From e13059658431d7f75ef75a9cc120b3377c3d024f Mon Sep 17 00:00:00 2001
From: Local Dev <dev@localhost>
Date: Sun, 20 Sep 2026 19:34:42 +0200
Subject: [PATCH] =?UTF-8?q?feat(gateway):=20clean=20URLs=20=E2=80=94=20ser?=
 =?UTF-8?q?ve=20<path>.html=20+=20strip=20.html=20from=20hrefs?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Two symmetric changes so a site published as flat .html files (sirius.x
has tld.html, portal.html, market.html, register.html, studio.html)
reads with clean URLs to the visitor.

1. S3 handler fallback order for extensionless requests:
     /tld  →  tld.html         (new — first)
     /tld  →  tld/index.html   (existing — second)
     /...  →  index.html       (existing — SPA last-resort)
   Requests with a real file extension (.js, .svg, .json) are never
   rescued — those must never be routed by a client-side router and
   swallowing them would hide real bugs.

2. transformHtml() gains stripHtmlExt(): rewrites href="foo.html"
   (relative) to href="foo", preserving anchor and query. Absolute
   URLs (https://silentmode.st/tlds/) are left alone — external
   references are none of our business.

Backwards-compatible: existing /bns/<name>/foo.html links still 200.
Verified end-to-end on sirius.x: /tld, /portal, /market, /register,
/studio all serve; assets under js/, assets/ unchanged; direct .html
requests still work.
---
 src/gateway/public-gateway.mjs | 42 +++++++++++++++++++++++++++-------
 1 file changed, 34 insertions(+), 8 deletions(-)

diff --git a/src/gateway/public-gateway.mjs b/src/gateway/public-gateway.mjs
index b6c45ff..c83eacd 100644
--- a/src/gateway/public-gateway.mjs
+++ b/src/gateway/public-gateway.mjs
@@ -317,11 +317,27 @@ function rewriteBnsLinks(html) {
     }
   );
 }
-// Apply both transforms (base first, then link rewrite) — order doesn't
-// matter because they touch disjoint constructs, but grouping them here
-// means every text/html response goes through the same path.
+// Strip `.html` from relative link targets in served HTML — so a link
+// written as href="tld.html" reads as href="tld" in the browser bar.
+// Only touches non-absolute URLs (leaves `https://…/x.html` alone —
+// those are external and often out of our control). Preserves anchor
+// and query, and skips the ".html" that's part of a filename like
+// "foo.htmlx" (regex requires end/? / #).
+function stripHtmlExt(html) {
+  return html.replace(
+    /\b(href\s*=\s*")([^"#?][^"#?]*)\.html(?=["#?])/gi,
+    (m, prefix, path) => {
+      if (/^\s*[a-z][a-z0-9+.-]*:/i.test(path)) return m; // absolute URL — leave alone
+      return prefix + path;
+    }
+  );
+}
+// Apply the three transforms (base first, link rewrite, extension strip).
+// Order doesn't matter across them because they touch disjoint constructs,
+// but grouping them here means every text/html response goes through the
+// same path.
 function transformHtml(html, baseHref) {
-  return rewriteBnsLinks(injectBase(html, baseHref));
+  return stripHtmlExt(rewriteBnsLinks(injectBase(html, baseHref)));
 }
 const nx = (h) => `<div style="font:16px system-ui;padding:2rem;color:#333"><h2>NXDOMAIN</h2><p><code>${h}</code> is not registered on the Bitcoin Cash chain.</p></div>`;
 
@@ -487,15 +503,25 @@ async function serve(res, name, reqPath, baseHref) {
         // NOT rescued — those should never be routed by a client-side
         // router and swallowing them would hide real bugs.
         if (up.status === 404 && !/\.[a-z0-9]+$/i.test(reqPath) && reqPath !== "/") {
-          // Try directory-index first
+          const cleanBase = base.endsWith("/") ? base : base + "/";
+          const cleanPath = reqPath.slice(1).replace(/\/+$/, "");
+          // 1) Extensionless-URL support: /tld -> tld.html. Lets sites
+          //    publish clean URLs without keeping mirror files around.
           if (!reqPath.endsWith("/")) {
-            const dirKey = (base.endsWith("/") ? base : base + "/") + reqPath.slice(1).replace(/\/+$/, '') + "/index.html";
+            const htmlKey = cleanBase + cleanPath + ".html";
+            const htmlHit = await aws.fetch(`${S3_ENDPOINT}/${htmlKey}`);
+            if (htmlHit.ok) { up = htmlHit; }
+          }
+          // 2) Directory-index: /docs -> docs/index.html
+          if (up.status === 404 && !reqPath.endsWith("/")) {
+            const dirKey = cleanBase + cleanPath + "/index.html";
             const dir = await aws.fetch(`${S3_ENDPOINT}/${dirKey}`);
             if (dir.ok) { up = dir; }
           }
-          // Then SPA root fallback
+          // 3) SPA root fallback — last resort so client-side routers
+          //    can pick up /katalogos, /anthemus, etc.
           if (up.status === 404) {
-            const fallbackKey = (base.endsWith("/") ? base : base + "/") + "index.html";
+            const fallbackKey = cleanBase + "index.html";
             const fb = await aws.fetch(`${S3_ENDPOINT}/${fallbackKey}`);
             if (fb.ok) { up = fb; }
           }