Am So., 22. Aug. 2021 um 09:33 Uhr schrieb Travis Cardwell via Haskell-Cafe :

With talk of retiring the 8.10 branch, I think it might be worthwhile to remind folks that GHC 8.10(.6) is still the most recent release that is suitable for production use. GHC 9.0.1 is *not* suitable for production use due to major issues, and GHC 9.0.2 is not ready yet. [...]

Exactly, The 8.10 series is the *only* one with full tool support at the moment, so retiring this seems to be a bit premature. The only recent LTS versions on Stackage are for 8.10, 9.0 is only available via nightly, and there is no 9.2 at all. The Haskell Language doesn't have support for 9.2, either.OK, there are only alphas for 9.2 up to now, nevertheless some alphas for the corresponding tools would be nice, too. I'm totally aware of the fact that there are different people working on the different tools/compilers/libraries, but this doesn't matter at all as an "end user". The Haskell ecosystem needs much more coordination to offer a smoother user experience. Other language ecosystems are *much* more comfortable to install and use. Cheers, S.

Travis Cardwell via Haskell-Cafe writes:

Hi Zubin,

Thank you for message!

On Sat, Aug 21, 2021 at 6:58 PM Zubin Duggal wrote:

...

As the oldest currently active branch, it is currently a priority to retire the 8.10 branch of GHC from active development as soon as possible.

With talk of retiring the 8.10 branch, I think it might be worthwhile to remind folks that GHC 8.10(.6) is still the most recent release that is suitable for production use. GHC 9.0.1 is *not* suitable for production use due to major issues, and GHC 9.0.2 is not ready yet.

Indeed 9.0.2 is not yet ready. However, we are actively working towards remedying this. Likewise, 9.2.1 is very close to being released. To clarify: by "retiring 8.10", Zubin means that the 8.10 series is moving into a truly "critical-bugfix-only" mode. This contrasts with our treatment of 8.10 thusfar, where we have been both trying to fix general bugs *and* backport rudimentary Darwin/ARM support. In this sense, 8.10 has been a rather odd release; in nearly all other releases the "feature-set" of a compiler release series is fixed after the N.N.1 release is cut. To stabilize Darwin support, our work has been focussed on 8.10. At this point we believe that 8.10.7 should be a very usable release and therefore we will focus our further efforts on 9.0 and 9.2. If after 8.10.7 we find an issue that renders the compiler unusable for a large numbers of users, we will of course make another release. However, we also very much hope that this won't be necessary. Cheers, - Ben