From: Eli Zaretskii <eliz@gnu.org>
Date: Fri, 28 Aug 2015 13:25:25 +0000 (+0300)
Subject: Fix a bug in recording a macro while flyspell-mode is active
X-Git-Tag: emacs-25.0.90~1232
X-Git-Url: http://git.eshelyaron.com/gitweb/?a=commitdiff_plain;h=fdd095d2fd79f73ec96b9d318248e6ad70931b4a;p=emacs.git

Fix a bug in recording a macro while flyspell-mode is active

* lisp/subr.el (sit-for): Don't call read-event when recording a
macro.  (Bug#21329)
---

diff --git a/lisp/subr.el b/lisp/subr.el
index 53dea3e2997..61b87065c8a 100644
--- a/lisp/subr.el
+++ b/lisp/subr.el
@@ -2277,7 +2277,18 @@ floating point support."
     t)
    ((input-pending-p t)
     nil)
-   ((<= seconds 0)
+   ((or (<= seconds 0)
+        ;; We are going to call read-event below, which will record
+        ;; the the next key as part of the macro, even if that key
+        ;; invokes kmacro-end-macro, so if we are recording a macro,
+        ;; the macro will recursively call itself.  In addition, when
+        ;; that key is removed from unread-command-events, it will be
+        ;; recorded the second time, so the macro will have each key
+        ;; doubled.  This used to happen if a macro was defined with
+        ;; Flyspell mode active (because Flyspell calls sit-for in its
+        ;; post-command-hook, see bug #21329.)  To avoid all that, we
+        ;; simply disable the wait when we are recording a macro.
+        defining-kbd-macro)
     (or nodisp (redisplay)))
    (t
     (or nodisp (redisplay))