* feat(backend): resize images with imagor
Open questions:
* Do we have external URLs for images? E.g. we have them for seeds. But
in production?
* Do we want to apply image transformations on these as well? My current
implementation does not apply image transformations as of now. If we
want to do that, we will also expose internal URLs in the kubernetes
Cluster to the S3 endpoint to the client.
TODOs:
* The chat component is using a fixed size for all avatars at the moment.
Maybe we can pair-program on this how to implement responsive images in
this component library.
Commits:
* do not replace upload domain url in the database
* fix all webapp specs
* refactor: remove behaviour we won't need
We don't want to apply image transformations on files, right?
* refactor: replace the domain on read not on write
* wip: webapp fixes
* refactor(backend): add another url to config
I've given up. There seems to be no nice way to tell the minio to return
a location which differs from it's host name.
* refactor: add test for s3Service
* refactor(backend): proxy minio via backend in local development
Commits:
* provide tests for message attachments
* remove S3_PUBLIC_URL config value
* refactor: follow @ulfgebhardt's review
* add missing environment variable
---------
Co-authored-by: Ulf Gebhardt <ulf.gebhardt@webcraft-media.de>
We kept this code for backwards compatibility but since we already deployed S3 to our kubernetes cluster and we're using it locally, let's remove this code. It will also make it easier to implement the image resize service as it reduces the total amount of code to maintain.
Co-authored-by: Ulf Gebhardt <ulf.gebhardt@webcraft-media.de>
This is a side quest of #8558. The motivation is to be able to do dependency injection in the tests without overwriting global data. I saw the first merge conflict from #8551 and voila: It seems @Mogge could have used this already.
refactor: follow @Mogge's review
See: https://github.com/Ocelot-Social-Community/Ocelot-Social/pull/8603#pullrequestreview-2880714796
refactor: better test helper methods
wip: continue refactoring
wip: continue posts
continue
wip: continue groups
continue registration
registration
continue messages
continue observeposts
continue categories
continue posts in groups
continue invite codes
refactor: continue notificationsMiddleware
continue statistics spec
followed-users
online-status
mentions-in-groups
posts-in-groups
email spec
finish all tests
improve typescript
missed one test
remove one more reference of CONFIG
eliminate one more global import of CONFIG
fix language spec test
fix two more test suites
refactor: completely mock out 3rd part API request
refactor test
fixed user_management spec
fixed more locatoin specs
install types for jsonwebtoken
one more fetchmock
fixed one more suite
fix one more spec
yet another spec
fix spec
delete whitespaces
remove beforeAll that the same as the default
fix merge conflict
fix e2e test
refactor: use single callback function for `context` setup
refactor: display logs from backend during CI
Because why not?
fix seeds
fix login
refactor: one unnecessary naming
refactor: better editor support
refactor: fail early
Interestingly, I've had to destructure `context.user` in order to make
typescript happy. Weird.
refactor: undo changes to workflows - no effect
We're running in `--detached` mode on CI, so I guess we won't be able to
see the logs anyways.
refactor: remove fetch from context after review
See:
refactor: found an easier way for required props
Co-authored-by: Max <maxharz@gmail.com>
Co-authored-by: Ulf Gebhardt <ulf.gebhardt@webcraft-media.de>
* copy from branding folder to backend public folder
provide default branding/public folder
* copy public folder correctly
* copy files again for providers.json
* copy more public folders
* more copy
* revert change
* fix naming of called script when using db:data:branding
* prod command for branding data
* close database connection
* lint fixes
* increase test timeout again
Ok, so here is the plan. Let's give both our cucumber features and your
cypress tests a prominent place to live. That would be the root level
folder of our application. Second, let's revive formerly dead code step
by step.
Ie. move code from the former location `backend/features/` to `features/`
when it is ready. All edge cases should be tested with unit tests in
`backend/`, see my `webfinger.spec.js` as an example.
@Mastercuber the justification for simply disabling the tests is that
the entire implemntation for the follow activities is wrong. It's still
using the `AddUserFollowedBy` auto-generated mutation. The mutation that
is actually used when you do a follow or unfollow are called `follow`
and `unfollow` - these are custom mutations implemented by ourselves.
So the proper fix for these tests would be to refactor the entire
ActivityPub implementation. Since there is a pending PR already I
decided not to go down the rabbit hole and simply tag the failing
sceario as "work-in-progress" (@wip).
