concat_idents in define_pins macro about atsamd HOT 4 CLOSED

Oooh. I'm a bit on the fence because the new dependency doesn't seem to add a huge amount of value and introduces some magic with the max recursion depth thingy... but it is nicer.
Does this make the compilation times dramatically longer?
Did it strike you as overly complicated to produce that macro vs. the vanilla rust macro experience?

from atsamd.

It didn't make compilation times noticeably longer.
I'd say it took me about an hour with the docs open to figure out a form that worked and how the mashup macro needed to be used in this case. However actually using the macro is less complicated now than it was before.

Having said that, honestly I'm kind of toying with just removing the macro completely despite it being a cool toy. It depends what we're optimising for - i.e. 1) somebody reading the source/ 2) our need to make future changes/ 3) other integrations taking less time.

The best things for those three cases respectively are 1) no macro (the source is entirely clear if boilerplatey), then 2) the old macro (less complicated to change but still halving the code you need to write/refactor), and then 3) the new macro form makes writing out the pin specifications quickest.

Honestly if it were just my crate I'd use the newest macro but the other options are reasonable too.

from atsamd.

let's use the new macro; I've a feeling that we're going to grow more board crates and it will help keep the boilerplate lower.

from atsamd.

Closed via #9

from atsamd.