Commit Graph

35 Commits

Author SHA1 Message Date
Hannah Wolfe
0fe0e09d62 Moved express init + sentry to a shared util
- added core/shared to watched folders in grunt
- moved sentry to shared
- moved express initialisation to a shared file
- always set trust proxy + sentry error handler
- use this new express init everywhere, and remove duplicate trust proxy and sentry error handler code
2020-04-27 18:17:50 +01:00
Hannah Wolfe
19dde146c1 Renamed redirect functions for clarity
- we had urlRedirects, urlRedirects.adminRedirect and adminRedirects
- all do kinda similar things, but for different contexts so for now I've done a minimal renaming for clarity
- and updated some comments!!
- also removed totally unnecessary if res.isAdmin clause, as we don't use that, and it was never true
2020-04-22 19:31:36 +01:00
Daniel Lockyer
7751e78c98 Integrated Sentry error tracking
no issue

- this allows tracking of application errors within Sentry
- only enabled for HTTP 500 errors for now
- it is disabled by default
2020-02-03 13:43:43 +00:00
Kevin Ansfield
68e2274d6d Reverted "Integrated Sentry error tracking"
This reverts commit 6e024331eb.

Temporarily reverting whilst we investigate an issue with Sentry and running Ghost via Ghost-CLI.

Ghost-CLI initiated boot was failing when Sentry was installed due to what appears to be `process.cwd()` returning `undefined` here https://github.com/TryGhost/Ignition/blob/master/lib/config/index.js#L26
2020-01-28 17:21:37 +00:00
Daniel Lockyer
6e024331eb Integrated Sentry error tracking
no issue

- this allows tracking of application errors within Sentry
- only enabled for HTTP 500 errors for now
- it is disabled by default
2020-01-22 12:15:10 +07:00
Fabien O'Carroll
2cb41dd8cd 🐛 Fixed CORS for errors from Content API
closes #10932

Previously we were only applying the cors middleware to the options
preflight request, which meant that if the request errored, the cors
headers would not be applied, resulting in the client being unable to
read response data. This applies the cors middleware to _all_ requests
to the Content API.
2019-07-22 12:57:41 +08:00
kirrg001
38e93c19b5 Reverted & Solved apiType differently
no issue

- was unable to revert 9dd7aff9c6, because it contains members changes
- functional calls did not work correctly, because the content and admin ctrl differentiation happend in the web layer
- `isContentAPI` returned true for `api.v2.settings.edit(data, {context: {internal:true{})`
- content & admin API are using different controllers
- we can just tell which ctrl is content API and which is not
- the direction fits for the content & admin API split
2019-02-26 08:33:10 +01:00
kirrg001
46d7a5666e Renamed authors ctrl to authors-public
no issue

- consistency
2019-02-26 08:33:10 +01:00
Katharina Irrgang
0a70226128 Separated pages & posts in Admin API v2 (#10494)
refs #10438, refs #10106

* Renamed existing pages ctrl
* Splitted posts & pages for Admin API v2
* Added pages JSON input schema for Admin API v2
* Removed single author for Content & Admin API v2
  - single author is not documented
  - single author usage is deprecated in v0.1
  - single author usage is removed in API v2
* Splitted posts & postsPublic controller for v2
* Removed requirement to send `status=all` from Admin API v2
* Removed `status` option from pages Content API v2
* Removed `status` options from Users Admin API v2
2019-02-22 10:17:14 +07:00
Fabien O'Carroll
6bdeeaba10
Added apiType property to frame for {Content,Admin} API (#10487)
no-issue

This sets the `apiType` property of the `frame` to 'content' and 'admin'
for the Content & Admin API respectively.
2019-02-13 16:59:10 +01:00
Fabien O'Carroll
9837df0b6f
Added brute protection to content api (#10362)
* Revert "Removed brute force middleware form content api (#10353)"

This reverts commit 63c8c310fb.

* Updated content api spam prevention to use memory store

* Used TooManyRequestsError instead of InternalServer
2019-01-14 14:11:48 +01:00
Fabien O'Carroll
63c8c310fb
Removed brute force middleware form content api (#10353)
* Removed brute force middleware form content api

no-issue

This is just temporary

* Skipped brute test for content api
2019-01-09 11:38:03 +01:00
Fabien O'Carroll
5cecf0cd61
Added brute protection to content api (#10343)
* Added spam config for content api key

no-issue

* Created contentApiKey spam prevention method

* Added contentApiKey brute middleware

no-issue

This middleware attaches a listener for when the request has completed,
if the request ends with a successful response code, we reset any spam
prevention data for that ip.

* Added contentApiKey brute middleware to the content api

* Multipled maxWait by 24, to 24 hours
2019-01-07 20:03:46 +07:00
Hannah Wolfe
5d977f23d4 Added Settings endpoint to V2 Content API
refs #10318

- This settings endpoint returns the commonly used, public information from our settings.
- The values are whitelisted each with a custom name for returning from the endpoint
2019-01-04 12:17:32 +00:00
Naz Gargol
d3f3b3dc20
Added plugin based author and public tag models in API v2 (#10284)
refs #10124

- Author model returns only users that have published non-page posts
- Added a public controller for tags (should be extracted to separate Content API controller https://github.com/TryGhost/Ghost/issues/10106)
- Made resource configuration dynamic based on current theme engine
- This needs a follow-up PR with fixes to the problems described in the PR
2019-01-03 20:30:35 +01:00
Fabien O'Carroll
dcfaecfe92
🐛 Added CORS support to the v2 Content API (#10257)
no-issue

When trying to use /api/v2/content from a different domain, the requests
were failing with CORS errors. This doesn't use the shared cors middleware,
because it should be open to all hosts, and not locked down via our
whitelist or trusted domains.
2018-12-11 11:44:12 +07:00
Fabien O'Carroll
2e922808e8 ♻ Updated naming for Content API specific middleware
no-issue

This is because the Content API will eventually be accessed not just
from Content API keys. The addition of a Content API specific
authorization middleware is because:
1. content api should not authorize based on req.user
2. content api will need separate authorization than admin api
2018-11-08 13:34:29 +07:00
Katharina Irrgang
ff6bf5f318 Renamed /users to /authors for Content API V2 (#10096)
refs #10061

- Made /authors endpoint available in Content API V2
2018-11-07 15:29:37 +01:00
kirrg001
65f6f64f91 Removed configuration endpoint from Content API V2
no issue

- we are not aware of a use case for this endpoint currently
2018-11-05 18:49:46 +01:00
kirrg001
cbf2817e39 Added missing read pages endpoint
refs #9866

- the endpoints were missing
- the site app needs pages.read for v2
2018-10-18 19:41:07 +02:00
Fabien O'Carroll
561c4b208d
Removed OPTIONS cors middleware from content api (#10013)
no-issue

The content API only supports GET requests so has no need for cors
middleware on OPTIONS. This also removes the router.del helper as it's
not used
2018-10-15 18:47:31 +07:00
Fabien O'Carroll
943e1c872c
Removed POST /subscribers from v2 content api (#10010)
no-issue

The content api is for serving read-only content only. This route does
not belong here.
2018-10-15 16:28:23 +07:00
Fabien O'Carroll
3db102a776
Added API Key auth middleware to v2 content API (#10005)
* Added API Key auth middleware to v2 content API

refs #9865

- add `auth.authenticate.authenticateContentApiKey` middleware
  - accepts `?key=` query param, sets `req.api_key` if it's a known Content API key
- add `requiresAuthorizedUserOrApiKey` authorization middleware
  - passes if either `req.user` or `req.api_key` exists
- update `authenticatePublic` middleware stack for v2 content routes

* Fixed functional content api tests

no-issue

This fixes the functional content api tests so they use the content api
auth.

* Fixed context check and removed skip

* Updated cors middleware for content api

* Removed client_id from frame.context

no-issue

The v2 api doesn't have a notion of clients as we do not use oauth for it

* Fixed tests for posts input serializer
2018-10-15 16:23:34 +07:00
Nazar Gargol
76f4a4bb03
Enforced non-page posts only to be returned by /posts endpoint from Content API (#10002)
refs #9866

- Added logic ensuring page filter is always set to false in posts endpoint for Content API
- Added functional tests to pages and posts
- Added absolute_url logic in pages controller
2018-10-13 00:48:49 +02:00
Katharina Irrgang
981ad28283
Changed tags content api to v2 (#10004)
refs #9866
2018-10-13 00:11:48 +02:00
Katharina Irrgang
c907237540
Changed users content api to v2 (#10003)
refs #9866
2018-10-13 00:11:16 +02:00
Fabien O'Carroll
2da74a614b
Removed clients endpoint from v2 content api (#9998)
refs #9865

The client resource is a v0.1 specific thing, we have no use for it on
the v2 content api
2018-10-12 18:18:00 +07:00
Katharina Irrgang
959912eca3
Added tiny framework to support multiple API versions (#9933)
refs #9326, refs #9866

**ATTENTION: This is the first iteration. Bugs are expected.**

Main Goals: 

- add support for multiple API versions.
- do not touch v0.1 implementation
- do not break v0.1

## Problems with the existing v0.1 implementation

1. It tried to be generic and helpful, but it was a mixture of generic and explicit logic living in basically two files: utils.js and index.js.

2. Supporting multiple api versions means, you want to have as less as possible code per API version. With v0.1 it is impossible to reduce the API controller implementation. 

----

This commit adds three things:

1. The tiny framework with well-defined API stages.
2. An example implementation of serving static pages via /pages for the content v2 API.
3. Unit tests to prove that the API framework works in general.

## API Stages

- validation
- input serialization
- permissions
- query
- output serialization

Each request should go through these stages. It is possible to disable stages, but it's not recommended.

The code for each stage will either live in a shared folder or in the API version itself. It depends how API specific the validation or serialization is. Depends on the use case.

We should add a specific API validator or serializer if the use case is API format specific.
We should put everything else to shared.

The goal is to add as much as possible into the shared API layer to reduce the logic per API version.

---

Serializers and validators can be added:

- for each request
- for specific controllers
- for specific actions

---

There is room for improvements/extensions:

1. Remove http header configuration from the API controller, because the API controller should not know about http - decouple.

2. Put permissions helpers into shared. I've just extracted and capsulated the permissions helpers into a single file for now. It had no priority. The focus was on the framework itself.

etc.

---

You can find more information about it in the API README.md (api/README.md)

- e.g. find more information about the structure
- e.g. example controllers

The docs are not perfect. We will improve the docs in the next two weeks.

---

Upcoming tasks:

- prepare test env to test multiple API versions
- copy over the controllers from v0.1 to v2
- adapt the v2 express app to use the v2 controllers
2018-10-05 00:50:45 +02:00
kirrg001
434a0435fd Optimised web/ debug logs
no issue

- optimised only for web/ folder, because it has used very general namespaces
- the debug namespace must be specific, otherwise i run `DEBUG=ghost:api:*` and i get web debug logs and api folder debug logs
- we can come up with a new namespace system, but for now it must be explicit enough
2018-10-04 17:43:08 +02:00
kirrg001
079b41e608 Added body parser to web/api/v2 express app
refs #9866

- req.body is undefined if we don't use the body parser
- the content API only offers "fetch" endpoints, but if a component/module in Ghost relies on req.body being present, it can crash
- e.g. the authentication service checks for the existence of client_id + client_secret in req.query or req.body
- we could theoretically change it from `if (!req.body.client_id` to `if (req.body && !req.body.client_id)`, but that makes the code very hard to read + maintain
- we will use the body parser for the content API now
- req.body will be {}
2018-10-03 00:47:03 +02:00
Katharina Irrgang
213474835b Refactored how we require shared middlewares from web/ (#9893)
refs #9866

- use package notation
- get rid of x requires for middlewares
- improved readability
- do not refactor web/api/v0.1
2018-09-21 16:17:11 +05:30
Rishabh Garg
fcd275f6c0 Refactored web/middleware and web/utils to web/shared (#9892)
refs #9866

- Moved web/middleware to web/shared/middlewares
- Moved util file to web/shared/utils
2018-09-20 20:04:34 +02:00
Nazar Gargol
006c83fbe4 ES6 migration: server/web/ (#9886)
refs #9589
2018-09-20 15:03:33 +02:00
Nazar Gargol
7e17c56feb Removed res.isAdmin usages from v2 express app (#9884)
refs #9866

- Removed `res.isAdmin` flag in v2 express app
- Did not touch v0.1 express app
- Separated url redirect middleware for admin and content API
2018-09-20 13:58:45 +02:00
Nazar Gargol
57271127f4 Added v2 api endpoints (#9874)
refs #9866

- Registered Content API under /ghost/api/v2/content/
- Registered Admin API under /ghost/api/v2/admin/
- Moved API v0.1 implementation to web/api/v0.1
- Created web/api/v2 for the new api endpoints
- Started with reducing the implementation for the new Content API (the Content api does not serve admin api endpoints, that's why it was reducible)
- Covered parent-app module with basic test checking correct applications/routes are being mounted
- Added a readme file, which contains a warning using v2, because it's under active development!
- This PR does only make the new endpoints available, we have not:
  - optimised the web folder (e.g. res.isAdmin)
  - started with different API controllers
  - reason: we want to do more preparation tasks before we copy the api controllers
2018-09-18 15:59:06 +02:00