Revert "build: add an initial CMake based build system for DocC (#818)"

[d-ronnqvist]

This reverts commit 7b74f5c.

This is failing builds and causing noise. It was my understanding that this wouldn't be blocking any work. If that's not correct I feel that it's premature to have it.

[d-ronnqvist]

I started a cross-repo test with this change here swiftlang/swift#84510 (comment)

[d-ronnqvist]

@swift-ci please test

[compnerd]

Reverting this is not really the right thing. CI is not something that can be setup outside of Apple - I cannot enable the GHA CI (and my PR has been pending review for a while now).

[etcwilde]

The CMake build was broken in #1331 because folders were moved around without moving the associated CMakeLists files. Pulling the entire build that Windows uses because you broke it seems pretty heavy-handed.

[d-ronnqvist]

The CMake build was broken in #1331 because folders were moved around without moving the associated CMakeLists files. Pulling the entire build that Windows uses because you broke it seems pretty heavy-handed.

The expectation for the approval of #818 was that the CMake build wouldn't be blocking or breaking anything.

I requested a CI setup as blocking feedback on #818 but allowed it to be approved/merged without it on the expectation that we wouldn't be responsible for it in the interim and that it wouldn't cause any breakages. Because those expectations turned out to not be true, I consider the approval of #818 to be invalidated.

[d-ronnqvist]

In my opinion, if we keep these files in the repo then there can't be anything else that depends on them until after there is a CI setup that verifies that they're correct before a PR is merged. In other words; I expect the order of events to be:

1. add the CMake files to the repo
2. set up a CI that verifies the CMake files' correctness
3. use the CMake files elsewhere

Right now it seems that 2 and 3 happened in the wrong order.