Better handling for msvc env when compiling cpp extensions

[peterjc123]

Fixes #38861 (comment).

1. Error out if msvc env is activated but DISTUTILS_USE_SDK is not set.
2. Attempt to activate msvc env before running ninja build

[dr-ci]

💊 CI failures summary and remediations

As of commit b0670df (more details on the Dr. CI page):

---

• 1/1 failures possibly* introduced in this PR
  • 1/1 non-CircleCI failure(s)

---

ci.pytorch.org: 1 failed

• Failed: pr/py3.6-clang7-rocmdeb-ubuntu16.04

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group.

See how this bot performed.

This comment has been revised 4 times.

[ezyang]

@peterjc123 I know this stuff is hard to test, but I guarantee you it will bitrot if we don't exercise this somehow in Windows CI

[facebook-github-bot]

@ezyang is landing this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

[peterjc123]

@ezyang I'll apply to the same logic to tools/build_pytorch_libs.py so that it will be easier to find out what's wrong there.

[facebook-github-bot]

@ezyang merged this pull request in a40049f.

[gchanan]

@peterjc123 is this fixing a regression (so we should add it to 1.5.1)? It doesn't seem like it, but I wanted to double check.

[peterjc123]

@gchanan Using ninja to build c++ extensions is added in v1.5.0 and this pr fixes the problem that users attempt to trigger a extension build without activating the VC environment, which is an allowed behavior in pre-v1.5.0 builds. So maybe it can be considered as a regression to some extent.