Great read. Did @rustlang people know about @swiftlang's overflow checking optimizations? A topic worth revisiting.https://twitter.com/johnregehr/status/718199181295362049 …
@CopperheadSec @rustlang ...it seems unlikely that significant optimizations would be made just for that non-default mode.
-
-
@BRIAN_____@CopperheadSec@rustlang I doubt we would ignore significant optimizations because we have an unwinding mode. -
@wycats@CopperheadSec@rustlang Good to hear. Looking forward to seeing the improvements. -
@BRIAN_____@wycats@CopperheadSec it just so happens that https://twitter.com/rustlang/status/718545850226589696 … just happened! -
@rustlang@wycats@CopperheadSec Good to hear. If I understand the RFC correctly, though, it requires substituting a custom libstd? -
@BRIAN_____@rustlang getting full benefit requires a std with special flags, but work on better cross-compilation makes this easy. -
@huon_w@BRIAN_____@rustlang in other words, we make progress incrementally :) slow and steady wins the race.
End of conversation
New conversation -
-
-
@BRIAN_____@rustlang If it was done well it'd benefit a lot more from optimizations implemented in LLVM. Ideally every fn marked nounwind. -
@BRIAN_____@rustlang However, even if it aborts, each code path is still going to be unique unless the debugging information is removed. -
@CopperheadSec@rustlang Do you have a concrete example handy that shows why Rust's panic/unwind inhibits these optimizations? -
@BRIAN_____@rustlang There are 2 separate issues: each error handling code path is unique, and then there's the unwinding control flow.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.