@geekity/cms 0.7.0 → 0.9.0

package/dist/webmention/reply-context.js

@@ -1,8 +1,7 @@
 import { discoverPostType } from '../content/post-type.js';
-import { WEBMENTION_USER_AGENT } from './discovery.js';
+import { fetchPublic, webUrl } from './fetch-public.js';
 import { elementsIn, parseHtml, textOf } from './html.js';
 import { citedEntry } from './microformats.js';
-import { publicHost } from './public-address.js';
 /**
  * Reply context: what a reply shows of the post it answers (TASK-123).
  *
@@ -16,8 +15,6 @@ import { publicHost } from './public-address.js';
 export const REPLY_CONTEXT_TIMEOUT_MS = 10_000;
 /** The most of a target page that is read. A bigger page is not read at all. */
 export const REPLY_CONTEXT_MAX_BYTES = 1_000_000;
-/** How many redirects are followed, each one checked like the first. */
-const MAX_REDIRECTS = 5;
 /** How many words of a target's text a preview keeps. */
 const EXCERPT_WORDS = 40;
 /** And how many characters, for a text with few spaces in it. */
@@ -30,49 +27,17 @@ const EXCERPT_CHARACTERS = 300;
  * over the whole exchange, and no page over the byte limit. Nothing throws.
  */
 export async function fetchReplyContext(target, options) {
-    const maxBytes = options.maxBytes ?? REPLY_CONTEXT_MAX_BYTES;
-    const signal = AbortSignal.timeout(options.timeoutMs ?? REPLY_CONTEXT_TIMEOUT_MS);
-    let at = target;
-    try {
-        for (let hop = 0;; hop += 1) {
-            const url = webUrl(at);
-            if (url === undefined)
-                return refuse('not an http or https URL');
-            if (!(await publicHost(url.hostname, options.lookup))) {
-                return refuse('not a public address');
-            }
-            const response = await fetch(url, {
-                headers: { accept: 'text/html, */*;q=0.8', 'user-agent': WEBMENTION_USER_AGENT },
-                redirect: 'manual',
-                signal,
-            });
-            const location = response.headers.get('location');
-            if (response.status >= 300 && response.status < 400 && location !== null) {
-                await response.body?.cancel();
-                if (hop === MAX_REDIRECTS)
-                    return refuse('too many redirects');
-                at = new URL(location, url).href;
-                continue;
-            }
-            if (!response.ok) {
-                await response.body?.cancel();
-                return refuse(`answered ${String(response.status)}`);
-            }
-            const type = response.headers.get('content-type') ?? '';
-            if (!/^\s*(text\/html|application\/xhtml\+xml)/i.test(type)) {
-                await response.body?.cancel();
-                return refuse('not an HTML page');
-            }
-            const body = await readWithin(response, maxBytes);
-            if (body === undefined)
-                return refuse(`larger than ${String(maxBytes)} bytes`);
-            const context = readReplyContext(body, target, url.href);
-            return context === undefined ? refuse('nothing to show') : { ok: true, context };
-        }
-    }
-    catch (thrown) {
-        return refuse(thrown instanceof Error ? thrown.message : String(thrown));
-    }
+    const fetched = await fetchPublic(target, {
+        lookup: options.lookup,
+        timeoutMs: options.timeoutMs ?? REPLY_CONTEXT_TIMEOUT_MS,
+        maxBytes: options.maxBytes ?? REPLY_CONTEXT_MAX_BYTES,
+        accept: 'text/html, */*;q=0.8',
+        contentType: { pattern: /^\s*(text\/html|application\/xhtml\+xml)/i, name: 'an HTML page' },
+    });
+    if (!fetched.ok)
+        return fetched;
+    const context = readReplyContext(new TextDecoder().decode(fetched.body), target, fetched.url);
+    return context === undefined ? refuse('nothing to show') : { ok: true, context };
 }
 /**
  * What a target page says about itself, or `undefined` when it says nothing a
@@ -112,35 +77,6 @@ export function readReplyContext(html, target, base = target) {
         return undefined;
     return { url: target, ...(name === '' ? {} : { name }), ...(text === '' ? {} : { text }) };
 }
-/** A response body when it is no bigger than the limit, else `undefined`. */
-async function readWithin(response, maxBytes) {
-    const declared = Number(response.headers.get('content-length') ?? '0');
-    if (declared > maxBytes) {
-        await response.body?.cancel();
-        return undefined;
-    }
-    const body = response.body;
-    if (body === null)
-        return '';
-    const reader = body.getReader();
-    const decoder = new TextDecoder();
-    const chunks = [];
-    let read = 0;
-    for (;;) {
-        const chunk = await reader.read();
-        if (chunk.done)
-            break;
-        const value = chunk.value;
-        read += value.length;
-        if (read > maxBytes) {
-            await reader.cancel();
-            return undefined;
-        }
-        chunks.push(decoder.decode(value, { stream: true }));
-    }
-    chunks.push(decoder.decode());
-    return chunks.join('');
-}
 /** The page's `<title>`, or empty. */
 function titleOf(root) {
     for (const element of elementsIn(root)) {
@@ -170,19 +106,6 @@ function excerpt(text) {
         cut = cut.slice(0, EXCERPT_CHARACTERS).trimEnd();
     return cut === normalized ? normalized : `${cut} …`;
 }
-/** A URL, when it is an http or https one with a host. */
-function webUrl(value) {
-    let url;
-    try {
-        url = new URL(value);
-    }
-    catch {
-        return undefined;
-    }
-    if (url.protocol !== 'http:' && url.protocol !== 'https:')
-        return undefined;
-    return url.hostname === '' ? undefined : url;
-}
 function refuse(reason) {
     return { ok: false, reason };
 }

package/dist/webmention/reply-context.js.map

@@ -1 +1 @@
-{"version":3,"file":"reply-context.js","sourceRoot":"","sources":["../../src/webmention/reply-context.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,yBAAyB,CAAC;AAC3D,OAAO,EAAE,qBAAqB,EAAE,MAAM,gBAAgB,CAAC;AACvD,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,MAAM,EAAE,MAAM,WAAW,CAAC;AAE1D,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAC/C,OAAO,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAGjD;;;;;;;;GAQG;AAEH,2DAA2D;AAC3D,MAAM,CAAC,MAAM,wBAAwB,GAAG,MAAM,CAAC;AAE/C,gFAAgF;AAChF,MAAM,CAAC,MAAM,uBAAuB,GAAG,SAAS,CAAC;AAEjD,wEAAwE;AACxE,MAAM,aAAa,GAAG,CAAC,CAAC;AAExB,yDAAyD;AACzD,MAAM,aAAa,GAAG,EAAE,CAAC;AAEzB,iEAAiE;AACjE,MAAM,kBAAkB,GAAG,GAAG,CAAC;AA+B/B;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACrC,MAAc,EACd,OAAiC;IAEjC,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,uBAAuB,CAAC;IAC7D,MAAM,MAAM,GAAG,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,SAAS,IAAI,wBAAwB,CAAC,CAAC;IAClF,IAAI,EAAE,GAAG,MAAM,CAAC;IAEhB,IAAI,CAAC;QACH,KAAK,IAAI,GAAG,GAAG,CAAC,GAAI,GAAG,IAAI,CAAC,EAAE,CAAC;YAC7B,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,CAAC;YACvB,IAAI,GAAG,KAAK,SAAS;gBAAE,OAAO,MAAM,CAAC,0BAA0B,CAAC,CAAC;YACjE,IAAI,CAAC,CAAC,MAAM,UAAU,CAAC,GAAG,CAAC,QAAQ,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC;gBACtD,OAAO,MAAM,CAAC,sBAAsB,CAAC,CAAC;YACxC,CAAC;YAED,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE;gBAChC,OAAO,EAAE,EAAE,MAAM,EAAE,sBAAsB,EAAE,YAAY,EAAE,qBAAqB,EAAE;gBAChF,QAAQ,EAAE,QAAQ;gBAClB,MAAM;aACP,CAAC,CAAC;YAEH,MAAM,QAAQ,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;YAClD,IAAI,QAAQ,CAAC,MAAM,IAAI,GAAG,IAAI,QAAQ,CAAC,MAAM,GAAG,GAAG,IAAI,QAAQ,KAAK,IAAI,EAAE,CAAC;gBACzE,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC;gBAC9B,IAAI,GAAG,KAAK,aAAa;oBAAE,OAAO,MAAM,CAAC,oBAAoB,CAAC,CAAC;gBAC/D,EAAE,GAAG,IAAI,GAAG,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC;gBACjC,SAAS;YACX,CAAC;YAED,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;gBACjB,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC;gBAC9B,OAAO,MAAM,CAAC,YAAY,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;YACvD,CAAC;YAED,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC;YACxD,IAAI,CAAC,2CAA2C,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC5D,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC;gBAC9B,OAAO,MAAM,CAAC,kBAAkB,CAAC,CAAC;YACpC,CAAC;YAED,MAAM,IAAI,GAAG,MAAM,UAAU,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;YAClD,IAAI,IAAI,KAAK,SAAS;gBAAE,OAAO,MAAM,CAAC,eAAe,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YAE/E,MAAM,OAAO,GAAG,gBAAgB,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,CAAC;YACzD,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC;QACnF,CAAC;IACH,CAAC;IAAC,OAAO,MAAM,EAAE,CAAC;QAChB,OAAO,MAAM,CAAC,MAAM,YAAY,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC;IAC3E,CAAC;AACH,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,gBAAgB,CAC9B,IAAY,EACZ,MAAc,EACd,IAAI,GAAW,MAAM;IAErB,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IAC7B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IAErC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,MAAM,KAAK,GAAG,gBAAgB,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,KAAK,SAAS,CAAC;QACxF,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,EAAE,GAAG,KAAK,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,GAAG,IAAI,EAAE,CAAC,CAAC;QAC3F,OAAO;YACL,GAAG,EAAE,MAAM;YACX,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YACtC,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;YAChC,GAAG,CAAC,KAAK,CAAC,MAAM,KAAK,SAAS;gBAC5B,CAAC,CAAC,EAAE;gBACJ,CAAC,CAAC;oBACE,MAAM,EAAE;wBACN,IAAI,EAAE,KAAK,CAAC,MAAM,CAAC,IAAI;wBACvB,GAAG,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,CAAC;qBAC5D;iBACF,CAAC;YACN,GAAG,CAAC,KAAK,CAAC,SAAS,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,KAAK,CAAC,SAAS,EAAE,CAAC;SACpE,CAAC;IACJ,CAAC;IAED,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,aAAa,CAAC,IAAI,MAAM,CAAC,IAAI,EAAE,gBAAgB,CAAC,CAAC,CAAC;IACpF,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,KAAK,EAAE;QAAE,OAAO,SAAS,CAAC;IAEjD,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;AAC7F,CAAC;AAED,6EAA6E;AAC7E,KAAK,UAAU,UAAU,CAAC,QAAkB,EAAE,QAAgB;IAC5D,MAAM,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,gBAAgB,CAAC,IAAI,GAAG,CAAC,CAAC;IACvE,IAAI,QAAQ,GAAG,QAAQ,EAAE,CAAC;QACxB,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC;QAC9B,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC;IAC3B,IAAI,IAAI,KAAK,IAAI;QAAE,OAAO,EAAE,CAAC;IAE7B,MAAM,MAAM,GAAI,IAAmC,CAAC,SAAS,EAAE,CAAC;IAChE,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAC;IAClC,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,IAAI,GAAG,CAAC,CAAC;IAEb,SAAS,CAAC;QACR,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QAClC,IAAI,KAAK,CAAC,IAAI;YAAE,MAAM;QACtB,MAAM,KAAK,GAAe,KAAK,CAAC,KAAK,CAAC;QACtC,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC;QACrB,IAAI,IAAI,GAAG,QAAQ,EAAE,CAAC;YACpB,MAAM,MAAM,CAAC,MAAM,EAAE,CAAC;YACtB,OAAO,SAAS,CAAC;QACnB,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACvD,CAAC;IAED,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAC9B,OAAO,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACzB,CAAC;AAED,sCAAsC;AACtC,SAAS,OAAO,CAAC,IAAiB;IAChC,KAAK,MAAM,OAAO,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;QACvC,IAAI,OAAO,CAAC,IAAI,KAAK,OAAO;YAAE,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC;IACvD,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,uEAAuE;AACvE,SAAS,MAAM,CAAC,IAAiB,EAAE,GAAW;IAC5C,KAAK,MAAM,OAAO,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;QACvC,IAAI,OAAO,CAAC,IAAI,KAAK,MAAM;YAAE,SAAS;QACtC,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC;QAC3E,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,EAAE,WAAW,EAAE,KAAK,GAAG,CAAC;YAAE,SAAS;QACjE,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC3E,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,8DAA8D;AAC9D,SAAS,OAAO,CAAC,IAAY;IAC3B,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IACpD,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACpC,IAAI,GAAG,GAAG,KAAK,CAAC,MAAM,GAAG,aAAa,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC;IAC9F,IAAI,GAAG,CAAC,MAAM,GAAG,kBAAkB;QAAE,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC,OAAO,EAAE,CAAC;IACtF,OAAO,GAAG,KAAK,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,GAAG,IAAI,CAAC;AACtD,CAAC;AAED,0DAA0D;AAC1D,SAAS,MAAM,CAAC,KAAa;IAC3B,IAAI,GAAQ,CAAC;IACb,IAAI,CAAC;QACH,GAAG,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;IACvB,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,IAAI,GAAG,CAAC,QAAQ,KAAK,OAAO,IAAI,GAAG,CAAC,QAAQ,KAAK,QAAQ;QAAE,OAAO,SAAS,CAAC;IAC5E,OAAO,GAAG,CAAC,QAAQ,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC;AAC/C,CAAC;AAED,SAAS,MAAM,CAAC,MAAc;IAC5B,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;AAC/B,CAAC"}
+{"version":3,"file":"reply-context.js","sourceRoot":"","sources":["../../src/webmention/reply-context.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,yBAAyB,CAAC;AAC3D,OAAO,EAAE,WAAW,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AACxD,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,MAAM,EAAE,MAAM,WAAW,CAAC;AAE1D,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAG/C;;;;;;;;GAQG;AAEH,2DAA2D;AAC3D,MAAM,CAAC,MAAM,wBAAwB,GAAG,MAAM,CAAC;AAE/C,gFAAgF;AAChF,MAAM,CAAC,MAAM,uBAAuB,GAAG,SAAS,CAAC;AAEjD,yDAAyD;AACzD,MAAM,aAAa,GAAG,EAAE,CAAC;AAEzB,iEAAiE;AACjE,MAAM,kBAAkB,GAAG,GAAG,CAAC;AA+B/B;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACrC,MAAc,EACd,OAAiC;IAEjC,MAAM,OAAO,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE;QACxC,MAAM,EAAE,OAAO,CAAC,MAAM;QACtB,SAAS,EAAE,OAAO,CAAC,SAAS,IAAI,wBAAwB;QACxD,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,uBAAuB;QACrD,MAAM,EAAE,sBAAsB;QAC9B,WAAW,EAAE,EAAE,OAAO,EAAE,2CAA2C,EAAE,IAAI,EAAE,cAAc,EAAE;KAC5F,CAAC,CAAC;IACH,IAAI,CAAC,OAAO,CAAC,EAAE;QAAE,OAAO,OAAO,CAAC;IAEhC,MAAM,OAAO,GAAG,gBAAgB,CAAC,IAAI,WAAW,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC;IAC9F,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC;AACnF,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,gBAAgB,CAC9B,IAAY,EACZ,MAAc,EACd,IAAI,GAAW,MAAM;IAErB,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IAC7B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IAErC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,MAAM,KAAK,GAAG,gBAAgB,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,KAAK,SAAS,CAAC;QACxF,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,EAAE,GAAG,KAAK,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,GAAG,IAAI,EAAE,CAAC,CAAC;QAC3F,OAAO;YACL,GAAG,EAAE,MAAM;YACX,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YACtC,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;YAChC,GAAG,CAAC,KAAK,CAAC,MAAM,KAAK,SAAS;gBAC5B,CAAC,CAAC,EAAE;gBACJ,CAAC,CAAC;oBACE,MAAM,EAAE;wBACN,IAAI,EAAE,KAAK,CAAC,MAAM,CAAC,IAAI;wBACvB,GAAG,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,CAAC;qBAC5D;iBACF,CAAC;YACN,GAAG,CAAC,KAAK,CAAC,SAAS,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,KAAK,CAAC,SAAS,EAAE,CAAC;SACpE,CAAC;IACJ,CAAC;IAED,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,aAAa,CAAC,IAAI,MAAM,CAAC,IAAI,EAAE,gBAAgB,CAAC,CAAC,CAAC;IACpF,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,KAAK,EAAE;QAAE,OAAO,SAAS,CAAC;IAEjD,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;AAC7F,CAAC;AAED,sCAAsC;AACtC,SAAS,OAAO,CAAC,IAAiB;IAChC,KAAK,MAAM,OAAO,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;QACvC,IAAI,OAAO,CAAC,IAAI,KAAK,OAAO;YAAE,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC;IACvD,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,uEAAuE;AACvE,SAAS,MAAM,CAAC,IAAiB,EAAE,GAAW;IAC5C,KAAK,MAAM,OAAO,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;QACvC,IAAI,OAAO,CAAC,IAAI,KAAK,MAAM;YAAE,SAAS;QACtC,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,OAAO,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC;QAC3E,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,EAAE,WAAW,EAAE,KAAK,GAAG,CAAC;YAAE,SAAS;QACjE,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC3E,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,8DAA8D;AAC9D,SAAS,OAAO,CAAC,IAAY;IAC3B,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IACpD,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACpC,IAAI,GAAG,GAAG,KAAK,CAAC,MAAM,GAAG,aAAa,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC;IAC9F,IAAI,GAAG,CAAC,MAAM,GAAG,kBAAkB;QAAE,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC,OAAO,EAAE,CAAC;IACtF,OAAO,GAAG,KAAK,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,GAAG,IAAI,CAAC;AACtD,CAAC;AAED,SAAS,MAAM,CAAC,MAAc;IAC5B,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;AAC/B,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@geekity/cms",
-  "version": "0.7.0",
+  "version": "0.9.0",
   "type": "module",
   "description": "A file-first CMS: Markdown content, a SQLite index, a Hono server that is both the editor and the public site.",
   "license": "MIT",

package/templates/site/content/_data/site.json

@@ -9,8 +9,14 @@
       { "label": "About", "url": "/about/" },
       { "label": "Search", "url": "/search/" }
     ],
-    "footer": [{ "label": "RSS", "url": "/feed/" }]
+    "footer": [
+      { "label": "RSS", "url": "/feed/" },
+      { "label": "Privacy", "url": "/privacy/" }
+    ]
   },
+  "commentEmailRetentionDays": 180,
+  "addressHashRetentionDays": 30,
+  "contactMessageRetentionDays": 365,
   "postsPerPage": 10,
   "feedSize": 20
 }

package/templates/site/content/pages/privacy.md

@@ -0,0 +1,127 @@
+---
+title: Privacy
+permalink: /privacy/
+description: What this site stores about the people who read it, write to it and comment on it, and for how long.
+---
+
+> **To the owner of this site.** `geekity init` wrote this page from what the
+> Geekity CMS does out of the box. It is a starting point, not legal advice,
+> and it knows nothing about your hosting, your analytics, your mail provider
+> or anything you added. Review it, correct it, and delete this note before
+> you rely on it. If you change a retention period on Settings > Discussion,
+> change the number here too. If you delete this page, delete the
+> `Privacy | /privacy/` line from the footer menu as well, so nothing links to
+> it.
+
+This page says what this site stores about you, where it keeps it, how long it
+keeps it, and who else sees it.
+
+## Reading the site
+
+This site sets no cookies on your browser while you read it, and it runs no
+analytics or advertising scripts. Its fonts, styles and scripts come from this
+site, not from another server.
+
+Pictures of other people, such as the avatar of somebody who replied from
+another site, are fetched and stored by this site and served from here. Your
+browser does not fetch them from the other server, so that server does not
+learn your IP address.
+
+When you follow a link from this site to another one, your browser tells that
+site only which site you came from, not which page.
+
+To slow down abuse, the site counts requests from each IP address for a few
+minutes. The counts are kept in memory only, and they are lost when the site
+restarts. The web server writes a log line for each request. By default the
+line does not contain your IP address. The owner can turn that on, and whoever
+hosts the site can keep its own logs.
+
+## Comments
+
+When you leave a comment, the site stores:
+
+- Your name, your website and the words you type. These are shown on the page
+  beside your comment, for as long as the comment is up.
+- Your email address, if you give one. It is optional, it is never shown, and
+  it is used to decide whether your comment needs a moderator and to email you
+  about replies if you ask for that. The site deletes it after 180 days.
+- A salted hash of your IP address, which lets the site spot a run of spam from
+  one place. The IP address itself is never stored. The site deletes the hash
+  after 30 days.
+
+If this site can send mail, a new comment waiting for a moderator is emailed
+to the people who moderate it.
+
+Comments are stored as files beside the posts, and the site may keep those
+files in version control. A copy deleted from the site can stay in that
+history until the owner rewrites it.
+
+**Reply emails (optional).** If this site can send mail and you tick "email me
+when somebody replies", the site emails you when a reply to your comment is
+approved. Each message has an unsubscribe link. If you unsubscribe, the site
+keeps your address on a list of addresses it must not email about replies.
+The mail provider the owner chose sees each address a message goes to.
+
+## The contact form (optional)
+
+If this site has a contact form and you send a message, the site stores your
+name, your email address, the subject, the message and a salted hash of your
+IP address, in a private file that is not published. If the site can send
+mail, it also emails the message to the owner, through the mail provider the
+owner chose. The site deletes the message after 365 days, and the hash after 30
+days.
+
+## Spam checking with Akismet (optional)
+
+Only if the owner turns it on, the site sends each new comment, webmention and
+contact message to Akismet, a spam-checking service run by Automattic. Akismet
+receives your IP address, your browser's user agent, the page you came from,
+your name, email address and website, and what you wrote. Akismet's own privacy
+policy covers what it does with them.
+
+## Replies from other sites
+
+**Webmentions.** When another site links to a post here and sends a
+webmention, this site fetches that page and stores its author's name, website
+and avatar address and the words of the page, and may show them as a comment.
+They are a copy of a page that is already public. When a post here links to
+another site, this site sends that site a webmention, which tells it only the
+address of the post.
+
+**The fediverse.** Posts here are published over ActivityPub, so they are
+copied to the servers of the people who follow this site, on Mastodon and
+similar services. If you follow this site from one of those services, the site
+stores your public profile: your handle, display name, avatar address and
+profile address, until you unfollow. That list is kept in the site's files
+beside the posts. When you reply to, like, boost or quote a
+post, the site stores the activity and who sent it, and may show replies under
+the post.
+
+## Accounts
+
+The people who write this site have accounts. The site stores each account's
+username, email address, a hash of its password and its public profile. A
+signed-in account gets one cookie, `__Host-geekity_session` (called
+`geekity_session` on a site without https), which holds a random session id and
+nothing else. It is deleted when the person signs out or the session expires.
+
+## How long the site keeps things
+
+| What                                             | How long           |
+| ------------------------------------------------ | ------------------ |
+| The email address on a comment                   | 180 days           |
+| The hash of the IP address on a comment or message | 30 days          |
+| A contact message                                | 365 days           |
+| Everything else above                            | Until it is deleted |
+
+The site checks these periods when it starts and every six hours after that.
+
+## Asking for your data to be removed
+
+Write to the owner of this site. The owner can find every comment and contact
+message that carries your email address and erase them together, on
+Tools > Personal data: your comments stay in their threads signed "Anonymous",
+without your email, website or address hash, your messages are deleted, and
+your address comes off the list of addresses not to email about replies.
+Comments left without an email address, webmentions and fediverse replies are
+removed one at a time.

package/themes/default/README.md

@@ -17,6 +17,8 @@ themes/default/
     author.njk   one person's archive, paginated
     search.njk   the search form and what it found, paginated
     404.njk      nothing at this URL
+    500.njk      the request failed on the server
+    503.njk      the site is in maintenance mode
   partials/
     post-list.njk     the h-feed a listing is made of
     pagination.njk    previous/next pager
@@ -373,6 +375,24 @@ alone.
 
 `layouts/404.njk` says `Content not found.` and links home and to the search.
 
+`layouts/500.njk` is the page a request gets when the server fails while
+answering it (TASK-129). It says `Something went wrong.` in an
+`article.server-error` and links home. The context is the 404's: `title`,
+`url` and `page.url`. The response is a 500 with `Cache-Control: no-store`, and
+the error itself goes to the server's log, never into the page. If this template
+fails too, the CMS answers with a plain HTML page of its own instead, so a theme
+cannot turn one failure into a crash. A request for the `.md` or `.json`
+representation gets a 500 in that format and never reaches this template.
+
+`layouts/503.njk` is the page a request gets while the site is in maintenance
+mode (TASK-130). It says `Down for maintenance.` in an `article.maintenance`.
+The context is the 404's, plus `until` (an ISO 8601 time) and `untilText` (the
+same time as an HTTP date) when the operator named when the site should be
+back. The response is a 503 with `Retry-After` and `Cache-Control: no-store`.
+The theme's files under `/theme/` are still served, so the page can use the
+stylesheet. If this template fails, the CMS answers with a plain HTML page of
+its own.
+
 ### Search
 
 `layouts/search.njk` is the page at `/search/` (TASK-22). It prints
@@ -1031,8 +1051,9 @@ Category archives get their own three the same way.
 ## The conversation
 
 `conversation` is what has been said about a post: the fediverse replies, likes
-and boosts its inbox was sent, the comments people left on the page itself, and
-the webmentions other pages sent it.
+and boosts its inbox was sent, the fediverse posts that quoted it with its
+author's approval, the comments people left on the page itself, and the
+webmentions other pages sent it.
 It is on the context of a rendered post **only when there is something in it**,
 so a post nobody has answered renders no empty section and a layout can simply
 ask:
@@ -1072,29 +1093,29 @@ the thread, because it is not an answer.
 | `replies`                                          | The replies to the post, oldest first, each carrying its own `replies`. See below.         |
 | `likes`                                            | The likes, oldest first, in the same shape.                                                |
 | `boosts`                                           | The boosts, oldest first, in the same shape. A webmention `repost` is one of them.         |
-| `mentions`                                         | The pages that linked here without answering, oldest first, in the same shape.             |
+| `mentions`                                         | The pages and fediverse quotes about this post, oldest first, in the same shape.           |
 | `counts.replies`                                   | How many replies, counted through the whole thread rather than the top of it.              |
 | `counts.likes`, `counts.boosts`, `counts.mentions` | How many of each.                                                                          |
 | `counts.total`                                     | All four added up. Zero never reaches a template: there would be no `conversation` at all. |
 
 Each entry — a reply, a like or a boost — is:
 
-| Key              | What it holds                                                                                                                                                                          |
-| ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `id`             | What it is called: a reply's own note id, which is what an answer to it names.                                                                                                         |
-| `source`         | `"activitypub"` for a fediverse reply, `"comment"` for one left on the page, `"webmention"` for another page linking here.                                                             |
-| `kind`           | `"reply"`, `"like"`, `"boost"`, `"repost"` or `"mention"`. The last two only ever come from a webmention.                                                                              |
-| `author.name`    | The best name available: their display name, else their handle, else their id. For a webmention, the source's `h-card` name, else its host.                                            |
-| `author.handle`  | `@user@host`, or `null`. Taken from the follower profile the site holds, else guessed from the actor URL. `null` for a native comment.                                                 |
-| `author.url`     | Their profile page, the website a commenter typed, or a webmention author's `u-url`. May be `null`, so guard the link.                                                                 |
-| `author.avatar`  | Their avatar, or `null`. The site knows one for an actor that follows it and for a webmention whose `h-card` carried a `u-photo`.                                                      |
-| `author.actorId` | Their id, which is what identifies them however they are named. `null` for a native comment.                                                                                           |
-| `url`            | Where it can be read: the remote note's `url` for a fediverse reply, the source page for a webmention, and `{permalink}#comment-{id}` — this page's own anchor — for a native comment. |
-| `content`        | What it says, **already sanitised**, so print it with `\| safe`. Empty for a like or a boost.                                                                                          |
-| `published`      | A `Date`: when it was published, or when it arrived if it did not say. Use the `date` filter.                                                                                          |
-| `inReplyTo`      | What it answers — the post's ActivityPub id, or another reply's — and `null` for a reaction.                                                                                           |
-| `status`         | `"published"`. On the record for the sources that moderate.                                                                                                                            |
-| `replies`        | The replies to this one, oldest first, nested as deep as the site has seen.                                                                                                            |
+| Key              | What it holds                                                                                                                                                                                                                     |
+| ---------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `id`             | What it is called: a reply's own note id, which is what an answer to it names.                                                                                                                                                    |
+| `source`         | `"activitypub"` for a fediverse reply, `"comment"` for one left on the page, `"webmention"` for another page linking here.                                                                                                        |
+| `kind`           | `"reply"`, `"like"`, `"boost"`, `"repost"` or `"mention"`. A `repost` comes from a webmention; a `mention` is a webmention or an approved fediverse quote.                                                                        |
+| `author.name`    | The best name available: their display name, else their handle, else their id. For a webmention, the source's `h-card` name, else its host.                                                                                       |
+| `author.handle`  | `@user@host`, or `null`. Taken from the follower profile the site holds, else guessed from the actor URL. `null` for a native comment.                                                                                            |
+| `author.url`     | Their profile page, the website a commenter typed, or a webmention author's `u-url`. May be `null`, so guard the link.                                                                                                            |
+| `author.avatar`  | Their avatar, or `null`. The site knows one for an actor that follows it and for a webmention whose `h-card` carried a `u-photo`. Always a same-origin `/_geekity/avatars/…` path, never the remote URL: see [Avatars](#avatars). |
+| `author.actorId` | Their id, which is what identifies them however they are named. `null` for a native comment.                                                                                                                                      |
+| `url`            | Where it can be read: the remote note's `url` for a fediverse reply or quote, the source page for a webmention, and `{permalink}#comment-{id}` — this page's own anchor — for a native comment.                                   |
+| `content`        | What it says, **already sanitised**, so print it with `\| safe`. Empty for a like or a boost.                                                                                                                                     |
+| `published`      | A `Date`: when it was published, or when it arrived if it did not say. Use the `date` filter.                                                                                                                                     |
+| `inReplyTo`      | What it answers — the post's ActivityPub id, or another reply's — and `null` for a reaction.                                                                                                                                      |
+| `status`         | `"published"`. On the record for the sources that moderate.                                                                                                                                                                       |
+| `replies`        | The replies to this one, oldest first, nested as deep as the site has seen.                                                                                                                                                       |
 
 Three rules decide what is in the thread, and they are the CMS's rather than a
 theme's: a reply whose author deleted it is gone, and its own answers move up to
@@ -1151,6 +1172,24 @@ and then loops over exactly the keys above. The slug is what names the post's
 comment file under `content/_data/comments/`; the format is documented in
 backlog doc-6.
 
+### Avatars
+
+A reader's browser never fetches a face from the server it lives on. The
+`author.avatar` a theme is handed is a path on this site, under
+`/_geekity/avatars/`, and the CMS fetches the picture, crops it to an
+80-pixel square WebP (twice the 40 pixels this theme shows a commenter at)
+and keeps it under `data/avatars/`. Print it as it is; it needs no filter.
+
+The route serves only the pictures the site shows: a follower who has
+liked, boosted, quoted or answered a post, and the author of an approved
+comment or webmention. Any other URL is a 404, and nothing is fetched for it.
+Pictures are fetched in the background when the site starts and every six
+hours after, refreshed once they are a week old, and forgotten once nothing
+shows them. A picture that cannot be fetched, is over 2 MB, or is not a JPEG,
+PNG, GIF, WebP or AVIF is served as a grey placeholder, and tried again an
+hour later. `data/avatars/` is a cache: delete it and every picture is fetched
+again.
+
 ## The comment form
 
 `commentForm` is on the context of a rendered post **only when that post is

package/themes/default/layouts/500.njk

@@ -0,0 +1,8 @@
+{% extends "layouts/base.njk" %}
+
+{% block content %}
+<article class="server-error">
+  <h1 class="page-title">Something went wrong.</h1>
+  <p>This page could not be shown because of a problem on the server. Try again in a moment, or go <a href="{{ "/" | url }}">home</a>.</p>
+</article>
+{% endblock %}

package/themes/default/layouts/503.njk

@@ -0,0 +1,8 @@
+{% extends "layouts/base.njk" %}
+
+{% block content %}
+<article class="maintenance">
+  <h1 class="page-title">Down for maintenance.</h1>
+  <p>This site is down for maintenance and will be back shortly.{% if until %} It should be back by <time datetime="{{ until }}">{{ untilText }}</time>.{% endif %} Please try again then.</p>
+</article>
+{% endblock %}