ref https://linear.app/tryghost/issue/CFR-31/
- when relations were included, erroneous logic resulted in the model's
default sort being applied
- the model default sort is not intended for the content API and
needlessly slowed down responses
- there's a change for users here that should be incredibly unlikely to
be hit; default sort is `published_at desc` which will be secondarily
sorted by `id desc` instead of `published_at desc, updated_at desc, id
desc`
This is a very significant performance improvement for content API
requests with includes for sites with a significant amount of data,
which will primarily impact those using Ghost as a CMS or theme {{#get}}
helpers.
fix https://linear.app/tryghost/issue/SLO-85/fix-http-500-on-contentposts
- in the event we give the incorrect format in a filter, MySQL will
throw an error and we'll throw a HTTP 500 error
- we can capture this error and return a more useful error to the user
- ideally we'd do this in a validation step before attempting the query,
but parsing this out of NQL and detecting which columns are DATETIME
could be quite tricky
closes ENG-627
We were using `cheerio` to parse+modify+serialize our rendered HTML to modify links for member attribution. Cheerio's serializer has a [long-standing issue](https://github.com/cheeriojs/cheerio/issues/720) (that we've [had to deal with before](https://github.com/TryGhost/SDK/issues/124)) where it replaces single-quote attributes with double-quote attributes. That was resulting in broken rendering when content used single-quotes such as in HTML cards that have JSON data inside a `data-` attribute or otherwise used single-quotes to avoid escaping double-quotes in an attribute value.
- swapped the implementation that uses `cheerio` for one that uses `html5parser` to tokenize the html string, from there we can loop over the tokens and replace the href attribute values in the original string without touching any other part of the content. Avoids a full parse+serialize process which is both more costly and can result unexpected content changes due to serializer opinions.
- fixes the quote change bug
- uses tokenization directly to avoid cost of building a full AST
- updated Content API Posts snapshot
- one of our fixtures has a missing closing tag which we're no longer "fixing" with a full parse+serialize step in the link replacer (keeps modified src closer to original and better matches behaviour elsewhere in the app / without member-attribution applied)
- the link replacer no longer converts `attr=""` to `attr` (these are equivalent in the HTML spec so no change in behaviour other than preserving the original source html)
- added a benchmark test file comparing the two implementations because the link replacer runs on render so it's used in a hot path
- new implementation has a 3x performance improvement
- the separate files with the old/new implementations have been cleaned up but I've left the benchmark test file in place for future reference
Benchmark results comparing implementations:
```
❯ node test/benchmark.js
LinkReplacer
├─ cheerio: 5.03K /s ±2.20%
├─ html5parser: 16.5K /s ±0.43%
Completed benchmark in 0.9976526670455933s
┌─────────────┬─────────┬────────────┬─────────┬───────┐
│ (index) │ percent │ iterations │ current │ max │
├─────────────┼─────────┼────────────┼─────────┼───────┤
│ cheerio │ '' │ '5.03K/s' │ 5037 │ 5037 │
│ html5parser │ '' │ '16.5K/s' │ 16534 │ 16534 │
└─────────────┴─────────┴────────────┴─────────┴───────┘
```
fixes PROD-61
This adds a new default plan setting. It defaults to yearly, which is
the current default selected interval in Portal.
Behind the new portal improvements feature flag, the default plan can be
changed. It will also change automatically if the available intervals
are changed.
This PR also wires up passing the new setting to the Portal preview.
fixes GRO-72
- added "default_email_address" and "support_email_address" to the
public settings
- when available, use these addresses in Portal. Otherwise, fallback to
current logic
refs https://github.com/TryGhost/Product/issues/4088
The Content API should not expose the lexical/mobiledoc source content because it's not membership-gated and although not used at the present time may in future contain additional internal metadata. We were handling this for the more-typical `?formats` param but it was still possible to access this data using the `?fields` param.
- updated post mapper used in our API output serializers to strip the `mobiledoc` and `lexical` fields ready for API output
- credits to Prathap Puthran for reporting
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
no issue
- Do not set ?ref in recommendations if analytics is disabled
- Do not send url_history if analytics is disabled
- Expose outboundLinkTagging as a public setting
fixes https://github.com/TryGhost/Product/issues/3851
- Order was not applied via the CRUD plugin
- Removed usage of CRUD findAll, and swapped to Bookshelf fetchAll
instead, to decrease dependencies of invisible Bookshelf plugins logic
- Reverted page and limit options possibility via findAll method
fixes https://github.com/TryGhost/Product/issues/3822
fixes https://github.com/TryGhost/Product/issues/3838
This PR became a bit big because it affected multiple parts of Ghost
that needed to be updated to prevent breaking anything.
### Backend
- Added pagination to the recommendations API's
- Updated BookshelfRepository template implementation to handle
pagination
- Allow to pass `page` and `limit` options to Models `findAll`, to allow
fetching a page without also fetching the count/metadata (=> in the
repository pattern we prefer to fetch the count explicitly if we need
pagination metadata)
- Added E2E tests for public recommendations API (content API)
- Extended E2E tests of admin recommendations API
### Portal
- Corrected recommendations always loaded in Portal. Instead they are
now only fetched when the recommendations page is opened.
### Admin-X
- Added `usePagination` hook: internally used in the new
`usePaginatedQuery` hook. This automatically adds working pagination to
a query that can be used to display in a table by passing the
`pagination` and `isLoading` results to the `<Table>`
- Added placeholder `<LoadingIndicator>` component
- Added a loading indicator to `<Table>`. This remembers the previous
height of the table, to avoid layout jumps when going to the next page.
fixes https://github.com/TryGhost/Product/issues/3820
- This adds a new public site endpoint in the members API to check if a
site can offer the one-click-subscribe feature
- This is implemented on the members API as a copy of the `site`
endpoint because the admin API site endpoint is protected by CORS and
mainly because it can be served on a different domain than the
recommended site and this is hard to detect reliably from the frontend
- Added a new calculated setting `allow_self_signup`, which can replace
the setting that is currently used in Portal (best to do this after a
release otherwise we risk creating issues if a patch release happens)
closes https://github.com/TryGhost/Arch/issues/76
- The posts test suite was failing when run in isolation. This was due to "collections" labs flag not being turned on, the events were not going through to collections service correctly
fixes https://github.com/TryGhost/Product/issues/3687
After this change, relative URLs in emails will be replaced with
absolute URLs using the post URL. Making relative Portal URLs possible
etc.
Updates the test data generator to fix invalid URL encoding (somehow a
backslash + escaped double quote was added when it wasn't required).
no issue
- Snapshot tests were incorrect but we weren't catching it because CI
was retrying them and yielding false passes
- This fix just fixes the broken tests, which will allow us to fix the
issue with CI yielding false passing results
closes https://github.com/TryGhost/Arch/issues/58
- Following assumptions were broken:
- Posts Admin API should include posts of all statuses when filtering by collection
- Posts Content API should not include any unpublished posts
- Updated the "status" filter which fixes the problem. We still disallow any custom filters to be applied on top of collections filter.
The only usecases we need to support at the moment are reading individual
collections by ID and by Slug. We can extend this API as we get more usescases
in future.
no issue
`show_title_and_feature_image` leads to more intuitive logic in themes and we can use `posts` rather than `posts_meta` as there are no longer row-length issues with MySQL 8.
- removed original add-column migration that was never in a release
- added new add-column migration that puts
`show_title_and_feature_image` column with a default of `true` on the `posts` table
- renamed property and default value everywhere
- bumped `@tryghost/admin-api-schema` to allow the new property through at the API level
closes https://github.com/TryGhost/Product/issues/3557
- bumped `@tryghost/admin-api-schema` to allow passthrough of the new property in API requests
- updated output mapper to ensure property always returns a boolean rather than `null` in the case where `posts_meta` doesn't exist for a page
- updated `PostsService.copyPost()` to include the new property when copying
- updated `checkResponse` test util and snapshots to expect `hide_title_and_feature_image` property in page API responses
- fixed pages e2e test so it doesn't inadvertently modify the match object breaking later tests
refs: https://github.com/TryGhost/Toolbox/issues/595
We're rolling out new rules around the node assert library, the first of which is enforcing the use of assert/strict. This means we don't need to use the strict version of methods, as the standard version will work that way by default.
This caught some gotchas in our existing usage of assert where the lack of strict mode had unexpected results:
- Url matching needs to be done on `url.href` see aa58b354a4
- Null and undefined are not the same thing, there were a few cases of this being confused
- Particularly questionable changes in [PostExporter tests](c1a468744b) tracked [here](https://github.com/TryGhost/Team/issues/3505).
- A typo see eaac9c293a
Moving forward, using assert strict should help us to catch unexpected behaviour, particularly around nulls and undefineds during implementation.
This prepares us to return a DTO rather than BookshelfModel to the serialiser
layer. When passing a BookshelfModel, the serialisation layer uses the model to
read from when building computed properties. By stripping values out in the
toJSON method it means that the DTO will be missing them and the computed
properties won't be able to be calculated. Instead we return ALL values to the
serialisation layer, and then strip out the ones that weren't requested in the
"clean" step.
This also inadvertently fixes the issue with `reading_time` requiring the
`html` field to be requested, we can now request just `reading_time`, as well
as have it included by default.
refs https://github.com/TryGhost/Team/issues/3248
The current test fixtures didn't include any hidden Tiers, so I've added
a new fixture to test the filtering of hidden Tiers. It's not enabled by
default to avoid breaking the existing tests.
- this cleans up all imports or variables that aren't currently being used
- this really helps keep the tests clean by only allowing what is needed
- I've left `should` as an exemption for now because we need to clean up
how it is used
fixes https://github.com/TryGhost/Team/issues/2666
- Somehow occurrences of `&map_` got replaced with `↦`
- Disables escaping &, ', " and other HTML characters when not needed
(escaping is already handled by mobiledoc/lexical)
- Bumps unit test coverage of link replacer to 100%
fixes https://github.com/TryGhost/Team/issues/2433
- Moved all outbound link tagging code to separate OutboundLinkTagger
- Because a site can easily enable/disable this feature, we don't store
the ?refs in the HTML but add them on the fly for now in the Content
API.
no issue
There are a couple of issues with resetting the Ghost instance between
E2E test files:
These issues came to the surface because of new tests written in
https://github.com/TryGhost/Ghost/pull/16117
**1. configUtils.restore does not work correctly**
`config.reset()` is a callback based method. On top of that, it doesn't
really work reliably (https://github.com/indexzero/nconf/issues/93)
What kinda happens, is that you first call `config.reset` but
immediately after you correcty reset the config using the `config.set`
calls afterwards. But since `config.reset` is async, that reset will
happen after all those sets, and the end result is that it isn't reset
correctly.
This mainly caused issues in the new updated images tests, which were
updating the config `imageOptimization.contentImageSizes`, which is a
deeply nested config value. Maybe some references to objects are reused
in nconf that cause this issue?
Wrapping `config.reset()` in a promise does fix the issue.
**2. Adapters cache not reset between tests**
At the start of each test, we set `paths:contentPath` to a nice new
temporary directory. But if a previous test already requests a
localStorage adapter, that adapter would have been created and in the
constructor `paths:contentPath` would have been passed. That same
instance will be reused in the next test run. So it won't read the new
config again. To fix this, we need to reset the adapter instances
between E2E tests.
How was this visible? Test uploads were stored in the actual git
repository, and not in a temporary directory. When writing the new image
upload tests, this also resulted in unreliable test runs because some
image names were already taken (from previous test runs).
**3. Old 2E2 test Ghost server not stopped**
Sometimes we still need access to the frontend test server using
`getAgentsWithFrontend`. But that does start a new Ghost server which is
actually listening for HTTP traffic. This could result in a fatal error
in tests because the port is already in use. The issue is that old E2E
tests also start a HTTP server, but they don't stop the server. When you
used the old `startGhost` util, it would check if a server was already
running and stop it first. The new `getAgentsWithFrontend` now also has
the same functionality to fix that issue.
fixes https://github.com/TryGhost/Team/issues/2432
Adds outbound_link_tagging setting (enabled by default and behind
feature flag). If the feature flag is enabled, and the setting is
disabled, we won't add ?ref to links in emails.
This includes new E2E tests for email click tracking, which were also
extended to check outbound link tagging (for both MEGA and the new email
stability flow).
Also fixes a test fixture for the comments_enabled setting.
refs https://github.com/TryGhost/Team/issues/2400
- we've deemed it useful to start to return `Content-Version` for all
API requests, because it becomes useful to know which version of Ghost
a response has come from in logs
- this should also help us detect Admin<->Ghost API mismatches, which
was the cause of a bug recently (ref'd issue)
- updated the cover image to be simpler
- made the change in text fixtures as well, just to keep the fixtures in sync
Co-authored-by: Hannah Wolfe <github.erisds@gmail.com>
refs https://github.com/TryGhost/Toolbox/issues/425
refs https://github.com/TryGhost/Toolbox/issues/280
- The versioned API responses vary based on requested version (passed in request's 'accept-version' header). shared caches that sit between Ghost's origin server and the browser would be putting responses with same Vary into the same caching bucket, which is incorrect.
- This change makes response's Vary more granular and tells caching mechanisms to take 'Accept-Version' request header into account when caching.
- Informative read on the topic - https://www.fastly.com/blog/getting-most-out-vary-fastly
refs https://github.com/TryGhost/Toolbox/issues/410
- Private cache control was preventing browser or shared caches from storing Content APIs response. The type of data served through the Content API is very much of a "public" nature, so should be cacheable.
- Right now the 'max-age' value of 'cache-control' header is hardcoded to '0', without 'must-revalidate' value, to allow browsers to cache content slightly more aggressively. In the future the 'max-age' value will most-likely become configurable to allow even more aggressive HTTP caching.
refs https://github.com/TryGhost/Toolbox/issues/410
- The 'private' value in 'Cache-Control' response header for all errors made it impossible for shared caches (e.g.: Fastly, Cloudflare) to cache 404 responses efficiently.
- The change substitutes 'max-age=0' which should not effect the browser cache behavior but would allow shared caches to process such requests efficiently.
- A more loose caching logic only applies to 404 responses from GET requests that are not user-specific (non-authenticated, non-cookie containing requests)
no issue
- similar to the `mobiledoc` field, the Content API should not return the source `lexical` field if requested via `?formats=`
- renamed `removeMobiledocFormat()` to `removeSourceFormats()` to better match it's behaviour
- This fixture is the main user fixture you'd want to use when testing staff roles
- At the moment it has a weird name that makes it less likely people will use it
- A tiny step in trying to make our fixture system make a tiny bit more sense
- added core and builtin integrations to test fixtures
- allowed passing a custom api key id to generate JWT
- updated admin key auth test to make successful request with a `core` integration, which doesn't work atm because relations are not returned
refs https://github.com/TryGhost/Team/issues/1695
This property can be used by theme developers to determine if comments
are available for the currently logged in member. It follows the same
logic as used internally in the comments helper, so that they can be
used interchangeably
refs https://github.com/TryGhost/Toolbox/issues/354
- this commit turns the Ghost repo into a monorepo so we can bring our
internal packages back in, which makes life easier when working on
Ghost
