PATCH: (perhaps) for use of _oldlist

PATCH: (perhaps) for use of _oldlist

[Peter Stephenson]

I was using Oliver's version of most-recent-file, which works as
follows (see zsh-users/6957, for example):

  zstyle ':completion:most-recent-file:*' match-original both
  zstyle ':completion:most-recent-file::::' completer _menu _files _match
  zstyle ':completion:most-recent-file:*' file-sort modification
  zstyle ':completion:most-recent-file:*' file-patterns '*(.):normal\ files'
  zstyle ':completion:most-recent-file:*' hidden all
  bindkey '^X.' most-recent-file
  zle -C most-recent-file menu-complete _generic

It allows you to cycle through the most recent matches in order.  I
found that this doesn't work with the _oldlist completer.  _oldlist
works the first time, preserving the list of matches, but on the second
TAB decides to generate a new list.  This appears to be because
_main_completer squirrels away compstate[nmatches] regardless of
_oldlist; the value is 0 after the first use of _oldlist because we
didn't generate any matches this time.  (I say "appears" because
although it certainly does do what I said it's possible there's
something deeper going on.)

If I've identified the real symptom and not a side effect, the following
patch will fix it.  I can well believe it's more involved, however.

It's offset a couple of lines because of my two-line change to give
progress reports on completion, which I'm happily using but no one else
liked.

Index: Completion/Base/Core/_main_complete
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Base/Core/_main_complete,v
retrieving revision 1.9
diff -u -r1.9 _main_complete
--- Completion/Base/Core/_main_complete	2 Mar 2004 12:52:06 -0000	1.9
+++ Completion/Base/Core/_main_complete	30 Sep 2004 11:26:51 -0000
@@ -167,7 +170,14 @@
 done
 
 curcontext="${curcontext/:[^:]#:/::}"
-nm=$compstate[nmatches]
+if [[ $compstate[old_list] = keep ]]; then
+  # We are keeping the old list of matches, so keep the
+  # number of matches we found last time rather than the
+  # number just generated.
+  nm=$_lastcomp[nmatches]
+else
+  nm=$compstate[nmatches]
+fi
 
 if [[ $compstate[old_list] = keep || nm -gt 1 ]]; then
   [[ _last_nmatches -ge 0 && _last_nmatches -ne nm ]] &&

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR Ltd., Science Park, Milton Road,
Cambridge, CB4 0WH, UK                          Tel: +44 (0)1223 692070


**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

This footnote also confirms that this email message has been swept by
MIMEsweeper for the presence of computer viruses.

www.mimesweeper.com
**********************************************************************

Re: PATCH: (perhaps) for use of _oldlist

[Bart Schaefer]

On Thu, 30 Sep 2004, Peter Stephenson wrote:

> _main_completer squirrels away compstate[nmatches] regardless of
> _oldlist; the value is 0 after the first use of _oldlist because we
> didn't generate any matches this time.  (I say "appears" because
> although it certainly does do what I said it's possible there's
> something deeper going on.)

I was having a similar problem with _oldlist and a custom completion 
context, but I thought it was due to the WIDGETSTYLE business in _oldlist 
itself, that you already patched.  So I think you probably have the right 
idea with this patch, if the problem persists.