The style I started with (putting each conformance on its own extension)
turned out to be pointlessly verbose in practice. It also separated the
protocol from the "implentation" in the sense that the property
declarations are what determine how the conformance works.
Factors the cursor_info requests into a single API to asynchronously
make the sourcekitd cursor_info request and wrap the result for easier
consumption from Swift (but not other processing).
Also, cleanup the now unused extensions on Hover that were superceded by
symbolInfo.
Clangd provides a symbolInfo request as an extension to LSP that is
equivalent to sourcekitd's cursor info. For now, just add the new types
and add plumb the request through so we can test it in isolation. In a
future commit this will be used to power jump-to-definition.
We are tied to using a SwiftPM that matches the toolchain, so upgrade
from 0.3.0 to .branch("master") and add a pins file to manage updating
that dependency.
This was driven by changes that broke loading packages being developed
with newer version of SwiftPM when using the 0.3.0 tag of libSwiftPM.
However, the changes seem to work when going in the other direction, so
using the newer libSwiftPM hasn't caused any known regressions for using
older toolchain versions (not guarantteed and not tested extensively
though).
This fixes using the latest (November 13) toolchain snapshot on macOS.
On Linux there are other issues not specific to LSP.