@CAD97_ @ekuber @rustlang - https://t.co/1lWTYKB3u2
- https://t.co/YlCOkxPt63
- https://t.co/eYOgWooghR
The last one is interesting because it is *not* a double-free bug.
@nick_r_cameron There's an unstable method in nightly:
https://t.co/4TwPu6jEvd
I wrote a similar thing for this benchmarksgame code:
https://t.co/C8sONsOaRy
@TedMielczarek @mycoliza @ekuber Only in cases where it has no effect, like trying to drop a reference. I think you are remembering this case which we talked about before: https://t.co/TZIF8Gw7UO
@nick_r_cameron I have seen a few real-world bugs caused by DST pointer comparison: https://t.co/9jztbGYQs2
It's especially tricky because it works in simple cases, to later fail unpredictably. And casting to thin pointer first fixes some bugs but introduces others.
@QEDunham “Ergonym” (from Greek, “work name”) would be good, but apparently that means something else. How about “merconym” from Latin “merces” (“wage” or “hire”)?
@fasterthanlime The worst thing is that there is one rule that applies to temporaries in `match` and `while let` expressions, and a different rule that applies to `if` and `if let` and `while`. https://t.co/AGSJy8z1kU
@ManishEarth King County Metro made a great train-style map of what the rapid lines will be in 2024, but they don’t seem to have an equivalent for the network today. https://t.co/dhu2EjvoAy
@JoshHaberman alloc_ref doesn’t have this problem because the return value doesn’t borrow from `self`, so `self` is only ever borrowed temporarily (for the duration of the alloc_ref call).
@JoshHaberman As an aside the GhostCell paper has a really mind-blowing way of solving some of the problems with arenas in Rust: https://t.co/AfdpLnzVCX
@JoshHaberman `alloc` taking `&mut self` would be a problem even outside the data structure case that you mention: Just calling `alloc` twice and then using both of the returned pointers would be disallowed.
@KromanDavid Not quite. To calculate efficacy, you would also need to know how much of the population was vaccinated vs. unvaccinated at the time, and control for factors like age. (Your calculation would be valid if every age group were 50% vaccinated and 50% unvax.)