Closed cartermp closed 6 years ago
Things like this seem awfully unnecessary if the SDK were a part of the VisualFSharp repo and built with our tasks and targets: https://github.com/fsprojects/FSharp.TypeProviders.SDK/blob/master/fsc.props
I think that file is now redundant now that .NET SDK F# compilers support type providers, e.g. see the comment at the top: https://github.com/fsprojects/FSharp.TypeProviders.SDK/blob/master/fsc.props#L3
It's plausible we could/should move the TPSDK into "visualfsharp" but not for that particular reason. However "visualfsharp" is already too big as a repo (especially the inclusion of vsintegration and VS setup) and still has a strong bias towards Visual Studio/Windows engineering issues. Those are very important and it's crucial to keep focus on those. At the moment I am loathe to increase the scope of what gets engineered there - we would trade one kind of efficiency for another.
Makes sense.
Things like this seem awfully unnecessary if the SDK were a part of the VisualFSharp repo and built with our tasks and targets: https://github.com/fsprojects/FSharp.TypeProviders.SDK/blob/master/fsc.props
Would it be preferable for this to be brought into VisualFSharp? We already build a type provider that we run through CI with the Type Provider SDK.