Evan Ovadia

> (@nmsmith) In fact, there's no reason that `__del__` even _needs_ to invoke `destroy`. `__del__` could very well just **transfer the value of `self` somewhere else.** In this sense, `__del__`...

If we were just trying to make a better Rust, or a better C++, I'd agree with everything you just said. Especially this part: > > As much as the...

+1 to pretty much everything you said. And thanks for bringing up how greppable `unsafe` is, we really should capture that benefit in Mojo. Your idea of having a compiler...

This sounds similar to what `_StringCapacityField.FLAG_IS_STATIC_CONSTANT` was meant for (looking at stdlib/collections/string/string.mojo), though that looks more like a run-time flag in the string itself. Though, I assume you're suggesting a...

Thanks for this bug report @saflota! And thank you for the investigation and fix @soraros =) Weiwei's conclusion was similar: > I think the problem here is that base case...

Thanks for the report! Reproduced with this program: ``` from gpu.host import DeviceContext, DeviceBuffer fn reduce_neighbored[n: UInt](mut in_data: LegacyUnsafePointer[Int32], out_data: LegacyUnsafePointer[Int32]): pass fn main() raises: var ctx = DeviceContext() alias...

Three things I want to think about: 1. Can we handle what the user wants to do in today's Mojo? 2. Do we want to support `mut` arguments in `enqueue_function_checked`...

Repro'd, thanks for the snippet! That should indeed be an error, even python doesn't allow that to span multiple lines. Moving to the backlog to be considered for next cycle.

Thanks for the report! Moving to the backlog to be considered for next cycle, but please DM me if this issue is time-sensitive.+1 to banning them in finally blocks, even...

Repro'd, thanks for the snippet! This might get fixed soon in some coming refactors in the parameter inference system, fingers crossed. Moving to the backlog to be considered for next...