Purging 0.4.0 in favour of simplicity and little bit of performance cost.

[00imvj00]

Purging 0.4.0 in favour of simplicity and little bit of performance cost.

[MathiasKoch]

I would be sad to see that go, as it is the only "feature" keeping me here. I need the no-alloc for my application. That said, i can always just fork. In an idea world, i would just lie to see less fragmentation of ecosystem, not more.

I am also not in a position currently where i have time to look into solving the painpoints, especially as i have still not been presented with the actual pain points, more than "it's a lifetime thing"

[MathiasKoch]

It could be worth looking into how smoltcp handles their lifetimes using the managed crate.
They have extensive async support, allows alloc free nostd usage, and is also widely used in full-blown std applications.

[00imvj00]

@MathiasKoch @vincentdephily ^^

[00imvj00]

Is there any other way we can resolve this then ? where we can do no-alloc without lifetimes?

[00imvj00]

@MathiasKoch can we use Pin instead of taking lifetime??

[MathiasKoch]

My familiarity with Pin is very limited, but i think the issue might be the inner type of said pin? String is not an option, and i don't think Pin<&str> removes the need for the lifetime?

[00imvj00]

Yeah, true. I tried that. Doesn't work. by the way, are you using this in prod? if yes, can i know what kind of prod environment this is being used?

[MathiasKoch]

Yeah, we are using it in prod ;)

It's an IIoT device based on STM32, sending data to AWS, with roughly a stack of

• https://github.com/BlackbirdHQ/ublox-cellular-rs
• https://github.com/BlackbirdHQ/mqttrust
• https://github.com/BlackbirdHQ/rustot

Most of it is still work in progress, while the lower level stuff is starting to settle in.

[00imvj00]

awesome. 👍🏽

[00imvj00]

Interesting. Sure I will check it out.