ref https://linear.app/tryghost/issue/DEV-20/faster-builds
- `build` should be the overall script to build the package
- `build:ts` should only build TypeScript
- by having them switched around, `build:ts` would call `yarn build`,
and `yarn` adds several hundreds of milliseconds of time to each build
ref https://linear.app/tryghost/issue/ENG-801/unable-to-recommend-sites-with-long-excerpts
- recommending a site with a long excerpt was being blocked by a
validation error
- with this change, we truncate the excerpt to 2000 characters max. and
avoid showing an error in the UI
- with this change, the description length validation is also now
stricter; 200 characters max, instead of 2000, to match the UI
ref https://linear.app/tryghost/issue/ENG-799
- recommendations were being stripped of query parameters and hash
fragments before save
- in particular, query parameters for attribution such as ?ref were not
being stored
ref https://linear.app/tryghost/issue/CFR-4/
- added request queueing middleware (express-queue) to handle high
request volume
- added new config option `optimization.requestQueue`
- added new config option `optimization.requestConcurrency`
- added logging of request queue depth - `req.queueDepth`
We've done a fair amount of investigation around improving Ghost's
resiliency to high request volume. While we believe this to be partly
due to database connection contention, it also seems Ghost gets
overwhelmed by the requests themselves. Implementing a simple queueing
system allows us a simple lever to change the volume of requests Ghost
is actually ingesting at any given time and gives us options besides
simply increasing database connection pool size.
---------
Co-authored-by: Michael Barrett <mike@ghost.org>
refs https://linear.app/tryghost/issue/ENG-750
- when adding a recommendation, we fetch the recommended site's metadata
- before this change, if the metadata fetch failed for some reason, we'd show an error and block the recommendation from being added
- after this change, we use fallback values if the metadata fails to fetch, instead of blocking the recommendation from being added. We use the site domain as the title and leave the rest empty (no favicon, no description)
- this change also means we are not checking whether a site exists or not for the publisher anymore. It’s then up to the publisher to make sure they don’t enter broken URLs
ref ENG-737
ref https://linear.app/tryghost/issue/ENG-737/http-500-errors-from-recommendations-check-endpoint
- it's still possible for `this.#externalRequest.get` to throw, like if
DNS resolution fails
- we want to try-catch this so we don't throw from this function and
return a HTTP 500 to the user
- instead, we can just return `undefined`, which is the fallback
- adds a breaking test too
- this version is written in TS, but was published a few months ago and
needs to be bumped here
- also updates a previous deep include into the library, which was
unnecessary anyway
ref PROD-233
- Stored whether Docker is used in the config files
- When running `yarn setup`, any existing Docker container will be
reset. Run with `-y` to skip the confirmation step.
- `yarn setup` will always init the database and generate fake data
- Increased amount of default generated data to 100,000 members + 500
posts.
- Made lots of performance improvements in the data generator so we can
generate the default data in ±170s
fixes https://github.com/TryGhost/Product/issues/4102
E.g. you recommend myghostsite.com, while that site redirects all
traffic to [www.myghostsite.com](#):
The redirect causes CORS issues, which means we cannot detect
one-click-subscribe support.
- This is fixed by moving the whole detection to the backend, which has
the additional benefit that we can update it in the background without
the frontend, and update it on every recommendation change.
- This change also fixes existing recommendations by doing a check on
boot (we can move this to a background job in the future).
fixes https://github.com/TryGhost/Product/issues/4005
We no longer use the 'reason' of a recommendation, but allow a flexible
description instead. Because this is a breaking change in the API, we do
this before making this feature GA.
- Added new database utils for renaming a column
- Added new migration to rename the column
- Updated all references in code
refs https://github.com/TryGhost/Product/issues/3978
- added "GET /incoming_recommendations/" browse endpoint to the Admin
API - we store incoming recommendations as mentions in the database. The
new endpoint reuses the Mentions API underneath to fetch verified
mentions of type recommendation - recommendation-specific attributes are
returned by the new endpoint, including calculated fields such as the
"RecommendingBack" boolean
- show "Recommend back" option for sites recommending me, only if I
haven't recommended the site already
refs https://github.com/TryGhost/Product/issues/3958
- `url:~a.com` matches a.com, sub.a.com and a.com/path
- added a stricter check based on hostname and pathname, but that
ignores www, protocol, query parameters and hash fragments
refs https://github.com/TryGhost/Product/issues/3957
This changes how we fetch recommendations:
- Relations can be included in one query instead of extra queries
- Sorting is now possible by click or subscriber counts
refs https://github.com/TryGhost/Product/issues/3818
- added a check to compare hostname and pathnames of URLs. Different
subdomain or different pathname = different URLs, but protocol, www,
query parameters and hash fragments are ignored.
refs https://github.com/TryGhost/Product/issues/3938
- Updated "Recommendation received" email design
- Slightly improved how dark favicons show in dark mode
- Added TabView with counter variant to Storybook
refs https://github.com/TryGhost/Product/issues/3958
- Disabled automatic network retries for external site lookups (=> timed
out to 5s in every situation because it returned 404 when a site doesn't
implement the Ghost api)
- Disabled representing a modal when it is already present on hash
changes
- Added support for search params in modals
- Handle `?url` search param in the addRecommendationModal
refs https://github.com/TryGhost/Product/issues/3940
- the {{recommendation}} helper fetches recommendations from the Content
API and renders a HTML template with pre-defined CSS classes
- the HTML template can be overridden in themes, by uploading a file
under partials/recommendations.hbs
- the CSS classes are not pre-defined, they need to be defined in
individual themes
- if there are no recommendations, nothing is rendered
- the {{recommendations}} helper currently accepts "page", "limit",
"filter", and "order" as options
closes https://github.com/TryGhost/Product/issues/3818
- instead of fetching all recommendations and matching URLs on the frontend, we now query the database directly to find an existing Recommendation by URL. When comparing URLs, we don't take into account the protocol, www, query parameters nor hash fragments
refs https://github.com/TryGhost/Product/issues/3850
- Added a recheck for recommendation related webmentions after boot (to
check missed webmentions during down time)
- Increased general timeouts to 15s for all webmention related HTTP
requests. Instead, increased retries to 3.
- Increased timeout for fetching webmention metadata from 2s to 15s
- Added more logging about verification and deletion status of
webmentions
fixes https://github.com/TryGhost/Product/issues/3900
1. The service never returns a Recommendation Entity, but always plain
objects (which for now is the same as Recommendation without the
methods).
2. Updated the controller to be more readable and minimal (we keep this
controller, in addition to the existing endpoints and serializers)
- The controller does minimal validation and allows for type checking
(so we get compile time errors in case the service expects new fields)
- The controller uses the `UnsafeData` class to easily validate the
input from requests, and throws appropriate errors (with correct field
descriptions — "Expected a string at recommendations.0.title") without
too much boilerplate code. In addition the interface is typed, so we get
compile errors if there are breaking changes in the service.
- Removed `EntityWithIncludes`, since we now use plain objects, we
inject the relations directly into those plain objects (with some new
types that add type support)
- Added new tests to make sure that edits only affect the given fields,
and never undefined fields
