Zsh Mailing List Archive Messages sorted by: Reverse Date, Date, Thread, Author

Re: zsh funcs

X-seq: zsh-workers 2220
From: Zoltan Hidvegi <hzoli@xxxxxxxxxx>
To: acs@xxxxxxxxxxxxx
Subject: Re: zsh funcs
Date: Mon, 14 Oct 1996 00:09:42 +0200 (MET DST)
Cc: zefram@xxxxxxxxxxxxxxxxx, zsh-workers@xxxxxxxxxxxxxxx
In-reply-to: <199610111549.LAA14209@xxxxxxxxxxxxxxxxxxxxxxxx> from Vinnie Shelton at "Oct 11, 96 11:49:34 am"

> zefram@xxxxxxxxxxxxxxxxx said:
> > This should be
> 
> > eval $1=\$PWD : ~$1
> 
> > -zefram 
> 
> Thanks, but why is this necessary?  Was the change in behavior intentional?

I would not say that it was intentional.  I always thought that eval is
necessary for that.  In addvars() there is an explicite singsub() call to
substitute the left-hand-side of an assignment which is now only used to
substitute the index of the assignment.  The change was that the lexer
gives the assignment token only if a word of the form number=value or
name=value or name[index]=value appears in command position where name has
only alphanumeric characters and underscores and starts with a letter or an
underscore.  This fixes problems with 1foo=bar assignments.

Zoltan

References:
- Re: zsh funcs
  - From: Vinnie Shelton

Messages sorted by: Reverse Date, Date, Thread, Author