refs #5942, #6150
There were a few key problems I was looking to solve with this:
- Introduce a single point of truth for what the URL for accessing the API should be
- Provide a simple way to configure the utility (much like a true SDK)
As of this commit, this utility is still automatically available in a Ghost theme.
To use it on an external site, the code would look like:
```
<script type="text/javascript" src="http://my-ghost-blog.com/shared/ghost-url.min.js"></script>
<script type="text/javascript">
ghost.init({
clientId: "<your-client-id>",
clientSecret: "<your-client-secret>"
});
</script>
```
To achieve this, there have been a number of changes:
- A new `apiUrl` function has been added to config, which calculates the correct URL. This needs to be unified with the other url generation functions as a separate piece of work.
- The serveSharedFile middleware has been updated, so that it can serve files from / or /shared and to substitute `{{api-url}}` as it does `{{blog-url}}`.
- ghost-url.js and ghost-url.min.js have been updated to be served via the serveSharedFile middleware
- ghost-url.js has been changed slightly, to take the url from an inline variable which is substituted the first time it is served
- `{{ghost_head}}` has been updated, removing the api url handling which is now in config/url.js and removing the configuration of the utility in favour of calling `init()` after the script is required
- `{{ghost_head}}` has also had the meta tags for client id and secret removed
- tests have been updated
no issue
- Cache the permalinks & postsPerPage settings on the config.theme object
- Use the config.theme cache to reference these items throughout the frontend of a blog
- Removes the need for workarounds and extra code to handle async fetches
- Makes these values accessible to all themes, which is very useful now we have the API stuff
refs #5091, refs #2263
- Move rss handling out of the frontend controller and into its own module
- Separate the code into logical blocks
- Wrap the generation code in a in-memory cache to prevent it being regenerated on every request
Closes#5033
- Added unit tests for the nav context of urlFor
- Fixed issue in the nav context of urlFor where subdomains of blog url were truncated
- Fixed issue in the nav context of urlFor where there was sometimes an extra preceding /
refs #4989
- this allows users to enter mailto and output links via the navigation UI
- the navigation validation/cleanup needs a bit of a refactor to handle other kinds or URI, so leaving #4989 open for now
needed for #4852
Before this, calling `{{url}}` with a nav context from #4541 would
output `/`. This adds a check in `urlFor` that looks for keys in a
nav context object, namely `slug`, `current`, `label`, & `url`.
This change allows for a url to pass through if used in a nav context.
* adds `schema.isNav()`
* adds tests to `url_spec.js`
* handles absolute urls correctly even if `absolute=true`
closes#4445
- post model gets permalink format
- post model queries urlPathForPost to return computed url
- url helper modified to use post url
- urlForPost method abolished and replaced where necessary
- updated tests
fixes#4555
- There's no easy way to declare an XSL with the node xml module, so I
needed to move the declarations to both be strings
- Ideally the code to serve the XSL would also be inside the sitemap
module, but I think we need to refactor a bit to get there easily
- Added the XSL from #4559, with minor amends to make the tables and urls
display correctly
No Issue
- Strip sub-directory from image paths before passing to
config.createUrl. Since images are stored with the sub-directory
and createUrl builds a URL with the sub-directory the result
would be a URL that contains the sub-directory twice.
Closes#3858
- urlPathForPost is now aware of the author’s slug, allowing for the
permalink setting to use :author
- Frontend controller only renders posts that have the specified
author, if one is given
- Extensive tests included
closes#2610, refs #2697
- cleanup API index.js, and add docs
- all API methods take consistent arguments: object & options
- browse, read, destroy take options, edit and add take object and options
- the context is passed as part of options, meaning no more .call
everywhere
- destroy expects an object, rather than an id all the way down to the model layer
- route params such as :id, :slug, and :key are passed as an option & used
to perform reads, updates and deletes where possible - settings / themes
may need work here still
- HTTP posts api can find a post by slug
- Add API utils for checkData
Ref #2061
- Add canThis permission checks to settings api calls
- Add strict rules about accessing core settings without internal: true
- Omit core settings in browse() call unless internal: true
- Update unit tests to call api.settings with contexts
- Add a couple unit tests for new scenarios
- Update all api.settings calls in the app to call with internal context
- Re-arrange permissions.init in server startup so config.theme.update
can access settings without permissions error
closes#2635
- add new logic to the api request handler to set a location header
when new objects are created
- added an api context to config.urlFor() to return the root url
- added functional tests for the affected routes
Closes#2606
- Refactor settings api responses to { settings: [ ] } format
- Update all code using api.settings to handle new response format
- Update test stubs to return new format
- Update client site settings model to parse new format into one object of key/value pairs
- Refactor to include all setting values
- Remove unused settingsCollection method
- Update settingsCache to store all attributes
- Update settingsResult to send all attributes
- Remove unnecessary when() wraps
- Reject if editing a setting that doesn't exist
- Reject earlier if setting key is empty
- Update tests with new error messages
- Use setting.add instead of edit that was incorrectly adding
- Update importer to properly import activePlugins and installedPlugins
- Update expected setting result fields
- Fix a weird situation where hasOwnProperty didn't exist 🤷
closes#1838
- adding `forceAdminSSL: {redirect: true/false}` option to allow 403 over non-SSL rather than redirect
- adding `urlSSL` option to specify SSL variant of `url`
- using `urlSSL` when redirecting to SSL (forceAdminSSL), if specified
- dynamically patching `.url` property for view engine templates to use SSL variant over HTTPS connections (pass `.secure` property as view engine data)
- using `urlSSL` in a "reset password" email, if specified
- adding unit tests to test `forceAdminSSL` and `urlSSL` options
- created a unit test utility function to dynamically fork a new instance of Ghost during the test, with different configuration options
fixes#2112
refs #1833
- modified config.urlFor to handle tag pages
- modified {{#tags}} handlebars helper to autolink to
tag pages. Additional autolink="false" parameter can
be used to deactivate autolinking
- modified url handlebars helper to handle tags
- added isTag function to schema
- added unit test for additional urlFor functionality
- added unit test for {{#tags}} helper modifications
- added unit test for url handlebards helper
addresses #1789, #1364
- Moves ./core/server/loader -> ./core/bootstrap.
The bootstrap file is only accessed once during startup,
and it’s sole job is to ensure a config.js file exists
(creating one if it doesn’t) and then validates
the contents of the config file.
Since this is directly related to the initializing
the application is is appropriate to have
it in the ./core folder, named bootstrap as that
is what it does.
This also improves the dependency graph, as now
the bootstrap file require’s the ./core/server/config
module and is responsible for passing in the validated
config file.
Whereas before we had ./core/server/config
require’ing ./core/server/loader and running its
init code and then passing that value back to itself,
the flow is now more straight forward of
./core/bootstrap handling initialization and then
instatiation of config module
- Merges ./core/server/config/paths into
./core/server/config
This flow was always confusing me to that some config
options were on the config object, and some were on
the paths object.
This change now incorporates all of the variables
previously defined in config/paths directly
into the config module, and in extension,
the config.js file.
This means that you now have the option of deciding
at startup where the content directory for ghost
should reside.
- broke out loader tests in config_spec to bootstrap_spec
- updated all relevant files to now use config().paths
- moved urlFor and urlForPost function into
./server/config/url.js
