On Tue, Feb 14, 2012 at 04:39:37PM -0600, Jeff Albert wrote:

On his windows machine it only plays backbeat (regardless of the order in which the Machine.adds appear). On my Mac it works just fine, both files play.

Ideas?

So, by it "not playing", do you mean ChucK simply doesn't indicate it's sporking the shred at all? Or that you don't hear a sound? Have you tried stuff like editing the offending file to start with a debug print? Assuming that the filenames indicated are the actual ones then that's not it, those should be legal under both, but Windows and Mac potentially use slightly different non-printable characters for stuff like the end of a line. You could try opening and re-saving them using a Windows native text editor or some tool to strip those characters, I forget the details (stuff like cariage-return....) but your favourite "serious programmer's editor" should have a tool to convert/strip those. That might help? These are just my first guesses, with more datapoints better guesses could be made. If the file in question isn't super-secrit I'd like to see a copy of it that's the most simple file that still causes this issue. Good luck, it sounds mysterious and I hope you'll share if/when you find what on earth (or elsewhere, as the case might be) is causing this. Yours, Kas.

On Tue, Feb 14, 2012 at 05:36:07PM -0600, Jeff Albert wrote:

Good point. Chuck says the shred is sparked, we just get no sound.

Up to here it sounds fairly normal....

Have you tried stuff like editing the offending file to start with a debug print?

If I run either file alone, they each work, individually.

Right, so here it gets truly interesting. So these somehow affect each other, perhaps using some common resource? Perhaps something like a system path is set? Is any sort of global name-space involved? It'd also be interesting to see whether they *finish* running without sound or whether they get stuck at some time before that?

This sounds like a good lead. I think he wrote the files in miniaudicle on Windows (although he is running them from the command line), but he may have copied/pasted some of it from some Mac generated files I gave him. I'll follow that up.

If they each run individually on both systems I suspect it's not this (but weirder things have happened). I suspect some sort of common resource is affected in some sort of unforeseen way. Probably some sort of bug in one of the places where the VM touches on the host OS. If the files are of a reasonable size you could try either stripping them of suspicious bits until they work or start with nothing and add parts of the file until it breaks. If there is time and space for that you could suggest that your student do that and that (s)he send a bug report, IMHO that's a skill that also deserves some time in a learning environment. Good luck, Kas.

Jeff, Is that the entire file or just a snippet? I've sometimes run into a problem where my last line is some tiny amount of time (e.g. 1::samp => now) and the previous line is a spork ~ someFun(); or a Machine.add() and the main code exits before the shred-spawning function ever has a chance to return. If your code is the entire file we're looking at, does it behave correctly when the last line is 10::ms => now; ? -Mike http://michaelclemow.com http://semiotech.org On Tue, Feb 14, 2012 at 7:12 PM, Hans Aberg wrote:

On 15 Feb 2012, at 00:18, Kassen wrote:

...

Assuming that the filenames indicated are the actual ones then that's not it, those should be legal under both, but Windows and Mac potentially use slightly different non-printable characters for stuff like the end of a line.

I recall there was a bug in miniAudicle at Mac where it choked at UTF-8 characters in the beginning, even though they were in a comment. Maybe fixed by now, but perhaps not on all platforms.

So try what happens in 'chuck' at the command line.

Hans

_______________________________________________ chuck-users mailing list chuck-users@lists.cs.princeton.edu https://lists.cs.princeton.edu/mailman/listinfo/chuck-users

On Tue, Feb 14, 2012 at 11:27:56PM -0500, mike clemow wrote: Mike,

Is that the entire file or just a snippet? I've sometimes run into a problem where my last line is some tiny amount of time (e.g. 1::samp => now) and the previous line is a spork ~ someFun(); or a Machine.add() and the main code exits before the shred-spawning function ever has a chance to return.

Quite so, and this is interesting. The same can (could?) happen for printing statements. With print I can understand it because (if I remember correctly) printing doesn't return anything. Machine.add() should be waited for (strictly speaking) because it will return a value, being a integer indicating the Shred's id. Waiting for such calls that depend on the OS (both involve the shell displaying some text) is risky though, because most OS's aren't realtime and where they are we could still advance time by less than the period in which the OS guarantees getting round to such tasks in. At any rate it's a interesting case. Kas.