search

lancaster-university / codal-microbit-v2

CODAL target for the micro:bit v2.x series of devices
MIT License
41 stars 50 forks source link

MakeCode simple parrot programme runs out of memory #375

Closed microbit-carlos closed 8 months ago

microbit-carlos commented 9 months ago

This simple MakeCode parrot programme from https://github.com/microsoft/pxt-microbit/issues/5312 runs out of memory consistently after a few recordings:

input.onButtonPressed(Button.A, function () {
    record.setSampleRate(11000)
    basic.showString("?")
    record.startRecording(record.BlockingState.Blocking)
    basic.showLeds(`
        . # . . .
        . # # . .
        . # # # .
        . # # . .
        . # . . .
        `)
    record.setSampleRate(18000)
    record.playAudio(record.BlockingState.Blocking)
    basic.clearScreen()
})

It might run out of memory with error 020 after a couple of recordings, or a few more. So far I've always managed to trigger it within 5 attempts.

❌ This can be replicated with MakeCode Live v6.0.18 with CODAL v0.2.57:

❌ I can also be replicated with MakeCode v6.0.17 with the same CODAL v0.2.57:

❌ I can also be replicated with MakeCode v6.0.16 with CODAL v0.2.56:

✅ Going back to version v6.0.15 with CODAL v0.2.55 and I cannot replicate anymore:

So it might have been introduced in v0.2.56: https://github.com/lancaster-university/codal-microbit-v2/compare/v0.2.55...v0.2.56

microbit-carlos commented 9 months ago

On Monday we should figure out if we can issue quick hotfix or if we need to ask the MakeCode team to revert back to CODAL v0.2.55.

JohnVidler commented 9 months ago

Just a note here to say I'm currently investigating this.

microbit-carlos commented 9 months ago

Tag v0.2.60 has increased the maximum size for stream recordings, as a temporary workaround to this issue while we continue investigating it: https://github.com/lancaster-university/codal-core/commit/a1163541d10c0a0d3dd9b2e8ac1b8906100449ad

It's been merged in pxt-microbit and we are hoping to be able to release to MakeCode Live asap:

JohnVidler commented 8 months ago

With the updated memory configuration in v0.2.61 Makecode seems much happier, and I'm now unable to reproduce the 020 crash.

There seems to be another issue with the sample rate being forgotten on the second run, but I don't believe this is related to this particular issue, so I'm going to close this one as done for now.

image

(Makecode versions for reference)