fix(step): Return round-tripped result on first run#543
Open
ayushiahjolia wants to merge 3 commits into
Open
Conversation
ayushiahjolia
force-pushed
the
fix/step-serdes-round-trip-first-run
branch
from
July 16, 2026 22:54
604e309 to
59835d5
Compare
ayushiahjolia
marked this pull request as ready for review
July 16, 2026 23:09
zhongkechen
approved these changes
Jul 17, 2026
| # which reconstructs the result by deserializing the checkpoint. | ||
| # A None payload is returned as-is, mirroring the replay path. | ||
| if serialized_result is None: | ||
| return None # type: ignore[return-value] |
Contributor
There was a problem hiding this comment.
curious why we didn't add Optional[T] to the return type
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Issue #, if available: #407
Description of changes:
A step serialized its result for the checkpoint but returned the raw in-memory value on the first run, while replay returned the value deserialized from the checkpoint. With a non-identity custom SerDes the first-run and replay results diverged, breaking the durability guarantee that a step yields the same value on every run.
Serialize then deserialize the result before returning it on the first run so the value matches what replay produces.
Added unit, integration, and example coverage for the first-run/replay equality with a non-identity SerDes.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.