GreySynth

[RJHollins]

Bugger!

That's something a little more serious, then. An 'Access Violation' means that FlowStone is trying to read or write to a memory address outside of the memory allocated to it by Windows (it quite often happens if code tries to read/write past the end of an array or string.) It might be possible to work out which component is causing it by a process of elimination, but really, I think MyCo is probably the only person with a good chance of getting to the bottom of it.

Thanks for even looking at his TROG.

I don't recall this ever happening with any other schematic. I was just curious to see BillV work.

oh well ...

still .... Thanks !

[billv]

Hey guys, thanks for troubleshooting the problem.
Since Trog has isolated the preset manager, makes sense to replace it.
RJ: here it is with the default stock preset manager.
The presets are still intact.

Grey Synth_v01_alt_RJHollins.fsm

(1.48 MiB) Downloaded 954 times

[RJHollins]

Hey guys, thanks for troubleshooting the problem.
Since Trog has isolated the preset manager, makes sense to replace it.
RJ: here it is with the default stock preset manager.
The presets are still intact.

Grey Synth_v01_alt_RJHollins.fsm

Hi BillV,

Wooo ... I just tried to DL and load ... but again hit with an Error message:

caused an Access Violation (0xc0000005)
in module FlowStone.exe at 0023:005416ae.

Exception handler called in ExceptionAttacher.cpp - AfxWinMain.
Error occurred at 8/15/2019 22:21:04.

I've no idea why this is happening ... and even less of a solution.

Are you using some specialized CODE in this schematic ???

[billv]

Shit RJ ..that's a bummer.
Ok.. so now we know its not the preset manager, and it's not Win 10(trog's on that).
There's no specialized code... the selectors are all ruby..but that's about it..
Its a very basic synth RJ...theres not that much in there.
I think now it looks like we have to wait for someone else to come up the same error.

[RJHollins]

Shit RJ ..that's a bummer.
Ok.. so now we know its not the preset manager, and it's not Win 10(trog's on that).
There's no specialized code... the selectors are all ruby..but that's about it..
Its a very basic synth RJ...theres not that much in there.
I think now it looks like we have to wait for someone else to come up the same error.

I'm equally perplex.

I've had the pleasure of some of your shared work [and help to me] over the years [and much appreciated].

To those with more insights, I'll mention I run FS 3.07 [without issue], on a Win7[64] system.

Regardless ... Thanks again BillV ... and the GANG here on the FS forum.

[trogluddite]

the selectors are all ruby...

There's a slim chance that it could be related to the Ruby components. FS 3.0.7 changed how the RubyEdit callback methods get called (e.g. event, draw, mouse methods, etc.) to support per-RubyEdit timeouts, and this is known to cause a memory leak and a lot of extra CPU load. This is the main reason that 3.0.6 is generally recommended over the later versions, despite the older 'global' Ruby timeouts being such a PITA.

I also noticed that the module "Ticker - Custom SEQ" uses 'scheduleMethod', and this too, has a bug which I reported a while ago which affects all of the FS 3.x.x versions since that method was introduced. If Ruby runs its 'garbage collection' routine while the scheduled method is pending, the arguments which are passed after the method name can get thrown away, so that they're not available when the method gets executed. It may be just a coincidence, but an access violation is exactly the kind of error which this might lead to. (NB: A workaround is to use the 'input' method with a time delay to schedule an event for a dummy input.)

It's a long-shot that either of those are the problem, as Ruby generally needs to be pretty low on memory before you'd notice any nasty side-effects from them; but given how many temporary drawing objects are created, it might be possible (NB: It's a lot better to create Pens/Brushes/Fonts etc. once at startup and re-use them rather than create them within the draw method - as I've said before, DSPr need a slap for providing such lousy code examples!)

[RJHollins]

As a test ... I reverted back to FS 3.06 ... had the same Error crash as with 3.07

[trogluddite]

Hmm, that's exhausted my ideas. My old XP box has departed for "silicon heaven" now (RIP), so I don't have any way to try replicating the problem any more; though I've used all of these FS versions in XP at one time or another without seeing anything quite like this (except for the dodgy Betas, of course!)

[RJHollins]

Hmm, that's exhausted my ideas. My old XP box has departed for "silicon heaven" now (RIP), so I don't have any way to try replicating the problem any more; though I've used all of these FS versions in XP at one time or another without seeing anything quite like this (except for the dodgy Betas, of course!)

This is a strange one ...

I wonder ... there was something about Audio In/Out modules being selected when saving/exporting a file.

Gosh ... I can't recall the specifics ... but I remember some issue that could happen ... anyone recall something of this nature.

I've no idea if this is possible here ... just stinkin' out loud.

[trogluddite]

Audio In/Out modules being selected when saving/exporting a file

Yes, you could be onto something there. I do usually advise people to turn audio off before uploading because it can sometimes cause startup problems (so I'm mystified why that never occurred to me - especially as I was advising djbrynte about that very thing earlier today - d'oh! I blame the porn-bots for distracting me! ).

I made sure to do that in my download, but maybe it can also happen even if they're turned off in some cases? Here's the schematic again, but with no audio or MIDI I/O primitives...

Grey Synth_v01_NO_AUDIO_TROG.fsm

(1.47 MiB) Downloaded 947 times

Fingers crossed!