patch 9.2.0050: WM_SETFOCUS not handled immediately
Commit:
https://github.com/vim/vim/commit/c4a6fa3eade9d30e0bda7e5b7077951da9e75279
Author: Muraoka Taro <
koron....@gmail.com>
Date: Tue Feb 24 21:30:43 2026 +0000
patch 9.2.0050: WM_SETFOCUS not handled immediately
Problem: In gvim on Windows, a certain problem can occur when the
WM_SETFOCUS event sent after an external command is not
processed immediately.
Solution: After posting WM_SETFOCUS, run the message loop to process it
as quickly as possible (Muraoka Taro).
The problem is that Test_normal11_showcmd may fail when running the
test_normal.vim test. Investigation revealed that the trigger was an
external command executed in the previous test,
Test_mouse_shape_after_failed_change, when two tests were executed
consecutively. In gvim on Windows, a WM_SETFOCUS event will be sent
when an external command finishes executing. This WM_SETFOCUS event is
not processed immediately, but rather by redraw, which is expected to
update showcmd. Because it is queued in typebuf at this time,
clear_showcmd(), which expects typebuf to be empty, cannot update
showcmd.
Also added a test that simulates the above problem.
closes: #19167
Signed-off-by: Muraoka Taro <
koron....@gmail.com>
Signed-off-by: Christian Brabandt <
c...@256bit.org>
diff --git a/src/os_win32.c b/src/os_win32.c
index edaf7cb30..6ccd185f0 100644
--- a/src/os_win32.c
+++ b/src/os_win32.c
@@ -4881,6 +4881,10 @@ mch_system_classic(char *cmd, int options)
// Try to get input focus back. Doesn't always work though.
PostMessage(hwnd, WM_SETFOCUS, 0, 0);
+ // To increase the chances that WM_SETFOCUS will work, run the message loop
+ // here. In addition, it prevents problems caused by delayed WM_SETFOCUS
+ // processing.
+ gui_mch_update();
return ret;
}
diff --git a/src/testdir/test_normal.vim b/src/testdir/test_normal.vim
index 58c1b9783..f41ae0552 100644
--- a/src/testdir/test_normal.vim
+++ b/src/testdir/test_normal.vim
@@ -4411,4 +4411,33 @@ func Test_pos_percentage_in_turkish_locale()
endtry
endfunc
+" This test simulates the problem with gvim on Windows, observed when
+" Test_normal11_showcmd in test_normal.vim is executed consecutively after
+" Test_mouse_shape_after_failed_change.
+"
+" The problem occurred because WM_SETFOCUS was processed slowly, and typebuf
+" was not empty when it should have been.
+func Test_win32_gui_setfocus_prevent_showcmd()
+ if !has('win32') || !has('gui_running')
+ throw 'Skipped: Windows GUI regression test'
+ endif
+
+ " WM_SETFOCUS event occurs when finish to execute filter command in gvim
+ exe 'silent !echo foo'
+
+ set showcmd
+ 10new
+ call setline(1, ['aaaaa', 'bbbbb', 'ccccc'])
+ call feedkeys("ggl\<C-V>lljj", 'xt')
+
+ " showcmd could not be updated because events originating from WM_SETFOCUS
+ " were stored in typebuf at here. clear_showcmd() executed from redraw,
+ " will not draw the selection information unless you are in visual mode and
+ " typebuf is empty.
+ redraw!
+
+ call assert_match('3x3$', Screenline(&lines))
+ call feedkeys("\<C-V>", 'xt')
+endfunc
+
" vim: shiftwidth=2 sts=2 expandtab nofoldenable
diff --git a/src/version.c b/src/version.c
index 620e1a5bd..076063b4c 100644
--- a/src/version.c
+++ b/src/version.c
@@ -734,6 +734,8 @@ static char *(features[]) =
static int included_patches[] =
{ /* Add new patch number below this line */
+/**/
+ 50,
/**/
49,
/**/