Re: PATCH: history improvements for 3.1.5-pws-17 and 3.0.6-pre-2

[Peter Stephenson <pws@xxxxxxxxxxxxxxxxx>]

Wayne Davison wrote:
> Appended is the history improvements for pws-17 that I mentioned in
> zsh-workers 6218.

Thanks, this will appear in pws-18.

One addition in the patch below:  share_history is set in ksh emulation.

Some comments on share_history (which is the one I've played about with
most): actually, it doesn't emulate ksh at the moment, because only a few
commands cause the history from any other instance of the shell to be used.
I know this is how it's documented, but I find the behaviour a bit
confusing, since after I've typed something in one shell, and hit return in
another shell, I expect to be able just to use up-line-or-history to get to
what's in the other shell --- otherwise I might just as well use an
explicit fc -RI and the option's not much use, or so I've been finding on
rather limited experience so far.

I presume that the real problem is numbering of bang-history references,
and I agree it's correct to be careful in that case.  However, I'd like to
suggest that any zle reference to the history (apart from expansion of
lines like !*, of course, but that's done internally via the lexer so
shouldn't cause problems as an exception) causes the newly imported stuff
to be visible.  I don't think this is likely to cause confusion, since you
always see the line before doing anything with it --- in fact, in my case
it would definitely help eliminate confusion, and also it would make the
behaviour more ksh-like.

Next point --- there is another slight confusion, partly at least as a
knock on effect of this.  If I type something in shell 1, I know it won't
be imported until after the next line in shell 2 is finished, and won't
show up until after the next history reference.  So I type `history'. As
explained above, I find this a bit clumsy anyway, but the point here is
that for some reason due to the internal handling it's not enough to make
what I just typed in shell 1 appear (even if I unset histnostore, which I
thought might have had an effect).  It only happens after I've typed a new
command in shell 2, which can't be a history command or a blank line, and
then after that I have to do something to get the commands to show up.

Ideally for sharing history, I'd like to be able just to hit return (to do
the reading) then up (to allow the new commands to show up) to get what I
just typed in the other shell, just as happens in ksh.  Hope that's
possible.

--- Src/options.c.newhist	Sat May  8 14:31:51 1999
+++ Src/options.c	Sat May  8 15:27:16 1999
@@ -181,7 +181,7 @@
 {NULL, "restricted",	      OPT_SPECIAL,		 RESTRICTED},
 {NULL, "rmstarsilent",	      OPT_BOURNE,		 RMSTARSILENT},
 {NULL, "rmstarwait",	      0,			 RMSTARWAIT},
-{NULL, "sharehistory",	      0,			 SHAREHISTORY},
+{NULL, "sharehistory",	      OPT_KSH,			 SHAREHISTORY},
 {NULL, "shfileexpansion",     OPT_EMULATE|OPT_BOURNE,	 SHFILEEXPANSION},
 {NULL, "shglob",	      OPT_EMULATE|OPT_BOURNE,	 SHGLOB},
 {NULL, "shinstdin",	      OPT_SPECIAL,		 SHINSTDIN},

-- 
Peter Stephenson <pws@xxxxxxxxxxxxxxxxx>       Tel: +39 050 844536
WWW:  http://www.ifh.de/~pws/
Dipartimento di Fisica, Via Buonarroti 2, 56127 Pisa, Italy