On Thu, Aug 2, 2018 at 9:28 AM, <[email protected]> wrote: > On Wednesday, August 1, 2018 at 10:17:38 AM UTC-4, Ian Lance Taylor wrote: >> >> [...] >> That said, the exact semantics of the sync/atomic operations are not >> written down (that is https://golang.org/issue/5045). > > > I really love go, but this is one thing that has annoyed me for years. This > issue is from 2013, and AFAICT, there is still no clear statement of what > guarantees the sync/atomic functions make. I know that they should be rarely > used. I know that the edge cases for atomic operations are complicated to > define, hence all the options for C++ atomics. But on the rare occasion that > I do need to use them, I always feel a lot of trepidation, because I can > really make few, if any guarantees based on the documentation.
Yes, the problem is that nobody is entirely happy with the memory model either. We need to figure out how to improve that first. Ian -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
