Increase asyncify stack size

[microbit-matt-hillsdon]

This fixes the example on #105.

I've added the debug script that I used, which I aspire to turn into an automated test in future.

The value does seem rather large but not really significant in a browser context. I've not explored whether it's possible to optimise our asyncify use.

I found an analogous issue in a JS interpreter here: justjake/quickjs-emscripten#112

[netlify]

✅ Deploy Preview for distracted-dubinsky-fd8a42 ready!

Name	Link
🔨 Latest commit	8802c14
🔍 Latest deploy log	https://app.netlify.com/sites/distracted-dubinsky-fd8a42/deploys/6512a245e5db2a00084c223d
😎 Deploy Preview	https://deploy-preview-106--distracted-dubinsky-fd8a42.netlify.app/demo
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[github-actions]

Preview build will be at
https://review-python-simulator.usermbit.org/asyncify-stack-size/

[dpgeorge]

I use the following test to determine maximum stack depth when calling a function:

def g():
    global depth
    depth += 1
    g()
depth = 0
try:
    g()
except RuntimeError:
    pass
print('maximum recursion depth g():', depth)

That works only when MICROPY_STACK_CHECK is enabled, which it's not for this simulator. It might be worth trying to enable that option here, which would protect against large call depths.

Otherwise, you can make a simple fixed-depth test like this:

def f(n):
    if n:
        f(n - 1)
f(168)

[microbit-matt-hillsdon]

I use the following test to determine maximum stack depth when calling a function:

def g():
    global depth
    depth += 1
    g()
depth = 0
try:
    g()
except RuntimeError:
    pass
print('maximum recursion depth g():', depth)

That works only when MICROPY_STACK_CHECK is enabled, which it's not for this simulator. It might be worth trying to enable that option here, which would protect against large call depths.

Otherwise, you can make a simple fixed-depth test like this:

def f(n):
    if n:
        f(n - 1)
f(168)

Yeah that makes sense. I got fixated on the structure of the original example. I've updated it to be a Python example alongside the others. Both examples you give work for me without enabling MICROPY_STACK_CHECK.

If I do enable MICROPY_STACK_CHECK (which is enabled for the CODAL port) I see worse behaviour:

FATAL: uncaught NLR 0x510d30 [firmware.js:1483:35](http://localhost:63072/build/firmware.js)
Aborted(Assertion failed: native function `stackRestore` called after runtime exit (use NO_EXIT_RUNTIME to keep it alive after main() exits)) [firmware.js:990:6](http://localhost:63072/build/firmware.js)
Aborted(Assertion failed: native function `mp_js_force_stop` called after runtime exit (use NO_EXIT_RUNTIME to keep it alive after main() exits)) [firmware.js:990:6](http://localhost:63072/build/firmware.js)
Simulator internal error: [demo:205:23](http://localhost:63072/demo)
Error: Aborted(Assertion failed: native function `mp_js_force_stop` called after runtime exit (use NO_EXIT_RUNTIME to keep it alive after main() exits))
    abort http://localhost:63072/build/firmware.js:1013
    assert http://localhost:63072/build/firmware.js:512
    createModule http://localhost:63072/build/firmware.js:1075
    forceStop http://localhost:63072/build/simulator.js:2201
    createRunningPromise http://localhost:63072/build/simulator.js:2470
    start http://localhost:63072/build/simulator.js:2440
    flash http://localhost:63072/build/simulator.js:2548
    createMessageListener http://localhost:63072/build/simulator.js:2751
    EventListener.handleEvent* http://localhost:63072/build/simulator.js:2826
    <anonymous> http://localhost:63072/build/simulator.js:2827
[demo:206:23](http://localhost:63072/demo)

Some of that (the mp_js_force_stop bit at least) seems like it might be a knock-on error from whatever originally goes wrong.

For the moment at least, I think it makes sense to go with the PR as I've just updated it.

[dpgeorge]

What does this do?

[microbit-matt-hillsdon]

Adds the example to the list in the demo

[dpgeorge]

Ah, very nice!