Re: PATCH: configurability of pattern characters, part 2

[Bart Schaefer <schaefer@brasslantern.com>]

[-- Attachment #1: Type: text/plain, Size: 1242 bytes --]

On Sat, Jun 1, 2013 at 4:09 PM, Peter Stephenson <
p.w.stephenson@ntlworld.com> wrote:

> Here's the basic disable -p / enable -p and emulation behaviour to be
> exposed to public view before the rest.
>
> Up to now, this has only had finger tests of some of the simpler forms,
> so there could well be bugs lurking in the disabling of certain
> patterns.
>
> If somebody can spot some effect I haven't documented, please shout.
>

Fiddling with this, the very first thing I tried gave me an unexpected
result.

schaefer<502> print *
aclocal.m4 aczsh.m4 autom4te.cache ChangeLog Completion Config config.guess
config.h config.h.in config.log config.modules
config.modules.shconfig.status config.sub configure
configure.ac Doc Etc FEATURES Functions INSTALL install-sh LICENCE MACHINES
Makefile Makefile.in META-FAQ Misc mkinstalldirs NEWS README Scripts Src
stamp-h stamp-h.in StartupFiles Test Util
schaefer<503> disable -p \*
schaefer<504> print *
zsh: no match
schaefer<505>

I was rather expecting to get a * as output rather than a no match error.
 It seems very odd to have a pattern that can't possibly match anything but
that is still treated as a pattern; this did not really disable *, it just
turned it into the equivalent of [*].