ref: 00febbc9618eb24e91e4f5b1d5f95cd5a88094be
parent: fc7d7eee3983036d2730145a47897f7f3f5fd539
author: cancel <cancel@cancel.fm>
date: Sat Jan 18 16:31:24 EST 2020
Update comment
--- a/tui_main.c
+++ b/tui_main.c
@@ -2936,10 +2936,14 @@
double secs_to_d = ged_secs_to_deadline(&t.ged);
int new_timeout;
// These values are tuned to work OK with the normal scheduling behavior
- // on Linux, Mac, and Windows. Of course, there's no guarantee about how
- // the scheduler will work so if you are using a modified kernel or
- // something, this might be sub-optimal. But there's not really much we
- // can do about it!
+ // on Linux, Mac, and Windows. All of the usual caveats of trying to
+ // guess what the scheduler will do apply.
+ //
+ // Of course, there's no guarantee about how the scheduler will work, so
+ // if you are using a modified kernel or something, or the measurements
+ // here are bad, or it's some OS that behaves differently than expected,
+ // this won't be very good. But there's not really much we can do about
+ // it, and it's better than doing nothing and burning up the CPU!
if (t.strict_timing) { if (secs_to_d < ms_to_sec(0.5)) {new_timeout = 0;