i wouldn't be surprised that the code that is hard to automatically translate into safe rust is *precisely* the code where memory safety bugs hide with highest density and helping focus developers on that (by handling the rest) could help in oxidation.

automated 'idiomitization' passes can be layered on as was the case in the original c2rust vision. but the reality (and where i agree with you), is that there will have to be deeper, human-led restructurings if we want to end with safe, idiomatic rust.

liked the article, but i disagree with the conclusion. while having formal semantics is a desirable project (and being worked on by the op-sem team), its not a blocker for c-to-rust work.

imo there's huge in translating c in a (informally) semantics-preserving way even if its unsafe.

a more "readable" and "useful" version of the same trace

I think it’s an American thing primarily in liberal arts / social sciences.

I'm trying to get shock diamonds going, that's how you find the *real* bugs

Also if you find your own pet cause or talking point reflected in this unhinged document, go within yourself, my brother. You are a fellow traveler to some of the dumbest enemies the academic enterprise has ever had.

You can just use your wallet app now, there’s a little bit of set up still but it’s a lot closer.

spanner? s3?

the internet?

It’s been a long held opinion of mine that the American system should rebalance the distinction between software *engineering* and computer *science* degrees. Trying to make CS a universal program means it fails at the science side of things.

Oh yeah I forgot that as well.

I wouldn’t go quite so far, typically includes discrete math, intro to probability & stats. You will find very few logic, semantics or other theoretical cs courses at the undergraduate level. That’s one of the reason I came it Paris for the MPRI