platformio-docs VS threadx

But ultimately platform.io it is an (ukrainian) open source project maintained by few and this most likely in free time. If you experience a lack of information, I would strongly suggest to contribute your experience to the project docs here and add the missing parts/hints to the documentation via pull request.

It is not open source. The source code is open but with an evaluation, i.e. proprietary license:

https://github.com/azure-rtos/threadx/blob/master/LICENSE.tx...

License looked alright, until I came across this: https://github.com/azure-rtos/threadx/blob/a8e5d0946c31385ff...

Portability, mainly. Zephyr runs on a wide variety of architectures and 450+ variations of popular boards are already supported upstream (adding more often being pretty easy). In your particular example, your own preferences with regards to building on top of open-source software may play a role too. For example, Azure RTOS is not open source, and its license will, among other things, prevent you to use it in production on other hardware than what's listed herehere. So easily moving from one hardware vendor / SoC to the other might be a problem, should this be a requirement (and it kind of becomes one for a lot of people in these times if silicon shortage)

Honestly as long as you use one of the supported MCU families ThreadX is probably a better choice these days given how much better NetX is over FreeRTOS’ TCP implementation.

Azure RTOS/ThreadX is only free on approved pre-licensed uCs, but with that said, there's a ton of devices covered by it so it's an option for a lot of projects.

I'm interested in something equivalent to tx_byte_pool but for non-embedded usage where I want a custom allocator for defined pools. Do you know if mimalloc can do this? I see it has custom heaps, but I didn't see any functions to create one from raw memory.

source is apparently available but no idea what licensing it uses