pages: wildcard zones — <sub>.zone serves dir/<sub>, apex /<sub>/ redirects there

pages/server.py

@@ -2,7 +2,9 @@
 
 SITES maps hostnames to `org/repo`, or `org/repo#branch/subdir`, on GITEA (public repos only; cloned over https, no
 token). A clone lives under /data/<org>__<repo>__<branch>; `POST /_hook/<org>/<repo>` (a Gitea webhook) or the
-periodic loop pulls it. Static files only:
+periodic loop pulls it. WILDCARDS maps a zone to a site whose top-level directories are its subdomains:
+`{"yolopreneur.lol": "teahubxyz/time#gh-pages/ventures"}` serves `<sub>.yolopreneur.lol` from `ventures/<sub>/`, and
+`yolopreneur.lol/<sub>/` redirects there. Static files only:
 a path maps to the file, a directory to its index.html, anything missing to 404.html or a plain 404. No listing, no
 execution, nothing writes except git. stdlib only.
 """
@@ -19,6 +21,7 @@ GITEA = os.environ.get("GITEA_URL", "https://teahub.xyz").rstrip("/")
 DATA = Path(os.environ.get("PAGES_DATA", "/data"))
 SITES = json.loads(os.environ.get("PAGES_SITES") or "{}")  # {"persimmons.app": "persimmons/site", ...}
 PULL_EVERY = int(os.environ.get("PAGES_PULL_SECONDS") or 300)
+WILDCARDS = json.loads(os.environ.get("PAGES_WILDCARDS") or "{}")  # {"zone": "org/repo#branch/dir"}: <sub>.zone -> dir/<sub>
 TYPES = {".html": "text/html; charset=utf-8", ".css": "text/css; charset=utf-8", ".js": "text/javascript; charset=utf-8",
          ".mjs": "text/javascript; charset=utf-8", ".json": "application/json", ".svg": "image/svg+xml", ".png": "image/png",
          ".jpg": "image/jpeg", ".jpeg": "image/jpeg", ".gif": "image/gif", ".webp": "image/webp", ".ico": "image/x-icon",
@@ -65,7 +68,7 @@ def sync(site: str) -> str:
 
 def loop() -> None:
     while True:
-        for repo in set(SITES.values()):
+        for repo in set(SITES.values()) | set(WILDCARDS.values()):
             print(repo, sync(repo), flush=True)
         time.sleep(PULL_EVERY)
 
@@ -76,13 +79,24 @@ class Handler(http.server.BaseHTTPRequestHandler):
     def log_message(self, fmt, *args):  # one line per request, no client address
         print(f"{self.headers.get('host', '-')} {self.command} {self.path.split('?')[0]} {args[1] if len(args) > 1 else ''}", flush=True)
 
-    def site(self) -> str | None:
-        host = (self.headers.get("host") or "").split(":")[0].lower()
-        return SITES.get(host) or SITES.get(host[4:] if host.startswith("www.") else "")
+    def host(self) -> str:
+        return (self.headers.get("host") or "").split(":")[0].lower()
+
+    def site(self) -> tuple[str, str] | None:
+        """(site, subdir) for this request's host: an exact site, its www, or a wildcard subdomain."""
+        host = self.host()
+        s = SITES.get(host) or SITES.get(host[4:] if host.startswith("www.") else "")
+        if s:
+            return s, ""
+        sub, _, zone = host.partition(".")
+        w = WILDCARDS.get(zone)
+        if w and sub and sub != "www" and "/" not in sub and ".." not in sub:
+            return w, sub
+        return None
 
     def do_POST(self):
         parts = urlsplit(self.path).path.strip("/").split("/")
-        hit = [s for s in SITES.values() if len(parts) == 3 and parts[0] == "_hook" and parse(s)[0] == f"{parts[1]}/{parts[2]}"]
+        hit = [s for s in [*SITES.values(), *WILDCARDS.values()] if len(parts) == 3 and parts[0] == "_hook" and parse(s)[0] == f"{parts[1]}/{parts[2]}"]
         if hit:
             self.rfile.read(int(self.headers.get("content-length") or 0))
             outs = [sync(s) for s in dict.fromkeys(hit)]  # every site on that repo (any branch)
@@ -104,12 +118,23 @@ class Handler(http.server.BaseHTTPRequestHandler):
             if not head:
                 self.wfile.write(body)
             return
-        repo = self.site()
-        if not repo:
+        found = self.site()
+        if not found:
             self.send_error(404, "no site for this host")
             return
+        repo, sub = found
         root = site_root(repo).resolve()
         rel = unquote(path).lstrip("/")
+        # on a wildcard zone's apex, /<sub>/... lives at <sub>.zone: send people there
+        head = rel.split("/", 1)[0]
+        if not sub and self.host() in WILDCARDS and head and (root / head).is_dir() and ".." not in rel:
+            self.send_response(301); self.send_header("location", f"https://{head}.{self.host()}/" + rel[len(head):].lstrip("/")); self.end_headers()
+            return
+        if sub:
+            root = (root / sub).resolve()
+            if not root.is_dir():
+                self.send_error(404, "no such venture")
+                return
         target = (root / rel).resolve() if rel else root
         if root not in target.parents and target != root:
             self.send_error(404)