merge: #3579's fused MoE chain under the hot-path contracts

Two conflicts in dasllama_common.das, both resolved to this branch:

  kv_ensure      — the growth loop is split into a [cold_path]
                   kv_grow_block_table here; master still had it inline.
                   Both sides had already moved to long_length, so the 64-bit
                   sweep merged silently — we ran the same fix independently.
  scratch_resize — the @scratch parameter. #3579 arrived with a new comment on
                   the same function reading "grow-only no-init: contents are
                   stage output (fully overwritten)", which is independently
                   the exact property @scratch declares.

Their new moe_experts_chain needed one annotation: the fourth
team_parallel_stages lambda capture, same dispatch-API cost as the three
already marked. Everything else in it passes untouched, because its only
sizing op is s.chain_stages, already @scratch on this branch.

Verified on the merged tree with a daslang rebuilt for #3579's lattice
operators (ubyte16 >> int needs its C++): engine compiles clean, 65 changed
files lint clean, ENVIRONMENT.md shows no drift, PERF fixtures hold, and
test_fused_decode is 13/13 under -jit with the gpt-oss arm skipped for want of
the gguf.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Author: borisbat

modules/dasLLAMA/dasllama/dasllama_common.das

@@ -8363,7 +8363,7 @@ def private moe_experts_chain(t : Model; var s : Session; l : int64) {
         let fbp = biased ? addr<float const?>(t.fblob[0]) : null
         let idxp = addr<int64 const?>(s.moe_idx[0])
         let wvp = addr<float const?>(s.moe_w[0])
-        team_parallel_stages(s.chain_stages) <| @(stage : int; jb : int; je : int) {
+        team_parallel_stages(s.chain_stages) <| @(stage : int; jb : int; je : int) {   // nolint:PERF026 — dispatch-API lambda capture, see attention_chain_decode
             unsafe {
                 if (stage == 0) {
                     // gate+up rows of expert j, then bias -> act -> requant on exactly these rows