LLVM 12.0 Should Be Released Soon Following RC3 Release

[phoronix]

Phoronix: LLVM 12.0 Should Be Released Soon Following RC3 Release

LLVM 12.0 was aiming for release at the start of March but it missed that goal post and a third release candidate was tagged on Wednesday. But, fortunately, it looks like the official release will be out soon...

https://www.phoronix.com/scan.php?page=news_item&px=LLVM-12.0-RC3-Released

[kylew77]

Darn shame it didn't release in time to make it into either FreeBSD 13.0 or OpenBSD 6.9 but should be the compiler of choice for both FreeBSD 13.1 and OpenBSD 7.0 assuming both release in October/November which means they will of course miss release 13.0 of LLVM. Seems like the OSes that use LLVM and the LLVM project could be more coordinated in releasing their new releases near each others with enough time for integration.

[oleid]

This release is especially interesting as they landed the fix for the (very longstanding) no-alias bug in LLVM.

https://www.reddit.com/r/rust/commen...anded_in_time/

The big question is: how long will it take this time until rust folks discover miscompilation -- or will they now be able to use it for real?

[bug77]

This release is especially interesting as they landed the fix for the (very longstanding) no-alias bug in LLVM.

https://www.reddit.com/r/rust/commen...anded_in_time/

The big question is: how long will it take this time until rust folks discover miscompilation -- or will they now be able to use it for real?

Thanks, reading the article I was thinking precisely about that.

[brad0]

Darn shame it didn't release in time to make it into either FreeBSD 13.0 or OpenBSD 6.9 but should be the compiler of choice for both FreeBSD 13.1 and OpenBSD 7.0 assuming both release in October/November which means they will of course miss release 13.0 of LLVM. Seems like the OSes that use LLVM and the LLVM project could be more coordinated in releasing their new releases near each others with enough time for integration.

Not really. That's just being overly dramatic. This nonsense of trying to align releases does not make sense.

[kassane]

This release is especially interesting as they landed the fix for the (very longstanding) no-alias bug in LLVM.

https://www.reddit.com/r/rust/commen...anded_in_time/

The big question is: how long will it take this time until rust folks discover miscompilation -- or will they now be able to use it for real?

True! That would be a big problem and hopefully the zig staff would keep a close eye on this so they don't make the same mistake.