Evaluate getting rid of gulp and npm

[chucker]

Best as I can see, we currently use the gulp (and npm) infrastructure for two reasons:

• it compiles SCSS to CSS, does some further jiggling, and copies the result to the two target projects
• it copies the JS* to the two target projects

*) it does not compile TS to JS: the MSBuild toolchain does that instead.

This works, but it's kind of an ugly and brittle fit. We might want to move to things like BuildWebCompiler.

[SQL-MisterMagoo]

We used to use (Build)WebCompiler but it has issues with @import when it is combining files (it doesn't move them to the top of the file) - and it is no longer maintained I believe. I did fork it and fix that to see how easy it would be (really very) but never submitted a PR because it's not being maintained - and I don't want to maintain it either!

[SQL-MisterMagoo]

It might be worth revisiting though because the build time is terrible with gulp

[chucker]

We used to use (Build)WebCompiler but [..] it is no longer maintained I believe.

Yeah, seems that way. I'm kind of hoping that Blazor will lead to a renaissance. Importing the npm ecosystem seems kind of ridiculous in this context.

the build time is terrible with gulp

I may have made it a lot slower with the change to npx (which first runs npm to check if the version of gulp is current before executing it, but does come with the advantage that you don't have to manually install anything).

[chucker]

https://github.com/excubo-ag/WebCompiler

This project is based on madskristensen/WebCompiler. However, the dependency to node and the node modules have been removed, to facilitate a pure dotnet core implementation.