refs f39d1d3aa3
- similar to the commit above, the JSON parser changed between Node 18
and Node 20, so the error message changed too
- we actually just want to check the error is forwarded to the user, so
we can do that by getting the error message from JSON.parse and check
against that
ref https://linear.app/tryghost/issue/MOM-48
This required some structural changes to our NestJS setup so that we can mount
it on multiple parts of the Ghost express app.
We've used the RouterModule to allow adding submodules that are mounted on
different paths, and we've had to be explicit about the base path for each
module. We've also had to switch back to using the Module decorator, because
RouterModule doesn't work with DynamicModule definitions.
Now that the NestJS app has knowledge of the full path, we need to "reset" the
url & baseUrl when passing the request into NestJS so that it can correctly
match the path. This is probably needed for the frontend too, for subdirs, but
that causes further issues - as this in prototype stage, we'll look later
Another issue is that NestJS replaces the express app instance with its own,
which isn't an issue for the Admin API (though we've fixed it anyway for
consistency), but did cause problems for the frontend, because the express app
is where view engine and directory information is stored.
The fix for this is to save a reference to the original ghost express
application, and reattach it to the request if it is not handled by Nest
Now that we have the Nest app mounted on the frontend, we're able to have it
handle the /.well-known/webfinger route with a proper controller, which is nice!
ref MOM-31
ref https://linear.app/tryghost/issue/MOM-31
We'll be building a lot of the new code for ActivityPub in Nest, so we'll need
to have it enabled in Ghost to work.
- The RSS cache has lived for a really long time, but I'm not sure it's useful
- Want to be able to determine if it gets used much, and if not, then we can remove it
ref 78311591d0
- updated tests to not click a button on the setup/done screen that is no longer shown
- fixed setup flow showing an alert bar due to not handling the `TransitionAborted` error that is thrown by the setup/done->dashboard redirect
ref https://linear.app/tryghost/issue/KTLO-1/members-spam-signups
- Some customers are seeing many spammy signups ("hundreds a day") — our
hypothesis is that bots and/or email link checkers are able to signup by
simply following the link in the email without even loading the page in
a browser.
- Currently new members signup by clicking a magic link in an email,
which is a simple GET request. When the user (or a bot) clicks that link, Ghost
creates the member and signs them in for the first time.
- This change, behind an alpha flag, requires a new member to click the
link in the email, which takes them to a new frontend route `/confirm_signup/`, then submit a form on the page which sends a POST request to the
server. If JavaScript is enabled, the form will be submitted
automatically so the only change to the user is an extra flash/redirect
before being signed in and redirected to the homepage.
- This change is behind the alpha flag `membersSpamPrevention` so we can
test it out on a few customer's sites and see if it helps reduce the
spam signups. With the flag off, the signup flow remains the same as
before.
ref https://linear.app/tryghost/issue/ENG-790/remove-use-of-sub-queries-in-email-analytics
- the `delivered_at` column is typically entirely/nearly entirely filled with values meaning the `IS NOT NULL` query matches a huge number of rows that MySQL has to fetch from the index to count
- using `IS NULL` switches that behaviour around as it will now match very few rows which has been shown in testing to be considerably quicker
- after switching to `IS NULL` the query returns an "undelivered" count rather than a "delivered" count, in order to keep the rest of the system behaviour the same we can calculate the delivered count by subtracting the query result from the total number of emails sent which we can fetch using a very fast primary key lookup query on the `emails` table
closes https://linear.app/tryghost/issue/ENG-790/remove-use-of-sub-queries-in-email-analytics
Avoiding sub queries means we don't have a process tied up for longer than necessary and we can more easily see if one of the queries is non-performant.
- extracted the count queries into separate queries and used the retrieved values in the final update query
- removed a query by moving the email open rate calculation into JS as we've already fetched the necessary data before that point
ref https://linear.app/tryghost/issue/CFR-13
- enabled saving traces on browser test failure; this makes troubleshooting a lot easier
- updated handling in offers tests to ensure the tier has fully loaded in the UI (not just `networkidle`)
- updated publishing test to examine the publish button reaction to the save action response instead of a 300ms pause
In general, our tests use a lot of watching for 'networkidle' - and sometimes just raw timeouts - which do not scale well into running tests on CI. In particular, 'networkidle' does not work if we're expecting to see React components' state updates propagate and re-render. We should always instead look to the content which encapsulates the response and the UI updates. This is something we should tackle on a larger scale.
ref ENG-774
ref https://linear.app/tryghost/issue/ENG-774
Staff Tokens will have both a `user` and an `apiKey` present on the
`loadedPermissions`.
The check here for `apiKey` was written when we could assume that an
`apiKey` was an Admin Integration - so it completely overwrote the
previous `allowed` list. When we added the concept of Staff Tokens -
this resulted in a privilege escalation.
This is a good lesson in not using proxies or indicators for data, as
changes elsewhere can invalidate them - if we had been specific and
checked the role of the current actor we wouldn't've had this bug!
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>
ref ENG-728
ref https://linear.app/tryghost/issue/ENG-728
This is not used anywhere, and makes the code more complicated, it's a good
step toward simplifying permissions and pulling them out of the database.
ref ENG-728
ref https://linear.app/tryghost/issue/ENG-728
This is NOT a functionality change. The Post#permissible method unit
tests have been updated to pass `true` as `hasUserPermission` and we can
see that the permission functionality remains the same.
The permissible method of the post model is responsible for removing
permission based on the data that is being modified, but the permissions
module is setup to allow the permissible method to grant permission -
this means that we call permissible, even if the current actor doesn't
have permission, this results in code that is hard to understand and
manage.
We are going to be instead returning early if an actor does not have
permission, this will allow permissible method signatures to be greatly
simplified (removing the need for hasUserPermission, hasApiKeyPermission
& hasMemberPermission arguments).
fixes https://linear.app/tryghost/issue/ENG-746/http-500-responses-when-handle-image-sizes-middleware-hits-missing
- in the event a request comes in for a resized image, but the source
image does not exist, we return a rendered 404 page
- we do this because we pass the NotFoundError to `next`, which skips
over the static asset code where we return a plaintext 404
- also included a breaking test that ensure we go to the next middleware
without an error
ref ENG-761
ref https://linear.app/tryghost/issue/ENG-761
Creating these pipelines is expensive, and we don't want to do it
repeatedly for the same controller. Adding caching should reduce the
amount of time spent setting up pipelines for each usage of the `get`
helper.
ref https://linear.app/tryghost/issue/IPC-66/onboarding-checklist-v1
- Adds a basic version of a new onboarding checklist behind the feature
flag, without incomplete/complete state logic
- Links to Design settings, Members screen and new post
- Opens amodal that we’ll use as Share modal
---------
Co-authored-by: Daniël van der Winden <danielvanderwinden@ghost.org>
ref [ENG-747](https://linear.app/tryghost/issue/ENG-747/)
ref https://linear.app/tryghost/issue/ENG-747
H'okay - so what we're trying to do here is make get helper queries more
cacheable. The way we're doing that is by modifying the filter used when
we're trying to remove a single post from the query.
The idea is that we can remove that restriction on the filter, increase
the number of posts fetched by 1 and then filter the fetched posts back
down, this means that the same query, but filtering different posts,
will be updated to make _exactly_ the same query, and so share a cache!
We've been purposefully restrictive in the types of filters we
manipulate, so that we only deal with the simplest cases and the code is
easier to understand.
ref ENG-742
ref https://linear.app/tryghost/issue/ENG-742
We don't do any parsing of layouts in gscan, which means themes can be
uploaded which use non-existent files for their layout.
We can catch the error in the res.render call, and wrap it, just like we
do for missing templates (e.g. the StaticRoutesRouter)
closes ENG-632
- This listens to a new property in the `milestones` config to set a minimum value of Milestones we wanna use the Slack notification service for
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-747
ref https://linear.app/tryghost/issue/ENG-747
H'okay - so what we're trying to do here is make get helper queries more
cacheable. The way we're doing that is by modifying the filter used when
we're trying to remove a single post from the query.
The idea is that we can remove that restriction on the filter, increase
the number of posts fetched by 1 and then filter the fetched posts back
down, this means that the same query, but filtering different posts,
will be updated to make _exactly_ the same query, and so share a cache!
We've been purposefully restrictive in the types of filters we
manipulate, so that we only deal with the simplest cases and the code is
easier to understand.
ref https://linear.app/tryghost/issue/ENG-721
- when changing the response to a `204` for requests with no cookie we'd lost the `Cache-Control: public, max-age: 0` header which meant some cache systems weren't caching as efficiently as possible
ref https://linear.app/tryghost/issue/TRI-65
In the context of referrals, we want to understand how useful our
“Powered by Ghost” badges are.
There are currently a few places where the “Powered by Ghost” badge can
be rendered:
- in newsletters (enabled/disabled by publisher, on a newsletter basis)
- in Portal popups, e.g. member signup/signin/account settings
- in the footer of some themes, including Source & Casper
We're adding the query param ?via to evaluate the usage of the badge in
newsletters.
no-issue
This adds the barebones of a NestJS application wired up to the Admin API
behind a feature flag, so that we can experiement with how to use Nest in the
context of Ghost
closes ENG-730
closes https://linear.app/tryghost/issue/ENG-730/
We've updated the input serializer to parse the filter, and responded
with an error if it cannot be parsed correctly.
Now that it's parsed, we can pass a mongo query object through the
stack, which will lend itself to better typing for this code, which is a
direction we want to go in anyway. We've had to update all the internal
usages of the `browse` method to use mongo query objects.
ref https://linear.app/tryghost/issue/ENG-740/http-500-error-when-image-processing-fails
refs 4aad551c72
- upon further discussion, we've decided it's better to throw an error
in this case because the uploaded image is deemed invalid and storing
it on the filesystem might cause more issues with resizing/further
processing in the future
- this commit implements that and alters the tests
fixes ENG-740
fixes https://linear.app/tryghost/issue/ENG-740/http-500-error-when-image-processing-fails
- in the event the image transform library throws (which can happen for
many reasons; sharp/libvips can come across a number of errors), we
currently return this as a HTTP 500 error to the user
- in this case, we should just try-catch the call and jump to the
non-processing flow where it just saves the original image
- also added breaking test
closes https://linear.app/tryghost/issue/ENG-721
ref https://linear.app/tryghost/issue/ENG-708
Comments-UI loads `/ghost/admin-frame/` in an iframe to check if a Staff User is authenticated in order to show moderation options. That iframe request loads a HTML page which in turn contains a script that fires off an API request that attempts to fetch the logged-in user details, resulting in a 403 "error" showing up when not authenticated. In the vast majority of cases there will be no staff user authenticated so lots of extra requests and "errors" are seen unnecessarily.
- adjusted the `/ghost/auth-frame/` endpoint to check if the request contains an Admin session cookie
- if it does, continue as before with rendering the HTML page so the script is loaded
- if it doesn't, return an empty 204 response avoiding the script request and subsequent 403-generating API request
- eliminates the 403 error being generated for all typical visitor traffic, the error should only be seen when an Admin was previously logged in but their cookie is no longer valid (either from logging out, or going past the 6month validity period)
fixes https://github.com/TryGhost/Product/issues/4237
- this fixes the fact that we return a HTTP 500 response when the oembed
library receives an error, such as a 401 or 403
- includes special handling for cases where we want to return a slightly
different error message
- also adds unit tests for @tryghost/oembed-service package
fixes ENG-733
ref https://linear.app/tryghost/issue/ENG-733/handle-image-uploads-where-name-is-too-long
- filesystems usually have a filename length limit; ie. on macOS it is
255 characters
- if a file is uploaded with a longer filename, we'll return a HTTP 500
- we shouldn't do this as it is user error, so we can just catch the
error code and return BadRequest
- this implements that, and adds a breaking test
- 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
refs INC-36
fixes https://github.com/TryGhost/Ghost/issues/19796
- The tiers-only paywall was incorrectly rendering "Free". Example:
"This post is for subscribers of the Free, Silver and Gold tiers only"
- Steps to reproduce the issue:
1. Create a post with public visibility, publish it
2. Then swap the visibility to specific tiers. The default selects all
paid tiers. Leave it like that
3. Update the post. The paywall show Free, even though it should be
showing only the paid tiers
- This fix filters out the "free" tier when visibility is set to tiers,
before updating a Post or a Page. The fix includes bulk updates from the
list of Posts and Pages (right-click on a Post/Page > Change Access).
refs INC-36
fixes https://github.com/TryGhost/Ghost/issues/19796
- The tiers-only paywall was incorrectly rendering "Free". Example:
"This post is for subscribers of the Free, Silver and Gold tiers only"
- Steps to reproduce the issue:
1. Create a post with public visibility, publish it
2. Then swap the visibility to specific tiers. The default selects all
paid tiers. Leave it like that
3. Update the post. The paywall show Free, even though it should be
showing only the paid tiers
- This fix filters out the "free" tier when visibility is set to tiers,
before updating a Post or a Page. The fix includes bulk updates from the
list of Posts and Pages (right-click on a Post/Page > Change Access).
no issue
Bumps `Comments-UI` app version that contains an improvement to data loading:
- within the comments block we only use Admin auth to show moderation options on each displayed comment but we were always pre-emptively loading the `admin-auth` frame and making the associated Admin API user request. That loading has now been deferred until at least one comment has been displayed cutting down unnecessary requests on each post view
no issue
Bumps `Comments-UI` app version that contains an improvement to data loading:
- within the comments block we only use Admin auth to show moderation options on each displayed comment but we were always pre-emptively loading the `admin-auth` frame and making the associated Admin API user request. That loading has now been deferred until at least one comment has been displayed cutting down unnecessary requests on each post view
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 │
└─────────────┴─────────┴────────────┴─────────┴───────┘
```
refs https://linear.app/tryghost/issue/ENG-712/
I don't think we ever need to respond with a 500 here, if the verify call
fails, we know that the token is unauthorized for use.
