Implement request idempotency (fixes #4735) #5329
Conversation
Nutomic
requested review from
dessalines,
phiresky,
dullbananas and
SleeplessOne1917
as code owners
There was a problem hiding this comment.
Nice, so then we just need to generate an idempotency UUID in the lemmy-js-client http code for puts and posts.
Should probably also do this for DELETE no?
I imagine CI is complaning about .expect, so you might need to either avoid those, or add a #[allow(clippy::expect_used)] to the functions.
| let set: Arc<RwLock<HashSet<Entry>>> = Default::default(); | ||
|
|
||
| let set_ = set.clone(); | ||
| tokio::spawn(async move { |
There was a problem hiding this comment.
Ah good, as long as there's only one of these per server, not per hash.
There was a problem hiding this comment.
Yes it must be shared across all server threads, otherwise it doesnt work.
|
The spec says that this only makes sense for CREATE and PUT. For example if you send the same "create post" multiple times, it will happily create multiple duplicate posts. But if you send the same "edit post", it will just set the same new text every time, which doesnt really matter. Same for delete, it would only throw an error that the item was already deleted which is fine. |
| // TODO: need to return LemmyError as well? | ||
| let response = HttpResponse::UnprocessableEntity() | ||
| .finish() | ||
| .map_into_right_body(); |
There was a problem hiding this comment.
What about this, is status 422 enough or do we need a json error?
|
Seems like this won't always work in a horizontally scaled setup |
If duplicate requests with the same
Idempotency-Keyare received, only the first one is processed and others receive an error.https://www.ietf.org/archive/id/draft-ietf-httpapi-idempotency-key-header-01.html