Revision history for Web::Mention

0.721 2021-06-17T22:42:54Z

    - `new_from_html` no longer throws an exception if any the URLs it
    targets can't support a new Web::Mention object. (This typically
    happens when a given source webpage contains a link to itself.)

0.720 2020-07-19T17:19:09Z

    - Improved methods for translating webmentions to and from JSON,
    including new `as_json` and `new_from_json` methods. Removing
    documentation for old and annoying serialization expectations.
    
    - The `author` method now always returns an object. (Its fields will
    be present but empty if a webmention's author is unknown or unset.)

0.712 2020-07-04T20:17:21Z

    - Explicitly setting the internal user agent object's User-Agent
    header to something other than 'libwww-perl/...', a red flag to
    many web servers, and thus an obstacle to webmention verification
    
0.711 2020-05-08T23:30:36Z

    - Fixing bug that would handle unprocessed webmentions poorly in
    certain circumstances.

0.710 2020-04-28T23:06:05Z

    - For the sake of faster load time, swapped out all use of Moose
    (and MooseX) for Moo (and MooX).

    - Improvements to the author-determination algorithm implementation.

    - The `send` method will clear the value of `response` if the object
    cannot determine a Webmention endpoint for the target.
    
    - Increasing the default value of `max_content_length` to 280.

    - Documentation improvements.

0.704 2020-01-08T21:13:41Z

    - Improved handling of targets that specify a `Link` HTTP header with a URL but no `rel` value. (Thanks to Tomaž Šolc for the patch.)

0.703 2019-06-23T17:07:03Z

    - Adding a 'time_published' read-only attribute.

0.702 2019-06-13T21:13:55Z

    - Adding a 'title' read-only attribute.

    - Fixing a bug that accidentally deleted structure from the
    source MF2 document representation in certain circumstances.

    - Fixing a bug where reverifying an already-verified webmention
    would not update its type.

0.701 2019-06-12T03:28:04Z

    - Improved handling for replies that encode their referent URLs as
    complete microformats (e.g. `h-cite`), rather than as simple
    property strings. This allows for proper handling of reply-context
    (https://indieweb.org/reply-context).

0.7 2018-10-06T21:09:00Z

    - Improved handling of source documents containing multiple URL
    microformats. This fixes a bug where mentions of a specific type
    (e.g. likes) would incorrectly have the generic "mention" type set.

    - Cleaner response to asking for the content of content-free
    webmentions.
    
0.6 2018-09-08T20:36:53Z

    - Adding support for RSVP webmention types.
    
    - Improving post type discovery, according to a W3C-published
    algorithm
    (https://www.w3.org/TR/post-type-discovery/#response-algorithm).
    This fixes a bug that would mis-type webmentions in source documents
    containing a mix of marked-up link types.

0.501 2018-05-29T13:32:33Z

    - Documentation cleanup.
    
    - Reducing the default maximum content length to 200 characters.

0.5 2018-05-28T22:33:03Z

    - Adding the from_html batch-creation class method.

    - Significant improvements to the content method.

    - Improvements to serialization, and adding tests and documentation for it.
    
    - Miscellaneous bugfixes and documentation improvements.

    - Adding Travis config.

0.4 2018-05-10T04:34:13Z

    - Adding support for webmention sending. (Passes all create / update
    / delete tests on http://webmention.rocks, as of May 9, 2018.)
    
    - Tidier construction of author objects.
    
    - Recognize a URI-escaped version of a the target URL on the source document as a valid mention.
    
    - Improved handling of source documents lacking any MF2 markup.
    
    - Improved handling of relative URLs found in source docs' MF2 markup.

0.3 2018-04-23T14:19:45Z

    - The internal author-object constructor no longer chokes if it
    encounters author metadata that doesn't define every possible
    property.

    - Constructor throws an exception if the source and target URLs are
    too similar. (Thanks to Christian Weiske for the heads-up.)

    - The cpanfile better reflects prerequisites. (Thanks to Mohammad S
    Anwar for the patch.)

    - Documentation updates.

0.2 2018-04-21T20:33:29Z

    - Made JSON de/serialization less fragile.

    - Serializing a webmention object no longer tries to verify it as a
    side effect.

    - Improved encoding-handling of web-fetched content.

0.1 2018-04-20T20:08:31Z

    - original version