On Thu, 4 Jan 2001, Dave Huang wrote:
> output would unblank the screen?) Would the solution be to not check the
> mtime of framebuffer devices, but do check the atime? I'll give that a
> try and see if it does the right thing :)

Okay, that does seem to work... (as long as something is actively
reading ttyE0... if I do a "sleep 1000" and wait for the screen to
blank, typing doesn't wake it up until the sleep is done, or I ^C the
sleep... but that's good enough for me :)

Index: screenblank.c
===================================================================
RCS file: /cvsroot/basesrc/usr.sbin/screenblank/screenblank.c,v
retrieving revision 1.10
diff -u -r1.10 screenblank.c
--- screenblank.c	1999/06/06 03:35:36	1.10
+++ screenblank.c	2001/01/05 04:11:58
@@ -226,16 +226,14 @@
 		change = 0;
 		for (dsp = ds_list.lh_first; dsp != NULL;
 		    dsp = dsp->ds_link.le_next) {
-			/* Don't check framebuffers. */
-			if (dsp->ds_isfb)
-				continue;
 			if (stat(dsp->ds_path, &st) < 0)
 				err(1, "stat: %s", dsp->ds_path);
 			if (st.st_atime > dsp->ds_atime) {
 				change = 1;
 				dsp->ds_atime = st.st_atime;
 			}
-			if (st.st_mtime > dsp->ds_mtime) {
+			/* Don't check a framebuffer's mtime. */
+			if (!dsp->ds_isfb && (st.st_mtime > dsp->ds_mtime)) {
 				change = 1;
 				dsp->ds_mtime = st.st_mtime;
 			}