ww+wcc: emit and lex // MODULE: <name> directive in combined.ww

selfhost/cmd/ww/main.combined.ww

@@ -1,3 +1,4 @@
+// MODULE: os
 // os — process and filesystem facade. The body of each call lands
 // either in libwwrt.a (rt_syscall trampoline) or libc bindings,
 // depending on how the program was linked.
@@ -207,6 +208,7 @@ export fn getdents64(fd: i32, buf: *u8, n: u64) i64 = {
 	return syscall3(SYS_GETDENTS64, fd: i64, buf: i64, n: i64);
 };
 
+// MODULE: wcc
 // selfhost/cmd/wcc/mem.ww — port of cmd/wcc/mem.c.
 //
 // Bump arena allocator. Backed by the runtime page allocator
@@ -314,6 +316,7 @@ export fn freearena(a: *arena) void = {
 	};
 };
 
+// MODULE: ww
 // selfhost/cmd/ww/main.ww — port of cmd/ww/main.c.
 //
 // The user-facing driver. Plan 9 cc(1) / Hare hare(1) analogue:
@@ -652,6 +655,50 @@ 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
+// 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) = {
+	if (plen == 0u64) { return nil, 0u64; };
+	// Find the last '/'.
+	let last: u64 = plen;
+	let i: u64 = plen;
+	for (i > 0u64) {
+		i -= 1u64;
+		if (path[i] == 47u8) { last = i; i = 0u64; }
+		else { if (i == 0u64) { last = plen; }; };
+	};
+	if (last == plen) {
+		// No '/' — path is a bare filename. Use its stem.
+		let n: u64 = plen;
+		if (n >= 3u64) {
+			if (path[n - 3u64] == 46u8) {
+				if (path[n - 2u64] == 119u8) {
+					if (path[n - 1u64] == 119u8) {
+						n = n - 3u64;
+					};
+				};
+			};
+		};
+		return path, n;
+	};
+	// Find the '/' before `last` — segment between is the dir basename.
+	let prev: u64 = 0u64;
+	let found: bool = false;
+	let j: u64 = last;
+	for (j > 0u64) {
+		j -= 1u64;
+		if (path[j] == 47u8) {
+			prev = j + 1u64;
+			found = true;
+			j = 0u64;
+		};
+	};
+	if (!found) { prev = 0u64; };
+	return path + prev, last - prev;
+};
+
 fn expand(c: *expctx, path_cs: *u8) void = {
 	let plen: u64 = cstrlen(path_cs);
 	let pathstr: str = astrndup(c.a, path_cs, plen);
@@ -687,7 +734,19 @@ fn expand(c: *expctx, path_cs: *u8) void = {
 		i = j + 1u64;
 	};
 
-	// Pass 2: emit our own bytes, then a trailing newline.
+	// Pass 2: prefix a `// MODULE: <name>` directive, then emit our
+	// bytes. The marker is a comment to the C-side wcc and any other
+	// reader; the ww-side wcc lexer recognises it and stamps each
+	// top-level decl with the originating module so cgen can mangle
+	// non-exported names by module.
+	let mp: *u8;
+	let mn: u64;
+	mp, mn = modulename(path_cs, plen);
+	if (mn > 0u64) {
+		os.writefull(c.out, "// MODULE: ".ptr, 11u64);
+		os.writefull(c.out, mp, mn);
+		os.writefull(c.out, "\n".ptr, 1u64);
+	};
 	os.writefull(c.out, bufp, blen);
 	os.writefull(c.out, "\n".ptr, 1u64);
 };

selfhost/cmd/ww/main.ww

@@ -336,6 +336,50 @@ 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
+// 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) = {
+	if (plen == 0u64) { return nil, 0u64; };
+	// Find the last '/'.
+	let last: u64 = plen;
+	let i: u64 = plen;
+	for (i > 0u64) {
+		i -= 1u64;
+		if (path[i] == 47u8) { last = i; i = 0u64; }
+		else { if (i == 0u64) { last = plen; }; };
+	};
+	if (last == plen) {
+		// No '/' — path is a bare filename. Use its stem.
+		let n: u64 = plen;
+		if (n >= 3u64) {
+			if (path[n - 3u64] == 46u8) {
+				if (path[n - 2u64] == 119u8) {
+					if (path[n - 1u64] == 119u8) {
+						n = n - 3u64;
+					};
+				};
+			};
+		};
+		return path, n;
+	};
+	// Find the '/' before `last` — segment between is the dir basename.
+	let prev: u64 = 0u64;
+	let found: bool = false;
+	let j: u64 = last;
+	for (j > 0u64) {
+		j -= 1u64;
+		if (path[j] == 47u8) {
+			prev = j + 1u64;
+			found = true;
+			j = 0u64;
+		};
+	};
+	if (!found) { prev = 0u64; };
+	return path + prev, last - prev;
+};
+
 fn expand(c: *expctx, path_cs: *u8) void = {
 	let plen: u64 = cstrlen(path_cs);
 	let pathstr: str = astrndup(c.a, path_cs, plen);
@@ -371,7 +415,19 @@ fn expand(c: *expctx, path_cs: *u8) void = {
 		i = j + 1u64;
 	};
 
-	// Pass 2: emit our own bytes, then a trailing newline.
+	// Pass 2: prefix a `// MODULE: <name>` directive, then emit our
+	// bytes. The marker is a comment to the C-side wcc and any other
+	// reader; the ww-side wcc lexer recognises it and stamps each
+	// top-level decl with the originating module so cgen can mangle
+	// non-exported names by module.
+	let mp: *u8;
+	let mn: u64;
+	mp, mn = modulename(path_cs, plen);
+	if (mn > 0u64) {
+		os.writefull(c.out, "// MODULE: ".ptr, 11u64);
+		os.writefull(c.out, mp, mn);
+		os.writefull(c.out, "\n".ptr, 1u64);
+	};
 	os.writefull(c.out, bufp, blen);
 	os.writefull(c.out, "\n".ptr, 1u64);
 };