The underlying issue is that, with finite resources, it is hard to support multiple releases long term. We can't really support 9.6, 9.8, 9.10, and HEAD. So the blog post was trying to be transparent, by saying that we'll focus efforts on 9.6, 9.10, and HEAD. Is 9.6 the right choice? We think so but, as Ben asys, it would be great to have some more objective basis on which to assess uptake. In particular, it'd be great to have some telemetry. Telemetry is a complex issue -- but we now have some very well-informed colleagues: the team at Scarf. So maybe we'll get some basic data before long, which would be a big step forward. (We'd expect to be transparent about that too of course!) Simon On Tue, 28 May 2024 at 17:37, Ben Gamari wrote:

Clinton Mead writes:

...

Hi All

The recent GHC blog post: https://www.haskell.org/ghc/blog/20240521-ghc-release-priorities.html states:

*"9.6.5 seems to be a relatively stable release so far and we plan to prioritise fixes given the relatively higher adoption of this branch"*

I was wondering where this information came from? Is this based on a number of downloads or some other metric? Number of companies that have adopted the version? How many stack/ghcup pulls? Or just a vibe/what people seem to be talking about?

If there are hard numbers, are they public?

Sadly we don't have hard numbers. However, we do try to keep an eye on a few "softer" metrics:

* which releases tend to see the most new tickets * which releases tend to come up in online discussion * how much of the ecosystem has adapted to which releases

I will admit that I'm not entirely happy with the amount of judgement that goes into determining support windows currently. My tick-tock release cadence proposal [1] was an attempt at proposing a more predictable schedule, but sadly it never gained much support.

...

The current project I'm working on is still on GHC 9.2.2, so I was considering an upgrade to GHC 9.4.8, but this blog post has made me lean towards the 9.6 series instead. This is just an upgrade for stability not features and so we don't get too far behind. But it would be good to have this popularity info generally instead of having to wait for a blog post.

9.4 likely isn't the best upgrade target at this point. We will likely [status] not produce any further releases in this series and our attention has turned primarily to 9.10 and 9.6.

[status]: https://gitlab.haskell.org/ghc/ghc/-/wikis/GHC-status [tick-tock]: https://github.com/haskellfoundation/tech-proposals/pull/34 _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs