Generic migrations

[jmatsushita]

In issue #30 @kosmikus you mention:

The unreleased work I'm doing on datatype-generic migrations using generics-sop is also depending on these features.

I just saw your talk on generic migrations (slides) and would love to be able to play with your code.

Any plans to make it available?

Slides

UPDATE: Added links to Youtube talk and Slides

[kosmikus]

I'm very sorry. I did unfortunately never complete this work, and did not need it urgently. I still have the half-finished code somewhere, but I have to actively look for it.

[jmatsushita]

Hi @kosmikus just a friendly bump that I've been watching your talk again last week and have been hoping for some code. In case it's helpful to find it (and if it has not vanished in the Void as code sometimes does) you mentioned in your talk that it was depending on a branch of generic-sop which included metadata about constructors.

[jmatsushita]

I was reminded of this thread after seeing on the readme of @stevana's https://github.com/stevana/hot-swapping-state-machines#contributing and @turion' s question on reddit about "changes of the state type"

Automatic state migration? C.f. essence-of-live-coding;

I think generic migrations could be a great framework for automatic or semi-automatic state migration!

I thought I'd share the link to the talk that motivated this issue https://www.youtube.com/watch?v=foF7Xx1K--U and the slides https://www.andres-loeh.de/migrations-sop-warsaw.pdf

[turion]

I think @kosmikus and me talked about this subject when I presented essence-of-live-coding at the Regensburg Haskell Meetup. In eolc's particular case, the old types are not available anymore, only the data. So one cannot use this approach (short of manually writing down the version history of the types, which is infeasible there).

But in cases where the state is given explicitly, this would be great!