…em release"

**What does this PR do?**

This reverts commit 551b5ae from #20.

See #20 for details. In that PR I mentioned:

So to break the loop, let's:

1. omit this test temporarily: ✅ done -> this was #20
2. then we can release this gem: ✅ done ->
   https://github.com/DataDog/datadog-ruby_core_source/actions/runs/20059401843,
   [version 3.4.2 is on rubygems.org](https://rubygems.org/gems/datadog-ruby_core_source/versions/3.4.2)
3. then we can undo the dd-trace-rb pinning: ✅ done ->
   DataDog/dd-trace-rb#5122
4. then we can re-enable the test

This PR is step 4 and gets us back to where we were before the
workaround.

**Motivation:**

Remove workaround now that it's not needed anymore!

**Additional Notes:**

N/A

**How to test the change?**

Validate that Ruby 4.0 + "corruption" test is running in CI and green.

[PROF-13115] Revert "Temporarily skip broken Ruby 4 test to unblock gem release"

This tiny change came in handy for testing, so "why not"?

The 4.0.0-preview2 headers are not compatible with -preview3 --
in particular there's a new field in the thread structure that changes
the offset of everything that comes after it, and thus makes the
profiler crash on preview3 when compiled with preview2 headers.

I followed the usual instructions from the `README.md` to extract
the headers.

The changelog is easy to forget, and anyway changes in this gem/repo
are easy to identify by looking at PRs/commits so this doesn't offer
much value IMHO.

…eaders

[PROF-13303] Add Ruby 4.0.0-preview3 headers and bump gem version to 3.4.3