django-wm 3.0.0 - beatonma.org

Upgrade warning

If upgrading from an older version please be aware of these changes:

Unused MentionableMixin.allow_incoming_webmentions field has been removed.
Any existing instances of PendingIncomingWebmention and PendingOutgoingContent will be deleted.
- These models have new constraints so it is necessary to recreate them.
- If this is problematic for you please don’t upgrade yet. Contact me or create an issue and I will make a tool to persist these between versions.

Changes

Thanks to @philgyford for reporting most of the issues referenced here.

Resolves #25.
- Added QuotableMixin.post_type field.
- Incoming webmentions are now checked for the following microformat properties that describe the type of mention they are:
  - u-bookmark-of
  - u-like-of
  - u-listen-of
  - u-in-reply-to
  - u-repost-of
  - u-translation-of
  - u-watch-of
- The /webmention/get endpoint serializes these values in the type field respectively as:
  - bookmark
  - like
  - listen
  - reply
  - repost
  - translation
  - watch
  - If no specific type is specified this defaults to webmention.
Resolves #30
- Added MentionableMixin.should_process_webmentions() -> bool method to enable custom logic.
Resolves #31
- Success message now rendered via template, enabling override by user.
Resolves #32
- Same-page #anchor links no longer treated as webmention targets.
Resolves #33, #34
- Much improved parsing of h-card.
  - Now tries to find anh-card that is directly related to the mention link (embedded in p-author of a parent h-entry or h-feed container).
  - If that doesn’t yield a result, try to find a top-level h-card on the page.
- Removed HCard.from_soup() classmethod. Parsing logic moved to tasks.incoming.parsing package.
Resolves #36
- Added constraints to PendingIncomingWebmention and PendingOutgoingContent to avoid duplication.
  - Warning: If updating from an older version of django-wm this will delete any existing Pending... model instances.
- Request timeouts are now handled gracefully.
- PendingIncomingWebmention and OutgoingWebmentionStatus now implement the new RetryableMixin.
- Reworked webmention processing tasks to allow failed webmentions to be retried periodically.
  - See settings WEBMENTIONS_MAX_RETRIES, WEBMENTIONS_RETRY_INTERVAL, WEBMENTIONS_TIMEOUT below for customisation details.
New dashboard/ view: a simple overview of recent mentions.
- Shows the latest instances of Webmention, OutgoingWebmentionStatus, PendingIncomingWebmention, PendingOutgoingContent and info on their current status.
- By default, restricted to users with mentions.view_webmention_dashboard permission.
- Can be made public via settings.WEBMENTIONS_DASHBOARD_PUBLIC = True.
New optional settings:
- settings.WEBMENTIONS_TIMEOUT [float | default=10] specifies the time (in seconds) to wait for network calls to resolve.
- settings.WEBMENTIONS_RETRY_INTERVAL [int | default=600] specifies the minimum time (in seconds) to wait before retrying to process a webmention.
- settings.WEBMENTIONS_MAX_RETRIES [int | default=5] specifies how many times we can attempt to process a mention before giving up.
- settings.WEBMENTIONS_DASHBOARD_PUBLIC [bool | default=False] specifies whether the the dashboard/ view can be viewed by anyone. If False (default), the dashboard/ view is only available to users with mentions.view_webmention_dashboard permission.
WebmentionHeadMiddleware
- Removed port from generated endpoint URL in HTTP headers as request.META.SERVER_PORT may not be reliable depending on reverse proxy configuration.
MentionableMixin:
- mentions() is now a method, not a property.
- Removed field allow_incoming_webmentions as it has never been used for anything.
Streamlined template tags
- {% load webmentions_endpoint %} replaced by {% load webmentions %}.
- {% webmention_endpoint %} replaced by {% webmentions_endpoint %} (used in HTML <head>).
- New tag {% webmentions_dashboard %} creates a <a>link</a> to your webmentions/dashboard/ view.

django-wm: 3.0.0

Upgrade warning

Changes