GNU bug report logs - #63949
30.0.50; `vc-print-log´ does not erase buffer when called from *vc-change-log* buffer, at least for CVS logs

Previous Next

Full log

Message #62 received at 63949 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dmitry <at> gutov.dev>
To: Jens Schmidt <jschmidt4gnu <at> vodafonemail.de>, Eli Zaretskii <eliz <at> gnu.org>
Cc: 63949 <at> debbugs.gnu.org
Subject: Re: bug#63949: 30.0.50; `vc-print-log´ does not erase buffer when called from *vc-change-log* buffer, at least for CVS logs
Date: Sun, 18 Jun 2023 05:42:44 +0300

On 16/06/2023 22:33, Jens Schmidt wrote:
> On 2023-06-10  19:23, Eli Zaretskii wrote:
> 
>> I'd also like to hear Dmitry's views on these issues.  He was until 
>> now silent in this discussion, AFAICT.
> 
> @Dmitry: What's your view on these issues?  Just making the handling of
> VC parent buffer more documented and "more consistent", whatever that
> means, or do you prefer some particular direction?

I think the appropriate thing here is to back out of the change that 
caused the regression (d494833d47968fcd97ba549654a259d6fb6c2eee, as 
we've found out) and then try to re-fix it some other way (in master?).

Or maybe adjust the current code such as when vc-deduce-fileset-1 does 
change the current buffer, vc-deduce-fileset retains that change.

For example, using this ugly-ish (100% untested) patch:

diff --git a/lisp/vc/vc.el b/lisp/vc/vc.el
index 91d3f6f70d3..91aae40a677 100644
--- a/lisp/vc/vc.el
+++ b/lisp/vc/vc.el
@@ -1121,10 +1121,13 @@ vc-deduce-fileset
 the returned list.

 BEWARE: this function may change the current buffer."
-  (with-current-buffer (or (buffer-base-buffer) (current-buffer))
-    (vc-deduce-fileset-1 not-state-changing
-                         allow-unregistered
-                         state-model-only-files)))
+  (let (new-buf)
+    (with-current-buffer (or (buffer-base-buffer) (current-buffer))
+      (vc-deduce-fileset-1 not-state-changing
+                           allow-unregistered
+                           state-model-only-files)
+      (setq new-buf (current-buffer)))
+    (set-buffer new-buf)))

 (defun vc-deduce-fileset-1 (not-state-changing
                             allow-unregistered

The fact that some backends do call vc-setup-buffer inside 
vc-xx-print-log and some dont', also bears investigation. But the 
question I would like to have answered is, can we drop this call from 
all of them? Rather than trying to add it everywhere.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.