Conversation
V2 styled string nodes are rendered by ArkUI Text internally. Registering the legacy foreground-draw callback for those nodes can route them through OH_Drawing_TypographyPaint before the paragraph is layout-ready, which spams paragraph-is-not-formatted warnings. Track the styled-string path on KRRichTextView, unregister/skip foreground draw for that path, and reset styled-string state when falling back to the legacy typography path.
Throttle Compose↔Native scroll sync (calc/expand deferral, contentSize dedup, scrollEnd finalize), quantize ArkUI onScroll in fling, and draw RichText when typography is ready during main-thread tasks. Co-authored-by: Cursor <cursoragent@cursor.com>
Quantize ArkUI scroll callbacks and SubcomposeLayout deltas at 0.5px, dedupe content size updates, and flush the final offset on scroll stop. Add OHOS demo cold-start page params and a CanScrollForward repro page. Co-authored-by: Cursor <cursoragent@cursor.com>
…ess test Add a 500-item repeated demo list with Card UI and inspector enabled to reproduce and profile HarmonyOS scroll performance locally. Co-authored-by: Cursor <cursoragent@cursor.com>
Clamp ArkUI scroll-frame-begin offset at nested list edges and skip compose-side expand/offset correction when nestedScroll is configured. Co-authored-by: Cursor <cursoragent@cursor.com>
Only recalculate native content size when contentView main-axis height changes during fling, avoiding ~650 redundant calc calls per gesture while keeping near-bottom and scrollEnd sync unchanged. Co-authored-by: Cursor <cursoragent@cursor.com>
Extend KuiklyScrollTrace with remeasure, frame timing, and offset/drag skip metrics for OHOS scroll stress validation; gate redundant contentOffset and isDragging writes; skip resetViewVisible when unchanged. Co-authored-by: Cursor <cursoragent@cursor.com>
bytemain
had a problem deploying
to
raft-artifacts-production
August 10, 2026 16:59 — with 
GitHub Actions
Failure
GitHub Actions
Failure
bytemain
had a problem deploying
to
raft-artifacts-production
August 10, 2026 17:04 — with
GitHub Actions
Failure
GitHub Actions
Failure
bytemain
temporarily deployed
to
raft-artifacts-production
August 10, 2026 17:08 — with
GitHub Actions
Inactive
GitHub Actions
Inactive
Signed-off-by: Argus <argus@mail.build> Co-authored-by: Argus <argus@mail.build>
bytemain
temporarily deployed
to
raft-artifacts-production
August 10, 2026 17:59 — with
GitHub Actions
Inactive
GitHub Actions
Inactive
Signed-off-by: Argus <argus@mail.build> Co-authored-by: Argus <argus@mail.build>
bytemain
temporarily deployed
to
raft-artifacts-production
August 11, 2026 12:13 — with
GitHub Actions
Inactive
GitHub Actions
Inactive
Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
[task #93] Add immutable Kuikly Raft release set Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build> Co-authored-by: BiSheng <raft-mobile-bisheng@mail.build>
Pin the protected Task93 writer to the exact terminal candidate run's five producer shards. Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
Restrict publish-mode Raft-exclusive Kotlin predecessors to the two public task #127 runtime modules, while compiler/toolchain dependencies retain the existing Tencent authority. Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
Skip producer and candidate jobs in writer mode, and start the publish job only after contract success plus exact skipped-result checks. The writer continues to use the frozen candidate run shards. Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
bytemain
had a problem deploying
to
raft-artifacts-production
August 12, 2026 12:42 — with
GitHub Actions
Failure
GitHub Actions
Failure
bytemain
had a problem deploying
to
raft-artifacts-production
August 12, 2026 12:46 — with
GitHub Actions
Failure
GitHub Actions
Failure
bytemain
had a problem deploying
to
raft-artifacts-production
August 12, 2026 12:48 — with
GitHub Actions
Error
GitHub Actions
Error
Signed-off-by: Lin Zhou <raft-mobile-linzhou@mail.build>
bytemain
had a problem deploying
to
raft-artifacts-production
August 12, 2026 20:35 — with
GitHub Actions
Failure
GitHub Actions
Failure
bytemain
had a problem deploying
to
raft-artifacts-production
August 12, 2026 20:37 — with
GitHub Actions
Failure
GitHub Actions
Failure
Signed-off-by: Lin Zhou <raft-mobile-linzhou@mail.build>
bytemain
had a problem deploying
to
raft-artifacts-production
August 12, 2026 20:55 — with
GitHub Actions
Error
GitHub Actions
Error
Signed-off-by: artin <artin@cat.ms>
bytemain
had a problem deploying
to
raft-artifacts-production
August 13, 2026 07:22 — with
GitHub Actions
Failure
GitHub Actions
Failure
Signed-off-by: artin <artin@cat.ms>
Task #141 follow-up to task #93. Freezes the 45-file closure for org.jetbrains.kotlinx:kotlinx-coroutines-test:1.8.0-KBA-002 and its available-at kotlinx-coroutines-test-ohosarm64 module, with a Hosted carrier that validates the contract and resolves the authority bytes from an empty cache. The authority supplies 50 files for these coordinates, not 45. The difference is the root -kotlin-tooling-metadata.json plus its 4 checksum sidecars, which take no part in dependency resolution and are absent from Raft for the already-published sibling family. This task keeps that family shape and mirrors 45; the gap is named in the manifest and bound by contract as authority_files=50, because a create-only Maven write can never be yanked while an absent file can always be added later. The mirrored root resolves for OHOS targets only; non-OHOS available-at platform modules are not mirrored here. Recorded, not incidental. kotlin-stdlib 2.0.21-KBA-001 stays a conditional closure member and is not published. Before the first PUT, a Raft-only empty-cache OHOS consumer resolution must record the real dependency graph and the stdlib finally selected. Landing this source grants no publication authority. Publication remains 0 PUT until that gate passes. Independent review: task #142 FINAL PASS / 0 blocker by CC-希乐 on exact 586df6f. Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
Lands the mirror publication process in the repo so the next coordinate does not have to rediscover it, per artin's request in #Kuiklybase:8d8d3ee7. RUNBOOK.md records the traps that actually cost CI rounds: the authority serves no directory listing and answers directory URLs with 200 plus a not-browseable notice rather than 404; Raft answers some absent paths with 401, so 404 is not the only absent signal; Raft serves no maven-metadata.xml, so only exact versions resolve; carriers must be exactly one commit; and expected_sha must be the landed staging3 tip, because squash-merging means the reviewed head is not on staging3. Two enumeration habits are recorded as actions rather than maxims: keep the probe grid in the manifest header, and pair every zero-hit claim with a known-present positive control. stage_manifest.py materialises a frozen manifest into a local Maven layout with per-byte verification, so a consumer can be exercised before an unyankable write. It ships no consumer-side version check: such a check reads a build file this repository does not own, and shipping it would imply coverage that does not exist. That contract belongs to Mobile. Independent review: task #143 FINAL PASS / 0 blocker by CC-希乐 on exact 0469c59, after they found and confirmed the fix for a blocker where the removed consumer check had taken the exception class with it, leaving every failure path raising NameError. No writer, manifest or workflow is touched, so the task #141 45-file publication receipt is unaffected. Signed-off-by: BiSheng <raft-mobile-bisheng@mail.build>
Ports two iOS crash guards that landed on staging2 and that the Kuikly 2.24.0 staging3 replay did not carry. Original authorship (metadata cannot hold it under the identity-diff gate, which requires exactly one commit with author == committer): 926462b fix(ios): guard optional render view method dispatch (#123) 07b14b3 fix(ios): guard gradient rich text method forwarding (#124) both authored by Jiacheng on staging2, 2026-08-07 Independent review: LiBai — SOURCE GO bound to exact 5a5022d (file set, line-for-line content, tree and DCO independently recomputed). EVIDENCE TIERS — do not collapse these: source transport fidelity verified (patches byte-identical to origin) artifact change confirmed and isolated by control: of 34 archives in the release candidate, 33 are content-identical to the published set (differing only in packaging timestamps); the sole content difference is 2 entries in core-render-ios xcframework.zip, the compiled OpenKuiklyIOSRender binaries — this PR's expected output runtime NOT RUN. No iOS build, simulator or device has exercised these guards. The crash exposure was established by reading the two call sites, never reproduced. There is currently no iOS physical execution seat. KNOWN COVERAGE GAP: The KRGradientRichTextViewMethodDispatchTest fixture covers the KRGradientRichTextView guard only. The KuiklyRenderLayerHandler.mm view dispatch guard has NO test — deleting it leaves CI fully green. The ios-renderer job's green does not speak to that guard. The fixture is also compiled for the macOS host (-DTARGET_OS_OSX=1 + AppKit), so its pass proves the host dispatch contract, not iOS device behaviour. The Task93 immutable-release-set workflow fails on this PR. That gate compares byte equality while POMs embed the source commit SHA and archives embed build timestamps, so it reports CONFLICT for any source change after publication and carries no information about correctness. artin ruled it does not gate source PRs (2026-08-21).
The candidate upload had no `if: always()`, so a failing "Assemble and preflight the source-review candidate" step skipped the artifact upload and discarded public-plan.json -- exactly in the case where it is the only thing that explains the failure. kuikly_maven_publish.py writes the full plan to disk and prints only a one-line summary. That line cannot distinguish the four causes of CONFLICT; the discriminating lists (unexpected / different / listingDisagreement) exist only in the uploaded plan. if-no-files-found relaxed to `warn` for the same step so that a failure early enough to abort before the candidate directory exists cannot manufacture a second, misleading failure on top of the real one. The release-receipt upload at the end of this workflow already carried `if: always()`. Verified on this PR's own failing run (32507129327): assemble-candidate failed with `publication plan: CONFLICT present=920/920`, the upload step reported success rather than skipped, and public-plan.json was retrievable from the artifact -- yielding unexpected=0, different=485, listingDisagreement=0 directly, instead of requiring the five stage artifacts to be re-downloaded and classify to be re-run by hand. Independent review: LiBai -- GO bound to exact 604bacb, including confirming that SOURCE_SHA is a workflow-level env and so is always populated on the newly-reachable early-failure paths. The Task93 workflow itself fails on this PR: it triggers on its own workflow file, and its byte-equality criterion cannot hold once POMs embed the source SHA. artin ruled (2026-08-21) that it does not gate source PRs. Not addressed here, both needing a design decision: the `core-render-ios/**` trigger pulling source PRs into a publication gate, and the byte-equality criterion itself.
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
11 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.
Purpose
Review-only Draft PR for task #73. This shows the complete net change required to move the bytemain fork from the current
mainbaseline to the auditedstaging3Kuikly 2.24.0 migration candidate.Do not merge this PR until the product cutover strategy is explicitly approved.
staging2and the current Mobile gitlink/HAR remain the production path.Exact coordinates
main@1048ad0189ee4579bd9585d00b9623e3d4d6a125staging3@d37094f3a7e2e105d1959e2509dd1d3c5828f76964d236099ddcaa2e405b98f137851fdc2fc1009fTencent-TDS/KuiklyUI 2.24.0@76a866795903014cfd2d0363ee31ccee7895775estaging2@6a531a57105ee453edcfdc07a54d1bb1b434843125cbc1fa25be4bba0ca800a67762a1f1317c5f25, with real parentsc8f92f053fb60d6fe515b5e986e1c6fe2db775a0and6a531a57105ee453edcfdc07a54d1bb1b4348431(not-s ours)Migration shape
The detailed carried-patch inventory, held commits, conflict decisions, cutover gates, and rollback plan are in
docs/migrations/staging3-2.24.0.md.Exact-head evidence
Dedicated run: https://github.com/bytemain/KuiklyUI/actions/runs/30790072289 — terminal SUCCESS for exact head
d37094f3….916115238459161152385691611523886— 162 / 85 / 2 tests, zero failures or errors91611523880— Xcode 16.2, locked Pods, Kotlin/Native main/tests, native fixtures, production renderer with warnings as errors91611523905— pinned container, fixtures, HAR assembly, fresh AArch64 production direct-link91612353738Review boundary
staging2change.Please review the migration diff and the carried/held decisions. Merge/cutover remains a separate explicit decision.