Re: PATCH Make org-open-at-point only ask once

[Anders Waldenborg <anders@0x63.nu>]

On Thu, Oct 06, 2011 at 10:00:38AM +0200, Carsten Dominik wrote:
> Hi Anders, Nicolas
> 
> I am wondering what the status of this patch is.  There was a discussion, but I am not sure about the conclusion...


Carsten,

I think that the simple patch still is valid, I'm including it again
below.

The discussion was about my more intrusive alternative patch, and how
links change their meaning depending on if the buffer is narrowed or
not. I did start to hack up a alternative link resolver as a proof of
concept to be able to better understand how links should work in case
there are multiple targets, but unfortunately I got busy with other
things.

 anders

--

commit 54702f063ae2df48dec7f9feb80859a6b64002a4
Author: Anders Waldenborg <anders@0x63.nu>
Date:   Sat Aug 27 21:18:46 2011 +0200

    Make org-open-at-point only ask once whether new header should be created.
    
    When following "thisfile" links org-open-at-point is kind enough to
    retry org-link-search again after widening the buffer it can't be
    found. However org-link-search also asks the question "No match -
    create this as a new heading? (y or n)" when target can't be
    found. This means that the question is asked twice when following a
    nonexistent link and answering no.
    
    This is fixed by setting org-link-search-inhibit-query in first try,
    so only second invocation asks the question.

diff --git a/lisp/org.el b/lisp/org.el
index d63b854..781de88 100644
--- a/lisp/org.el
+++ b/lisp/org.el
@@ -9537,7 +9537,8 @@ application the system uses for this file type."
 			      ((equal arg '(16)) ''org-occur)
 			      (t nil))
 		       ,pos)))
-	    (condition-case nil (eval cmd)
+	    (condition-case nil (let ((org-link-search-inhibit-query t))
+				  (eval cmd))
 	      (error (progn (widen) (eval cmd))))))
 
 	 (t