Viktor Vilhelm Sonesten

Engine error if mjcb_passive callback installed before model instantiation

2

On x86 Linux with mujoco 3.1.1, in a REPL, a passive force callback can be installed after model instantiation without issue: ```python >>> import mujoco >>> def callback(model, data): ......

Why does the flake integration wrapper just offer a subset?

2

Installed via flake, I try ``` $ devenv container shell https://devenv.sh (version 1.0): Fast, Declarative, Reproducible, and Composable Developer Environments This is a flake integration wrapper that comes with a...

Request for new release

2

I was trying to package the latest release for NixOS but stumbled upon some portability issues with tup, which has since then been removed as a build dependency. Packaging of...

Crate reimplemented with more features. Replace?

2

I recently published [itm-decode](https://github.com/tmplt/itm-decode) before stumbling upon this crate just now. Comparing the two after a cursory dive, `itm` seems a bit more general with its application of traits; `itm-decode`...

Add iterator for instrumentation/extension packet arbitration and decoding

2

Refer to https://github.com/rtic-scope/itm-decode/issues/19.

Review relationship between local and global timestamps

1

Current implementation assumes that local and global timestamps are sourced from the same clock. This is not neccesarily the case, and is platform dependent: global timestamps are sourced from a...

`itm-decode` aims to incorporate the feature set of https://github.com/japaric/itm-tools. A `pcsampl` equivalence in `itm-decode` can be achieved with [addr2line](https://crates.io/crates/addr2line).

[As documented](https://github.com/rtic-scope#why-am-i-getting-an-erroneous-trace), configuring a GPIO pin for SWO induces transient noise. Noise is not deterministic, and will be different across platforms. Recording the SWO pin from target start will then...

The *Embedded Trace Macrocell* (ETM) appears to be some superset of tracing features above ITM. Documentation suggests that this trace stream is merged with the ITM trace stream. It would...

As discussed in https://github.com/rust-embedded/itm/pull/41, it may be apt to adopt an asynchronous API. This could potentially simplify the implementation: the first EOF condition would be signalled with a `Poll::Pending`.