Repository navigation
Conversation
A stored LINK names its target with the Jetty API type, so its bytes stay the same in every stage.
A link names only a Workflow's stream here. The batch digest takes each link field, and a record with no link keeps its digest.
A client or an Activity gets one order across a stream and its links, and resumes from the answer's state.
The follow cases use a child and a grandchild Workflow, and every conformance lane reads a LINK back.
1 of 3 tasks
A late OPEN from an older attempt opened a segment again after its JOIN. The child's records then read twice, and the next turn had no gate.
A retry that writes the same OPEN again showed each child record twice, or the withdrawn child read on. Now the SUPERSEDED pauses the segment and the same OPEN resumes it at its cursor.
A resumed child could wait on a JOIN of a grandchild that the list had dropped first.
A resumed descendant could read a Stream that was open again, and a resume could pass max_open. The list also kept a parent after its descendants.
A failed resume opened the child from after, so its earlier Records showed twice. Now a top entry past the limit is skipped, and a descendant that can't come back stays ended and gives up its JOIN.
An entry under a descendant that stays ended was left with no way to resume. Descendants now come back in source number order.
moetemp
changed the title
Owner
Author
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What changed?
This PR adds Link Records and
Followto Brook. A producer writes a LINK record with aStreamLinkthat names a Workflow's stream.Followreads a stream and the streams it links in one order, and itsstateresumes the follow. The crate copies onlyStreamReferenceandStreamOwnerKindfrom the stream API, at the paths Jetty uses, so a stored LINK has the same bytes in every stage. The Brook messages are a subset of Jetty's at the same field numbers. There is noFollowRange, nofloor_holder, no subscriber and no close outcome onDONE.ReadResponse.start_cursorcomes along, since a$endread cut at a gate reads again from it.Why?
Moe moved Link Records and the client Link Follower from Jetty into Brook, so a client or an Activity gets one order across a stream and its links without a merge of its own.
How did you test it?
Link to a test plan if any -