revert: #506 bounds-check elimination (main CI red since merge)

[cs01]

Main CI has been failing since #506 merged at 22:40 UTC. Stage1 self-hosting test semantic/binary-type-in-function fails because checkBinaryTypesDeep stops catching "hello" - 5 inside function bodies when chad is self-hosted. Stage0 (node-driven) is unaffected.

Local verification: building stage1 chad from this revert branch and running it directly against the fixture correctly emits cannot use '-' between 'string' and 'number' and exits 1.

Reverting to unblock all work on main. The perf work from #506 (bounds-check elimination) is a real improvement and should be re-applied, but it needs a proper self-hosting-safe implementation and a CI gate that waits for stage1 to finish before declaring green.

Followups needed (separate PRs)

1. Re-implement bounds-check elimination with full stage0→stage1→stage2 self-hosting verification before merge. The root cause of the miscompilation isn't yet understood — needs deeper investigation into why stage1 chad's checkBinaryTypesDeep silently drops its binary-op type check when compiled through perf: bounds-check elimination in loop-indexed array access #506's codegen path.

2. Fix the pre-push auto-merge hook (~/.claude/hooks/pre-push-rebase.sh). It currently evaluates all_passed on the PR's statusCheckRollup at push time and triggers auto-merge if everything currently visible is green. GitHub populates the rollup incrementally as workflows start, so PRs with slow build-linux-glibc / build-macos jobs get merged after the fast checks (format-check, build, benchmarks) report success but before the slow ones have started. The hook should verify all expected jobs from the CI workflow file are COMPLETED before merging.

Diff

8 files changed, 6 insertions(+), 557 deletions(-)

[github-actions]

Benchmark Results (Linux x86-64)

Benchmark	C	ChadScript	Go	Node	Place
Binary Trees	1.345s	1.253s	2.723s	1.159s	🥈
Cold Start	0.9ms	0.8ms	1.1ms	27.5ms	🥇
Fibonacci	0.814s	0.763s	1.559s	3.166s	🥇
File I/O	0.117s	0.090s	0.084s	0.198s	🥈
JSON Parse/Stringify	0.004s	0.005s	0.017s	0.015s	🥈
Matrix Multiply	0.444s	0.652s	0.611s	0.364s	#4
Monte Carlo Pi	0.389s	0.410s	0.405s	2.249s	🥉
N-Body Simulation	1.667s	2.117s	2.202s	2.387s	🥈
Quicksort	0.215s	0.246s	0.213s	0.262s	🥉
SQLite	0.347s	0.368s	—	0.441s	🥈
Sieve of Eratosthenes	0.015s	0.026s	0.020s	0.038s	🥉
String Manipulation	0.008s	0.018s	0.016s	0.036s	🥉

CLI Tool Benchmarks

Benchmark	ChadScript	grep	node	xxd	Place
Hex Dump	0.426s	—	1.037s	0.135s	🥈
Recursive Grep	0.019s	0.010s	0.097s	—	🥈