feat: split Lean.Kernel.Environment from Lean.Environment

[Kha]

This PR clarifies Lean's Trusted Code Base (TCB) by splitting the data used by the kernel from the data used by general elaboration and code generation, and in the future asynchronously elaborated declarations.

• kernel diagnostics are moved from an environment extension to a direct environment as they are the only extension used by the kernel
• initQuot is moved from an environment header field to a direct environment as it is the only header field used by the kernel; this also makes the remaining header immutable after import
• the status of native reduction as a TCB extension is made explicit in the code by moving it into an optional callback the kernel uses to call into the richer environment with access to the interpreter. This would also allow us to correctly forward Options to the interpreter such as preferNative (not implemented yet).

[tobiasgrosser]

Hi @Kha, this sounds exciting. I don't understand the appreciation (TCB); spelling it out may help others here. Most importantly, let me not distract you from your mission towards more parallelism in the lean compilation. It is exciting to see progress here.

[tobiasgrosser]

I got it. TCB = Trusted Code Base

[Kha]

!bench

[leanprover-bot]

Here are the benchmark results for commit 4ba7bdc.
There were significant changes against commit c779f3a:

  Benchmark             Metric          Change
  =====================================================
- big_omega.lean MT     task-clock       13.7% (12.0 σ)
- big_omega.lean MT     wall-clock       13.7% (12.4 σ)
- bv_decide_mul         branch-misses     2.0% (12.3 σ)
- bv_decide_realworld   branch-misses     1.2% (31.5 σ)
- rbmap_fbip            task-clock        4.1% (18.2 σ)
- rbmap_fbip            wall-clock        4.1% (18.2 σ)

[leodemoura]

@Kha This PR has update stage0 commits. So, squash should not be used, and the commit messages should be fixed.

[leanprover-community-bot]

Mathlib CI status (docs):

• 💥 Mathlib branch lean-pr-testing-5145 build failed against this PR. (2024-11-19 16:36:31) View Log
• 💥 Mathlib branch lean-pr-testing-5145 build failed against this PR. (2024-11-19 16:58:36) View Log
• 💥 Mathlib branch lean-pr-testing-5145 build failed against this PR. (2024-11-20 09:51:38) View Log
• 💥 Mathlib branch lean-pr-testing-5145 build failed against this PR. (2024-11-20 11:09:59) View Log
• 💥 Mathlib branch lean-pr-testing-5145 build failed against this PR. (2024-11-20 12:12:46) View Log
• ✅ Mathlib branch lean-pr-testing-5145 has successfully built against this PR. (2024-11-20 14:31:01) View Log
• ✅ Mathlib branch lean-pr-testing-5145 has successfully built against this PR. (2024-11-20 17:35:18) View Log
• ✅ Mathlib branch lean-pr-testing-5145 has successfully built against this PR. (2024-11-25 17:49:26) View Log

[leanprover-bot]

Here are the benchmark results for commit e0353ed.
There were significant changes against commit 4600bb1:

  Benchmark       Metric          Change
  ===============================================
- bv_decide_mul   branch-misses     1.0% (13.8 σ)