Suggestion to close issues (issue triage)

[BioPhoton]

Dear @benlesh @JWO719 @kwonoj @cartant

I collected these issues as in my humble opinion they could be closed, or their discussion stopped a while ago. All listed issues are suggestions. I don't want to make people upset that opened issues a while ago and did not get a fully satisfying solution yet! So please consider these lists as a suggestion, not a definite thing! I will add more issues in the future here. Just that you know it is a growing list.

To Contributors If any contributors see this and maybe have some more information on why their issue is still relevant, please add the information to the issue as a comment. I will do my best to help you further.

I hope this can help to get some outdated or already solved issues closed. :)

I'll use the pathfinder approach and step by step ask people to close an issue if I pass one that could be closed.

Benefits of having it as an issue in the repo:

• Solved issues get checked and therefore we have a good overview of the progress
• Mouse hover shows overview data of the issue
• As issues are referenced by id, the title of the issue shows up in other issues comments and informs them to check if they could maybe close it

Suggested issues to close

Should stay open: ✔ Got closed: ❌

Opened/Related to Core-Team

• [x] ❌Where to put the unhandled error event hook? Looking for ideas/thoughts can be closed by @benlesh as the PR is merged

• [x] ❌Remove secondary mappings for flattening operators in v6? @benlesh we can close this issue as already merged.

• [x] ❌Eliminate @reactivex/rxjs publish in 6.0? @benlesh we can close this as it’s already implemented

• [x] ❌Add close state to finally operator? poke @benlesh

• [x] ❌Docs on Observable methods are all duplicated can be closed as docs are updated. Double check with @JWO719

• [x] ❌‘distinct’ can't work as the description in tutorials Confusing wording used in docs. @JWO719 we could rethink the used words in the documentation

• [x] ❌Autopublish nightly package ask @kwonoj to close it

• [x] ❌Discuss module structure for 6.0 could most likely be closed by @benlesh Opened by contributors from here on

• [x] ❌Node Require() Support

• [x] ❌RxJS v6 TimeoutError is missing name property can be closed because it’s already implemented. See: https://github.com/ReactiveX/rxjs/pull/3656

• [x] ❌Remove static create methods and provide importable functions instead can be closed as docs are updated.

• [x] ❌BehaviorSubject created with "create" does not have getValue() method can be closed because it’s obsolete because we dropped getValue

• [x] ❌Get response url from response.xhr on ie11 See: https://github.com/ReactiveX/rxjs/issues/3472#issuecomment-379518028

• [x] ❌Split creation functions from creation functions with Schedulers We can close this as it is already done.

• [x] ❌Transducer Support can be closed as related to RxJs4 and just a question that is already answered.

• [x] ❌Add flush functionality to distinctUntilChanged can be closed as it is already answered and the creator reconsidered his proposal

• [x] ❌Documentation missing for operators in typescript declaration files

• [x] ❌bufferTime and bufferCount will mutate current buffer if a source is a subject that emits in the subscription for that buffer can be closed as it is by design

• [x] ❌Handle stream like maybe monad can be closed, if the creator does not do it, as it is answered

• [x] ❌Lettable operators: scope hoisting prevents dead code removal quite sure this can be closed

• [x] ❌Add selector to publishReplay can be closed as already implemented

WebSocket

• [x] WebSocketCtor in WebSocketSubject must allow more params in new instance => poke @benlesh how to proceed
• [x] ❌Guidance on improving WebSocketSubject multiplex method
• [x] ❌WebSocketSubjectConfig's resultSelector is accepting a not necessary generic type T
• [ ] Observable multiplexed from WebSocketSubject sends subMsg/unsubMsg to the server per every subscription
• [ ] Documentation (in particular Observable.webSocket)

AJAX :D

• [x] ❌AJAX response with status 0 must throw an error can be closed because it’s already implemented in 6.4.*. See: https://github.com/ReactiveX/rxjs/issues/3606#issuecomment-488614083
• [x] ❌AJAX Content-Type Handling consider to close in favour of fromFetch
• [x] AJAX responseType Handling consider to close in favour of fromFetch
• [x] ❌Observable AJAXPOST with wrong Content-Type after PUT
• [x] ❌Rx.Observable.AJAX onProgress only set for Upload consider to close in favour of fromFetch. Other suggestion, switch to immutable header object
• [ ] AJAX config & intercept
• [ ] errors in .ajax are not propagated to .catch
• [x] ❌Observable.ajax halts when error callback is not given to progressSubscriber
• [x] ❌Ajax methods not parsing settings objext
• [ ] Missing response value on ajax result.
• [x] ❌Observable.ajax throws exception when it receives HTTP 405 Method not allowed The discussion here is pretty much dead. Also we could consider to close in favour of fromFetch

Lost bug

• [x] ❌MS Edge fails with SCRIPT1028: Expected identifier, string or number => I have to investigate more
• [x] ❌concatAll() unexpectedly overlaps inner Observable subscriptions => We could look into triggering the next subscription in concat, merge, etc after the unsubscribe of the previous observable. That seems reasonable.
• [x] ❌Standardize ReplaySubject to behave like Subject related: https://github.com/ReactiveX/rxjs/pull/3074

Needs poke (may stay open but needs further, mostly minimal, conversation for a decision)

Opened/Related to Core-Team

• [x] ✔lift called with incorrect context => bug still there
• [x] ❌question about buffer operators =>
  Poke @mattpodwysocki
• [x] ❌[Proposal] New flatten map operators implementation => Poke @benlesh
• [x] ❌Using AbortSignal and AbortController => poke @benlesh

Opened by contributors

• [x] ❌Mutable-reductions friendly version of reduce (something similar to transduce / RxJava's collect) => Poke author
• [x] ❌CHANGELOG changes link problem => Bug still exists. From my current investigation this is very interesting. It’s not always wrong. As cz-conventional-changelog is saved in package.json under a herd version I have no clue how this can happen.
• [x] ❌No way to tell user they are using wrong version of RxJS (i.e v4 vs v5)

Breaking changes

• [x] ❌Discussing the future of schedulers => As the refactoring is already in progress It would be good to know when this

Needs collection work

• [x] ✔Switch to config objects for operators with a lot of arguments => Answered with a well-thought procedure

Module bundles

• [x] ❌UMD and/or FESM bundles - now tractable with 6.0.0
• [ ] import operator' code differs: UMD vs. node_modules
• [x] ❌Path Mapping is not properly work

Typings

• [x] types inferring problem with scan operator
• [x] ❌Pipe operator cannot infer return type as ConnectableObservable => can likely be closed - after discussion - in favour #3833
• [x] ❌Subject.prototype.next() should only accept T, not T | undefined

New Features

• [x] ❌Add selector to publishBehavior => poke @benlesh on how to proceed. This will not introduce any breaking changes
• [x] ❌Support for spaces in marbles strings => check out @kwonoj lib to see if could replace the current ACSII marble parsing => use TestScheduler#run
• [x] ❌Add concurrency parameter to switch and switchMap very old already closed request but reopened by @benlesh 2017

Needs more investigation

• [x] ❌throttleTime(n, undefined, {leading: true, trailing: true}) ... double output
• [x] ❌auditTime does not emit last value => Duplicates: #3325, #3430, #2859 Related: #3351

Most probably outdated

• [x] ❌SafeSubscriber does not unsubscribe itself => Pinged author. No answer.
• [x] ❌Make fromPromise support promise cancellation?
• [x] Strongly type Observable Error
• [x] ❌TypeScript 2.1+ and Errors
• [ ] bufferTime minBufferSize => IMHO to be closed because we have a solution and the operator extension is maybe a bigger discussion as It might include all other buffer maybe also windowing operators.
• [x] ❌Automated Publish => could be closed as the current process is different and already includes a publishing strategy.
• [ ] Different behavior with Rx 4 and buffer
• [x] ❌Implement join operator
• [x] ❌BehaviorSubject created with "create" does not have getValue() method

NEEDS DETAILED CHECK AND CATEGORISATION

• [x] ❌fromEvent() should emit error when NodeJS EventEmitter emits error event
• [x] ❌TestScheduler seems to use bogus values for frameTimeFactor
• [x] ❌first operator does not accept default value only signature with null-value predicate
• [x] ❌subscribe ignores second argument of event emitter
• [x] ❌A minor documentation bug in tutorial
• [ ] reactivex.io documentation needs to reference v5+
• [x] ❌Documentation from contributor days
• [x] ❌Document special cases of reduce operator usage for certain pattern
• [x] ❌Observable constructor shouldn't give people a Subscriber<T> type.
• [x] ❌Run browser testing via global builds
• [x] ❌Record perf results vs RxJS 4, then drop RxJS 4 from perf testing
• [x] ❌Errors thrown in nextHandler passed to subscribe should not tear down observable
• [x] ❌Runtime validation API in "development mode"
• [x] ❌ Subject.create vs AnonymousSubject.create
• [x] ❌ QueueScheduler schedule method: this.SchedulerAction is not a constructor
• [x] ❌Bug in buffer or bufferCount operation when used with subject
• [x] ❌Allow creation of Observable from EventSource (server-sent events): enhancement
• [x] ❌Better cache implementation
• [x] ❌Proposal: Changes to retryWhen
• [x] ❌Can't repeat or retry Observable.bindCallback and Observable.bindNodeCallback

[benlesh]

This is AWESOME! Thanks for doing this, @BioPhoton!

[benlesh]

@kwonoj can you please look at the ajax related stuff above? I think that you were working on something around that.

[BioPhoton]

Welcome!

[llorenspujol]

@BioPhoton I think this issue/feature: https://github.com/ReactiveX/rxjs/issues/3430 is interesting to solve. I have added a 'mini-explanation' on the issue about it.

[KeijiBranshi]

Am seeing several open docs issues with merged PRs associated with them. Just wondering if they should also be closed out or not.

• [docs]: Operator toArray is missing documentation #4613
• [docs]: Operator timestamp is missing the short description, description, and examples #4615
• [docs]: Operator single is missing the description, and examples #4616
• [docs]: Operator ignoreElements is missing the description, and examples #4618
• [docs]: Operator generate need a total over whole of the documentation #4672

[niklas-wortmann]

thanks for the reminder @KeijiBranshi