Shift `clang-tidy` to github action

[phlax]

Currently clang-tidy is not working very well and need some attention

I am gradually picking off various bits of ci for shifting azp -> gh

This one seems like quite a good candidate to start on the checks with.

[phlax]

cc @wbpcode @yanavlasov

[phlax]

clang-tidy is currently disabled pending resolution here

[github-actions]

This issue has been automatically marked as stale because it has not had activity in the last 30 days. It will be closed in the next 7 days unless it is tagged "help wanted" or "no stalebot" or other activity occurs. Thank you for your contributions.

[mathetake]

can we prioritize this? I genuinely think this is really bad for the long term as it directly affects the code quality plus security risk

[phlax]

its half complete - atm you can run:

./ci/do_ci.sh clang-tidy

and it kinda works

the reason i didnt land in ci is that it is not 100% - i took someone elses (aspect-based) clang-tidy implementation and hacked a few things to make it work to my expectations

i havent had too much time to do it recently - but in the past i ran it and fixed any issues - probably that would be a good place to start - the tool could do with some non-me testing and feedback

[mathetake]

so yeah one thing I noticed is

xternal/com_google_absl/absl/log/absl_check.h:47:3: note: expanded from macro 'ABSL_DCHECK'
  ABSL_LOG_INTERNAL_DCHECK_IMPL((condition), #condition)
  ^
external/com_google_absl/absl/log/internal/check_impl.h:40:3: note: expanded from macro 'ABSL_LOG_INTERNAL_DCHECK_IMPL'
  ABSL_LOG_INTERNAL_CHECK_IMPL(condition, condition_text)
  ^
external/com_google_absl/absl/log/internal/check_impl.h:26:3: note: expanded from macro 'ABSL_LOG_INTERNAL_CHECK_IMPL'
  ABSL_LOG_INTERNAL_CONDITION_FATAL(STATELESS,                        \
  ^
external/com_google_absl/absl/log/internal/conditions.h:123:3: note: expanded from macro 'ABSL_LOG_INTERNAL_CONDITION_FATAL'
  ABSL_LOG_INTERNAL_##type##_CONDITION(                                    \

can we exclude external?

[mathetake]

I think that would save tons of CPU times in CI