From 162fd02f2110b76d79ef1882625fa9778094e754 Mon Sep 17 00:00:00 2001
From: Albert Graef <aggraef@gmail.com>
Date: Mon, 13 Feb 2017 11:30:29 +0100
Subject: [PATCH] Make sure that inmidi_clk() is called in the right place in
the MIDI back-end. Fixes #255.
---
pd/src/s_midi.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/pd/src/s_midi.c b/pd/src/s_midi.c
index d1a20a003..50f5ed7e9 100644
--- a/pd/src/s_midi.c
+++ b/pd/src/s_midi.c
@@ -291,6 +291,7 @@ typedef struct midiparser
/* functions in x_midi.c */
void inmidi_realtimein(int portno, int cmd);
+void inmidi_clk(double timing);
void inmidi_byte(int portno, int byte);
void inmidi_sysex(int portno, int byte);
void inmidi_noteon(int portno, int channel, int pitch, int velo);
@@ -313,7 +314,17 @@ static void sys_dispatchnextmidiin( void)
outlet_setstacklim();
if (byte >= 0xf8)
+ {
inmidi_realtimein(portno, byte);
+ if (byte == 0xf8) {
+ // AG: Not sure what the timebase for the right outlet is supposed
+ // to be. I'm using msecs right now, which is in line with the other
+ // tempo-related objects such as metro. Uncomment the .001 factor to
+ // get seconds instead.
+ double timing = /*.001 * */clock_gettimesince(sys_midiinittime);
+ inmidi_clk(timing);
+ }
+ }
else
{
inmidi_byte(portno, byte);
--
GitLab