test(gateway): unit regressions for _split_text_chunks under a custom length unit - #55843
Open
MaxFreedomPollard wants to merge 1 commit into
Open
Conversation
Contributor
Author
|
Issue with reproduction: #55844 |
alt-glitch
added
type/bug
Collaborator
Competing fix cluster for the |
7 tasks
Contributor
|
Thanks for isolating the fallback chunking defect; the diagnosis is correct on current main. Problems
Suggested changes
Automated hermes-sweeper review. |
teknium1
added
the
sweeper:blast-moderate
MaxFreedomPollard
force-pushed
the
fix/stream-chunk-utf16-oversize
branch
from
July 25, 2026 04:17
3083e76 to
a1b23aa
Compare
MaxFreedomPollard
changed the title
alt-glitch
added
type/test
Collaborator
Re-triage correction: the current head is a test-only follow-up after the fallback implementation landed elsewhere. It retains direct UTF-16 chunking coverage; the remaining progressive existing-message overflow concern is a separate code-path gap noted in review. Related to #42523 and #55844, not a competing implementation duplicate. |
MaxFreedomPollard
force-pushed
the
fix/stream-chunk-utf16-oversize
branch
from
July 25, 2026 05:10
a1b23aa to
49c0116
Compare
alt-glitch
added
the
area/streaming
MaxFreedomPollard
force-pushed
the
fix/stream-chunk-utf16-oversize
branch
3 times, most recently
from
July 25, 2026 06:09
0782064 to
1126e24
Compare
alt-glitch
removed
the
sweeper:risk-message-delivery
MaxFreedomPollard
force-pushed
the
fix/stream-chunk-utf16-oversize
branch
from
July 31, 2026 03:52
1126e24 to
c0071b7
Compare
… custom length unit The code fix (fallback split honoring the platform length unit, and the progressive existing-message overflow path) has landed on main via the competing fix; the suite still has no direct unit tests of _split_text_chunks with a non-codepoint len_fn. Add the emoji/UTF-16 regression cases: no-newline fallback, newline-in-first-half, and default-len behaviour unchanged.
MaxFreedomPollard
force-pushed
the
fix/stream-chunk-utf16-oversize
branch
from
July 31, 2026 04:27
c0071b7 to
8608203
Compare
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
3 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.
Rebased and rescoped: the code fix landed on main via the competing fix in this cluster (the fallback split now uses the codepoint budget, and the progressive existing-message overflow path was corrected as well — both verified on current main).
What remains from this PR is direct unit coverage: the suite exercises the UTF-16 paths at integration level but has no direct unit tests of
_split_text_chunkswith a non-codepointlen_fn. This adds the three regression cases (emoji no-newline fallback, newline-in-first-half, default-lenbehaviour unchanged) from the original report (#55844).