Re: potential regression on .. completion on FreeBSD a least

[Peter Stephenson <pws@xxxxxxx>]

On Wed, 19 Nov 2008 09:23:56 -0800
Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx> wrote:
> On linux, I get this:
> 
> schaefer<501> cd ..
> Completing directory in cdpath
> .gimp-2.0/         .gstreamer-0.8/    zsh-3.0.5/         zsh-3.0.8-pre/   
> .rhopenoffice1.1/  .emacs.d/          zsh-3.1.6/
> 
> (The cursor backs up to be located between the two dots.)
> 
> In _path_files, it appears that the change from
> 
>       # `clever' filing systems where names pop into existence
>       # when referenced.
>       if (( ! $#tmp1 )); then
> 
> to
> 
>       # we only do this once for completion of each path segment.
>       if (( ! $#tmp1 && npathcheck == 0 )); then
> 
> is causing the existence check
> 
> 	  if [[ -e "$tmp3${(Q)PREFIX}${(Q)SUFFIX}" ]] then
> 
> to be skipped.  Therefore ".." is never added as a possible completion.
> It appears we may need another special case.

Thanks for looking.

The correct fix is

  zstyle ':completion:*' special-dirs true

else the completion system is correct to ignore ".." as a completion.

The previous behaviour was a bug---as you pointed out before, we were
ignoring the result of compfiles and completing files that existed even if
they didn't fulfill the match conditions we were carefully constructing.
The special-dirs behaviour is documented in zshcompsys.  The confusion
is my fault for not realising before that we needed to be more careful
with matching files.

-- 
Peter Stephenson <pws@xxxxxxx>                  Software Engineer
CSR PLC, Churchill House, Cambridge Business Park, Cowley Road
Cambridge, CB4 0WZ, UK                          Tel: +44 (0)1223 692070