|
|
@@ -59,25 +59,27 @@ Suite **204/204**; FIXPOINT OK (3,405,813 bytes).
|
|
|
`ProcLink` fails after scopes close.
|
|
|
3. Definition-module proc replay (done), nested globals (done).
|
|
|
|
|
|
-## Per-unit emission attempted — blocked by parse-time state
|
|
|
-
|
|
|
-Tried plan item 3 (Lower emits each unit at its END, with the legacy
|
|
|
-output suppressed via `SetNoEmit`). Small sessions and every `lower_ok`
|
|
|
-single-file test byte-match, **but the compiler's session hangs/differs**:
|
|
|
-
|
|
|
-- The grammar's inline `QbeGen` calls still run during parse; `noEmit`
|
|
|
- only suppresses the *output*, not the **state**. So `nTemp` (stuck at
|
|
|
- 211), `inBody` (stuck TRUE, so `main`'s header is dropped), `nLoc`,
|
|
|
- `funcDepth`, … are all corrupted before Lower's per-unit emit.
|
|
|
-- `EndModule` had to respect `noEmit` too, or it closed the session.
|
|
|
-
|
|
|
-Conclusion: per-unit Lower needs the grammar's parse-time `QbeGen`
|
|
|
-**state** neutralised — which is exactly the flip. Options:
|
|
|
-1. **Remove/replace the grammar's 533 `QbeGen.*` calls.** The blockers
|
|
|
- are the ~20 semantic checks that read `QbeGen` values (`IsImm`,
|
|
|
- `ConstInt`, `Fold`); they must be reworked to use `SymTab`/the AST.
|
|
|
-2. **Two `QbeGen` states** — but a second 4 MiB `sessBuf` pushes the
|
|
|
- compiler image past its own buffer, so this does not self-host.
|
|
|
-
|
|
|
-So the flip is a grammar rewrite of the emit actions, gated by the
|
|
|
-fixpoint. Reverted to the green session-re-emit state (`ce2fc67`).
|
|
|
+## The "hang" was a Lower bug, not a miscompile
|
|
|
+`LowerDecls` emitted a procedure's parameter loop even when `ProcInfo`
|
|
|
+failed, using an uninitialised `np` — so the loop ran ~forever inside
|
|
|
+`FindProc`. (`MaxProc = 64` is far too small for a whole session, so
|
|
|
+`ProcInfo` failed for most of the compiler's procs.) **Fixed**
|
|
|
+(`a9a8ecd`): the proc is skipped unless `ProcInfo` succeeds, and each
|
|
|
+`ParamInfo` is guarded. Also grew the AST arena/pool (`caeeeb5`) — the
|
|
|
+arena/pool data lines are compact `z` runs, so this is nearly free.
|
|
|
+
|
|
|
+## What remains for the session re-emit
|
|
|
+- **Correctness:** `MaxProc`/`MaxSym` must be ≥ the session's proc/var
|
|
|
+ counts (the compiler has hundreds of procs; 64 is too small), so
|
|
|
+ procs beyond the cap are currently skipped.
|
|
|
+- **Performance:** `FindSym`/`FindProc` are linear scans, and `Lower`
|
|
|
+ looks up per node — O(nodes × table size). A whole-session re-emit is
|
|
|
+ very slow. A hash index (or reusing `SymTab` for module-level names)
|
|
|
+ would fix it.
|
|
|
+- **Size-dependent behaviour:** with `MaxProc = 1024` the same run
|
|
|
+ **segfaults** (in `CopySpan` via `ParamInfo`) — a size/layout-sensitive
|
|
|
+ latent bug, the same family as the earlier `ASTDUMP` fragility.
|
|
|
+
|
|
|
+So the session re-emit is capacity-fixable but needs (a) a fast lookup
|
|
|
+and (b) the size-dependent bug isolated. The grammar rewrite remains the
|
|
|
+alternative flip path.
|
