Just realized we could consider this, but this is mainly a call for opinions. I can see some pros/cons.
#Should we .gitignore dagger.gen.go?
outer kraken
What triggered this: I got tired of seeing results from dagger.gen.go show up in my editor's fuzzy search. fzf will automatically ignore .gitignored files, but right now the pattern is not to ignore them, and commit them to the repo.
Typically I've erred on the side of committing all generated files, mostly as a courtesy; not doing so makes it frustrating for others to build your repo.
But one difference here is that the entrypoint for this code is always going to be Dagger.
manic stirrup
if the tooling gets good / smooth enough, we could keep those files out of your repo altogether
outer kraken
Yeah, that's what I'm thinking
manic stirrup
plus a dagger mod eject for escape hatch
outer kraken
What would eject do? Clear your local tree of generated files?
manic stirrup
I may be using the term wrong. it’s a react term andrea explained to me
No Lock-In
Under the hood, we use webpack, Babel, ESLint, and other amazing projects to power your app. If you ever want an advanced configuration, you can ”eject” from Create React App and edit their config files directly.
not exactly the same thing (not codegen) but pretty damn close
bonus: would make our module repos look better
outer kraken
yeah this idea is growing on me quickly
theoretically you'd never even have to run sync
oh, i guess you'd want it so your editor knows wtf you're talking about
manic stirrup
That is a major downside - code that doesn’t build in the usual way - so IDE needs a plugin to be 100% ok with it
That downside could perhaps become an upside though, if said ide plugin can be excellent
now you have a happy IDE and a dagger plugin which can do many other things 😁
I like it too
outer kraken
the return of the Dagger LSP?