Add CI pipeline step and script that validates all files in the repo contain ASCII-only bytes and no UTF-8 BOM

[ahsonkhan]

The goal is to detect issues like this and remain green by making it a part of CI:
#4130

This is similar to what we do in Embedded C:
https://github.com/Azure/azure-sdk-for-c/blob/cd62e2b78a396c3d5478d3b60c046adb72687274/eng/pipelines/templates/jobs/archetype-sdk-client.yml#L369-L390

[ahsonkhan]

/azp run cpp - core - ci

[azure-pipelines]

Azure Pipelines failed to run 1 pipeline(s).

[ahsonkhan]

/azp run cpp - core - ci

[azure-pipelines]

Azure Pipelines failed to run 1 pipeline(s).

[antkmsft]

LibcurlTransportAdapter.md has the ’ character. If this catches it, we're good.

[antkmsft]

It did catch ’ in the LibcurlTransportAdapter.md, great!

As for the errors in unit tests like test/nlohmann-json-test/src/unit-inspection.cpp:301: CHECK(json("❤️").dump() == "\"❤️\"");, if it was in our tests, I'd say we should rewrite string literals using escape characters (for example, "❤️" can be written as "\u2764\uFE0F" or as "\xE2\x9D\xA4\xEF\xB8\x8F" - they all produce the very same string with exactly the same character: strcmp("\u2764\uFE0F", "\xE2\x9D\xA4\xEF\xB8\x8F") == 0 && strcmp("❤️", "\u2764\uFE0F") == 0).

But given that it is in 3rd party code, CI probably should be ignoring errors in these files.

[ahsonkhan]

Yep, that's the plan. Will filter out test recordings and nlohmann json. The main focus is product/SDK code.

[antkmsft]

@ahsonkhan, if this PR is close to achieving its goal, can we get to merge it soon?