* Relax timeout for sending activities
Lemmy considers timeouts during activity sending as retryable errors.
While it is frequently enough to retry sending the same activity again after
the original submission attempt resulted in a timeout, allowing the receiving
side to use more time for synchronous processing should reduce the number of
retries needed overall and improve overall compatibility.
Some ActivityPub software, such as Mastodon, implements a queue for processing
received activities asynchronously, which allows immediately returning a
response for activity submissions. Other software, such as Lemmy or Hubzilla
implement synchronous processing of activities before returning a response.
ActivityPub does not specify specific timeouts to be used:
https://github.com/w3c/activitypub/issues/365
* Simplify usage of federation_sender_config Option
* federation: parallel sending
* federation: some comments
* lint and set force_write true when a request fails
* inbox_urls return vec
* split inbox functions into separate file
* cleanup
* extract sending task code to separate file
* move federation concurrent config to config file
* off by one issue
* improve msg
* fix both permanent stopping of federation queues and multiple creation of the same federation queues
* fix after merge
* lint fix
* Update crates/federate/src/send.rs
Co-authored-by: dullbananas <dull.bananas0@gmail.com>
* comment about reverse ordering
* remove crashable, comment
* comment
* move comment
* run federation tests twice
* fix test run
* prettier
* fix config default
* upgrade rust to 1.78 to fix diesel cli
* fix clippy
* delay
* add debug to make localhost urls not valid in ap crate, add some debug logs
* federation tests: ensure server stop after test and random activity id
* ci fix
* add test to federate 100 events
* fix send 100 test
* different data every time so activities are distinguishable
* allow out of order receives in test
* lint
* comment about https://github.com/LemmyNet/lemmy/pull/4623#discussion_r1565437391
* move sender for clarity, add comment
* move more things to members
* update test todo comment, use same env var as worker test but default to 1
* remove else below continue
* some more cleanup
* handle todo about smooth exit
* add federate inboxes collector tests
* lint
* actor max length
* don't reset fail count if activity skipped
* fix some comments
* reuse vars
* format
* Update .woodpecker.yml
* fix recheck time
* fix inboxes tests under fast mode
* format
* make i32 and ugly casts
* clippy
---------
Co-authored-by: dullbananas <dull.bananas0@gmail.com>
* Switch PostgreSQL container to use pgautoupgrade
This handles automatically upgrading the data files to newer versions of PostgreSQL.
* Fixing other uses of postgres:16-alpine image.
* Simplifying upgrade scripts.
---------
Co-authored-by: Dessalines <tyhou13@gmx.com>
* Adding checks for higher admin and mod.
* Adding admin/mod checks for ban and comment removal.
* Combining mod or admin check with an SQL union.
* Making community ban or add mod also allow higher admins.
* Making sure remove post also checks higher mods or admins.
* Add unit test for is_higher_mod_or_admin_check
* Fixing comment.
* Addressing PR comments.
* Get rid of pointless wrapper functions, return lemmyresult directly.
* Fixing not being able to create comments on local community posts.
- This was caused by not passing my_person_id into various
`CommentView::read` functions.
- Fixes#4853
* Refactoring views to use local_user, rather than person
* Addressing PR comments.
* Fixing API tests.
* Relax timeout for sending activities
Lemmy considers timeouts during activity sending as retryable errors.
While it is frequently enough to retry sending the same activity again after
the original submission attempt resulted in a timeout, allowing the receiving
side to use more time for synchronous processing should reduce the number of
retries needed overall and improve overall compatibility.
Some ActivityPub software, such as Mastodon, implements a queue for processing
received activities asynchronously, which allows immediately returning a
response for activity submissions. Other software, such as Lemmy or Hubzilla
implement synchronous processing of activities before returning a response.
ActivityPub does not specify specific timeouts to be used:
https://github.com/w3c/activitypub/issues/365
* Simplify usage of federation_sender_config Option
* Adding a show_read override to GetPosts.
- If show_read is true, it overrides the local user show_read
setting.
- Fixes#4124
* Addressing PR comments.
* Update crates/db_views/src/post_view.rs
Co-authored-by: dullbananas <dull.bananas0@gmail.com>
* Fixing formatting.
---------
Co-authored-by: dullbananas <dull.bananas0@gmail.com>
* Adding an image_details table to store image dimensions.
- Adds an image_details table, which stores the height,
width, and content_type for local and remote images.
- For LocalImages, this information already comes back with
the upload.
- For RemoteImages, it calls the pictrs details endpoint.
- Fixed some issues with proxying non-image urls.
- Fixes#3328
- Also fixes#4703
* Running sql format.
* Running fmt.
* Don't fetch metadata in background for local API requests.
* Dont export remote_image table to typescript.
* Cleaning up validate.
* Dont proxy url.
* Fixing tests, fixing issue with federated thumbnails.
* Fix tests.
* Updating corepack, fixing issue.
* Refactoring image inserts to use transactions.
* Use select exists again.
* Fixing imports.
* Fix test.
* Removing pointless backgrounded metadata generation version.
* Removing public pictrs details route.
* Fixing clippy.
* Running prettier.
* A few more fixes.
* Moving diesel schema check back down.
* Addressing PR comments.
* Changing back request head to get.
* Fixing lockfile.
---------
Co-authored-by: SleeplessOne1917 <28871516+SleeplessOne1917@users.noreply.github.com>
