@geekity/cms 0.3.0 → 0.4.0

package/dist/webmention/discovery.js

@@ -0,0 +1,176 @@
+import { elementsIn, hasRel, parseHtml } from './html.js';
+/**
+ * Finding out where to send a webmention.
+ *
+ * The W3C spec fixes the order and it matters, because a page may advertise
+ * two: the HTTP `Link` header first, then the first `<link>` **or** `<a>`
+ * carrying `rel="webmention"` in document order — one search over both, not
+ * one search each. Whatever is found is resolved against the URL the request
+ * actually ended at, so a target that redirects to its canonical home
+ * advertises that home's endpoint rather than a path under the old one.
+ */
+/** How long one discovery request is given before it is abandoned. */
+export const DISCOVERY_TIMEOUT_MS = 10_000;
+/** How much of a target page is read while looking for its endpoint. */
+export const DISCOVERY_MAX_BYTES = 1_000_000;
+/** What a sender says it is. */
+export const WEBMENTION_USER_AGENT = 'Geekity/1.0 (+https://github.com/andrewshell/geekity)';
+/**
+ * The Webmention endpoint a page advertises, or `undefined` when it advertises
+ * none — which is most pages, and is not an error.
+ *
+ * Nothing throws: a target that times out, refuses the connection or answers
+ * an error is a target with no endpoint as far as the sender is concerned, and
+ * the caller records that rather than stopping.
+ */
+export async function discoverEndpoint(target) {
+    let response;
+    try {
+        response = await fetch(target, {
+            headers: { accept: 'text/html, */*;q=0.8', 'user-agent': WEBMENTION_USER_AGENT },
+            redirect: 'follow',
+            signal: AbortSignal.timeout(DISCOVERY_TIMEOUT_MS),
+        });
+    }
+    catch {
+        return undefined;
+    }
+    if (!response.ok)
+        return undefined;
+    // Where the request ended up, which is what a relative endpoint is relative
+    // to. `Response.url` is empty for a response nothing redirected, in which
+    // case the URL asked for is the URL arrived at.
+    const arrivedAt = response.url === '' ? target : response.url;
+    const header = endpointInHeader(response.headers.get('link'), arrivedAt);
+    if (header !== undefined) {
+        // The body is not read, but it has to be released or the socket is held
+        // until the process notices.
+        await response.body?.cancel();
+        return header;
+    }
+    const type = response.headers.get('content-type') ?? '';
+    if (!/^\s*(text\/html|application\/xhtml\+xml)/i.test(type)) {
+        await response.body?.cancel();
+        return undefined;
+    }
+    return endpointInHtml(await readCapped(response), arrivedAt);
+}
+/**
+ * The endpoint one or more `Link` headers name, or `undefined`.
+ *
+ * A header may carry several links separated by commas and each may name
+ * several relations, so both are split. Splitting on commas that are not
+ * inside angle brackets or quotes is what keeps a URL holding a comma — which
+ * is legal — from being torn in half.
+ */
+export function endpointInHeader(header, base) {
+    if (header === null)
+        return undefined;
+    for (const value of splitLinkHeader(header)) {
+        const match = /^\s*<([^>]*)>\s*(.*)$/.exec(value);
+        if (match === null)
+            continue;
+        const relations = /(?:^|;)\s*rel\s*=\s*("([^"]*)"|'([^']*)'|([^;\s]*))/i.exec(match[2] ?? '');
+        const rel = relations?.[2] ?? relations?.[3] ?? relations?.[4] ?? '';
+        if (!rel.split(/\s+/).some((token) => token.toLowerCase() === 'webmention'))
+            continue;
+        return resolve(match[1] ?? '', base);
+    }
+    return undefined;
+}
+/**
+ * The endpoint a page's markup names, or `undefined`.
+ *
+ * One walk over every element rather than one for `<link>` and another for
+ * `<a>`: the spec asks for the first of either in document order, and two
+ * walks would answer with whichever kind was searched for first.
+ */
+export function endpointInHtml(html, base) {
+    for (const element of elementsIn(parseHtml(html))) {
+        if (element.name !== 'link' && element.name !== 'a')
+            continue;
+        if (!hasRel(element, 'webmention'))
+            continue;
+        // An `href` that is there but empty means the page itself, which is what
+        // an endpoint on the same URL as the target looks like.
+        return resolve(element.attributes['href'] ?? '', base);
+    }
+    return undefined;
+}
+/** One `Link` header value split into its links, commas inside a URL kept. */
+function splitLinkHeader(header) {
+    const values = [];
+    let start = 0;
+    let inBrackets = false;
+    let quote;
+    for (let at = 0; at < header.length; at += 1) {
+        const character = header[at];
+        if (quote !== undefined) {
+            if (character === quote)
+                quote = undefined;
+            continue;
+        }
+        if (character === '"' || character === "'")
+            quote = character;
+        else if (character === '<')
+            inBrackets = true;
+        else if (character === '>')
+            inBrackets = false;
+        else if (character === ',' && !inBrackets) {
+            values.push(header.slice(start, at));
+            start = at + 1;
+        }
+    }
+    values.push(header.slice(start));
+    return values;
+}
+/** An advertised endpoint as an absolute URL, or `undefined` when it is not one. */
+function resolve(href, base) {
+    try {
+        const url = new URL(href.trim(), base);
+        return url.protocol === 'http:' || url.protocol === 'https:' ? url.href : undefined;
+    }
+    catch {
+        return undefined;
+    }
+}
+/**
+ * A response body, up to {@link DISCOVERY_MAX_BYTES}.
+ *
+ * A cap rather than `response.text()`, because the page at the other end is a
+ * stranger's and nothing stops it being a gigabyte. What is read is more than
+ * enough markup for a `<link>` in the head.
+ */
+export async function readCapped(response) {
+    const body = response.body;
+    if (body === null)
+        return '';
+    const reader = body.getReader();
+    const decoder = new TextDecoder();
+    const chunks = [];
+    let read = 0;
+    try {
+        for (;;) {
+            const chunk = await reader.read();
+            if (chunk.done)
+                break;
+            const value = chunk.value;
+            read += value.length;
+            chunks.push(decoder.decode(value, { stream: true }));
+            if (read >= DISCOVERY_MAX_BYTES)
+                break;
+        }
+    }
+    catch {
+        // A body that stops mid-stream is whatever arrived before it stopped.
+    }
+    finally {
+        await reader.cancel().catch(ignore);
+    }
+    chunks.push(decoder.decode());
+    return chunks.join('');
+}
+function ignore() {
+    // Deliberately empty: a body that will not close is not worth an error.
+}
+//# sourceMappingURL=discovery.js.map

package/dist/webmention/discovery.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"discovery.js","sourceRoot":"","sources":["../../src/webmention/discovery.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AAE1D;;;;;;;;;GASG;AAEH,sEAAsE;AACtE,MAAM,CAAC,MAAM,oBAAoB,GAAG,MAAM,CAAC;AAE3C,wEAAwE;AACxE,MAAM,CAAC,MAAM,mBAAmB,GAAG,SAAS,CAAC;AAE7C,gCAAgC;AAChC,MAAM,CAAC,MAAM,qBAAqB,GAAG,uDAAuD,CAAC;AAE7F;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,MAAc;IACnD,IAAI,QAAkB,CAAC;IACvB,IAAI,CAAC;QACH,QAAQ,GAAG,MAAM,KAAK,CAAC,MAAM,EAAE;YAC7B,OAAO,EAAE,EAAE,MAAM,EAAE,sBAAsB,EAAE,YAAY,EAAE,qBAAqB,EAAE;YAChF,QAAQ,EAAE,QAAQ;YAClB,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,oBAAoB,CAAC;SAClD,CAAC,CAAC;IACL,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,IAAI,CAAC,QAAQ,CAAC,EAAE;QAAE,OAAO,SAAS,CAAC;IAEnC,4EAA4E;IAC5E,0EAA0E;IAC1E,gDAAgD;IAChD,MAAM,SAAS,GAAG,QAAQ,CAAC,GAAG,KAAK,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC;IAE9D,MAAM,MAAM,GAAG,gBAAgB,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,SAAS,CAAC,CAAC;IACzE,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;QACzB,wEAAwE;QACxE,6BAA6B;QAC7B,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC;QAC9B,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC;IACxD,IAAI,CAAC,2CAA2C,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC5D,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC;QAC9B,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,OAAO,cAAc,CAAC,MAAM,UAAU,CAAC,QAAQ,CAAC,EAAE,SAAS,CAAC,CAAC;AAC/D,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,gBAAgB,CAAC,MAAqB,EAAE,IAAY;IAClE,IAAI,MAAM,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,eAAe,CAAC,MAAM,CAAC,EAAE,CAAC;QAC5C,MAAM,KAAK,GAAG,uBAAuB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAClD,IAAI,KAAK,KAAK,IAAI;YAAE,SAAS;QAE7B,MAAM,SAAS,GAAG,sDAAsD,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAC9F,MAAM,GAAG,GAAG,SAAS,EAAE,CAAC,CAAC,CAAC,IAAI,SAAS,EAAE,CAAC,CAAC,CAAC,IAAI,SAAS,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACrE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,YAAY,CAAC;YAAE,SAAS;QAEtF,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,IAAI,CAAC,CAAC;IACvC,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,cAAc,CAAC,IAAY,EAAE,IAAY;IACvD,KAAK,MAAM,OAAO,IAAI,UAAU,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QAClD,IAAI,OAAO,CAAC,IAAI,KAAK,MAAM,IAAI,OAAO,CAAC,IAAI,KAAK,GAAG;YAAE,SAAS;QAC9D,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,YAAY,CAAC;YAAE,SAAS;QAE7C,yEAAyE;QACzE,wDAAwD;QACxD,OAAO,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,EAAE,EAAE,IAAI,CAAC,CAAC;IACzD,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,8EAA8E;AAC9E,SAAS,eAAe,CAAC,MAAc;IACrC,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,UAAU,GAAG,KAAK,CAAC;IACvB,IAAI,KAAyB,CAAC;IAE9B,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,MAAM,CAAC,MAAM,EAAE,EAAE,IAAI,CAAC,EAAE,CAAC;QAC7C,MAAM,SAAS,GAAG,MAAM,CAAC,EAAE,CAAC,CAAC;QAC7B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,IAAI,SAAS,KAAK,KAAK;gBAAE,KAAK,GAAG,SAAS,CAAC;YAC3C,SAAS;QACX,CAAC;QACD,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG;YAAE,KAAK,GAAG,SAAS,CAAC;aACzD,IAAI,SAAS,KAAK,GAAG;YAAE,UAAU,GAAG,IAAI,CAAC;aACzC,IAAI,SAAS,KAAK,GAAG;YAAE,UAAU,GAAG,KAAK,CAAC;aAC1C,IAAI,SAAS,KAAK,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;YAC1C,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,CAAC;YACrC,KAAK,GAAG,EAAE,GAAG,CAAC,CAAC;QACjB,CAAC;IACH,CAAC;IAED,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC;IACjC,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,oFAAoF;AACpF,SAAS,OAAO,CAAC,IAAY,EAAE,IAAY;IACzC,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,IAAI,CAAC,CAAC;QACvC,OAAO,GAAG,CAAC,QAAQ,KAAK,OAAO,IAAI,GAAG,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC;IACtF,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,QAAkB;IACjD,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,IAAI,CAAC;QACH,SAAS,CAAC;YACR,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;YAClC,IAAI,KAAK,CAAC,IAAI;gBAAE,MAAM;YACtB,MAAM,KAAK,GAAe,KAAK,CAAC,KAAK,CAAC;YACtC,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC;YACrB,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;YACrD,IAAI,IAAI,IAAI,mBAAmB;gBAAE,MAAM;QACzC,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,sEAAsE;IACxE,CAAC;YAAS,CAAC;QACT,MAAM,MAAM,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACtC,CAAC;IAED,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAC9B,OAAO,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACzB,CAAC;AAED,SAAS,MAAM;IACb,wEAAwE;AAC1E,CAAC"}

package/dist/webmention/html.d.ts

@@ -0,0 +1,71 @@
+/**
+ * Somebody else's page, as a tree small enough to ask questions of.
+ *
+ * A webmention is a claim about a page this site does not control, and both
+ * halves of checking it are questions about that page's markup: does it link
+ * to the target, and what do its microformats say. Neither can be answered by
+ * a regular expression over the source — a `class` list, an `href` and the
+ * text of an element are all nested facts — so the source is read once into a
+ * tree and asked afterwards.
+ *
+ * It is deliberately not a conforming HTML parser and never will be. It has no
+ * table foster-parenting, no adoption agency, no character encoding detection:
+ * what it has is elements, attributes and text, nested the way a page that is
+ * not broken nests them, which is what {@link elementsIn} and
+ * {@link textOf} need. `web/sanitize.ts` scans markup with the same shape of
+ * scanner for a different job — it rebuilds output rather than building a
+ * tree — and the two are kept apart rather than shared because one of them is
+ * a security boundary and should not grow features for the other's sake.
+ */
+/** One element of a parsed page. */
+export interface HtmlElement {
+    /** Its tag name, lower-cased. */
+    readonly name: string;
+    /** Its attributes, names lower-cased and values with entities resolved. */
+    readonly attributes: Readonly<Record<string, string>>;
+    /** What is inside it, in document order. */
+    readonly children: HtmlNode[];
+}
+/** A run of text between tags, with its character references resolved. */
+export interface HtmlText {
+    /** The text itself. */
+    readonly text: string;
+}
+/** Either of the two things a tree is made of. */
+export type HtmlNode = HtmlElement | HtmlText;
+/** Whether a node is an element rather than a run of text. */
+export declare function isElement(node: HtmlNode): node is HtmlElement;
+/**
+ * A page as a tree, under one synthetic root standing for the document.
+ *
+ * Nothing throws. Markup this scanner cannot make sense of contributes
+ * whatever it can and no more, because the alternative — refusing a page — is
+ * refusing a webmention over somebody else's typo.
+ */
+export declare function parseHtml(source: string): HtmlElement;
+/** Every element under `root`, itself excluded, in document order. */
+export declare function elementsIn(root: HtmlElement): Generator<HtmlElement>;
+/**
+ * The text an element holds, everything nested in it included, with runs of
+ * whitespace collapsed.
+ *
+ * Collapsed because it is read as a name or a title, and a page that wrote one
+ * across three indented lines meant one line.
+ */
+export declare function textOf(node: HtmlNode): string;
+/** The text an element holds, exactly as the page spelled it. */
+export declare function rawTextOf(node: HtmlNode): string;
+/**
+ * An element's inner markup, rebuilt from the tree.
+ *
+ * Rebuilt rather than sliced out of the source because the tree is what this
+ * module trusts: an unclosed element in the page cannot leave a stray closing
+ * tag in the result. It goes through `sanitizeCommentHtml` before it is stored
+ * either way, so this only has to be close enough for that to read.
+ */
+export declare function innerHtmlOf(element: HtmlElement): string;
+/** The classes an element carries, in the order it wrote them. */
+export declare function classesOf(element: HtmlElement): string[];
+/** Whether an element's `rel` names this link relation. */
+export declare function hasRel(element: HtmlElement, relation: string): boolean;
+//# sourceMappingURL=html.d.ts.map

package/dist/webmention/html.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html.d.ts","sourceRoot":"","sources":["../../src/webmention/html.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,oCAAoC;AACpC,MAAM,WAAW,WAAW;IAC1B,iCAAiC;IACjC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,2EAA2E;IAC3E,QAAQ,CAAC,UAAU,EAAE,QAAQ,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC;IACtD,4CAA4C;IAC5C,QAAQ,CAAC,QAAQ,EAAE,QAAQ,EAAE,CAAC;CAC/B;AAED,0EAA0E;AAC1E,MAAM,WAAW,QAAQ;IACvB,uBAAuB;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;CACvB;AAED,kDAAkD;AAClD,MAAM,MAAM,QAAQ,GAAG,WAAW,GAAG,QAAQ,CAAC;AAE9C,8DAA8D;AAC9D,wBAAgB,SAAS,CAAC,IAAI,EAAE,QAAQ,GAAG,IAAI,IAAI,WAAW,CAE7D;AA4DD;;;;;;GAMG;AACH,wBAAgB,SAAS,CAAC,MAAM,EAAE,MAAM,GAAG,WAAW,CA8DrD;AAED,sEAAsE;AACtE,wBAAiB,UAAU,CAAC,IAAI,EAAE,WAAW,GAAG,SAAS,CAAC,WAAW,CAAC,CAMrE;AAED;;;;;;GAMG;AACH,wBAAgB,MAAM,CAAC,IAAI,EAAE,QAAQ,GAAG,MAAM,CAE7C;AAED,iEAAiE;AACjE,wBAAgB,SAAS,CAAC,IAAI,EAAE,QAAQ,GAAG,MAAM,CAGhD;AAED;;;;;;;GAOG;AACH,wBAAgB,WAAW,CAAC,OAAO,EAAE,WAAW,GAAG,MAAM,CAExD;AAaD,kEAAkE;AAClE,wBAAgB,SAAS,CAAC,OAAO,EAAE,WAAW,GAAG,MAAM,EAAE,CAGxD;AAED,2DAA2D;AAC3D,wBAAgB,MAAM,CAAC,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,GAAG,OAAO,CAItE"}

package/dist/webmention/html.js

@@ -0,0 +1,285 @@
+/**
+ * Somebody else's page, as a tree small enough to ask questions of.
+ *
+ * A webmention is a claim about a page this site does not control, and both
+ * halves of checking it are questions about that page's markup: does it link
+ * to the target, and what do its microformats say. Neither can be answered by
+ * a regular expression over the source — a `class` list, an `href` and the
+ * text of an element are all nested facts — so the source is read once into a
+ * tree and asked afterwards.
+ *
+ * It is deliberately not a conforming HTML parser and never will be. It has no
+ * table foster-parenting, no adoption agency, no character encoding detection:
+ * what it has is elements, attributes and text, nested the way a page that is
+ * not broken nests them, which is what {@link elementsIn} and
+ * {@link textOf} need. `web/sanitize.ts` scans markup with the same shape of
+ * scanner for a different job — it rebuilds output rather than building a
+ * tree — and the two are kept apart rather than shared because one of them is
+ * a security boundary and should not grow features for the other's sake.
+ */
+/** Whether a node is an element rather than a run of text. */
+export function isElement(node) {
+    return 'name' in node;
+}
+/** Elements with no content of their own, so nothing ever closes them. */
+const VOID_ELEMENTS = new Set([
+    'area',
+    'base',
+    'br',
+    'col',
+    'embed',
+    'hr',
+    'img',
+    'input',
+    'link',
+    'meta',
+    'param',
+    'source',
+    'track',
+    'wbr',
+]);
+/**
+ * Elements whose content is not markup.
+ *
+ * A `<` inside one of them is a character in a program or a stylesheet, so
+ * everything up to the closing tag is skipped rather than parsed. Their text
+ * is dropped as well: an author name taken out of a script is not an author
+ * name.
+ */
+const RAW_TEXT = new Set(['script', 'style']);
+/**
+ * Elements an opening tag of the same name implicitly closes.
+ *
+ * Without this, a page written as `<p>one<p>two` would nest, and an `h-entry`
+ * looking for the properties directly inside it would find the second
+ * paragraph inside the first.
+ */
+const CLOSED_BY_ITSELF = new Set([
+    'p',
+    'li',
+    'dt',
+    'dd',
+    'option',
+    'tr',
+    'td',
+    'th',
+]);
+/**
+ * A page as a tree, under one synthetic root standing for the document.
+ *
+ * Nothing throws. Markup this scanner cannot make sense of contributes
+ * whatever it can and no more, because the alternative — refusing a page — is
+ * refusing a webmention over somebody else's typo.
+ */
+export function parseHtml(source) {
+    const root = { name: '#document', attributes: {}, children: [] };
+    const open = [root];
+    let at = 0;
+    /** Where a node goes: inside the innermost element still open. */
+    function current() {
+        return open[open.length - 1] ?? root;
+    }
+    /** Close open elements down to and including `name`, if it is open at all. */
+    function closeThrough(name) {
+        const index = open.findLastIndex((element) => element.name === name);
+        // Never the root, which nothing closes.
+        if (index > 0)
+            open.length = index;
+    }
+    while (at < source.length) {
+        const next = source.indexOf('<', at);
+        if (next < 0) {
+            pushText(current(), source.slice(at));
+            break;
+        }
+        if (next > at)
+            pushText(current(), source.slice(at, next));
+        at = next;
+        if (source.startsWith('<!--', at)) {
+            at = after(source, '-->', at);
+            continue;
+        }
+        if (source.startsWith('<!', at) || source.startsWith('<?', at)) {
+            at = after(source, '>', at);
+            continue;
+        }
+        const tag = readTag(source, at);
+        if (tag === undefined) {
+            // A `<` that begins no tag is a character somebody typed.
+            pushText(current(), '<');
+            at += 1;
+            continue;
+        }
+        at = tag.end;
+        if (tag.closing) {
+            closeThrough(tag.name);
+            continue;
+        }
+        if (RAW_TEXT.has(tag.name)) {
+            at = afterElement(source, tag.name, at);
+            continue;
+        }
+        if (CLOSED_BY_ITSELF.has(tag.name) && current().name === tag.name)
+            closeThrough(tag.name);
+        const element = { name: tag.name, attributes: tag.attributes, children: [] };
+        current().children.push(element);
+        if (!VOID_ELEMENTS.has(tag.name))
+            open.push(element);
+    }
+    return root;
+}
+/** Every element under `root`, itself excluded, in document order. */
+export function* elementsIn(root) {
+    for (const child of root.children) {
+        if (!isElement(child))
+            continue;
+        yield child;
+        yield* elementsIn(child);
+    }
+}
+/**
+ * The text an element holds, everything nested in it included, with runs of
+ * whitespace collapsed.
+ *
+ * Collapsed because it is read as a name or a title, and a page that wrote one
+ * across three indented lines meant one line.
+ */
+export function textOf(node) {
+    return rawTextOf(node).replace(/\s+/g, ' ').trim();
+}
+/** The text an element holds, exactly as the page spelled it. */
+export function rawTextOf(node) {
+    if (!isElement(node))
+        return node.text;
+    return node.children.map(rawTextOf).join('');
+}
+/**
+ * An element's inner markup, rebuilt from the tree.
+ *
+ * Rebuilt rather than sliced out of the source because the tree is what this
+ * module trusts: an unclosed element in the page cannot leave a stray closing
+ * tag in the result. It goes through `sanitizeCommentHtml` before it is stored
+ * either way, so this only has to be close enough for that to read.
+ */
+export function innerHtmlOf(element) {
+    return element.children.map(markupOf).join('');
+}
+/** One node as markup. */
+function markupOf(node) {
+    if (!isElement(node))
+        return escapeText(node.text);
+    const attributes = Object.entries(node.attributes)
+        .map(([name, value]) => ` ${name}="${escapeText(value)}"`)
+        .join('');
+    if (VOID_ELEMENTS.has(node.name))
+        return `<${node.name}${attributes}>`;
+    return `<${node.name}${attributes}>${innerHtmlOf(node)}</${node.name}>`;
+}
+/** The classes an element carries, in the order it wrote them. */
+export function classesOf(element) {
+    const value = element.attributes['class'];
+    return value === undefined ? [] : value.split(/\s+/).filter((name) => name !== '');
+}
+/** Whether an element's `rel` names this link relation. */
+export function hasRel(element, relation) {
+    const value = element.attributes['rel'];
+    if (value === undefined)
+        return false;
+    return value.split(/\s+/).some((token) => token.toLowerCase() === relation);
+}
+/** Add a run of text, unless it is nothing at all. */
+function pushText(parent, text) {
+    if (text === '')
+        return;
+    parent.children.push({ text: decodeEntities(text) });
+}
+/** Where the source continues after the next `marker`, or its end. */
+function after(source, marker, at) {
+    const end = source.indexOf(marker, at);
+    return end < 0 ? source.length : end + marker.length;
+}
+/** Where the source continues after this element's closing tag, or its end. */
+function afterElement(source, name, at) {
+    const closing = new RegExp(`</${name}\\s*>`, 'i');
+    const match = closing.exec(source.slice(at));
+    return match === null ? source.length : at + match.index + match[0].length;
+}
+/** The tag beginning at `at`, or `undefined` when nothing there is a tag. */
+function readTag(source, at) {
+    const match = /^<(\/?)([A-Za-z][A-Za-z0-9-]*)/.exec(source.slice(at));
+    if (match === null)
+        return undefined;
+    const name = (match[2] ?? '').toLowerCase();
+    const closing = match[1] === '/';
+    const attributes = {};
+    let cursor = at + match[0].length;
+    for (;;) {
+        while (cursor < source.length && /\s/.test(source[cursor] ?? ''))
+            cursor += 1;
+        if (cursor >= source.length)
+            return { name, closing, attributes, end: source.length };
+        if (source.startsWith('/>', cursor))
+            return { name, closing, attributes, end: cursor + 2 };
+        if (source[cursor] === '>')
+            return { name, closing, attributes, end: cursor + 1 };
+        const attribute = /^([^\s/>=]+)(\s*=\s*("[^"]*"|'[^']*'|[^\s>]*))?/.exec(source.slice(cursor));
+        if (attribute === null) {
+            // Nothing that could be an attribute and no `>`: the tag is malformed,
+            // so it ends where the source does.
+            return { name, closing, attributes, end: source.length };
+        }
+        const raw = attribute[3];
+        attributes[(attribute[1] ?? '').toLowerCase()] =
+            raw === undefined ? '' : decodeEntities(unquote(raw));
+        cursor += attribute[0].length;
+    }
+}
+/** An attribute value with its quotes, if any, taken off. */
+function unquote(value) {
+    const first = value[0];
+    return (first === '"' || first === "'") && value.endsWith(first) ? value.slice(1, -1) : value;
+}
+/** The character references this module resolves. */
+const ENTITIES = {
+    amp: '&',
+    lt: '<',
+    gt: '>',
+    quot: '"',
+    apos: "'",
+    nbsp: ' ',
+};
+/** Character references resolved; one this module does not know is left alone. */
+function decodeEntities(value) {
+    if (!value.includes('&'))
+        return value;
+    return value.replace(/&(#[0-9]+|#[xX][0-9a-fA-F]+|[A-Za-z][A-Za-z0-9]*);/g, (whole, body) => {
+        if (body.startsWith('#x') || body.startsWith('#X')) {
+            return codePoint(Number.parseInt(body.slice(2), 16)) ?? whole;
+        }
+        if (body.startsWith('#'))
+            return codePoint(Number.parseInt(body.slice(1), 10)) ?? whole;
+        return ENTITIES[body] ?? whole;
+    });
+}
+/** One code point as a string, or `undefined` when it names no character. */
+function codePoint(value) {
+    if (!Number.isInteger(value) || value < 0 || value > 0x10ffff)
+        return undefined;
+    if (value >= 0xd800 && value <= 0xdfff)
+        return undefined;
+    try {
+        return String.fromCodePoint(value);
+    }
+    catch {
+        return undefined;
+    }
+}
+/** Text as markup again, so a rebuilt fragment says what the page said. */
+function escapeText(value) {
+    return value
+        .replaceAll('&', '&amp;')
+        .replaceAll('<', '&lt;')
+        .replaceAll('>', '&gt;')
+        .replaceAll('"', '&quot;');
+}
+//# sourceMappingURL=html.js.map

package/dist/webmention/html.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html.js","sourceRoot":"","sources":["../../src/webmention/html.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAqBH,8DAA8D;AAC9D,MAAM,UAAU,SAAS,CAAC,IAAc;IACtC,OAAO,MAAM,IAAI,IAAI,CAAC;AACxB,CAAC;AAED,0EAA0E;AAC1E,MAAM,aAAa,GAAwB,IAAI,GAAG,CAAC;IACjD,MAAM;IACN,MAAM;IACN,IAAI;IACJ,KAAK;IACL,OAAO;IACP,IAAI;IACJ,KAAK;IACL,OAAO;IACP,MAAM;IACN,MAAM;IACN,OAAO;IACP,QAAQ;IACR,OAAO;IACP,KAAK;CACN,CAAC,CAAC;AAEH;;;;;;;GAOG;AACH,MAAM,QAAQ,GAAwB,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC;AAEnE;;;;;;GAMG;AACH,MAAM,gBAAgB,GAAwB,IAAI,GAAG,CAAC;IACpD,GAAG;IACH,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,QAAQ;IACR,IAAI;IACJ,IAAI;IACJ,IAAI;CACL,CAAC,CAAC;AAcH;;;;;;GAMG;AACH,MAAM,UAAU,SAAS,CAAC,MAAc;IACtC,MAAM,IAAI,GAAgB,EAAE,IAAI,EAAE,WAAW,EAAE,UAAU,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;IAC9E,MAAM,IAAI,GAAkB,CAAC,IAAI,CAAC,CAAC;IACnC,IAAI,EAAE,GAAG,CAAC,CAAC;IAEX,kEAAkE;IAClE,SAAS,OAAO;QACd,OAAO,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC;IACvC,CAAC;IAED,8EAA8E;IAC9E,SAAS,YAAY,CAAC,IAAY;QAChC,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;QACrE,wCAAwC;QACxC,IAAI,KAAK,GAAG,CAAC;YAAE,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC;IACrC,CAAC;IAED,OAAO,EAAE,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QAC1B,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACrC,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YACb,QAAQ,CAAC,OAAO,EAAE,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC;YACtC,MAAM;QACR,CAAC;QACD,IAAI,IAAI,GAAG,EAAE;YAAE,QAAQ,CAAC,OAAO,EAAE,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC,CAAC;QAC3D,EAAE,GAAG,IAAI,CAAC;QAEV,IAAI,MAAM,CAAC,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC,EAAE,CAAC;YAClC,EAAE,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;YAC9B,SAAS;QACX,CAAC;QACD,IAAI,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,EAAE,CAAC,EAAE,CAAC;YAC/D,EAAE,GAAG,KAAK,CAAC,MAAM,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC;YAC5B,SAAS;QACX,CAAC;QAED,MAAM,GAAG,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;QAChC,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACtB,0DAA0D;YAC1D,QAAQ,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC;YACzB,EAAE,IAAI,CAAC,CAAC;YACR,SAAS;QACX,CAAC;QACD,EAAE,GAAG,GAAG,CAAC,GAAG,CAAC;QAEb,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC;YAChB,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACvB,SAAS;QACX,CAAC;QAED,IAAI,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3B,EAAE,GAAG,YAAY,CAAC,MAAM,EAAE,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACxC,SAAS;QACX,CAAC;QAED,IAAI,gBAAgB,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,OAAO,EAAE,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI;YAAE,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAE1F,MAAM,OAAO,GAAgB,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,UAAU,EAAE,GAAG,CAAC,UAAU,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;QAC1F,OAAO,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACjC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC;YAAE,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACvD,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,sEAAsE;AACtE,MAAM,SAAS,CAAC,CAAC,UAAU,CAAC,IAAiB;IAC3C,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC;YAAE,SAAS;QAChC,MAAM,KAAK,CAAC;QACZ,KAAK,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;IAC3B,CAAC;AACH,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,MAAM,CAAC,IAAc;IACnC,OAAO,SAAS,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACrD,CAAC;AAED,iEAAiE;AACjE,MAAM,UAAU,SAAS,CAAC,IAAc;IACtC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC,IAAI,CAAC;IACvC,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC/C,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,WAAW,CAAC,OAAoB;IAC9C,OAAO,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACjD,CAAC;AAED,0BAA0B;AAC1B,SAAS,QAAQ,CAAC,IAAc;IAC9B,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC;QAAE,OAAO,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAEnD,MAAM,UAAU,GAAG,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC;SAC/C,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,IAAI,IAAI,KAAK,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC;SACzD,IAAI,CAAC,EAAE,CAAC,CAAC;IACZ,IAAI,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,IAAI,CAAC,IAAI,GAAG,UAAU,GAAG,CAAC;IACvE,OAAO,IAAI,IAAI,CAAC,IAAI,GAAG,UAAU,IAAI,WAAW,CAAC,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,GAAG,CAAC;AAC1E,CAAC;AAED,kEAAkE;AAClE,MAAM,UAAU,SAAS,CAAC,OAAoB;IAC5C,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;IAC1C,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC;AACrF,CAAC;AAED,2DAA2D;AAC3D,MAAM,UAAU,MAAM,CAAC,OAAoB,EAAE,QAAgB;IAC3D,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;IACxC,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACtC,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,QAAQ,CAAC,CAAC;AAC9E,CAAC;AAED,sDAAsD;AACtD,SAAS,QAAQ,CAAC,MAAmB,EAAE,IAAY;IACjD,IAAI,IAAI,KAAK,EAAE;QAAE,OAAO;IACxB,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACvD,CAAC;AAED,sEAAsE;AACtE,SAAS,KAAK,CAAC,MAAc,EAAE,MAAc,EAAE,EAAU;IACvD,MAAM,GAAG,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IACvC,OAAO,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC;AACvD,CAAC;AAED,+EAA+E;AAC/E,SAAS,YAAY,CAAC,MAAc,EAAE,IAAY,EAAE,EAAU;IAC5D,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,KAAK,IAAI,OAAO,EAAE,GAAG,CAAC,CAAC;IAClD,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC;IAC7C,OAAO,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;AAC7E,CAAC;AAED,6EAA6E;AAC7E,SAAS,OAAO,CAAC,MAAc,EAAE,EAAU;IACzC,MAAM,KAAK,GAAG,gCAAgC,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC;IACtE,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAErC,MAAM,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IAC5C,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC;IACjC,MAAM,UAAU,GAA2B,EAAE,CAAC;IAC9C,IAAI,MAAM,GAAG,EAAE,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAElC,SAAS,CAAC;QACR,OAAO,MAAM,GAAG,MAAM,CAAC,MAAM,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;YAAE,MAAM,IAAI,CAAC,CAAC;QAC9E,IAAI,MAAM,IAAI,MAAM,CAAC,MAAM;YAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,CAAC;QACtF,IAAI,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,MAAM,CAAC;YAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3F,IAAI,MAAM,CAAC,MAAM,CAAC,KAAK,GAAG;YAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,GAAG,CAAC,EAAE,CAAC;QAElF,MAAM,SAAS,GAAG,iDAAiD,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;QAC/F,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACvB,uEAAuE;YACvE,oCAAoC;YACpC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,CAAC;QAC3D,CAAC;QACD,MAAM,GAAG,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;QACzB,UAAU,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;YAC5C,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,cAAc,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC;QACxD,MAAM,IAAI,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAChC,CAAC;AACH,CAAC;AAED,6DAA6D;AAC7D,SAAS,OAAO,CAAC,KAAa;IAC5B,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACvB,OAAO,CAAC,KAAK,KAAK,GAAG,IAAI,KAAK,KAAK,GAAG,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;AAChG,CAAC;AAED,qDAAqD;AACrD,MAAM,QAAQ,GAAqC;IACjD,GAAG,EAAE,GAAG;IACR,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;IACP,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;CACV,CAAC;AAEF,kFAAkF;AAClF,SAAS,cAAc,CAAC,KAAa;IACnC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IAEvC,OAAO,KAAK,CAAC,OAAO,CAClB,qDAAqD,EACrD,CAAC,KAAK,EAAE,IAAY,EAAE,EAAE;QACtB,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YACnD,OAAO,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,KAAK,CAAC;QAChE,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,OAAO,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,KAAK,CAAC;QACxF,OAAO,QAAQ,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC;IACjC,CAAC,CACF,CAAC;AACJ,CAAC;AAED,6EAA6E;AAC7E,SAAS,SAAS,CAAC,KAAa;IAC9B,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,GAAG,QAAQ;QAAE,OAAO,SAAS,CAAC;IAChF,IAAI,KAAK,IAAI,MAAM,IAAI,KAAK,IAAI,MAAM;QAAE,OAAO,SAAS,CAAC;IACzD,IAAI,CAAC;QACH,OAAO,MAAM,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC;IACrC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC;AAED,2EAA2E;AAC3E,SAAS,UAAU,CAAC,KAAa;IAC/B,OAAO,KAAK;SACT,UAAU,CAAC,GAAG,EAAE,OAAO,CAAC;SACxB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC;SACvB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC;SACvB,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;AAC/B,CAAC"}

package/dist/webmention/index.d.ts

@@ -0,0 +1,12 @@
+export { DISCOVERY_MAX_BYTES, DISCOVERY_TIMEOUT_MS, discoverEndpoint, endpointInHeader, endpointInHtml, readCapped, WEBMENTION_USER_AGENT, } from './discovery.ts';
+export { classesOf, elementsIn, innerHtmlOf, isElement, parseHtml, rawTextOf, textOf, } from './html.ts';
+export type { HtmlElement, HtmlNode, HtmlText } from './html.ts';
+export { externalLinks } from './links.ts';
+export { itemsIn, linksTo, sourceEntry } from './microformats.ts';
+export type { MicroformatItem, MicroformatValue, SourceEntry, WebmentionKind, } from './microformats.ts';
+export { checkWebmentionRequest, VERIFY_TIMEOUT_MS, verifyWebmention } from './receive.ts';
+export type { CheckWebmentionOptions, IncomingWebmention, VerifyWebmentionOptions, WebmentionOutcome, WebmentionRequest, } from './receive.ts';
+export { mountWebmentions, WEBMENTION_FIELDS, WEBMENTION_PATH, webmentionEndpointFor, } from './routes.ts';
+export { createWebmentionService, SEND_TIMEOUT_MS } from './service.ts';
+export type { CreateWebmentionServiceOptions, WebmentionLogger, WebmentionReport, WebmentionService, } from './service.ts';
+//# sourceMappingURL=index.d.ts.map

package/dist/webmention/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/webmention/index.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,mBAAmB,EACnB,oBAAoB,EACpB,gBAAgB,EAChB,gBAAgB,EAChB,cAAc,EACd,UAAU,EACV,qBAAqB,GACtB,MAAM,gBAAgB,CAAC;AACxB,OAAO,EACL,SAAS,EACT,UAAU,EACV,WAAW,EACX,SAAS,EACT,SAAS,EACT,SAAS,EACT,MAAM,GACP,MAAM,WAAW,CAAC;AACnB,YAAY,EAAE,WAAW,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,WAAW,CAAC;AACjE,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAClE,YAAY,EACV,eAAe,EACf,gBAAgB,EAChB,WAAW,EACX,cAAc,GACf,MAAM,mBAAmB,CAAC;AAC3B,OAAO,EAAE,sBAAsB,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAC3F,YAAY,EACV,sBAAsB,EACtB,kBAAkB,EAClB,uBAAuB,EACvB,iBAAiB,EACjB,iBAAiB,GAClB,MAAM,cAAc,CAAC;AACtB,OAAO,EACL,gBAAgB,EAChB,iBAAiB,EACjB,eAAe,EACf,qBAAqB,GACtB,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,uBAAuB,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AACxE,YAAY,EACV,8BAA8B,EAC9B,gBAAgB,EAChB,gBAAgB,EAChB,iBAAiB,GAClB,MAAM,cAAc,CAAC"}

package/dist/webmention/index.js

@@ -0,0 +1,8 @@
+export { DISCOVERY_MAX_BYTES, DISCOVERY_TIMEOUT_MS, discoverEndpoint, endpointInHeader, endpointInHtml, readCapped, WEBMENTION_USER_AGENT, } from './discovery.js';
+export { classesOf, elementsIn, innerHtmlOf, isElement, parseHtml, rawTextOf, textOf, } from './html.js';
+export { externalLinks } from './links.js';
+export { itemsIn, linksTo, sourceEntry } from './microformats.js';
+export { checkWebmentionRequest, VERIFY_TIMEOUT_MS, verifyWebmention } from './receive.js';
+export { mountWebmentions, WEBMENTION_FIELDS, WEBMENTION_PATH, webmentionEndpointFor, } from './routes.js';
+export { createWebmentionService, SEND_TIMEOUT_MS } from './service.js';
+//# sourceMappingURL=index.js.map

package/dist/webmention/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/webmention/index.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,mBAAmB,EACnB,oBAAoB,EACpB,gBAAgB,EAChB,gBAAgB,EAChB,cAAc,EACd,UAAU,EACV,qBAAqB,GACtB,MAAM,gBAAgB,CAAC;AACxB,OAAO,EACL,SAAS,EACT,UAAU,EACV,WAAW,EACX,SAAS,EACT,SAAS,EACT,SAAS,EACT,MAAM,GACP,MAAM,WAAW,CAAC;AAEnB,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAOlE,OAAO,EAAE,sBAAsB,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAQ3F,OAAO,EACL,gBAAgB,EAChB,iBAAiB,EACjB,eAAe,EACf,qBAAqB,GACtB,MAAM,aAAa,CAAC;AACrB,OAAO,EAAE,uBAAuB,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC"}

package/dist/webmention/links.d.ts

@@ -0,0 +1,18 @@
+/**
+ * Which pages a post talks about.
+ *
+ * A webmention is "I wrote about you", so what goes out is decided by what the
+ * post links to. Only the links pointing somewhere else count: telling this
+ * site about its own pages would be a loop, and the conversation under a post
+ * is not the place for it.
+ */
+/**
+ * Every external page a rendered body links to, in the order it links to them
+ * and each named once.
+ *
+ * A fragment is dropped. `#section` says where on a page to look and the page
+ * is what the webmention is about, so two links into one article are one
+ * target rather than two — and the `target` a receiver verifies is a page.
+ */
+export declare function externalLinks(html: string, baseUrl: string): string[];
+//# sourceMappingURL=links.d.ts.map

package/dist/webmention/links.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"links.d.ts","sourceRoot":"","sources":["../../src/webmention/links.ts"],"names":[],"mappings":"AAEA;;;;;;;GAOG;AAEH;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,MAAM,EAAE,CAcrE"}