PATCH: make get-line widget update history number

PATCH: make get-line widget update history number

[Oliver Kiddle]

After accept-line-and-down-history, a get-line does not set the history
number (HISTNO).

I override history-incremental-search-backward to use
narrow-to-region when invoked on a non-empty buffer. This is useful for
retrieving history into, e.g the middle of a for loop.

With this change, if the recursive-edit was finished with an
accept-line-and-down-history, it can now call zle get-line, loop
back around and subsequent accept-line-and-down-historys will insert
subsequent history lines. This is probably also possible by manually
editing PRE/POSTDISPLAY and HISTNO but it's much easier this way.

I'm trying to think of things that this might possibly break but I
can't. I don't think I've ever used get-line before. Anyone else see a
problem?

diff --git a/Src/Zle/zle_hist.c b/Src/Zle/zle_hist.c
index bd5bc36..44b39d1 100644
--- a/Src/Zle/zle_hist.c
+++ b/Src/Zle/zle_hist.c
@@ -890,6 +890,10 @@ zgetline(UNUSED(char **args))
 	free(s);
 	free(lineadd);
 	clearlist = 1;
+	if (stackhist != -1) {
+	    histline = stackhist;
+	    stackhist = -1;
+	}
     }
     return 0;
 }

Re: PATCH: make get-line widget update history number

[Bart Schaefer]

On Mar 1,  1:20am, Oliver Kiddle wrote:
}
} I'm trying to think of things that this might possibly break but I
} can't. I don't think I've ever used get-line before. Anyone else see a
} problem?

No; this looks like something that always should have been there.

The only possible gotcha I can think of would be if a widget invoked
from inside recursive-edit did something with "fc -p ..." without -a,
or "fc -P", but I haven't really traced through what's happening with
stackhist there.

Re: PATCH: make get-line widget update history number

[Peter Stephenson]

On Fri, 28 Feb 2014 18:30:37 -0800
Bart Schaefer <schaefer@brasslantern.com> wrote:
> On Mar 1,  1:20am, Oliver Kiddle wrote:
> }
> } I'm trying to think of things that this might possibly break but I
> } can't. I don't think I've ever used get-line before. Anyone else see a
> } problem?
> 
> No; this looks like something that always should have been there.

...and given this appears to be a fairly obscure feature, "suck it and
see" seems to be about the best way of proceeding anyway.

-- 
Peter Stephenson <p.w.stephenson@ntlworld.com>
Web page now at http://homepage.ntlworld.com/p.w.stephenson/