"Init midi driver failed" when loading the program

• Aug 15, 2018 - 06:03
Reported version
3.0
Type
Functional
Severity
S4 - Minor
Status
by design
Project
MuseScore

GIT commit : c8199ac / Windows 7 and 10

Specific Windows issue when loading (and exiting) the currents 3.0 dev?

init.jpg

Anyway, as said here: https://musescore.org/en/node/275108, the problem goes back August 3/4

The cause maybe here, in one of these 5 commits ? https://github.com/musescore/MuseScore/pull/3796/commits

Comments

• Aug 15, 2018 - 12:35

As mentioned in the forum topic:
Debug: Init midi driver failed (...\MuseScore\mscore\pa.cpp:145, virtual bool Ms::Portaudio::init(bool))
So we know where that message comes from

Reply

• Aug 17, 2018 - 09:19
Status (old) active ⇒ closed
Status active ⇒ closed

cadiz1, I see this message every time I run MuseScore. It exists in 2.1, 2.2, 2.3 and 3.0. This message doesn't show something bad, it just says that "Midi driver initialization failed", which is true if you don't have MIDI device connected.
I agree, that message could be improved, but there are no errors here.

Reply

• Aug 17, 2018 - 09:44
Status (old) closed ⇒ needs info
Status closed ⇒ needs info

"I see this message every time I run MuseScore. It exists in 2.1, 2.2, 2.3 and 3.0"

Sorry, never in the past here, it's the very first time I observe this. And I speak knowingly: I check and test the nightlies since the Beta versions of the version 2.X, ie more than 4 years!
And until proof to the contrary (or oblivion on a period), I never see this in my archives.

So, something has changed (in any case, for Windows) since the five mentioned commits in beginning August. And I guess it may, I would not say improved, but fixed, by someone?

When you open several times the nigthlies over several days, months and years, it is more pleasant to find the expected behaviour :)
Thanks.

Reply

• Aug 17, 2018 - 09:55

I've just check and built 2.3.2 once again myself: Indeed that message shows there too.
Not in the development builds from the download page though, as those don't show the command window at all. I guess there it really started with the MSVC change

Reply

• Aug 17, 2018 - 10:09

Then again I think it OK for development builds to produce debug output like this, it rather was a bug that they didn't before (and didn't only on Windows)

It currently shows:

Internal error: shortcut <no-horizontal-stretch> not found
cannot find shortcut <no-horizontal-stretch>
Internal error: shortcut <no-vertical-stretch> not found
cannot find shortcut <no-vertical-stretch>
Internal error: shortcut <show-segment-shapes> not found
cannot find shortcut <show-segment-shapes>
Internal error: shortcut <show-measure-shapes> not found
cannot find shortcut <show-measure-shapes>
Internal error: shortcut <show-bounding-rect> not found
cannot find shortcut <show-bounding-rect>
Internal error: shortcut <show-corrupted-measures> not found
cannot find shortcut <show-corrupted-measures>
Internal error: shortcut <relayout> not found
cannot find shortcut <relayout>
Internal error: shortcut <autoplace-slurs> not found
cannot find shortcut <autoplace-slurs>
InitTime : multimedia timer resolution set to 1 milliseconds
Cannot connect to named pipe after wait = \\.\pipe\server_jack_default_0 err = 2
 
Cannot connect to server request channel
Init midi driver failed
Internal error: shortcut <relayout> not found

Reply

• Aug 17, 2018 - 10:16
Status (old) by design ⇒ needs info
Status by design ⇒ needs info

For the record and to be clear: I use only and exclusively the development builds from the download page.

Reply

• Aug 17, 2018 - 10:19
Status (old) needs info ⇒ by design
Status needs info ⇒ by design

development builds are for finding bugs, for that those debug messages can help, so it is even good design, even if this fix got added by accident.
The only other option to see those debug message is to build it yourself (or have some developer do it and reproduce your issue), nothing we'd want to impose on our testers (or developers ;-))

Reply