Re: [Denemo-devel] Crash on Windows (SHOWSTOPPER)

[Richard Shann]

On Mon, 2013-03-11 at 17:37 +0000, Richard Shann wrote:
> I wish I hadn't started trying to fix this. I am pretty sure my fixes
> are good, but my build set-up is generating bizarre effects -
> denemo-console.exe no longer prints out messages and my MIDI controller
> is no longer found.
This seemed to clear up after I did a rm -rf target/mingw/*/denemo-* in
my gub. I've tested it now on windows vista and it seems good. The audio
backend output device is now showing as "default" under windows and it
works. The MIDI keyboard is found too, on my machine.
I'll update the translations we have updates for (French and Italian)
tomorrow and we can then make a final build to test before we upload it
and announce.
I have made a new video, this time with two recording devices, the
screen recorder and a camera capturing the pc & MIDI keyboard action. I
just have to figure out how to composite these two together. Any
suggestions welcome.

Richard
BTW Jeremiah - what were the key presses you were using to input the
notes on your Demo video?


> Jeremiah - can you build from git master for windows and we can start
> testing again?
> (The fixes I have done stop the audio output defaulting to the first
> location which is usually an input on windows - portaudio is correctly
> picking the default on windows - I haven't checked if it is doing so on
> GNU/Linux as I have only altered the windows case in these fixes. It was
> this bad behavior on windows that was causing there to be a failure to
> create the audio output which in turn meant that fluidsynth was never
> initialized, even though it will still be called - it was this that
> caused the crash changing tuning).
> 
> Richard
> 
> 
> On Mon, 2013-03-11 at 15:21 +0000, Richard Shann wrote:
> > I have noticed that the release crashes on windows if you try and change
> > the tuning. It happens when the initialization of the audio backend has
> > failed. What happens is that the fluidsynth initialization is being
> > skipped.
> > I am checking a fix in to master as this bug could cause loss of data
> > for users. I can only test the fix once it is in master because of the
> > way my gub is set up.
> > 
> > Richard
>