Beware!

As stated in #351 and the documentation, the Message.time parameter unit is ill defined and using it in this way may break things if another unit is expected down the processing line.

Do you know what’s the time unit provided by rtmidi?
Is it the same across operating systems APIs?

Well, this change won't "break" anything, because old code won't be trying to get that non-existent value, I think?

If you mean that it would be possible to write new code that used this time and that worked in some cases and not others, perhaps so.

This could be documented?

Having a way to measure the time between MIDI events would be really useful.

Having to compute that time in Python at the end after the event has come through a cascade of Python code just savages both the accuracy and the precision of the result.

Actually, calling it msg.rtmidi_time seems like a significantly better idea.

1. It makes it clearer that it's something peculiar to rtmidi.
2. You can immediately guess where to look for documentation.
3. Other projects are probably already be hanging a time onto their messages, and msg.time would conflict.

I'm going to make this change and rebase it.

I too would love to have that information available.
I was worried it would add to the confusion if someone tried to convert the messages with this time information to a MIDI file which expects time to be in ticks…

Thanks for understanding !

The build failure seems to be some configuration issue where it can't even find Python 2.7 and nothing to do with my delta.

Hey, I have fallen into so many such traps before. :-D

Honestly, I'd rather the name were a bit too long, but clear to "the next guy". Typing doesn't slow me down, mistakes do!

Is there some way to discover the units of the time field at this point in the code, now you mention it?

Not that I’m aware but I’m quite new here.
I’m willing to dig and contribute since this lib is a huge asset for a project of mine and I need it to be reliable and spec compliant ;)

Upon looking at the rtmidi code, what they provide is a time delta since the previously received message (aka relative time) and its unit is in milliseconds.

I too would like to be able to retrieve midi timestamps.
Is there a status update on this?

@stephen322 No ETA yet but I intend to ship this feature with MIDO 2.0. Got distracted implementing a native port backend for Windows in the meantime though. Once I finish debugging it, I'll resume work on version 2.
If you really need this now, I'd be happy to accept a PR against 1.x with the issues highlighted in this thread fixed ;)