Timing for new elm-lang/core

[Evan Czaplicki]

There are some small improvements in core that it'd be cool to release.

The trouble is that it will be a major version bump, but nothing is really changing. It's just that the 0.15.1 release found a lot of sneaky stuff in the documentation and exports!

Does it make sense to go to 3.0.0 now to get the nicer docs and such? Should we wait til there's real stuff changing?

[Texas Toland]

I've personally been anticipating your effects/components work. It sounds like your progress as well as Pete and Joey's is nearing readiness. Why not wait? Or release an alpha tag.

--
You received this message because you are subscribed to the Google Groups "Elm Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elm-discuss...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jeff Smits]

I'd say just release. New people don't care, they care more about the 0.15.1 version number and just get the core that elm-package provides. Package authors can see nothing really changed and update. Why wait?

--

[Evan Czaplicki]

Texas, that work is independent of core. My stuff is a package, Pete's needs a platform release.

Jeff, yeah, that makes sense. I just didn't want to cause the ripple through the packages of folks needing to widen their version bounds, but that's probably a smaller deal than bad docs though. I'll do the last little bits of clean up needed there and maybe put it out tomorrow if there are no big concerns that come up.

[Texas Toland]

that work is independent of core. 

Apology I misread. Agreed unless there are more changes queued up.

[Evan Czaplicki]

Texas, no worries, lots of stuff that's not communicated in a clear way :)

The general goal is to try to get core to be relatively minimal. I suspect there will be many approaches to rendering and components. Already we have Graphics, elm-webgl, and elm-html. I imagine there will be others, so I already regret that Graphics is built into core and maybe want to move it out someday. Point is, I do as much outside of core as possible so the release cycles can be decoupled. This is one of the benefits of the *-extra stuff as well: add as much as you want as fast as you want.

On Wed, Jul 29, 2015 at 5:29 PM, Texas Toland <te...@appship.it> wrote:

that work is independent of core. 

Apology I misread. Agreed unless there are more changes queued up.

[Janis Voigtländer]

Maybe first address the part of https://github.com/elm-lang/core/issues/288 that relates to Set? Uploading the current master branch of core will lead to quite unexpected outcomes (and different ones than before, which were also strange OTOH) if people apply toString (or show) on a Set-typed thing.

Also, I am wondering whether it’s worth to address https://github.com/elm-lang/core/issues/283 on this occasion as well. It is yet another case like the other “type exposure” issues fixed with this release. So it would force another major version bump “despite no true changes internally” later on if addressed then.

(Finally, speaking of doing this release for the sake of improved docs, maybe merge https://github.com/elm-lang/core/pull/237 as well?)

​

--