refs 91efa4605c
- When the instance is booted without any redirects files configured it's not supposed to error but rather default to an "empty" [] redirects configuration.
- Ideally the logic shoudl not contain try/catch block at all and fail as soon as there's any error during the initialization. This wasn't changed at this time due to possible break of existing Ghost instances
refs https://linear.app/tryghost/issue/CORE-84/have-a-look-at-the-eggs-redirects-refactor-branch
- The problem this change is addressing is inability to override config values once the code is extracted into a class+DI pattern
- The work around is restarting the instance with the configuration testing expected behavior - in this case missing or existing types of redirects files
refs https://github.com/TryGhost/Members/commit/9e59f5a9
Since we have a DynamicRedirectManager for handling adding/removing
redirects at runtime, we no longer need the custom-redirects middleware.
The redirects service does however need an init method now to add the
custom redirects at Ghost boot, so it's been refactored into our Class &
DI pattern.
refs https://linear.app/tryghost/issue/CORE-121/create-a-video-storage-adapter
- Uploading large files is costly and might cause DDoS. Limits would allow fair use of the site instance.
- The configureation in hostSettings to enable an "uploads" limit would look like following:
```
"hostSettings": {
"limits": {
"uploads": {
"max": 5,
"error": "Your plan supports uploads of max size up to {{max}}MB. Please upgrade to reenable uploading."
}
}
}
```
- Read more at ba37890be4/packages/limit-service (usage)
refs https://linear.app/tryghost/issue/CORE-121/create-a-video-storage-adapter
- This is an experimental implementation of video file upload support (audio is yet to follow)
- The storage adapter still needs more thinking as it's almost the same as the "LocalStorgeAdapter" that stores images.
- Also the output serializer skipped use of url utils in favor of inline implementatoin - this should almost certainly be it's own package
refs https://github.com/TryGhost/Team/issues/1197
We were always sending the 'signin' email, rather than respecting the
email_type param passed to the API. This updates our email sending when
creating members to force the requested type if it's present.
refs https://github.com/TryGhost/Team/issues/1178
The "up" migration that this util generates correctly throws if the
pre-requisite data cannot be found in the database. The "down" migration
however was incorrectly mirroring this behaviour of throwing - which
meant that it wasn't idempotent, as it does not require a permission or
role to existing if it wants to move relations between them.
refs https://github.com/TryGhost/Ghost/pull/10534
- The original PR has no explanation around the purpose of the ref field and it's easy to forget without knowing a wider context. Documented it to remember next time we come around working on this part of the codebase!
no-issue
This removes the logic to check if stripe connect is allowed into the
stripe connect service, which makes the feature easier to maintain, as
well as fixes the v3 API - which previously did not have this check.
no-issue
This removes logic from the Members API controller, and into the Members
BREAD service, this allows our controllers to be simpler and easier to
maintain, as well as keeping the important logic all together.
no issue
The way GA flags were introduced means that they stop existing in the `'labs'` setting in the db and are instead forced to always return `true` when checking the flag in the labs service. However, Admin which uses the flags fetches them via the `/settings/` API endpoint which was only returning the raw labs setting db value meaning GA flags appeared to be disabled unless the flag had previously been enabled and no settings save had occured.
- updated the settings bread service to replace the labs setting value with the JSON stringified output of `labs.getAll()` which is the ultimate source-of-truth for a feature being enabled/disabled
- extracted `browse()` behaviour to an internal `_formatBrowse()` method so we can apply the same filtering/modification for output of `browse()` and `edit()`
Co-authored-by: Fabien O'Carroll <fabien@allou.is>
refs https://github.com/TryGhost/Team/issues/1115
This allows users to create Offers for Tier/Cadence pairs in order to
provide discounted subscriptions to Members! We have support for
percentage based discounts & fixed price discounts, either for the first
payment, all payments, or a number of monthly payments.
Offers also have a code, which can be used as an easy way to share them,
as visiting https://site.com/offer-code will automatically open Portal
with the Offer prepopulated.
no issue
- author line was appearing as "Author One,Author Two,..." with no space
- fixed the generation of `post.authors` when serializing a post before it's rendered for email
no issue
- if the verification threshold is 0, `_.get(..)` becomes falsy so we
fallback to Infinity
- this is not correct - we only want to default to Infinity if the value
is not set
- this commit explicitly compares the config value to `undefined` and sets
the fallback accordingly
refs https://github.com/TryGhost/Team/issues/1175
We found the ETag header sent when serving the Admin template for /ghost/ was not changing between versions which after an upgrade could result in out of date cached content being served containing links to JS/CSS files that no longer existed.
The culprit is weak etags served by Node's `send` package, coupled with Admin template filesize not changing between versions and `npm pack` setting a fixed modification date for every file. See https://github.com/pillarjs/send/issues/176 for more details.
- updated the Admin app's controller to read the template and generate an md5 hash of the contents so we can serve a strong ETag header value when serving the `/ghost/` html
refs https://github.com/TryGhost/Team/issues/1163
We want to make the title for Offers optional, our nullable validation
means that we cannot store an empty string, so we must remove the NOT
NULL constraint from the column if we want to store either an empty
value or null.
There is a bug with editing columns in SQLite with `knex` which strips all
the indexes, so we have to manually add them afterwards.
no issue
- following on from f4fb0fcbaa,
this commit moves around some package requires in Ghost
- these are often niche packages that do something in a subsystem of
Ghost, and are not necessarily be needed to boot the application
- these packages use non-negligible CPU and memory when they are
required, so it makes sense to lazy-require them
- the concern here is that we obscure the code too much by moving
random requires further into code, but the changes are small and the
improvements big
- this commit bring the boot time since 4.19.0 down ~31% and initial
memory usage down by a total of ~12%
- we're slowly trying to draw the lines between the backend and the frontend correctly
- these files deal only with serving the frontend so they should live there
- there are lots of mixed requires in these files, so having them in the right place makes that clear
- this has been a niggle for ages, we shouldn't need to care what order our models are loaded in
- this is stopping us from having a built-in model loader and "frameworkizing" models
closes https://github.com/TryGhost/Team/issues/1172
Custom theme settings sync and cache population had been left to complete in the background as it wasn't essential for it to be complete for the front-end to start. However that was causing problems for the API where theme activation and custom theme settings list requests happen very close together, with the latter often not containing the theme settings data when it is expected to.
- changed `activationBridge.*` methods to `async` so they can `await` the completion of custom theme settings sync before activating a theme
refs https://linear.app/tryghost/issue/CORE-1/multiple-adapters-per-type
- Having this preemptive change allows to separate implementation of "image" storage from future usecases like "videos", "audios" etc. Even if the "image" adapter is not configured the default behavior will fallback to use the "active" storage adapter. If there's a need to handle "images" differently through a custom apapter that'll work out of the box ;)
refs https://linear.app/tryghost/issue/CORE-1/multiple-adapters-per-type
- When the storage is requested the caller can now specify a "feature" they want to use the storage for. For example there could be different configurations for "images" or "vidoes" storages and the caller would not necessarily have to know about the details of how the feature is implemented.
refs https://linear.app/tryghost/issue/CORE-1/multiple-adapters-per-type
- There's a need to support multiple adapter variations per given adapter type (storage, sso, etc.)
- With the introduced changes we can specify a version of an adapter that should be fetched based on `:feature` postfix. For example:
`adapterManager.getAdapter('storage')` - would return the default adapter listed under "active" configuration
`adapterManager.getAdapter('storage:videos') - would return an adapter configured for videos *feature*
- Here's an example configuration for a custom video storage:
```
"storage": {
"active": "LocalFileStorage",
"videos": "ghost-storage-custom-video",
"ghost-storage-custom-video": {
"custom": "configHere"
}
}
```
- This is a minor refactor to reorder functions, use async/await & fat arrows
- It will make the coming bigger refactor slightly easier to read :D
- This should not make any functional changes
closes https://github.com/TryGhost/Team/issues/1150
Our override of the base Bookshelf `insert` operation so that our own `formatOnWrite()` method is called on attributes was working on a false assumption that an `attrs` attribute is passed in as it is for the `update` operation. Instead Bookshelf's base update uses the `model.attributes` values to create an `attrs` object that is then passed through the usual `.format()` method meaning that our `insert` override was not actually doing anything.
- added a failing regression test for the `formatOnWrite()` override behaviour
- adjusted our insert/update overrides to set an internal `_isWriting` property on the model, then if that property is true our `.format()` override (which is called by Bookshelf on a generated `attrs` object during inserts) we manually call our `.formatOnWrite()` method
- updated both overrides even though `update` was working for consistency and less cognitive overhead for reasoning between two different approaches
refs https://github.com/TryGhost/Team/issues/1149
Ghost allows different themes to have the same `name` value in `package.json` but gives them a custom name based on the zip file used to upload it. The custom theme settings service however was always using the package.json name meaning that when syncing a theme with a duplicate name it was overwriting the settings for all other themes.
- bumped `@tryghost/custom-theme-settings-service` which changes the `.activateTheme()` signature to accept a theme name as the first argument
- updated our calls to `.activateTheme()` to pass Ghost's known theme name as the first argument so we're always syncing with the correct `theme` field value in the `custom_theme_settings` table
no-issue
* Removed NOT_NULL constraint from stripe_coupon_id
When handling disconnecting from Stripe - we remove all Stripe data from
our database to ensure we do not have bad/invalid data stored. Removing
this constraint will allow us to set the value to NULL.
* Added created_at column to offer_redemptions
Offer Redemptions are not just a joining table, but an event. A created_at
date allows them to be ordered
Because this is in alpha it is simpler to just drop the tables and
re-add them, due to offer_redemptions depending on offers, we also drop
this table and re-add it.
refs https://linear.app/tryghost/issue/CORE-1/multiple-adapters-per-type
- While digging around this area refactored the code to use more readable async/await syntax instead of chaining then's
- Dropped unneeded "catch" block in save method as all it was doing was rethrowing an error with out additional handling
