ww: lift introspection files to lib/ww/ (ast, lex, tok, parse, typ, sym)

Makefile

@@ -101,39 +101,42 @@ $(OBJ)/w6l/%.o: cmd/w6l/%.c cmd/w6l/l.h | $(OBJ)/w6l
 	$(CC) $(CFLAGS) -Icmd/w6l -c -o $@ $<
 
 # ---- ww-side wwdump (the lexer port, exercised by 990_selfhost) --------
-# Built via the user-facing ww driver, with -I selfhost/cmd/wcc so it
-# can find the lex/tok/mem ports. Output named wwdump_ww to avoid
-# colliding with the C-side wwdump in $(BIN).
+# Built via the user-facing ww driver. The introspection ports (lex,
+# tok, ast, parse, typ, sym) live in lib/ww/; the compiler-only bits
+# (mem, check, cgen*) stay in selfhost/cmd/wcc/. Output named wwdump_ww
+# to avoid colliding with the C-side wwdump in $(BIN).
 $(BIN)/wwdump_ww: selfhost/cmd/wwdump/main.ww \
-		selfhost/cmd/wcc/lex.ww selfhost/cmd/wcc/tok.ww \
-		selfhost/cmd/wcc/mem.ww selfhost/cmd/wcc/ast.ww \
-		selfhost/cmd/wcc/parse.ww selfhost/cmd/wcc/typ.ww \
-		selfhost/cmd/wcc/sym.ww selfhost/cmd/wcc/check.ww \
+		lib/ww/lex.ww lib/ww/tok.ww lib/ww/ast.ww \
+		lib/ww/parse.ww lib/ww/typ.ww lib/ww/sym.ww \
+		selfhost/cmd/wcc/mem.ww selfhost/cmd/wcc/check.ww \
 		selfhost/cmd/wcc/cgen.ww selfhost/cmd/wcc/cgen_expr.ww \
 		selfhost/cmd/wcc/cgen_stmt.ww selfhost/cmd/wcc/cgen_util.ww \
 		selfhost/cmd/wcc/cgen_decl.ww \
 		lib/os/os.ww lib/strconv/strconv.ww \
 		$(BIN)/ww $(BIN)/w6c $(BIN)/w6a $(BIN)/w6l \
 		$(LIB)/libwwrt.a | $(BIN)
-	cd $(BIN) && ./ww build -I $$PWD/../../selfhost/cmd/wcc \
+	cd $(BIN) && ./ww build \
+		-I $$PWD/../../lib/ww \
+		-I $$PWD/../../selfhost/cmd/wcc \
 		$$PWD/../../selfhost/cmd/wwdump/main.ww
 	mv $(BIN)/main $@
 
 # ---- ww-side w6c (compiler port, exercised by 994_w6c_ww) -------------
-# Thin driver: parse + cgen, both already living in selfhost/cmd/wcc.
-# Build through `ww build` like the other wwstage tools.
+# Thin driver: parse + cgen. The frontend (lex/parse/ast/...) lives in
+# lib/ww/; cgen + check live in selfhost/cmd/wcc/.
 $(BIN)/w6c_ww: selfhost/cmd/w6c/main.ww \
-		selfhost/cmd/wcc/lex.ww selfhost/cmd/wcc/tok.ww \
-		selfhost/cmd/wcc/mem.ww selfhost/cmd/wcc/ast.ww \
-		selfhost/cmd/wcc/parse.ww selfhost/cmd/wcc/typ.ww \
-		selfhost/cmd/wcc/sym.ww selfhost/cmd/wcc/check.ww \
+		lib/ww/lex.ww lib/ww/tok.ww lib/ww/ast.ww \
+		lib/ww/parse.ww lib/ww/typ.ww lib/ww/sym.ww \
+		selfhost/cmd/wcc/mem.ww selfhost/cmd/wcc/check.ww \
 		selfhost/cmd/wcc/cgen.ww selfhost/cmd/wcc/cgen_expr.ww \
 		selfhost/cmd/wcc/cgen_stmt.ww selfhost/cmd/wcc/cgen_util.ww \
 		selfhost/cmd/wcc/cgen_decl.ww \
 		lib/os/os.ww \
 		$(BIN)/ww $(BIN)/w6c $(BIN)/w6a $(BIN)/w6l \
 		$(LIB)/libwwrt.a | $(BIN)
-	cd $(BIN) && ./ww build -I $$PWD/../../selfhost/cmd/wcc \
+	cd $(BIN) && ./ww build \
+		-I $$PWD/../../lib/ww \
+		-I $$PWD/../../selfhost/cmd/wcc \
 		$$PWD/../../selfhost/cmd/w6c/main.ww
 	mv $(BIN)/main $@
 
@@ -392,7 +395,9 @@ nocc:
 	@# The driver's default lib path is $self_dir/../../lib — under
 	@# $(NOCC_BIN) that resolves to $(NOCC_OUT)/lib (libwwrt only).
 	@# Pass -I $(CURDIR)/lib so `use os` etc. resolve to source.
-	@cd $(NOCC_BIN) && ./ww build -I $(CURDIR)/selfhost/cmd/wcc \
+	@cd $(NOCC_BIN) && ./ww build \
+		-I $(CURDIR)/lib/ww \
+		-I $(CURDIR)/selfhost/cmd/wcc \
 		-I $(CURDIR)/lib $(CURDIR)/selfhost/cmd/w6c/main.ww && mv main w6c_ww1
 	@cd $(NOCC_BIN) && ./ww build -I $(CURDIR)/selfhost/cmd/w6a \
 		-I $(CURDIR)/selfhost/cmd/wcc -I $(CURDIR)/lib \

cmd/ww/main.c

@@ -115,7 +115,7 @@ locate_import(const char *dirs, const char *name, char *out, size_t outsz)
 
 /* module_of — pick the source's containing-directory basename.
  * lib/os/os.ww      -> "os"
- * selfhost/cmd/wcc/sym.ww -> "wcc"
+ * lib/ww/sym.ww      -> "ww"
  * bare "main.ww"    -> "main"
  * Writes into dst (size ≥ 1); always NUL-terminates. */
 static void

lib/ww/ast.ww

@@ -1,4 +1,4 @@
-// selfhost/cmd/wcc/ast.ww — port of cmd/wcc/ast.c (Node defs + printer).
+// lib/ww/ast.ww — port of cmd/wcc/ast.c (Node defs + printer).
 //
 // Status: AST printer is fully ported. Constructor `newnode` is here.
 // The parser (parse.ww) is currently minimal — see its file header.

lib/ww/lex.ww

@@ -1,4 +1,4 @@
-// selfhost/cmd/wcc/lex.ww — port of cmd/wcc/lex.c.
+// lib/ww/lex.ww — port of cmd/wcc/lex.c.
 //
 // The DFA, the helpers, and the order of decisions all mirror the C
 // version exactly. The 990_selfhost test diffs the resulting token

lib/ww/parse.ww

@@ -1,4 +1,4 @@
-// selfhost/cmd/wcc/parse.ww — port of cmd/wcc/parse.c.
+// lib/ww/parse.ww — port of cmd/wcc/parse.c.
 //
 // Status: GROWING stub. Currently handles top-level `use IDENT;`,
 // `def NAME: TYPE = LIT;`, `type NAME = TYPE;`, and `fn NAME(params)

lib/ww/sym.ww

@@ -1,4 +1,4 @@
-// selfhost/cmd/wcc/sym.ww — port of cmd/wcc/sym.c.
+// lib/ww/sym.ww — port of cmd/wcc/sym.c.
 //
 // Per-scope hashtable, chained to the parent. Lookup walks up.
 // Plan 9 / Hare flavoured. Duplicate definitions in the same scope

lib/ww/tok.ww

@@ -1,4 +1,4 @@
-// selfhost/cmd/wcc/tok.ww — port of cmd/wcc/tok.c plus the Tkind /
+// lib/ww/tok.ww — port of cmd/wcc/tok.c plus the Tkind /
 // Tok / Pos shapes from cmd/wcc/ww.h.
 //
 // Token kind values must stay numerically equal to the C side: the

lib/ww/typ.ww

@@ -1,4 +1,4 @@
-// selfhost/cmd/wcc/type.ww — port of cmd/wcc/type.c.
+// lib/ww/typ.ww — port of cmd/wcc/type.c.
 //
 // Status: full structural port. The C version uses module-globals for
 // the primitive types (tyvoid, tyi32, …); ww doesn't have writable

selfhost/cmd/w6c/main.combined.ww

@@ -437,8 +437,8 @@ export fn parseu64(s: str) (u64 | str) = {
 	return v;
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/tok.ww — port of cmd/wcc/tok.c plus the Tkind /
+// MODULE: ww
+// lib/ww/tok.ww — port of cmd/wcc/tok.c plus the Tkind /
 // Tok / Pos shapes from cmd/wcc/ww.h.
 //
 // Token kind values must stay numerically equal to the C side: the
@@ -927,8 +927,8 @@ export fn toupper(c: u8) u8 = {
 	return c;
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/lex.ww — port of cmd/wcc/lex.c.
+// MODULE: ww
+// lib/ww/lex.ww — port of cmd/wcc/lex.c.
 //
 // The DFA, the helpers, and the order of decisions all mirror the C
 // version exactly. The 990_selfhost test diffs the resulting token
@@ -1616,8 +1616,8 @@ export fn lexnext(l: *lex, out: *tok) void = {
 	out.text = astrndup(l.a, one.ptr, 1u64);
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/ast.ww — port of cmd/wcc/ast.c (Node defs + printer).
+// MODULE: ww
+// lib/ww/ast.ww — port of cmd/wcc/ast.c (Node defs + printer).
 //
 // Status: AST printer is fully ported. Constructor `newnode` is here.
 // The parser (parse.ww) is currently minimal — see its file header.
@@ -1954,8 +1954,8 @@ export fn astprint(fd: i32, n: *node) void = {
 	pr(fd, n, 0);
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/parse.ww — port of cmd/wcc/parse.c.
+// MODULE: ww
+// lib/ww/parse.ww — port of cmd/wcc/parse.c.
 //
 // Status: GROWING stub. Currently handles top-level `use IDENT;`,
 // `def NAME: TYPE = LIT;`, `type NAME = TYPE;`, and `fn NAME(params)
@@ -2942,8 +2942,8 @@ export fn parsefile(p: *parser) *node = {
 	return f;
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/type.ww — port of cmd/wcc/type.c.
+// MODULE: ww
+// lib/ww/typ.ww — port of cmd/wcc/type.c.
 //
 // Status: full structural port. The C version uses module-globals for
 // the primitive types (tyvoid, tyi32, …); ww doesn't have writable
@@ -3273,8 +3273,8 @@ export fn typeeq(a: *tinfo, b: *tinfo) bool = {
 	return true;	// primitives match by kind alone
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/sym.ww — port of cmd/wcc/sym.c.
+// MODULE: ww
+// lib/ww/sym.ww — port of cmd/wcc/sym.c.
 //
 // Per-scope hashtable, chained to the parent. Lookup walks up.
 // Plan 9 / Hare flavoured. Duplicate definitions in the same scope

selfhost/cmd/ww/main.combined.ww

@@ -656,7 +656,7 @@ fn scanuse(src: *u8, len: u64) (*u8, u64) = {
 // Recursively expand `path` into c.out. Imported files are emitted
 // before their importer; cycles are broken via the visited set.
 // modulename — pick the source's containing-directory basename. So
-// `lib/os/os.ww` → "os"; `selfhost/cmd/wcc/sym.ww` → "wcc". Falls back
+// `lib/os/os.ww` → "os"; `lib/ww/sym.ww` → "ww". Falls back
 // to the file's own basename (sans .ww) when there is no parent dir.
 // Returns ("",0) if `path` ends in a '/' (degenerate).
 fn modulename(path: *u8, plen: u64) (*u8, u64) = {

selfhost/cmd/ww/main.ww

@@ -337,7 +337,7 @@ fn scanuse(src: *u8, len: u64) (*u8, u64) = {
 // Recursively expand `path` into c.out. Imported files are emitted
 // before their importer; cycles are broken via the visited set.
 // modulename — pick the source's containing-directory basename. So
-// `lib/os/os.ww` → "os"; `selfhost/cmd/wcc/sym.ww` → "wcc". Falls back
+// `lib/os/os.ww` → "os"; `lib/ww/sym.ww` → "ww". Falls back
 // to the file's own basename (sans .ww) when there is no parent dir.
 // Returns ("",0) if `path` ends in a '/' (degenerate).
 fn modulename(path: *u8, plen: u64) (*u8, u64) = {

selfhost/cmd/wwdump/main.combined.ww

@@ -437,8 +437,8 @@ export fn parseu64(s: str) (u64 | str) = {
 	return v;
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/tok.ww — port of cmd/wcc/tok.c plus the Tkind /
+// MODULE: ww
+// lib/ww/tok.ww — port of cmd/wcc/tok.c plus the Tkind /
 // Tok / Pos shapes from cmd/wcc/ww.h.
 //
 // Token kind values must stay numerically equal to the C side: the
@@ -927,8 +927,8 @@ export fn toupper(c: u8) u8 = {
 	return c;
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/lex.ww — port of cmd/wcc/lex.c.
+// MODULE: ww
+// lib/ww/lex.ww — port of cmd/wcc/lex.c.
 //
 // The DFA, the helpers, and the order of decisions all mirror the C
 // version exactly. The 990_selfhost test diffs the resulting token
@@ -1616,8 +1616,8 @@ export fn lexnext(l: *lex, out: *tok) void = {
 	out.text = astrndup(l.a, one.ptr, 1u64);
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/ast.ww — port of cmd/wcc/ast.c (Node defs + printer).
+// MODULE: ww
+// lib/ww/ast.ww — port of cmd/wcc/ast.c (Node defs + printer).
 //
 // Status: AST printer is fully ported. Constructor `newnode` is here.
 // The parser (parse.ww) is currently minimal — see its file header.
@@ -1954,8 +1954,8 @@ export fn astprint(fd: i32, n: *node) void = {
 	pr(fd, n, 0);
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/parse.ww — port of cmd/wcc/parse.c.
+// MODULE: ww
+// lib/ww/parse.ww — port of cmd/wcc/parse.c.
 //
 // Status: GROWING stub. Currently handles top-level `use IDENT;`,
 // `def NAME: TYPE = LIT;`, `type NAME = TYPE;`, and `fn NAME(params)
@@ -2942,8 +2942,8 @@ export fn parsefile(p: *parser) *node = {
 	return f;
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/type.ww — port of cmd/wcc/type.c.
+// MODULE: ww
+// lib/ww/typ.ww — port of cmd/wcc/type.c.
 //
 // Status: full structural port. The C version uses module-globals for
 // the primitive types (tyvoid, tyi32, …); ww doesn't have writable
@@ -3273,8 +3273,8 @@ export fn typeeq(a: *tinfo, b: *tinfo) bool = {
 	return true;	// primitives match by kind alone
 };
 
-// MODULE: wcc
-// selfhost/cmd/wcc/sym.ww — port of cmd/wcc/sym.c.
+// MODULE: ww
+// lib/ww/sym.ww — port of cmd/wcc/sym.c.
 //
 // Per-scope hashtable, chained to the parent. Lookup walks up.
 // Plan 9 / Hare flavoured. Duplicate definitions in the same scope

selfhost/test/uses.ww

@@ -3,7 +3,7 @@
 // `type NAME = TYPE;` (alias + struct), top-level `let NAME: TYPE = LIT;`.
 //
 // Function declarations are still recovered past — the body parser
-// is the next major chunk. See selfhost/cmd/wcc/parse.ww header.
+// is the next major chunk. See lib/ww/parse.ww header.
 
 use os;
 use mem;

test/wcc/990_selfhost.c

@@ -171,10 +171,10 @@ probe_dump_diff(const char *bin)
 	const char *tok_inputs[] = {
 		"selfhost/cmd/wcc/mem.ww",
 		"selfhost/cmd/wcc/err.ww",
-		"selfhost/cmd/wcc/lex.ww",
-		"selfhost/cmd/wcc/tok.ww",
-		"selfhost/cmd/wcc/ast.ww",
-		"selfhost/cmd/wcc/parse.ww",
+		"lib/ww/lex.ww",
+		"lib/ww/tok.ww",
+		"lib/ww/ast.ww",
+		"lib/ww/parse.ww",
 		"selfhost/cmd/wwdump/main.ww",
 		"selfhost/test/smoke.ww",
 		NULL,
@@ -188,12 +188,12 @@ probe_dump_diff(const char *bin)
 		"selfhost/test/uses.ww",
 		"selfhost/cmd/wcc/err.ww",
 		"selfhost/cmd/wcc/mem.ww",
-		"selfhost/cmd/wcc/lex.ww",
-		"selfhost/cmd/wcc/tok.ww",
-		"selfhost/cmd/wcc/ast.ww",
-		"selfhost/cmd/wcc/parse.ww",
-		"selfhost/cmd/wcc/typ.ww",
-		"selfhost/cmd/wcc/sym.ww",
+		"lib/ww/lex.ww",
+		"lib/ww/tok.ww",
+		"lib/ww/ast.ww",
+		"lib/ww/parse.ww",
+		"lib/ww/typ.ww",
+		"lib/ww/sym.ww",
 		"selfhost/cmd/wcc/check.ww",
 		"selfhost/cmd/wcc/cgen.ww",
 		"selfhost/cmd/wwdump/main.ww",
@@ -435,7 +435,7 @@ probe_resolve(const char *bin)
 	const char *complete[] = {
 		"selfhost/cmd/wcc/err.ww",
 		"selfhost/cmd/wcc/mem.ww",
-		"selfhost/cmd/wcc/tok.ww",
+		"lib/ww/tok.ww",
 		"selfhost/test/smoke.ww",
 		NULL,
 	};
@@ -656,8 +656,8 @@ probe_ww_links(const char *bin)
 		runwait(cmd);
 		/* ww build to get the .combined.ww as a side effect. */
 		snprintf(cmd, sizeof cmd,
-		    "cd %s && %s/ww build -I %s/selfhost/cmd/wcc %s >/dev/null 2>&1",
-		    tmpdir, bin, cwd, tmpsrc);
+		    "cd %s && %s/ww build -I %s/lib/ww -I %s/selfhost/cmd/wcc %s >/dev/null 2>&1",
+		    tmpdir, bin, cwd, cwd, tmpsrc);
 		if (runwait(cmd) != 0) {
 			fprintf(stderr, "ww-links FAIL: ww build %s\n", fix);
 			fail++;
@@ -717,7 +717,7 @@ probe_cgen_match(const char *bin)
 	const char *files[] = {
 		"selfhost/cmd/wcc/mem.ww",
 		"selfhost/cmd/wcc/err.ww",
-		"selfhost/cmd/wcc/tok.ww",
+		"lib/ww/tok.ww",
 		"selfhost/test/smoke.ww",
 		NULL,
 	};

test/wcc/993_ww_ww.c

@@ -152,8 +152,8 @@ main(void)
 	 * regardless of the per-driver workdir. */
 	static char wwdump_src[2048], wwdump_inc_a[2048], wwdump_inc_b[2048];
 	snprintf(wwdump_src,   sizeof wwdump_src,   "%s/selfhost/cmd/wwdump/main.ww", cwd);
-	snprintf(wwdump_inc_a, sizeof wwdump_inc_a, "%s/selfhost/cmd/wcc", cwd);
-	wwdump_inc_b[0] = '\0';
+	snprintf(wwdump_inc_a, sizeof wwdump_inc_a, "%s/lib/ww", cwd);
+	snprintf(wwdump_inc_b, sizeof wwdump_inc_b, "%s/selfhost/cmd/wcc", cwd);
 	cases[1].src   = wwdump_src;
 	cases[1].inc_a = wwdump_inc_a;
 	cases[1].inc_b = wwdump_inc_b;

test/wcc/995_self_rebuild.c

@@ -59,10 +59,11 @@ slurp_eq(const char *a, const char *b)
 	return rc;
 }
 
-/* Each tool builds via `ww_ww build -I <local> -I selfhost/cmd/wcc src`.
- * Some tools have a local module dir (w6a, w6l with sibling .ww files);
- * wwc-only tools (w6c, ww, wwdump) just need the wwc -I. inc_local is
- * "" for those.
+/* Each tool builds via `ww_ww build -I <local> -I lib/ww -I selfhost/cmd/wcc src`.
+ * lib/ww holds the language introspection (lex/tok/ast/parse/typ/sym);
+ * selfhost/cmd/wcc holds the compiler internals (mem/check/cgen*).
+ * Some tools have a local module dir (w6a, w6l with sibling .ww files).
+ * inc_local is "" for tools without one (w6c, ww, wwdump).
  */
 static int
 rebuild_one(const char *bin, const char *cwd, const char *tool,
@@ -76,14 +77,14 @@ rebuild_one(const char *bin, const char *cwd, const char *tool,
 
 	if (inc_local && inc_local[0]) {
 		snprintf(cmd, sizeof cmd,
-		    "cd %s && %s/ww_ww build -I %s/%s -I %s/selfhost/cmd/wcc "
+		    "cd %s && %s/ww_ww build -I %s/%s -I %s/lib/ww -I %s/selfhost/cmd/wcc "
 		    "%s/%s >/dev/null 2>&1",
-		    workdir, bin, cwd, inc_local, cwd, cwd, src_rel);
+		    workdir, bin, cwd, inc_local, cwd, cwd, cwd, src_rel);
 	} else {
 		snprintf(cmd, sizeof cmd,
-		    "cd %s && %s/ww_ww build -I %s/selfhost/cmd/wcc "
+		    "cd %s && %s/ww_ww build -I %s/lib/ww -I %s/selfhost/cmd/wcc "
 		    "%s/%s >/dev/null 2>&1",
-		    workdir, bin, cwd, cwd, src_rel);
+		    workdir, bin, cwd, cwd, cwd, src_rel);
 	}
 	if (runwait(cmd) != 0) {
 		fprintf(stderr, "self-rebuild FAIL: ww_ww build errored on %s\n", tool);