Unverified Commit 14a349ff authored by Hong Minhee's avatar Hong Minhee
Browse files

Recognize `alternate` AS objects in <link>/<a>

parent fb33da0e

CHANGES.md

+11 −2
Original line number Diff line number Diff line
@@ -10,12 +10,21 @@ To be released. 


 -  Removed `expand` option of `Object.toJsonLd()` method, which was deprecated

    in version 0.14.0.  Use `format: "expand"` option instead.



 -  Added `Context.lookupObject()` method.

 -  Default document loaders now recognize `alternate` ActivityStreams objects

    in the `Link` header.



 -  Default document loaders now recognize ActivityStream objects in more ways:



     -  Loaders now recognize `alternate` ActivityStreams objects in the `Link`

        header.

     -  Loaders now recognize `alternate` ActivityStreams objects in

        the `<link>`/`<a>` HTML elements.



 -  Added `allowPrivateAddress` option to `CreateFederationOptions` interface.



 -  Renamed the short option `-c` for `--compact` of `fedify lookup` command to

    `-C` to avoid conflict with the short option `-c` for `--cache-dir`.



 -  Added `-r`/`--raw` option to `fedify lookup` command to output the raw JSON

    object.

src/runtime/docloader.test.ts

+62 −0
Original line number Diff line number Diff line
@@ -106,6 +106,68 @@ test("fetchDocumentLoader()", async (t) => { 
    });

  });



  mf.mock("GET@/html-link", (_req) =>

    new Response(

      `<html>

        <head>

          <meta charset=utf-8>

          <link

            rel=alternate

            type='application/activity+json'

            href="https://example.com/object">

        </head>

      </html>`,

      {

        status: 200,

        headers: { "Content-Type": "text/html; charset=utf-8" },

      },

    ));



  await t.step("HTML <link>", async () => {

    assertEquals(await fetchDocumentLoader("https://example.com/html-link"), {

      contextUrl: null,

      documentUrl: "https://example.com/object",

      document: {

        "@context": "https://www.w3.org/ns/activitystreams",

        id: "https://example.com/object",

        name: "Fetched object",

        type: "Object",

      },

    });

  });



  mf.mock("GET@/html-a", (_req) =>

    new Response(

      `<html>

        <head>

          <meta charset=utf-8>

        </head>

        <body>

          <a

            rel=alternate

            type=application/activity+json

            href=https://example.com/object>test</a>

        </body>

      </html>`,

      {

        status: 200,

        headers: { "Content-Type": "text/html; charset=utf-8" },

      },

    ));



  await t.step("HTML <a>", async () => {

    assertEquals(await fetchDocumentLoader("https://example.com/html-a"), {

      contextUrl: null,

      documentUrl: "https://example.com/object",

      document: {

        "@context": "https://www.w3.org/ns/activitystreams",

        id: "https://example.com/object",

        name: "Fetched object",

        type: "Object",

      },

    });

  });



  mf.mock("GET@/404", (_req) => new Response("", { status: 404 }));



  await t.step("not ok", async () => {

src/runtime/docloader.ts

+35 −0
Original line number Diff line number Diff line
@@ -146,6 +146,41 @@ async function getRemoteDocument( 
      }

    }

  }

  if (

    !jsonLd &&

    (contentType === "text/html" || contentType?.startsWith("text/html;") ||

      contentType === "application/xhtml+xml" ||

      contentType?.startsWith("application/xhtml+xml;"))

  ) {

    const p = /<(a|link)((\s+[a-z][a-z:_-]*=("[^"]*"|'[^']*'|[^\s>]+))+)\/?>/ig;

    const p2 = /\s+([a-z][a-z:_-]*)=("([^"]*)"|'([^']*)'|([^\s>]+))/ig;

    const html = await response.text();

    let m: RegExpExecArray | null;

    const rawAttribs: string[] = [];

    while ((m = p.exec(html)) !== null) rawAttribs.push(m[2]);

    for (const rawAttrs of rawAttribs) {

      let m2: RegExpExecArray | null;

      const attribs: Record<string, string> = {};

      while ((m2 = p2.exec(rawAttrs)) !== null) {

        const key = m2[1].toLowerCase();

        const value = m2[3] ?? m2[4] ?? m2[5] ?? "";

        attribs[key] = value;

      }

      if (

        attribs.rel === "alternate" && "type" in attribs && (

          attribs.type === "application/activity+json" ||

          attribs.type === "application/ld+json" ||

          attribs.type.startsWith("application/ld+json;")

        ) && "href" in attribs

      ) {

        logger.debug(

          "Found alternate document: {alternateUrl} from {url}",

          { alternateUrl: attribs.href, url: documentUrl },

        );

        return await fetch(attribs.href);

      }

    }

  }

  logger.debug(

    "Fetched document: {status} {url} {headers}",

    {