Why work only with tty: true?

[Foldren]

I tried to start the application aioclock in the docker container but it doesn't start without tty: true

[ManiMozaffar]

Can you share the logs of container when you run without tty?
I’m using the library without using tty. It’s working fine, I guess the issue is somewhere else

[ManiMozaffar]

@Foldren
I think I understand why
So aioclock serve function basically exit if there’s no more task to be running, why keep running a forever idle application?

for that to happen, all your trigger conditions should no longer ever satisfy. For example having loop_count=1 ensures all tasks are executed only once. If all tasks are executed, then aioclock quit as mentioned.

There is 2 possible scenarios:

1. Either it correctly determines that no longer task happened
2. Or it incorrectly determined that and you actually had some tasks to be running.

i can improve logging a bit to make this more clear,
But if you share your triggers that you use, that would reproduce what happened and we can find out it was a bug or not.

[Foldren]

On start in lifespan i get message, but it's doesn't work without tty

[ManiMozaffar]

Can you provide a snippet code that reproduce the issue? @Foldren
I'm 99% sure the root of cause is aioclock quitting for some reasons. that's why tty true would fix it.

[Foldren]

from asyncio import run
from contextlib import asynccontextmanager
from aioclock import AioClock, At, Every, Once
from tortoise import Tortoise
from config import TORTOISE_CONFIG, APP_NAME
from modules.logger import Logger
from modules.playwright.maybank import Maybank


@asynccontextmanager
async def lifespan(_: AioClock):
    await Tortoise.init(TORTOISE_CONFIG)
    await Tortoise.generate_schemas(safe=True)
    await Logger(APP_NAME).success(msg="Started.", func_name="startup")
    yield _

app = AioClock(lifespan=lifespan)


@app.task(trigger=At(tz="Europe/Moscow", hour=3, minute=5))
async def load_stats():
    await Maybank().load_stats()


if __name__ == '__main__':
    run(app.serve())

[Foldren]

by the way, it would be cool to add your own logger to aioclock

[ManiMozaffar]

@Foldren I don't understand what do you mean by own logger, we're using python stdlib logger which is highly configurable and should be compatible with any other logging libraries.
If you raise Exception in your lifespan, it would raise th exception all the way up. I can also document this, doesn't matter also it should be case with Start and End event as well.
I think the setup you do with tortoise sometimes fail and raise exception, can you put a try catch to see if there's any exception on your side? This should likely to solve your problem.

P.S: you can run this code few time to understand what I mean :)

import asyncio
import random
from contextlib import asynccontextmanager

from aioclock import AioClock, Once


@asynccontextmanager
async def lifespan(app: AioClock):
    print("Hey")
    if random.choice([True, False]):
        raise Exception("lifespan")
    yield app
    print("Bye")


app = AioClock(lifespan=lifespan)


@app.task(trigger=Once())
async def foo():
    print("foo")


if __name__ == "__main__":
    asyncio.run(app.serve())

[ManiMozaffar]

Closing due to inactivity.

[Foldren]

Can you try launch this in docker and send container log?

[Foldren]

Well, the problem has not been solved