get no matches even though "King Crimson" shows up in the
completions window if I type King<TAB>.  (If I pick some name that doesn't
have many possible completions then one of them is inserted into the
buffer; but if there are many, there is currently no way for me to finish
the match manually so that the rest will complete.)

Also, this fails:

    ~ $ ls '/com/mp3/albums/King<TAB>
    ->  (file-error "Opening directory" "no such file or directory" "/home/friedman/'/com/mp3/albums/")

In the old comint-dynamic-complete world, the leading quote enables
comint-dynamic-complete to insert filenames without backslashifying all the
spaces or other shell metacharacters.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: more shell-pcomplete UI regressions

[John Wiegley]

>>>>> On Tue Mar 12, Noah writes:

> For instance this used to work:

> ~ $ ls /com/mp3/albums/King\ Cr<TAB>
>     -| ~ $ ls /com/mp3/albums/King\ Crimson/

> ~ $ ls '/com/mp3/albums/King<TAB>

> In the old comint-dynamic-complete world, the leading quote enables
> comint-dynamic-complete to insert filenames without backslashifying
> all the spaces or other shell metacharacters.

Ah, this means the argument parsing function being used by pcomplete
for shell-mode is not good enough.  Since that is entirely
customizable, it could be fixed without touching pcomplete.

I guess it depends on whether RMS wants to try to make pcomplete work
or not.  If so, then I should probably use the old completion logic's
argument parsing function for use with pcomplete.

John

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: more shell-pcomplete UI regressions

[Miles Bader]

John Wiegley <johnw@gnu.org> writes:
> I guess it depends on whether RMS wants to try to make pcomplete work
> or not.  If so, then I should probably use the old completion logic's
> argument parsing function for use with pcomplete.

Well, I vote for fixing up pcomplete to address all the issues that
have been raised, and then giving it another try.

-Miles
-- 
P.S.  All information contained in the above letter is false,
      for reasons of military security.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel