refs #9601, refs #9744
- the express router reference wasn't updated fully
- the stack was the old router stack when you have uploaded a new routes.yaml file
- this has caused e.g. that new redirects for channels/collections didn't work after the upload
refs #9601
- when using the short form `data: tag.welcome` the redirect is enabled by default
- /tag/welcome/ will redirect to the channel/collection which makes use of the data key
- you can disable the redirect by using the long form
e.g. data:
tag:
resource: tags
type: read
slug: welcome
redirect: false
closes#9774, refs #9742
- added new fixture posts for Ghost 2.0
- added migration file to remove old fixture posts
- only remove them if they are owned by the Ghost author and if they are tagged with getting-started
- added new fixture posts if you had all (!) old fixture posts
- ensure on rollback we remove the new fixture posts again
- updated default settings
refs #9742
- removed usage of single permalink setting
- with dynamic routing this configuration does no longer makes sense
- because you can configure your permalinks in the routes.yaml
- furthermore you can have multiple collections with multiple permalinks
- removed @blog.permalinks
- do not export permalink setting
- do not import permalink setting
- permalink setting UI will be removed soon
- get rid of {globals.permalink} completely
- remove yaml in-built migration
- do not expose settings.permalinks via the private API
- do not allow to edit this setting
- keep phyiscal value in case a blog needs to rollback from v2 to v1
- sorted out when the routers should be created
- ensure routes.yaml file doesn't get validated before Ghost is fully ready to start
refs #9601, refs #9742
- Upgraded NQL to 0.1.0
- The new version of NQL supports aliases e.g. `tag: tags.slug`, which makes it possible to define `filter=tag:support`
- Furthermore, this allows us to support advanced filtering like tag:[a,b]
- In dynamic routing, we use mingo via NQL which has a slightly different feature set to GQL in the API:
- AND NOT, OR and other advanced logic combos DO work on joined tables
- Counts are not yet supported
- The Dynamic Routing beta docs should describe that API filtering and Dynamic Routing filtering is different
refs #9742, refs #9724
- handle König Editor format for 2.0
- adapted importer to be able to import 1.0 and 2.0 exports
- added migration scripts
- remove labs flag for Koenig
- migrate all old editor posts to new editor format
- ensure we protect the code against mobiledoc or html field being null
- ensure we create a blank mobiledoc structure if mobiledoc field is null (model layer)
- ensure you can fully rollback 2.0 to 1.0
- keep mobiledoc/markdown version 1 logic to be able to rollback (deprecated code)
refs #9744
- added two new endpoints to upload/download routes.yaml
- reload site express app on successful/valid upload
- reload url service on sucessfuly upload
- force clear cache of pages
- ensure we keep a backup of the routes.yaml file
- this feature was mostly tested manually
- @TODO: i have to write unit tests - will do later
- @TODO: do a memory test to ensure we haven't introduced any memory leaks with this feature
refs #9742
- rebase against master updated some docs links again
- go over code base again and double check that all docs links are correct
- 2.0 will become the latest version on our readme pages
refs #9742, refs https://github.com/TryGhost/Ghost-CLI/issues/759
- required a reordering of Ghost's bootstrap file, because:
- we have to ensure that no database queries are executed within Ghost during the migrations
- make 3 sections: check if db needs initialisation, bootstrap Ghost with minimal components (db/models, express apps, load settings+theme)
- create a new `migrator` utility, which tells you which state your db is in and offers an API to execute knex-migrator based on this state
- ensure we still detect an incompatible db: you connect your 2.0 blog with a 0.11 database
- enable maintenance mode if migrations are missing
- if the migration have failed, knex-migrator roll auto rollback
- you can automatically switch to 1.0 again
- added socket communication for the CLI
refs #9742
- rename column from amp -> comment_id
- iterate over all posts and ensure we use the resource id or the original amp value
- provide down hook to undo this change
no issue
- these tests were a mess
- we had many duplicated tests
- it was very hard to work with the exporter files
- i have tidied up the whole file
- first section: all general importer tests based on the current Ghost version
- second section: 1.0 tests
- everything is now JSON based (much easier to control)
refs 9742
- when we've introduced Ghost 1.0, we have noticed that we broke Disqus comments
- Disqus comments use a unique identifier, which is the post id
- that means if you have exported your LTS content and imported it into 1.0, all resource identifiers are regenerated
- but the Disqus must use the original post resource id
- that's why we have imported the old post id and remembered it in the `amp` field 🤠
- that was the only field which was available and un-used
- now in Ghost 2.0, we would like to rename the `amp` field to `comment_id`
- string length: 50 (i thought it might be safer to not use 24 characters, because who knows if support other comment id's in the future)
refs https://github.com/TryGhost/Ghost/issues/9724
- perform a HEAD request on a url if we don't find a matching provider, following any redirects until we hit success response before looking up providers for the resulting url
refs #9601
- on resource update/add we have cached mobiledoc, html etc
- we have to ensure we exclude the fields (same procdure happens on bootstrap)
- these excluded fields don't have to be cached
- otherwise memory usage is higher in general
- ensure we cache relations with a minimal field set on resource update/add
closes#9587
- when `useNullAsDefault` was invented in knex 0.10, it was a breaking change, that update/insert etc no longer set's null as default if a field is missing
- at this time we thought it only affects our test env, because the test generator doesn't generate all fields
- but turned out the importer is affected as well e.g. you import a post with missing fields
- the importer doesn't iterate over all fields and checks if the field is present or not
- as this only happens with SQlite3, we should enable `useNullAsDefault` by default
- you can still disable this option if you want, but not recommended
- the reason why knex added this breaking change was that some applications want "undefined" as value
- this is not the case in Ghost, so it's fine to make use of the default null behaviour
refs #8719
- initial commit: 40c8eacd44
- we have forgotten that there is another endpoint which triggers an export (the backup endpoint)
- this endpoint needs to accept the new `include` query param as well (was missing)
no issue
- discovered while testing
- SQLite3 has a restriction that you cannot query more than 999 SQL variables (e.g. fetch all tag relations for all posts)
- if you have more than 999 posts in your database, Ghost wasn't able to finish the bootstrap
- the url service tried to fetch all posts & relations, but could not complete, because SQlite3 has thrown
-> "too many SQL variables"
- i have added a recursive query strategy only for SQLite3
- use offset/limit
refs #9742, refs #8719
- make it possible to import more tables (optional)
- available tables: clients, trusted domains
- by default we won't import these tables, you have to tell Ghost using `include` (same syntax on export)
- we won't announce this ability for now (stays hidden)
refs #9742, refs #8719
- you can now use `include` to export extra tables e.g. `include=clients`
- admin client won't make use of this option yet, maybe later and optional
- we won't announce this new ability for now (stays hidden)
refs #9751
- the mobiledoc field can be null
- e.g. if you import a JSON with no markdown/mobiledoc or html field
The migration script for 1.25 had only the purpose to migrate existing Koenig Beta posts.
no issue
- add a new migration for 1.25 to insert the draft demo post for existing blogs
- ensure new blogs get the draft demo post as well
- tested on sqlite3 + mysql
- added handling if Ghost Author user doesn't exist anymore (fallback to owner user)
refs https://github.com/TryGhost/Ghost/issues/9742
We've identified some changes we need to make to the HTML output of the [new Koenig editor](
https://forum.ghost.org/t/koenig-editor-beta-release/1284/102) for future proofing and consistency across cards.
- the `<div class="kg-post">` wrapper around post content has been removed
- for image cards the `.kg-image-wide` and `.kg-image-full` classes have been changed to `.kg-width-wide` and `.kg-width-full` and applied to the `<figure>` element rather than the `<img>` element
Before:
```html
<div class="kg-post">
<figure class="kg-image-card">
<img class="kg-image kg-image-wide" src="...">
<figcaption>example wide image</figcaption>
</figure>
</div>
```
After:
```html
<figure class="kg-image-card kg-width-wide">
<img class="kg-image" src="...">
<figcaption>example wide image</figcaption>
</figure>
```
refs #9742
- Ghost 2.0 is coming
- all doc links in 1.0 must use concrete links e.g. docs.ghost.org/v1 or themes.ghost.org/v1.23.0/
- if we release Ghost 2.0, docs.ghost.org will show 2.0 docs
closes#9727
- this is a short term fix to proof that Disqus comments on preview pages no longer appear on other threads
- this is not a full solution to the problem
- the private API still returns /404/, which is right now inconsistent, but not critical in any way
- the url helper will now output the post preview url if you serve a draft/scheduled post
- this should register unique page urls at Disqus and ensure uniquness for threads
- i still don't understand why the cross posting happens at all, because we also pass an unique identifier to Disqus (the post ID)
- it could be that comments, which are added on the preview page, won't appear on the published urls, because
the published url !== preview url. I wasn't able to figure this out via testing or reading their docs
refs https://github.com/TryGhost/Ghost/issues/9623
- added `DomModifier` class to walk a SimpleDom document and modify as needed
- adds `id` attributes to `h1`, `h2`, etc heading tags
- converts H* tag content to a dasherized string for the id attribute (dasherized id's are different to the smushed ids that are generated by our markdown converted but there are no backwards-compatibility concerns here)
- if a duplicate id is detected then add a `-1`, `-2`, etc suffix to the id
- use `DomModifier` after converting mobiledoc to SimpleDom but before serialising to html
- switched top-level var declarations to es6
closes#9715
- changed the `urlWithoutSubdirectoryWithoutAmp` variable in the amp
router to only match /amp or /amp/ at the end of the url string, instead
of just matching any occurrence of /amp in the url string
no issue
- replaced `querySelector` with `querySelectorAll` and a loop so that all subscribe form inputs have their values updated rather than only the first form on the page
- made the selector more specific so that it only updates `<input>` elements
- switched to a template string so it's easier to read/write
refs #9601
- this was already working for collections or channels
- but the `routeName` was not parsed for static routes
- ensure we push the route name into the context object
e.g. /about/: about
-> name of the route is "about"
no issue
- discovered while testing (issue doesn't exist)
- with dynamic routing we have introduced a bug
- we pre-generate urls and if your permalink contains a date (dated permalink),
we have to regenerate the urls, otherwise the dated urls do not respect your blog timezone
- collection router has to subscribe to the timezone event
- collection router must trigger an update on it's url generator if the timezone changes and the it's permalink is dated
- ensure we also update the urls on import
no issue
- reported in the forum: https://forum.ghost.org/t/publishing-with-a-single-post-request-to-posts/1648
- the defaults are defined in two places
1. on the schema level (defaults for the database)
2. on the ORM (model layer)
- the defaults on the db layer are set correctly when inserting a new resource
- but if we don't apply all defaults on the model layer, it will happen that model events are emitted without the correct defaults
- see comment in code base
- it's caused by the fact that knex only returns the inserted resource id (probably caused by the fact knex has to support x databases)
- components/modules are listening on model events and expect:
1. a complete set of attributes
2. a complete set of defaults
3. sanitized values e.g. bool, date
- this commit fixes:
1. added missing defaults for user & post model
2. sanitize booleans (0|1 => false|true)
3. added tests to ensure this works as expected
4. clarfies the usage of `defaults`
Regarding https://forum.ghost.org/t/publishing-with-a-single-post-request-to-posts/1648:
- the post event was emitted with the following values {page: undefined, featured: undefined}
- the urlservice receives this event and won't match the resource against collection filters correctly
- NOTE: the post data in the db were correct
- it’s good practice with password reset emails to offer peace of mind
to the user if they didn’t initiate the request and to let them know
that if they didn’t, it’s safe to ignore
refs #9601
- replace jsonpath with [NQL](https://github.com/NexesJS/NQL)
- jsonpath was just a temporary solution (a short-term fix)
- with NQL we are able to filter collections more powerful in the near future
- NQL is not feature complete
- we still support `featured:true` for collections
refs #9601
- you can now use `rss:false`
- ability to define a custom rss url with a target template (+ content_type)
- ability to disable rss for channel or collection
refs #9601
- refactor architecture of routing so you can define a channel
- a channel is a different way of looking at your posts (a view)
- a channel does not change the url of a resource
Example channel
```
routes:
/worldcup-2018-russia/:
controller: channel
filter: tag:football18
data: tag.football18
```
- added ability to redirect resources to a channel/static route
- support templates for channels
- ensure we still support static routes (e.g. /about/: home)
- ensure pagination + rss works out of the box
refs #9601
- you can define a redirect in your routes.yaml
e.g. from `page.home` to /about/
- we have to check for a possible redirect before rendering the target static page
refs #9601
- support data, limit and order for collections
- limit definition in routes.yaml is stronger than theme package.json limit configuration
- ensure we update hbs template options
refs #9601
- the home.hbs behaviour for the index collection (`/`) is hardcoded in Ghost
- we would like to migrate all existing routes.yaml files
- we only replace the file if the contents of the routes.yaml file equals the old routes.yaml format (with home.hbs as template)
- updated README of settings folder
- if we don't remove the home.hbs template from the default routes.yaml file, home.hbs will be rendered for any page of the index collection
- the backwards compatible behaviour was different
- only render home.hbs for page 1
- remember: the default routes.yaml file reflects how Ghost was working without dynamic routing
refs #9601
Example:
```
collections:
/podcast/:
permalink: /{slug}/
```
- the name of the collection is remembered as `routerName` (in the case above: "podcast")
- the name of the collection is important for two things
1. context value
2. template name
- the context value is available for specific theme helpers e.g. is helper, body_class helper
- we auto-lookup the collection name in your theme e.g. podcast.hbs
- this logic does not apply to static routes
- if you define templates on your collection, they are stronger than the collection name
refs #9601
- implementation of resource listener updated
- if you define two collections: `featured:true` (1) and `page:false` (2) you can run into the following bug:
- you create a published post (owned by (2))
- you change the status to featured
- still owned by (2), because the filter still matches (it's still not a static page)
- this adaption fixes the behaviour
- less smart logic, but less error prone
closes https://github.com/TryGhost/Ghost/issues/9674
- with dynamic routing the first collection get's the "index" context attached
- the index context signalises the main post listening route (first collection)
- this behaviour was present < 1.24 - we have to keep this behaviour
refs #9601
- if you call `express.Router()`, the router's name is always "router"
- that is caused by the closure behaviour in express:
- https://github.com/expressjs/express/blob/4.16.3/lib/router/index.js#L46
- Ghost creates a couple of express routers for dynamic routing
- it depends how much you configure in your routes.yaml file
- but every router is called "router"
- this is hard to work with
- with this router wrapping logic, we are able to give each router an exact name
If you enable `DEBUG=ghost:services:routing:*`, you have seen this before
> ghost:services:routing:ParentRouter site: mountRouter: router +0ms
With the wrapper logic, you will see:
> ghost:services:routing:ParentRouter site: mountRouter: StaticPagesRouter +0ms
- furthermore, if you have to access the router stack (`app.router.stack`), you can easily identify and find router instances by name
no issue
- if you define no collections, but a static route, it can happen that the target template to render
makes use of the {{ghost_head}} helper
- the {{ghost_head}} helper tries to create the primary rss feed url
- at the moment: no collections, no primary rss feed url
- if we offer the option to define custom rss rules, this function might need an extension
closes#9675
- with dynamic routing we have introduced a breaking change, which we have overseen
- Ghost does not return absolute urls, that's why the clients need to concat the blog url and the resource url
- with 1.24.0 Ghost returned resource urls including the subdirectory
- this caused trouble for e.g. zapier or the preview feature in the admin client
- revert breaking change and ensure we only expose resource urls without subdirectory
refs https://github.com/TryGhost/Ghost/issues/9623
- add `oembed-parser` module for checking provider availability for a url and fetching data from the provider
- require it in the `overrides.js` file before the general Promise override so that the `promise-wrt` sub-dependency doesn't attempt to extend the Bluebird promise implementation
- add `/oembed` authenticated endpoint
- takes `?url=` query parameter to match against known providers
- adds safeguard against oembed-parser's providers list not recognising http+https and www+non-www
- responds with `ValidationError` if no provider is found
- responds with oembed response from matched provider's oembed endpoint if match is found
no issue
- was introduced with dynamic routing beta: https://github.com/TryGhost/Ghost/releases/tag/1.24.0
- the slug param wasn't forwarded correctly
- you were not able to render a custom tag or author template e.g. `tag-news.hbs`
refs #9681
- we already had a protection against these situations when serving the site (theme)
- it can happen that we have to initialise the express engine in the error handler in case the first request to /ghost produces an error (e.g. 503)
- otherwise the underlying error message is hidden and Ghost doesn't render the error html template correctly
closes#9674
- the collection router had a hardcoded default context "home"
- this is wrong
- the context array get's automatically filled for the collection
- if you are serving a page e.g. /page/2/ -> it's "paged"
- if you are serving / -> it's "home"
- same for {{body_class}}, it outputs "home-template" on "/"
- this is the same behaviour as in 1.23.x
no issue
- reverse must happen once in the constructor
- otherwise we reverse the array on each request
- Ghost would randomly pick the first and then the second template
no issue
- from now on: you have to manually reconfigure your slack hook after importing your data
- we were running into trouble that Ghost had import slack hooks, because it can happen very fast
that you are importing someone's slack hook
no issue
- there was a timing bug in Ghost
- we do operations in parallel on bootstrap
- 1) we fetch the resources as early as possible
- 2) we do all the rest (express bootstrapping, theme loading, router registration) etc.
- it can happen that (2) happens too slow and ends in the situation that the queue, which is responsible
to handle both parallel actions, does not wait for the routers and closes the event
- this is a short term fix
- i need to reconsider if there is a better long term fix
refs #9601
### Dynamic Routing
This is the beta version of dynamic routing.
- we had a initial implementation of "channels" available in the codebase
- we have removed and moved this implementation
- there is now a centralised place for dynamic routing - server/services/routing
- each routing component is represented by a router type e.g. collections, routes, static pages, taxonomies, rss, preview of posts
- keep as much as possible logic of routing helpers, middlewares and controllers
- ensure test coverage
- connect all the things together
- yaml file + validation
- routing + routers
- url service
- sitemaps
- url access
- deeper implementation of yaml validations
- e.g. hard require slashes
- ensure routing hierarchy/order
- e.g. you enable the subscriber app
- you have a custom static page, which lives under the same slug /subscribe
- static pages are stronger than apps
- e.g. the first collection owns the post it has filtered
- a post cannot live in two collections
- ensure apps are still working and hook into the routers layer (or better said: and register in the routing service)
- put as much as possible comments to the code base for better understanding
- ensure a clean debug log
- ensure we can unmount routes
- e.g. you have a collection permalink of /:slug/ represented by {globals.permalink}
- and you change the permalink in the admin to dated permalink
- the express route get's refreshed from /:slug/ to /:year/:month/:day/:slug/
- unmount without server restart, yey
- ensure we are backwards compatible
- e.g. render home.hbs for collection index if collection route is /
- ensure you can access your configured permalink from the settings table with {globals.permalink}
### Render 503 if url service did not finish
- return 503 if the url service has not finished generating the resource urls
### Rewrite sitemaps
- we have rewritten the sitemaps "service", because the url generator does no longer happen on runtime
- we generate all urls on bootstrap
- the sitemaps service will consume created resource and router urls
- these urls will be shown on the xml pages
- we listen on url events
- we listen on router events
- we no longer have to fetch the resources, which is nice
- the urlservice pre-fetches resources and emits their urls
- the urlservice is the only component who knows which urls are valid
- i made some ES6 adaptions
- we keep the caching logic -> only regenerate xml if there is a change
- updated tests
- checked test coverage (100%)
### Re-work usage of Url utility
- replace all usages of `urlService.utils.urlFor` by `urlService.getByResourceId`
- only for resources e.g. post, author, tag
- this is important, because with dynamic routing we no longer create static urls based on the settings permalink on runtime
- adapt url utility
- adapt tests
no issue
- the dot notation only works if you install a single lodash dependency e.g. `yarn install lodash.get`
- otherwise we have to use `lodash/get`
no issue
- discovered while coding
- the value was always false, because we've tried to read the value from the config object
- the value lives in the database and is accessible via the labs service
no issue
- we sanitise any incoming slug on the model layer e.g uppercase -> lowercase
- and when importing e.g. an uppercase slug, the importer was trying to compare the uppercase slug with the sanitised slug
closes https://github.com/TryGhost/Ghost/issues/9620
- adjust the `deduplicateSubDir` function's regex to only match duplicate subdirectories when the `url` is only a path rather than full url or the duplicate match starts with a `/`
closesTryGhost/Support#426
refs TryGhost/gscan#106
needs TryGhost/gscan#107
GScan can return errors, which was not handled in our theme validator and caused Ghost to crash completely. GScan will now return an Ignition error when its not able to read the `.zip` file.
e. g.: `{"errors":[{"message":"Failed to read zip file","context":"tife.zip","errorType":"ValidationError","errorDetails":"invalid relative path: ../tife/"}]}`
refs https://github.com/TryGhost/Ghost/issues/9505
- remove requirement for the `enableDeveloperExperiments` flag to be able to use Koenig
- it's now possible to enable as a standard Labs beta feature
refs https://github.com/TryGhost/Ghost/issues/9505
- updates mobiledoc converter's `render` method to accept a `version` argument
- `1` === Ghost 1.0's markdown-only renderer output
- `2` === Koenig's full mobiledoc renderer output
- switch between mobiledoc renderer versions in Post model's `onSaving` hook
- version 1 by default
- version 2 if Koenig is enabled (currently behind dev experiments config + labs flag)
- version 2 if the post's mobiledoc is not compatible with the markdown-only renderer
- "version 2" full-Koenig mobiledoc renderer output
- wraps content in a `.kg-post` div
- removes wrapper around markdown and html card output
- adds classes to image card output including selected image size/style
- standardises es6 usage across mobiledoc related files
refs #9601
- while i was testing different collections and different filters, i somehow thought that the default
collection does not contain featured posts 😀🙊
- this is wrong (!!!!)
- the url service is not yet connected
- so: this is not a bug
refs https://github.com/TryGhost/Ghost/pull/9592
- we add bookshelf-relations step by step if we need it
- with https://github.com/TryGhost/Ghost/pull/9592 we have rewritten the test env to use Bookshelf
- this is important for our new url service
- because the service is listening on model updates and updates the urls based on the model events
- so with moving to Bookshelf, we need any easy way to add relations
- the test env inserts test fixtures
- it adds permissions and each permission get's roles attached
- `models.Permission.add({roles: [...]})
no issue
- replaced token creation by `lib.common.security`
- added unit tests for adding invites
- allow a different invite status for internal access
refs https://github.com/TryGhost/Team/issues/65
- it's easier for the architecture if we read the setting files synchronously,
because the dynamic routing component is part of the express bootstrap and
the whole routing bootstrap is synchronously
- for now: we only read one file anyway
- it's for now easier to read the file synchronously, then i don't have to change
any existing express bootstrap architecture
refs https://github.com/TryGhost/Team/issues/65
- this is just the first optimisation regarding relative/absolute urls
- the full strike will happen when i start with the url utility re-write
- for now: there will be only one subscriber of url events -> the sitemaps service
- the sitemaps service outputs absolute urls
- we don't want to receive an url event and ask the url service again to get an absolute version of the url
closes#9569
- Removed the `<1 min read` time clause, effectively making `1 min read` the minimum reading time
- Removed the `seconds` option for i18n strings, which contained the less than one minute display string
- Kept the other i18n string options the same
- Amended and improved tests for new functionality
refs #9584
- object by reference 🎡
- we modify the object later on
- this has resulted in a wrong context output in the admin client
- e.g. we've output an updated user reference
refs https://github.com/TryGhost/Team/issues/65
- currently we generate a relative resource url
- if you configure a subdirectory, the urls have to respect that
- e.g. you configure `localhost:2368/blog`, your url results in e.g. `/blog/my-post/`
- this is not yet a critical bug, because the url service is not connected yet
- @TODO: consider absolute vs. relative urls in the url service
no issue
- removed the `routeKeywords` property from the config and used hard coded keywords.
- removed `routeKeywords` from public configuration API endpoint, as it's no longer used in the Admin.
refs https://github.com/TryGhost/Team/issues/65
We are currently work on dynamic routing (aka channels).
An important piece of this feature is the url service, which always knows the url of a resource at any time.
Resources can belong to collections or taxonomies, which can be defined in a [routing yaml file](https://github.com/TryGhost/Ghost/issues/9528). We are currently shipping portions, which will at end form the full dynamic routing feature.
### Key Notes
- each routing type (collections, taxonomies, static pages) is registered in order - depending on the yaml routes file configuration
- static pages are an internal concept - they sit at the end of the subscriber queue
- we make use of a temporary [`Channels2`](https://github.com/TryGhost/Ghost/pull/9550/files#diff-9e7251409844521470c9829013cd1563) file, which simulates the current static routing in Ghost (this file will be modified, removed or whatever - this is one of the next steps)
- two way binding: you can ask for a resource url based on the resource id, you can ask for the resource based on the url
- in theory it's possible that multiple resources generate the same url: we don't handle this with collision (because this is error prone), we handle this with the order of serving content. if you ask the service for a resource, which lives behind e.g. /test/, you will get the resource which is served
- loose error handling -> log errors and handle instead of throw error and do nothing (we log the errors with a specific code, so we can react in case there is a bug)
- the url services fetches all resources on bootstrap. we only fetch and keep a reduced set of attributes (basically the main body of a resource)
- the bootstrap time will decrease a very little (depending on the amount of resources you have in your database)
- we still offer the option to disable url preloading (in your config `disableUrlPreload: true`) - this option will be removed as soon as the url service is connected. You can disable the service in case you encounter a problem
- **the url service is not yet connected, we will connect the service step by step. The first version should be released to pre-catch bugs. The next version will add 503 handling if the url service is not ready and it will consume urls for resources.**
----
- the url service generates urls based on resources (posts, pages, users, tags)
- the url service keeps track of resource changes
- the url service keeps track of resource removal/insert
- the architecture:
- each routing type is represented by a url generator
- a routing type is a collection, a taxonomiy or static pages
- a queue which ensures that urls are unique and can be owned by one url generator
- the hierarchy of registration defines that
- we query knex, because bookshelf is too slow
- removed old url service files + logic
- added temp channels alternative (Channels2) -> this file will look different soon, it's for now the temporary connector to the url service. Also the name of the file is not optimal, but that is not really important right now.
closes#9528
These code changes introduce a YAML parser which will load and parse YAML files from the `/content/settings` directory. There are three major parts involved:
1. `ensure-settings.js`: this fn takes care that on bootstrap, the supported files are present in the `/content/settings` directory. If the files are not present, they get copied back from our default files. The default files to copy from are located in `core/server/services/settings`.
2. `loader.js`: the settings loader reads the requested `yaml` file from the disk and passes it to the yaml parser, which returns a `json` object of the file. The settings loader throws an error, if the file is not accessible, e. g. because of permission errors.
3. `yaml-parser`: gets passed a `yaml` file and returns a `json` object. If the file is not parseable, it returns a clear error that contains the information, what and where the parsing error occurred (e. g. line number and reason).
- added a `get()` fn to settings services, that returns the settings object that's asked for. e. g. `settings.get('routes').then(()...` will return the `routes` settings.
- added a `getAll()` fn to settings services, that returns all available settings in an object. The object looks like: `{routes: {routes: {}, collections: {}, resources: {}}, globals: {value: {}}`, assuming that we have to supported settings `routes` and `globals`.
Further additions:
- config `contentPath` for `settings`
- config overrides for default `yaml` files location in `/core/server/services/settings`
**Important**: These code changes are in preparation for Dynamic Routing and not yet used. The process of copying the supported `yaml` files (in this first step, the `routes.yaml` file) is not yet activated.
no issue
- required for model events
- otherwise you won't receive a full data set
- in worst case you have to re-fetch the post
- required for the url service
- the url service always needs relations (authors,tags) to be able to generate the url properly
@IMPORTANT
- no API change, we still return what you are asking for
- we first edit/add the resource
- then we fetch the data with the API options
- @TODO: this can be optimised and will improve performance
picking/selecting it from the insert/update response
- this is an internal change
closes#9547
- you setup a blog with the following owner:
- email: test@ghost.org
- name: test
- slug: test
- now you import a JSON db file, which holds the exact same owner
- this owner won't be imported, because it's a duplicate
- but the slug is different (!)
- the importer tries to find a matching existing user, but won't find anything
- the importer then send an empty authors array `post.authors=[]` into the model layer
- this is not allowed -> this would mean, you are actively trying to unset all authors
no issue
- if multiple queries run in a transaction, the model events are triggered before the txn finished
- if the txn rolls back, the events are anyway emitted
- the events are triggered too early
- solution:
- `emitChange` needs to detect that a transaction is happening
- it listens on a txn event to determine if events should be triggered
no issue
- if you delete all content, we expect two events
- `post.deleted` and `post.unpublished`
- `post.unpublished` was never triggered, because the api implementation made use of `collection.invoke(`destroy`)`
- what happened?
- you fetch all posts (columns:id)
- you destroy the post (only id column is available)
- the model events are triggered
- but you have no access to a default set of data
- the result is that the event handler can't even tell if this is a post or a page
- added a proper test to ensure which events are triggered
no issue
- add a big comment
- describe:
- how this works
- why this is in place
- what does currently not work
- and why it will work with channels
- @TODO:
- figure out how to disallow:
- `models.Post.findAll({columns: id})`
- `post.save(data)`
- this will trigger bookshelf events and model events
- url generation currently needs a set of attributes (e.g. slug, published_at)
- will be auto-fixed with channels, because you can call `urlService.getUrl(post.id)`
- but what doesn't get solved is our model events
- e.g. `emitChange` needs `post.get('page')` to determine if it's a page
no issue
- this is no longer needed for now
- it was anyway a little bit ugly to modify bookshelf's `changed` object
- if we want to change something about figuring out if a model has changed (including relations)
-> we probably need to override bookshelf
refs #9548
- we always receive date strings from the client in ISO format
- we ensure that we transform these strings into JS dates for comparison
- when the client sends relations, we need to ensure that relations are checked as well
- will only work for the post model for now, because this is the only model which uses `bookshelf-relations`
- added unit tests
- removed some model tests, which do the same
refs #9548
- do not forward `tag.parent` to the model layer
- the model layer should only know `tag.parent_id`
- and the API should only expose `tag.parent` (this is an API feature)
- currently Ghost has a mixture of using `toJSON` and the API validation layer for this
- we just continue with this for now (no time to fix this)
- disallow sending nested-nested relations
- unsupported
- see comment for more information
- this can cause problems with calling `hasChanged` on relations
- add unit tests
no issue
- the previous commit will insert two post author relations if the author id of a post is invalid
- if a blog has an invalid author_id (which should be an edge case), we update the author id to the owner id
- `posts_authors` are auto inserted in this case
no issue
This PR adds the server side logic for multiple authors. This adds the ability to add multiple authors per post. We keep and support single authors (maybe till the next major - this is still in discussion)
### key notes
- `authors` are not fetched by default, only if we need them
- the migration script iterates over all posts and figures out if an author_id is valid and exists (in master we can add invalid author_id's) and then adds the relation (falls back to owner if invalid)
- ~~i had to push a fork of bookshelf to npm because we currently can't bump bookshelf + the two bugs i discovered are anyway not yet merged (https://github.com/kirrg001/bookshelf/commits/master)~~ replaced by new bookshelf release
- the implementation of single & multiple authors lives in a single place (introduction of a new concept: model relation)
- if you destroy an author, we keep the behaviour for now -> remove all posts where the primary author id matches. furthermore, remove all relations in posts_authors (e.g. secondary author)
- we make re-use of the `excludeAttrs` concept which was invented in the contributors PR (to protect editing authors as author/contributor role) -> i've added a clear todo that we need a logic to make a diff of the target relation -> both for tags and authors
- `authors` helper available (same as `tags` helper)
- `primary_author` computed field available
- `primary_author` functionality available (same as `primary_tag` e.g. permalinks, prev/next helper etc)
closes#9520
- it contains a dependency bump of the latest Bookshelf release
- Bookshelf introduced a bug in the last release
- see https://github.com/bookshelf/bookshelf/pull/1583
- see https://github.com/bookshelf/bookshelf/pull/1798
- this has caused trouble in Ghost
- the `updated_at` attribute was not automatically set anymore
---
The bookshelf added one breaking change: it's allow to pass custom `updated_at` and `created_at`.
We already have a protection for not being able to override the `created_at` date on update.
We had to add another protection to now allow to only change the `updated_at` property.
You can only change `updated_at` if you actually change something else e.g. the title of a post.
To be able to implement this check i discovered that Bookshelfs `model.changed` object has a tricky behaviour.
It remembers **all** attributes, which where changed, doesn't matter if they are valid or invalid model properties.
We had to add a line of code to avoid remembering none valid model attributes in this object.
e.g. you change `tag.parent` (no valid model attribute). The valid property is `tag.parent_id`.
If you pass `tag.parent` but the value has **not** changed (`tag.parent` === `tag.parent_id`), it will output you `tag.changed.parent`. But this is wrong.
Bookshelf detects `changed` attributes too early. Or if you think the other way around, Ghost detects valid attributes too late.
But the current earliest possible stage is the `onSaving` event, there is no earlier way to pick valid attributes (except of `.forge`, but we don't use this fn ATM).
Later: the API should transform `tag.parent` into `tag.parent_id`, but we are not using it ATM, so no need to pre-optimise.
The API already transforms `post.author` into `post.author_id`.
closes#9507
- Changed the utils.wordCount implementation to the one used by simpleMDE
- Added extra À-ÿ to the regex to support diacritics characters
- Added corresponding text with Chinese text mentioned in the issue
refs #9519
- `errors.models.posts.postNotFound` -> wrong
- `errors.models.post.postNotFound` -> correct
- the i18n lib just logs the error and falls back to a valid error key
- wrong i18n keys will never break Ghost
closes#9495
- Added a clause for amp being disabled
- In this clause, we strip the final 'amp/' part of the url, and redirect
- Changed corresponding test in frontend_spec.js
- Used `urlService.utils.redirect301()` instead of `res.redirect()`
refs https://github.com/TryGhost/Ghost/issues/9311
- very basic implementation, still needs proper classes and default stylesheet implementation
- change image card output to a `<figure>` with optional `<figcaption>`
- add optional `<p>` caption output to the html card
refs #9200
- We have not yet counted the images within your html, this commit counts images based on the this algorithm: https://blog.medium.com/read-time-and-you-bc2048ab620c
- Added imageCount utility, which counts images using an img-tag regex, amended from the general tag-regex found in wordCount
- Added this imageCount to the {{reading_time}} helper, adding 12 seconds to the reading time for every image
- The feature image is still counted as before
- The first image adds 12 seconds, the second 11, the third 10, and so on
- Images from the tenth onwards add 3 seconds to the reading time
closes#9085
Fixes an issue, where the client sets image properties to `""` after deleting the image. This causes problems with the query filter (see https://github.com/TryGhost/GQL/issues/24), as they have to be `null`.
Added a check in the model layer saving method to set value to `null`, when the property is empty.
Affected models and properties:
- `posts`:
- `feature_image`
- `og_image`
- `twitter_image`
- `users`:
- `profile_image`
- `cover_image`
- `tags`:
- `feature_image`
no issue
- currently if you would like to edit a resource (e.g. post) and you pass an invalid model id, the following happens
- permission check calls `Post.permissible`
- the Post could not find the post, but ignored it and returned `userPermissions:true`
- then the model layer is queried again and figured out that the post does not exist
- A: there is no need to query the model twice
- B: we needed proper error handling for post and role model
no issue
- replace logic for preparing nested tags
- if you have nested tags in your file, we won't update or update the target tag
- we simply would like to add the relationship to the database
- use same approach as base class
- add `posts_tags` to target post model
- update identifiers
- insert relation by foreign key `tag_id`
- bump bookshelf-relations to 0.1.10
no issue
- change behaviour from updating user references after the actual import to update the user reference before the actual import
- updating user references after the import is way less case intense
- that was the initial decision for updating the references afterwards
- but that does not play well with adding nested relations by identifier
- the refactoring is required for multiple authors
- if we e.g. store invalid author id's, we won't be able to add a belongs-to-many relation for multiple authors
- bookshelf-relations is generic and always tries to find a matching target before attching a model
- invalid user references won't work anymore
- this change has a very good side affect
- 17mb takes on master ~1,5seconds
- on this branch it takes ~45seconds
- also the memory usage is way lower and stabler
- 40mb takes 1,6s (times out on master)
no issue
- otherwise we will have trouble in the future fetching relations by foreign key
- e.g. `tag_id: {id}`
- this won't work if we don't explicitly define the name of the keys
- bookshelf can't fulfil the request
- this does not change any behaviour, it just makes use of the ability to define the names of your foreign keys
refs https://github.com/TryGhost/Ghost/issues/3658
- the `validateSchema` helper was a bit broken
- if you add a user without email, you will receive a database error
- but the validation error should catch that email is passed with null
- it was broken, because:
- A: it called `toJSON` -> this can remove properties from the output (e.g. password)
- B: we only validated fields, which were part of the JSON data (model.hasOwnProperty)
- we now differentiate between schema validation for update and insert
- fixed one broken import test
- if you import a post without a status, it should not error
- it falls back to the default value
- removed user model `onValidate`
- the user model added a custom implementation of `onValidate`, because of a bug which we experienced (see https://github.com/TryGhost/Ghost/issues/3638)
- with the refactoring this is no longer required - we only validate fields which have changed when updating resources
- also, removed extra safe catch when logging in (no longer needed - unit tested)
- add lot's of unit tests to proof the code change
- always call the base class, except you have a good reason
no issue
- `isNew` does not work in Ghost, because Ghost does not use auto increment id's
- see https://github.com/bookshelf/bookshelf/issues/1265
- see https://github.com/bookshelf/bookshelf/blob/0.10.3/src/base/model.js#L211
- we only had one occurance, which was anyway redundant
- if you add a user, `hasChanged('password') is true
- if you edit a user and the password has changed, `hasChanged('password')` is true as well
NOTE #1:
1. We can't override `isNew` and throw an error, because bookshelf makes use of `isNew` as well, but it's a fallback if `options.method` is not set.
2. It's hard to re-implement `isNew` based on `options.method`, because then we need to ensure that this value is always set (requires a couple of changes)
NOTE #2:
If we need to differentiate if a model is new or edited, we should manually check for `options.method === insert`.
NOTE #3:
The unit tests are much faster compared to the model integration tests.
I did a comparision with the same test assertion:
- unit test takes 70ms
- integration test takes 190ms
no issue
- move password hashing and password comparison to lib/security/password
- added two unit test
- FYI: password hashing takes ~100ms
- we could probably mock password hashing in certain cases when unit testing
no issue
- this commit cleans up the usages of `include` and `withRelated`.
### API layer (`include`)
- as request parameter e.g. `?include=roles,tags`
- as theme API parameter e.g. `{{get .... include="author"}}`
- as internal API access e.g. `api.posts.browse({include: 'author,tags'})`
- the `include` notation is more readable than `withRelated`
- and it allows us to use a different easier format (comma separated list)
- the API utility transforms these more readable properties into model style (or into Ghost style)
### Model access (`withRelated`)
- e.g. `models.Post.findPage({withRelated: ['tags']})`
- driven by bookshelf
---
Commits explained.
* Reorder the usage of `convertOptions`
- 1. validation
- 2. options convertion
- 3. permissions
- the reason is simple, the permission layer access the model layer
- we have to prepare the options before talking to the model layer
- added `convertOptions` where it was missed (not required, but for consistency reasons)
* Use `withRelated` when accessing the model layer and use `include` when accessing the API layer
* Change `convertOptions` API utiliy
- API Usage
- ghost.api(..., {include: 'tags,authors'})
- `include` should only be used when calling the API (either via request or via manual usage)
- `include` is only for readability and easier format
- Ghost (Model Layer Usage)
- models.Post.findOne(..., {withRelated: ['tags', 'authors']})
- should only use `withRelated`
- model layer cannot read 'tags,authors`
- model layer has no idea what `include` means, speaks a different language
- `withRelated` is bookshelf
- internal usage
* include-count plugin: use `withRelated` instead of `include`
- imagine you outsource this plugin to git and publish it to npm
- `include` is an unknown option in bookshelf
* Updated `permittedOptions` in base model
- `include` is no longer a known option
* Remove all occurances of `include` in the model layer
* Extend `filterOptions` base function
- this function should be called as first action
- we clone the unfiltered options
- check if you are using `include` (this is a protection which could help us in the beginning)
- check for permitted and (later on default `withRelated`) options
- the usage is coming in next commit
* Ensure we call `filterOptions` as first action
- use `ghostBookshelf.Model.filterOptions` as first action
- consistent naming pattern for incoming options: `unfilteredOptions`
- re-added allowed options for `toJSON`
- one unsolved architecture problem:
- if you override a function e.g. `edit`
- then you should call `filterOptions` as first action
- the base implementation of e.g. `edit` will call it again
- future improvement
* Removed `findOne` from Invite model
- no longer needed, the base implementation is the same
no issue
- Date comparisons are possible via API, but there's no way to inject a valid date into the get helper
- JavaScript's Date.toString() function outputs dates in a useless format
- Swap to using Date.toISOString() and now the format can be understood anywhere!
- {{#get "posts" filter="published_at:<='{{published_at}}'"}}{{/get}} works now as expected
refs #6103
- simplify `toJSON`
- `baseKey` was not used - have not find a single use case
- all the functionality of our `toJSON` is offered in bookshelf
- `omitPivot` does remove pivot elements from the JSON obj (bookshelf feature)
- `shallow` allows you to not return relations
- make use of `serialize`, see http://bookshelfjs.org/docs/src_base_model.js.html#line260
- fetching nested relations e.g. `users.roles` still works (unrelated to this refactoring)
> pick('shallow', 'baseKey', 'include', 'context')
We will re-add options validation in https://github.com/TryGhost/Ghost/pull/9427, but then with the official way: use `filterOptions`.
---
We return all fetched relations (pre-defined with `withRelated`) by default.
You can disable it with `shallow:true`.
closes#9445
- redirects all asset requests if https is configured (theme, core, images)
- re-use and extend our url-redirect middleware
- add proper integration tests for our express site app (no db interaction, component testing required for such important use cases)
- i added some more general tests
- should avoid mixed content warnings in the browser
no issue
- discovered while testing
- the fixture utility needed a protection against non existent roles in the database
- it tries to fetch the contributor role from the database, which does not exist yet
closes#9314
* added fixtures for contributor role
* update post api tests to prevent contributor publishing post
* update permissible function in role/user model
* fix additional author code in invites
* update contributor role migration for knex-migrator v3
* fix paths in contrib migration
* ensure contributors can't edit or delete published posts, fix routing tests [ci skip]
* update db fixtures hash
* strip tags from post if contributor
* cleanup post permissible function
* excludedAttrs to ignore tag updates for now (might be removed later)
* ensure contributors can't edit another's post
* migration script for 1.21
no issue
- all of the error message keys were unused
- the only html anchor i found was for mail, but this doesn't change anything, because the admin does only show the message and not the context at the moment
no issue
- returning and remembering the data, which was imported, is...
- not required when using the API
- not required when importing via script
- required for tests
- added an option to have control over it
- make more usage of local variables
- the GC cannot tidy up variables, which are defined outside of a loop, but used in the loop
- try to keep less memory in process
- reduce the number of properties we have to remember
no issue
- if you import a JSON file with a post, which has an unknown author,
the target user was removed from the blog
- Ghost can handle this case and still succeeds with import
- but we have stored an `author_id` in the database, which does not map to any user and won't map in the future
- this can trouble if we add support for multiple authors
- currently, we only return the `author_id` to the client and the client can map with `author_id` with users fetched by the API
- if it does not find a user, it just falls back to a different user
- but multiple authors have to be included explicit (`include=authors`) and we will return a mapped (author_id => user) result
- it won't be able to find the user, because we lookup the database
- this would result in an error
- there is in general no reason to import (or store) an unknown/invalid `author_id` into the database
- on import, we show you a warning and you can choose a different author if you want
- solution: fallback to owner user and extend warning
- it's not a behaviour change, you still can import unknown author id's and the import won't fail
- but we ensure valid author id's
- updated test
- further more: returning `author={}` when requesting `include=author` could trouble with ember currently
- it expects the author to be returned
no issue
- the warning is "Transaction was already complete"
- destroying a user happens in a transaction, but the event is not asynchronous
- so we have to ensure that we don't operate on a finished transaction
