[clang] [llvm] [clang][ci] Move libc++ testing into the main PR pipeline (PR #93318)

[Mehdi Amini via llvm-commits]

================
@@ -231,6 +245,10 @@ linux_projects_to_test=$(exclude-linux $(compute-projects-to-test ${modified_pro
 linux_check_targets=$(check-targets ${linux_projects_to_test} | sort | uniq)
 linux_projects=$(add-dependencies ${linux_projects_to_test} | sort | uniq)
 
+linux_runtimes_to_test=$(compute-runtimes-to-test ${linux_projects_to_test})
----------------
joker-eph wrote:

> I believe Flang was tested on Clang changes to make sure Clang didn't make any incompatible changes to the driver, which Flang reuses. This is a different use case from testing Clang itself,

I see exactly the same kind of transitive dependency issue: one project depends on another and running **all** the possible tests for the dependent project on a change to the upstream project is deemed "not worth it".


> Additionally, "don't run those Clang tests if Clang is tested because of an LLVM change" kind of logic make it harder to reason about CI. 

Referring to "clang CI" is a rather vague statement to me: testing in generally goes into "tiers" and the "premerge" tiers is about tradeoff in terms of value/cost, while you're making a completeness arguments. There are tons of things that we don't test in pre-merge.


Also [the buildbot](https://lab.llvm.org/buildbot/#/builders/153) does not seem to cover this, which is quite annoying because now pre-merge is broken (see https://buildkite.com/llvm-project/github-pull-requests/builds/104995#019239b1-1aaa-411a-9fd2-9dcfdce25c3d ) by libcxx while the post merge is green, which should only ever happen if the PR is responsible for the breakage.


https://github.com/llvm/llvm-project/pull/93318