@jamesmunns.com Where do we best spool these RE'd parallels between all those who do this?

I started something for pointers to board pinouts, stock bootloaders and that called an embedded encyclopedia (not public yet), that could also collect those chip details.

Or a structured wiki, something […]

@nytpu Sounds awesome for applications like "The hardware team bought PC based digital signage, but actually we want it to be up as fast as the display gets when power comes, and be able to audit all our deployed codebase".
(No sarcasm: As someone using embedded Rust a lot, this is tempting to try.)

@jschauma I get that this is a problematic trend, but I don't understand why mitigation is difficult. Someone submits this, it's suspicious, they get asked to explain it, they fail to, and get a strike for cheating. Depending on school policy, they're out for good on the n'th strike. No […]

@hvdsomp @jyasskin I was using alternate as a strawman; the actual relations probably need to be custom for this purpose, except for the RFC7089 mementos that IIRC describe that precise "A was archived at B" relation needed here in terms of web links.

@jyasskin My kneejerk reaction: -- but probably there are more thought-through approaches around.
Having something specified and supported would be great.

To clarify, by "fully resolve symlinks" I mean that if the symlink target, or any directory in its path is, the original symlink would be updated to the target instead. Possibly with allow/disallow-rules that allow leaving a single level of symlinks in.
Undoing that operation would mean having a […]

@asonix Not 100% sure if it really works that way, though: There are a lot of other friction points with serde, eg. that it doesn't support integer keys serialized structs, so maybe those newtypes just work when going through minicbor's own derives rather than through serde's.

@asonix The closest example I can give is 8 RFC numbers off what you're doing:
In CBOR, there exists both a byte-string type, and an array type (that can, among other things, contain integers). So any CBOR serde binding, when faced with a &[u8], has to make a choice between "it's obviously […]