Keep track of which variables are section variables and use this info to fix bugs

[SkySkimmer]

Named declarations now contain an extra field status = SecVar | ProofVar. SecVar means it is a section variable.

To work this requires the invariant that if a variable is marked SecVar in an evar's context then that evar is always instantiated with that section variable.

Fix #18858
Fix #12304
Fix #11487
Fix #6773

Overlays (.v):

• Adapt to rocq-prover/rocq#21987 (clear checks that ids are bound even when from match goal) AbsInt/CompCert#582
• Adapt to rocq-prover/rocq#21987 (clear checks ids are bound even when from match goal) charguer/tlc#23
• Adapt to rocq-prover/rocq#21987 (destruct clears non section variable even when reusing a section variable's name) HoTT/Coq-HoTT#2363

Overlays (.ml):

• Adapt to rocq-prover/rocq#21987 (secvar status) equations#719
• Adapt to rocq-prover/rocq#21987 (secvar status) LPCIC/coq-elpi#1026

[SkySkimmer]

hott failure is because they workaround #18858 by explcitly clearing

[SkySkimmer]

@coqbot ci minimize ci-stdlib

[silene]

Section variables have almost nothing in common with variable do they?

• clear maybe although should probably be done via using

Note that there are proofs where you want to do using H (because you need H in the proof) but also clear H (because you want to temporarily hide H to prevent some disaster, e.g., clear H ; nia). So, their roles are somewhat orthogonal. The case where they overlap is when the call to clear is the very first step of a proof (in which case using is a much better choice, indeed).

[SkySkimmer]

@coqbot bench

[SkySkimmer]

bug_4095 is failing because we push an already used name in

rocq/pretyping/evarconv.ml

Line 1821 in f018a77

let env_rhs' = push_named (NamedDecl.LocalAssum (id,idty)) env_rhs in

IDK WTF is going on here