>>94His wording about heap allocation is pretty weird. What he's trying to say is you have to be able to avoid calling malloc() ever. I agree with that.
I'm not sure if he's written code for one of these microcontrollers. I haven't, but writing code for an ARM chip with a few kb of TCM worked out pretty similarly to what he describes. I had malloc() but it was used to request allocations in EWRAM by carrier pigeon so I could later deliver data there by horse-drawn carriage, or more likely, send carrier pigeons to the other components asking them to send carriages with data directly to EWRAM, or vice-versa.
>>95What's off base about that? I mean sure you can write FFI, but nobody is going to accept that for a language deemed suitable for embedded programming. Except the gophers.
I don't know what's better, the sweet sweet tears of gopher's failure to write an OS kernel, or Rob Pike's admission that he should never have described Go as a systems language.