That LLVM bug I was complaining about? Turns out that qemu developers discovered it back in 2011. And it's still broken! #notimpressed
-
-
Replying to @johnregehr
@johnregehr@cperciva even SSA construction needs to be disabled (which is what caused the bug Colin filed).1 reply 0 retweets 0 likes -
Replying to @SCombinator
@SCombinator@cperciva ugh! bogus part of the standard then I guess1 reply 0 retweets 0 likes -
Replying to @johnregehr
@johnregehr@SCombinator I have trouble accepting that the standard is bogus here. How would you fix it?2 replies 0 retweets 0 likes -
Replying to @cperciva
@cperciva@johnregehr "all nonvolatile locals have indeterminate values on longjmp"? Or will that break too much legacy code?2 replies 0 retweets 0 likes -
Replying to @SCombinator
@SCombinator@johnregehr LLVM doesn't obey that either, because it turns variables *into* locals.2 replies 0 retweets 0 likes -
Replying to @cperciva
@cperciva@SCombinator what about making setjmp a macro that does an asm volatile (“” : : : “memory”) before calling the real setjmp?4 replies 0 retweets 0 likes
Brian Smith Retweeted John Regehr
Brian Smith added,
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.