00:05 <rafael[m]> <rafael[m]> "BTW there seem to be registers..." <- barafael: now I may have reached way outside of what I do understand properly - but hey, motion detction apparently works. I am kind of sorry I have spammed You with PR 🙂 I hope You don`t mind.

05:59 <barafael[m]> @rafael all good! I'm glad someone with actual hardware is tinkering with the driver and verifying it works. I have a long history with the mpu6050 and even though it's an old part I hold it dearly. I respect your drive to improve on it and will take a look soon

06:16 <barafael[m]> I had a brief look. Tremendous work. After your changes, let's add conditional opt-out derive for defmt format.

08:36 <rafael[m]> <barafael[m]> "I had a brief look. Tremendous..." <- I was missing that in the new motion.rs, right?

09:00 <barafael[m]> I mean because of Debug2Format

09:01 <rafael[m]> barafael[m]: saw it, I managed to not see the defmt-03 feature in the driver, so that is missing in the motion.rs as well as in all examples. Want that as a separate PR or stuffed into existing No. 11?

09:09 <barafael[m]> Ah right. No.11 is fine

09:21 <rafael[m]> done

10:38 <chomp4570[m]> I can't believe you and I are the first people to need this, putting a custom codec on an underlying transport must be very common, even in no_std world.

10:38 <chomp4570[m]> I'm looking for this exactly, have you found something in the meantime? I know there is futures-codec, but it allocates. Then there is async-codec but it doesn't support a no_std stream/sink impl because it uses std::io::Error.

10:39 <chomp4570[m]> Or am I the weird one and people are using the allocating futures-codecs and then just being very careful not to overflow their tiny custom heaps? I'm not experienced in embedded and don't know the state of the art.

10:42 <dirbaio[m]> chomp4570[m]: you're not the first one to need this. people just write the decoding loop themselves: read some bytes from the stream, feed them to a COBS (or whatever) decoder. You don't need a Codec abstraction to read COBS.

10:45 <dirbaio[m]> (in fact i've never understood the point of Codec... If you're implementing protocol X and it uses COBS or whatever, you just write code to decode that. you never need a "codec" abstraction so you can easily swap between COBS and line-delimited and length-delimited, no protocol out there ever needs that!)

10:48 <chomp4570[m]> For me it's about only needing to write only the "frame definition" and then getting a nice stream of messages from an asyncread, without having to implement scary unsafe pin things that I would most likely get wrong.

10:48 <chomp4570[m]> s/only//

10:54 <dirbaio[m]> you shouldn't need to use Pin to do that

11:27 <chomp4570[m]> I'm sure you're right, I was saying this after seeing the implementation in futures_codec and async-codec, which do use pin

11:56 <barafael[m]> In my case it was like LinesCodec from Tokio Util. I hand rolled it.

12:24 <barafael[m]> I also had the feeling that I can't be the first one needing this, though. Could embassy-io benefit from a few basic codecs?

12:26 <chomp4570[m]> I would like that

12:33 <barafael[m]> (not sure embassy-io is a good place for that)

12:34 <barafael[m]> For reference: https://github.com/barafael/tracker/tree/main/lines-codec

12:57 <mchodzikiewicz[m> hi, do you know if it is possible to query features in `target.`cfg()`` in `.cargo/config.toml`?... (full message at <https://catircservices.org/_irc/v1/media/download/AQMtx8ztrCVFN2W40PWRb40GJJcJNsZB5Cei2ub0GzWvN5fg5NqtJV97oHAqStdUjyCX676oz6c-IZsz6euccA-_8AAAAAAAAGNhdGlyY3NlcnZpY2VzLm9yZy9zTVJlT3lKdW91emtPYVlEWER3S0ZwbnQ>)

13:11 <JamesMunns[m]> docs say it should: https://doc.rust-lang.org/cargo/reference/config.html#targetcfgrunner

13:11 <JamesMunns[m]> but it also says:

13:11 <JamesMunns[m]> > If several <cfg> and <triple> entries match the current target, the flags are joined together.

13:11 <JamesMunns[m]> you might need to explicitly `not(feature = "println")` in the first example

13:11 <JamesMunns[m]> so:

13:11 <JamesMunns[m]> ```

13:12 <JamesMunns[m]> so:... (full message at <https://catircservices.org/_irc/v1/media/download/AeeTl9yJ4jYjQc5O7wDuKSSNuYH74-1EQSRotdr_y8YJkNRlxol6g5FGw1gi3ON2vmLq7VlYPNrowdyLZThn4CK_8AAAAAAAAGNhdGlyY3NlcnZpY2VzLm9yZy9QTHRNUVlxcFlzZkFuTnFSdVZjb2pGY3E>)

13:12 <JamesMunns[m]> s/printn/println

13:12 <JamesMunns[m]> but yeah, I think the issue is that you have multiple cfg matches

13:13 <JamesMunns[m]> Though it does say:

13:13 <JamesMunns[m]> > It is an error if more than one <cfg> runner matches the current target.

13:13 <JamesMunns[m]> dunno if that should print an error proactively? might be worth reporting to `cargo` as a bug

13:14 <JamesMunns[m]> and no idea if you can use crate features in the cfg statement either.

13:15 <JamesMunns[m]> (but it seems like you should be able to?)

13:17 <JamesMunns[m]> you could probably test it pretty easily by making your runner like echo or something

13:17 <mchodzikiewicz[m> adding explicit nots like this:... (full message at <https://catircservices.org/_irc/v1/media/download/AR9sPekfhawfWdeBxtp2hJOKTpN3ZRWYqGymCrA0fS2PcffrmNyHXt67Nd96ydlei9M1vdIkXQBlERY_sb68CKS_8AAAAAAAAGNhdGlyY3NlcnZpY2VzLm9yZy9MV1VPcGpQVFNmSnpEQXZ6YWhtTEJTaGs>)

13:18 <JamesMunns[m]> yeah, idk then. the docs seem to imply that should be allowed, I'd open a github issue on cargo with a mininal repro (one that just uses echo as a runner on a mostly empty repo)

13:19 <mchodzikiewicz[m> thanks, that's what I'll do

13:22 <mchodzikiewicz[m> * adding explicit nots like this:... (full message at <https://catircservices.org/_irc/v1/media/download/AUTWqR_-DgwuWfTsS5lkhxEJVUzfM-JpoiPN1AIVV6dA3t6mur3cLI16nlkbfLTTcqlJzHu6KPRJBBFFFAHkv76_8AAAAAAAAGNhdGlyY3NlcnZpY2VzLm9yZy9aUUhwdXlSa2JVb1NDT0V1R0V2T2tGSEI>)

13:23 <mchodzikiewicz[m> just fyi, my snippet above had non-matched braces but I fixed that and it didn't help - beside question if it supports crate features, it's a shame that it is a silent drop instead of error

13:34 <mchodzikiewicz[m> there's this issue possibly related but if I understand correctly issue have not been directly acknowledged https://github.com/rust-lang/cargo/issues/13737#issuecomment-2050052195

13:34 <mchodzikiewicz[m> multiple `config.toml`s suggested in the comment is a good way to proceed for me - in fact it will allow me to get rid of some other args I need to provide atm, namely to support `imc` and `imac` targets

22:23 <laenzi[m]> I am struggling with using embedded_hal_bus in an RTIC application. How can I store an AtomicDevice in the Local struct to access it from the task?