Backport of master commit 0592cefd03f1de2f04b721d07a16e6e0a9e48f73.
This should fix the bug fixed by Mike Crowe's patch in:
https://lists.gnu.org/archive/html/emacs-devel/2015-07/msg00106.html
A problem in this area has been reported by several users; see
Bug#16737, Bug#17101, Bug#17026, Bug#17172, Bug#19320, Bug#20283.
This fix differs from Mike Crowe's patch in that it should avoid a
race condition that could lose SIGIO signals.  ignore_sigio dates
back to the 1980s when some platforms couldn't block signals, and
could only ignore them, which led to races when signals arrived
while being ignored.  We shouldn't have to worry about those old
platforms now.
* src/dispextern.h, src/sysdep.c (ignore_sigio): Remove.
* src/emacs.c (shut_down_emacs):
Don't call ignore_sigio; unrequest_sigio should suffice.
* src/keyboard.c (kbd_buffer_store_buffered_event):
Use unrequest_sigio, not ignore_sigio.
(kbd_buffer_get_event):
Call request_sigio when getting the ball rolling again.

* gtkutil.c (xg_display_open):
* xterm.c (x_display_ok, x_term_init): Block SIGIO when opening
a display.

* gtkutil.c (update_theme_scrollbar_width): Don't round up
scroll bar width with GTK3 (Bug#20182).

No need to merge to trunk, since it corrects a backported entry that
should not appear in trunk.

Fixes: debbugs:20156

* src/fontset.c (fontset_pattern_regexp): Backport: Replace + 1 with
+ 3 in alloca (tiny change)

* src/frame.h (x_set_bitmap_icon): Don't set the icon if icon-type is
nil/not present in the parameter alist.

Fixes: debbugs:19680

 src/search.c (find_newline): Avoid assertion violations in
 CHAR_TO_BYTE when a portion of the buffer was deleted and we look
 for newlines near the end of the buffer.  This happens in Rmail
 hen JIT font-lock fontifies a newly displayed portion of the
 buffer.