[mew-int 01184] Re: limitations on character usage in config names?

[sen_ml at example.com]

Hi,

Thanks for your response.

From: Kazu Yamamoto
Date: Wed, 04 Dec 2002 15:11:59 +0900 (JST)

> From: sen_ml at example.com
> Subject: [mew-int 01175] limitations on character usage in config names?
> 
> > I have not been able to access $inbox for one of my cases, while I am
> > able to access $inbox for a different case.  FWIW, I am able to invoke
> > "i" for both cases and I was able to use +mdrop for both immediately
> > prior to upgrading to 3.1.
> 
> Did you specify the case like "case:$inbox"?
> 
> i.e. "g" + "case:$inbox"

For the case that doesn't use a @ character, the sequence of keystrokes:

  g
  $
  <TAB>

provides me w/ a fully complete string that allows me access to $inbox
for that case.  I.e. the string looks like "case:$inbox".

For the case that does use a @ character, the sequence of keystrokes:

  g
  $
  <TAB>

gives a message in the minibuffer like:

  No matching case

In both cases, after typing '$', the string after the prompt in the
minibuffer is replaced w/ a string like 'case:$'.

> > At this point, the only idea that comes to mind is that the case that
> > doesn't work uses the character "@" in the case name, while the one
> > for which it does work does not use "@" in the case name.  A wild
> > guess, perhaps (-;
> 
> With the currect spec, case should be in the form of
> "[-a-zA-Z0-9_][-a-zA-Z0-9_]+".

I've been using cases w/ the "@" in them in Mew 2.2 (and may be
before) -- some of my case names are email addresses so they also
include period signs.  Has the spec changed since 2.2?

I guess I failed to locate documentation stating the limitations of
case name strings -- if it exists somewhere other than the source,
would you mind telling me where it is?

> Some people want to loose this spec. More discussions are necessary I
> guess.

Hmmm...

As a practical matter for me, do you recommend I try to downgrade to
an older version of Mew?