[Fix 11.4.0] 11.4 RC4 "SCS Thread Assertion Failed"

[rborsey]

Hi Mike - sorry but . . .

Having done some serious programming with the wonderful call cues I am sadly having frequent crashes of RC4 with the message for example :

SCS Thread Assertion Failed
openMidiFile[104RL] expected thread #SCS_THREAD_MAIN but current thread is #SCS_THREAD_CONTROL

The message stays a while if I let it play on and then the software dies. There is also an OK button - if I press on that the software dies immediately.

It always happens when I have a sequence of cues such as follows which start an audio and midi file pair playing. The audio file goes into a loop. There is a manual cue to release the loop and stop the midi file and which also fires off some cue send messages and then starts a second midi file playing. This all worked fine in 11.3.8 when I used cue sends instead of call cues - for my 11.4 version I have created some simple callable cues to do my cue sends. So the typical sequence is :

104 manual start song cue with sub cues for audio file with a loop), a cue send or two, and now a few Call Cue sub cues
104L a midi file cue which auto starts with the start of 104
104R a manual start release cue which has sub cues to release the loop of 104 and stop immediately 104L, and sub cues calling cue send callable cues
104RL an auto start midi file cue which starts at the end of 104R (ie when the loop releases)

The error seems to mostly occur on loop release although sometimes on starting the release cue. Although the example quoted above has an error referring to 104RL which is the directly related midi file cue, sometimes it refers to a later midi file cue in the file - perhaps it is trying to get the cue ready to go??

Last feature possibly Worth mentioning is that some of my callable cues themselves call other callable cues. SCS seems to allow this although when you initially proposed a different call/return kind of implementation you did say there would be no multi-level calling possible.

Not sure if I can get a diagnostic file as the error doesn't always occur - and once it has happened it is too late of course

Hope this makes sense

Cheers

Richard

[rborsey]

It occurs to me that the trace version you sent me for the cue send midi blocking problem mights provide some useful information so just in case I'll try and get a log up to the point of failure . . .

Cheers

Richard

(I know this is an RC version but if we can make it work I want to use it Sunday evening for a show . . .)

[rborsey]

Thanks Mike for tracking this down. I am a happy boy!

[Mike Daniell]

Fixed in RC5.