docs: refresh Blaise phase map vs 11-phase plan + V3 comparison

docs/blaise-phases-map.md

@@ -1,77 +1,84 @@
-# Blaise phase sync map (checked 2026-09-20 against graemeg/blaise master)
+# Blaise phase sync map (refreshed 2026-09-23 vs graemeg/blaise master)
 
-Blaise's 10 phases (README status: 1–5 + 7 complete, 6 ongoing,
-8–10 planned). This file tracks where m2compiler-V3 stands against
-each, and records deliberate divergences so they aren't
-rediscovered. Refresh after every major step.
+Blaise's plan is now **11 phases**; it is fully self-hosting (byte-
+for-byte fixpoint, FPC no longer required after stage 1), ships two
+backends (native x86-64 default, QBE opt-in), two targets (Linux +
+FreeBSD x86_64, bidirectional cross-compile via internal assembler/
+linker), and 4977+ tests. This file compares m2compiler-V3.
 
 ## Phase table
 
-| Blaise | Status there | Us | Status here |
+| Blaise | Blaise status | V3 | V3 status |
 |---|---|---|---|
-| 1 Bootstrap pipeline | ✅ | Step 1 (M2→QBE hello) | ✅ done (`v3-step1`) |
-| 2 Type system (classes, records, ARC, exceptions) | ✅ | Steps 2–4.3 (scalars ✅, composites ✅, calls/frames/recursion ✅, nested + display ✅, separate compilation ✅) | ✅ complete modulo deliberate outs |
-| 3 Generics + interfaces | ✅ | Explicitly OUT (R10 blueprints rejected) | ❌ deliberate |
-| 4 OPDF debug info | ✅ | Deferred (listing only) | ❌ deferred |
-| 5 Self-hosting | ✅ | Step 8 (last) | ⚠️ same mechanism, later slot |
-| 6 RTL/StdLib expansion | 🔄 ongoing | Step 5 (Benjamin stdlib) | Not started |
-| 7 Native backend | ✅ | Not planned (QBE is terminal) | ❌ deliberate |
-| 8 Windows/macOS ARM64 | Planned | Out of scope (amd64_sysv) | ❌ |
-| 9 LSP + VS Code | Planned | Deferred vaguely | ❌ |
-| 10 Migration analyser | Planned | 1.12 TopSpeed grammar, 662-file corpus parse-clean | ✅ convergent |
+| 1 Bootstrap pipeline | ✅ | step 1 | ✅ |
+| 2 Type system (classes, records, ARC, exceptions) | ✅ | steps 2–4.3 | ✅ modulo deliberate outs (no ARC, no exceptions) |
+| 3 Generics + zero-GUID interfaces | ✅ | — | ❌ deliberate (R10 blueprints rejected) |
+| 4 OPDF debug info | ✅ | — | ❌ deferred (listing only) |
+| 5 Self-hosting | ✅ | step 8 | 🔄 in progress (V3 cannot yet compile its own sources; gaps below) |
+| 6 Language + expand RTL & StdLib | 🔄 ongoing | step 5 | 🄾 partial (`stdlib/`: SysIO, TextIO, Strings, WholeIO, Math, Files, Storage; `runtime/syslib`: SysShim, FileIO) |
+| 7 Native backend + internal asm/linker | ✅ | — | ❌ deliberate (QBE is terminal) |
+| 8 FreeBSD x86_64 target | ✅ | — | ❌ out of scope (Linux/amd64_sysv) |
+| 9 Windows + macOS ARM64 | ⏳ planned | — | ❌ |
+| 10 LSP + VS Code extension | ⏳ planned | — | ❌ |
+| 11 Migration analyser (FPC/Delphi) | ⏳ planned | 1.12 | ~ convergent (TopSpeed-V3 grammar, 662-file corpus parse-clean) |
 
-## What "in sync" means here
+## Bottom line
+
+V3 corresponds to **Blaise phases 1–2 done**, **phase 5 (self-hosting)
+in progress**, **phase 6 (RTL/StdLib) partial**; phases 3, 4, 7–10 are
+**deliberate divergences or deferrals**; phase 11 has a convergent
+sidecar. In raw maturity Blaise is far ahead (4977+ tests, self-
+hosting, two backends/targets); V3 is 103 tests and has not yet
+closed the self-hosting loop.
 
-Process, not phases: phased steps with per-step design docs (their
-`design.adoc` rhythm → our `summary_stepN.md`), test-guarded
-progress, QBE-first backend for speed-to-hosting (their stated
-reason, ours too), fixpoint verification (their 3-stage
-byte-identical bootstrap → our planned `bootstrap/fixpoint.sh`
-over `.ssa`).
+## What "in sync" means here
 
-Two places we're ahead of their learning curve, per graemeg's own
-v0.12.0 retrospective: our harness has been end-to-end
-(`qbe→cc→run`) since step 1 — exactly the gap (IR-only tests
-missing link/runtime bugs) he's now closing. Keep it that way:
-no step lands on IR inspection alone.
+Process, not phases: phased landings with per-step design docs (their
+`docs/design.adoc` → our `docs/summary_stepN.md`), test-guarded
+progress, QBE-first (their historical reason, ours by choice), and
+fixpoint verification (their 3-stage byte-identical bootstrap → our
+planned `bootstrap/fixpoint.sh` over `.ssa`).
 
 ## Warnings adopted from Blaise experience
 
-- **Rolling bootstrap.** Their `scripts/rolling-bootstrap.sh` exists
-  because new parser features break old binaries: once M2 compiles
-  itself (step 8), a grammar extension can strand the stage-1
-  binary. Our `fixpoint.sh` must rebuild commit-by-commit from the
-  last green binary, not just diff three stages. (Recorded pre-emptively;
-  hits at step 8.)
-- **Stabilisation cycles.** Their v0.12.0 goes thin-on-features to
-  harden the suite. After step 3.4 (composites complete), do the
-  same before step 4 calls: a hardening pass over the 50+ suite
-  (fuzz bounds, cross-span copies, trap paths) instead of new syntax.
-- **Turnkey gap.** Self-hosting ≠ turnkey: they still need GCC/FPC
-  scaffolding. Our step 8 must define "done" as: pinned `qbe`+`cc`
-  plus documented host-gm2 bootstrap, nothing else.
+- **Rolling bootstrap** (`scripts/rolling-bootstrap.sh`): new parser
+  features strand old binaries. Our `fixpoint.sh` must rebuild
+  commit-by-commit from the last green binary, not just diff stages.
+- **Stabilisation cycles** (their v0.12.0): go thin-on-features and
+  harden the suite periodically. We did this once (step 3.7); repeat
+  before the fixpoint.
+- **End-to-end tests, not IR-only**: our harness has run
+  `qbe → cc → run` since step 1 — exactly the gap graemeg closed in
+  v0.12.0. Keep it.
+- **Turnkey gap**: self-hosting ≠ turnkey. V3's "done" = pinned
+  `qbe`+`cc` plus a documented host-gm2 bootstrap, nothing else.
 
 ## Deliberate divergences (do not relitigate without cause)
 
-- No ARC/GC, no exceptions, no generics, no native backend, no
-  second platform: Modula-2 scope discipline, not Pascal parity.
-- Self-hosting last, not middle: full-language-first was locked;
-  the cost is accepted (hygiene rules from step 1 pay it down).
+- No ARC/GC, no exceptions, no generics, no native backend, no second
+  platform, no LSP: Modula-2 scope discipline, not Pascal parity.
+- Self-hosting last, not middle: full-language-first was locked.
 - Two grammars, not one: modern `M2.atg` + legacy
-  `TopSpeed-V3-M2.atg` (1.12). This mirrors their phase-10
-  split (modern core + migration sidecar) and quarantines the
-  opposite dialect decisions (`_`, `<>`, `H`/`C`/`B` suffixes).
-  Never merge them; the modern grammar stays Redux-clean.
+  `TopSpeed-V3-M2.atg` (1.12) — mirrors their phase-11 split and
+  quarantines the opposite dialect decisions.
+
+## Remaining to reach V3's phase 5
+
+1. Named-CONST constant folding (biggest unblocker) and type-operand
+   `TSIZE`/`SIZE` so V3 can compile `SymTab`/`QbeGen`.
+2. Cross-stage `FileIO` selection (stage 1 host library, stage 2
+   `runtime/syslib/FileIO`).
+3. `bootstrap/fixpoint.sh` (Stage1 gm2 → Stage2 V3 → Stage3 V3,
+   byte-identical `.ssa`, suite green on all stages).
 
 ## Drift log
 
-- 2026-09-20: map created. `plan.md` refreshed (0x-only,
-  no-AST-module reality, two-grammar strategy, step-3 substeps).
-- 2026-09-22: 3.5 showcase8, 3.6 pointers, 3.7 hardening (68/68);
-  Phase-2 row → near-complete (calls remaining = step 4).
-- 2026-09-22: 4.1 call emission (79/79); Phase-2 row →
-  complete modulo deliberate outs (ARC/exceptions/generics).
-- 2026-09-22: 4.2 nested procedures + static-link display (81/81).
-- 2026-09-22: 4.3 separate compilation, single-image session (87/87).
-  Step 4 complete; Phase-2 row closed.
-- Next refresh: start of step 5 (Benjamin stdlib).
+- 2026-09-20: map created.
+- 2026-09-22: steps 3.5–4.3 (composites, calls, nesting, separate
+  compilation) — Phase-2 row closed.
+- 2026-09-22: steps 5.1–5.7 (FFI, stdlib, module-qualified symbols,
+  ADDRESS/Storage) — phase 6 partial.
+- 2026-09-22/23: steps 6.1–6.2 (built-ins, 64-bit longs), 7–7.1
+  (syslib port analysis + self-hosted `SysShim`/`FileIO`), 8.1
+  (Redux sweep + self-compile gap analysis) — phase 5 in progress.
+- 2026-09-23: refreshed against Blaise's 11-phase plan (was 10).