21:57 <thejpster[m]> Over in the rp-rs room, this came up:... (full message at <https://catircservices.org/_irc/v1/media/download/AaXObS6S8EYrTEnNdr7yOa0J2BPsDiDJ2ygX3okiTmvq863CmOBY9AEq2jKEGOWnjduV0o718nJ6_wWRQmJcYzq_8AAAAAAAAGNhdGlyY3NlcnZpY2VzLm9yZy9mVE9mQUJMZVlKT2RMZ3dkRmtWYUpUVHY>)

21:58 <JamesMunns[m]> In .text

21:58 <JamesMunns[m]> With the pointer living happily in .data

21:58 <thejpster[m]> Yeah :/

21:59 <whitequark[cis]> not .rodata?

21:59 <thejpster[m]> static slices are probably an anti-pattern.

21:59 <thejpster[m]> mmm, fair. Likely rodata. But in Flash, not SRAM as intended.

22:00 <whitequark[cis]> what could possibly be the benefit of placing a &[u8] in SRAM?

22:02 <whitequark[cis]> oh, i guess you could have a SoC where that gives you single-cycle access

22:02 <thejpster[m]> Yeah it’s very unlikely to be useful. Slice references are Copy anyway.

22:03 <whitequark[cis]> you definitely can't modify it, so it has to be something where the reads are different in some useful way

22:03 <whitequark[cis]> (and you can't have a &mut anything in a static)

22:03 <JamesMunns[m]> https://play.rust-lang.org/?version=stable&mode=debug&edition=2024&gist=413f8e53c07595044a571a3c0bafa051

22:03 <JamesMunns[m]> you can do fun things for the cost of a small macro tho

22:03 <thejpster[m]> The author clearly wanted the [u8] in RAM to avoid XIP cache contention and messed up latency

22:04 <thejpster[m]> whitequark[cis]: I misunderstood you here

22:04 <whitequark[cis]> whitequark[cis]: yeah, figured

22:04 <whitequark[cis]> you could have a #[link_section = ".data"] static DATA: [u8; N] = [...] like in James' example

22:05 <thejpster[m]> That is what was intended in the code example

22:06 <thejpster[m]> And confusion arose when core 0 activity affected video rendering on core 1.

22:06 <thejpster[m]> I wonder if clippy should lint for it

22:11 <thejpster[m]> As an aside, I call &[T] a slice when teaching but often later have to explain that [T] is actually the slice and &[T] is really a shared reference to a slice that knows the length of what it references.

23:33 <Lumpio[m]> Would anybody happen to have any ARM embedded elf binaries built with the current rustc to do a sanity check?... (full message at <https://catircservices.org/_irc/v1/media/download/AXYBtPBDaXzuyCZlSC_qrt9Ric9raO-bXN54t0WTMhujobC6Fv999_pSJw46RYMFpOc8xpqzP4g804HmUGj3r-2_8AAAAAAAAGNhdGlyY3NlcnZpY2VzLm9yZy9JRllBQ2tLQUxJQmJJVUhlQVljcFhmeVk>)

23:34 <Lumpio[m]> Not that it matters a lot but elf2uf2-rs is not liking it - i patched out the check and the binary seems fine but what has changed exactly here

23:34 <Lumpio[m]> Has the ABI field changed for any of you people..?

23:51 <KevinPFleming[m]> Yes, this has been discussed in this channel quite a few times in the past week or two, always in the context of elf2uf2-rs, and the recommendation is to use picotool instead.

23:58 <Lumpio[m]> Oh, well, serves me right for not searching!

23:59 <Lumpio[m]> Thanks