ComfyUI-H3-Continuum 3.9.9 — public KJ / Review continuation CPU 2075 passed / 2 skipped / 2 known xfailed; upstream KJ GPU and installed local-KJ browser Continue PASS. 既存Take・Workflowは維持。更新後は再起動・ブラウザー再読込。旧unsafe/v5は自動変換しません。 対応条件と検証範囲: docs/RELEASE_NOTES_V399.md / docs/KJ_PREVIEW_OVERRIDE_V398.md --- Historical package record (unchanged) --- ComfyUI-H3-Continuum 3.9.8 / KJ Model Preview Override continuation Full real-Core CPU gate: 1,995 passed / 3 skipped / 2 existing Inspector expected failures / zero unexpected failures and CUDA initialization attempts; the 48 dedicated KJ tests passed without skips. Normal ComfyUI GPU acceptance passed four Core-API Queues: PyTorch and Comfy Kitchen INT8 + the audited local Spectrum implementation, each with separate-Queue Chunk 1 to 2 continuation and held preview images. Both second Queues reused one chunk and generated one; original Chunk 1 bytes and contracts were unchanged. All four video/audio outputs fully decoded. See docs/KJ_PREVIEW_OVERRIDE_V398.md for exact support and limits. Support: explicitly audited provider only; Preview Override VAE/Audio VAE disconnected, Tiny VAE none, Core NoPreviews/Auto/Latent2RGB. Restart and start a new Run. Old unsafe/v5 Takes and workflow bytes remain protected. --- Earlier package validation records (historical) --- ComfyUI-H3-Continuum 3.9.7 / External-input regeneration cache CPU: 1,948 passed / 2 skipped / 2 known Inspector expected failures; zero unexpected failures and zero CUDA initialization attempts. Targeted Core-API GPU queues verified external Chunks=2 / Nonce=0, repeated identical automatic variations 1 then 2, and 1 reused / 1 generated chunk per variation. Browser save/reload and Run-button submissions were checked separately. GPU scope: PyTorch attention, Spectrum OFF, 512x512, 8-step Turbo, two 5-second chunks. This is a targeted cache repair, not general image-quality or all-accelerator acceptance. Existing workflow and Take files were hash-protected. No storage/schema migration. --- Previous 3.9.6 package record (historical) --- ComfyUI-H3-Continuum 3.9.6 / Regeneration intent and automatic variations Full CPU gate: 1,908 passed / 2 skipped / 2 known Inspector expected failures; zero unexpected failures and CUDA initialization attempts. Normal ComfyUI browser save/reload retained Chunk 2 / nonce 5 and Chunk 1 / nonce 0 after history loading. Five new-Run GPU queues passed continuation, Chunk 2 regeneration and two identical Full Run / nonce 0 submissions; earlier Chunk 1 bytes were preserved. GPU scope: PyTorch, Spectrum OFF, 512x512, 8-step Turbo; all five media outputs fully decoded. See docs/REGENERATE_INTENT_V396.md. Update: restart ComfyUI and refresh the browser; select and save old unmarked Chunk 1 / nonce 0 again. Do not delete Runs/Takes. GitHub Release/Registry upload, service scan approval and Manager installation are separate publication checks. --- Previous 3.9.5 package record (historical) --- Previous 3.9.5 validation: 1,893 CPU passed / 2 skipped / 2 existing Inspector xfailed / failures 0; CUDA initialization 0. Public Spectrum 0.2.29 PyTorch/Kitchen Chunk 1→2 GPU continuation and full 10-second AV decode PASS. Details: docs/ISSUE31_SAMPLING_SIGNATURES.md ComfyUI-H3-Continuum 3.9.5 / Issue #31 Attention and Spectrum identity - 2026-10-15 Package metadata: 3.9.5. Public registration: twelve nodes. Existing V3.9 and separate V3.8X2 workflows remain supported without schema migration. Release artifacts and integrity (2026-10-13) **Latest package: 3.9.5.** Use the fixed-version GitHub downloads below. Registry publication is verified separately. | Version | Release | Custom-node ZIP | SHA-256 checksum | | --- | --- | --- | --- | | 3.9.5 | [Release](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/tag/v3.9.5) | [ZIP](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.5/ComfyUI-H3-Continuum-v3.9.5.zip) | [SHA256SUMS.txt](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.5/SHA256SUMS.txt) | | 3.9.4 | [Release](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/tag/v3.9.4) | [ZIP](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.4/ComfyUI-H3-Continuum-v3.9.4.zip) | [SHA256SUMS.txt](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.4/SHA256SUMS.txt) | | 3.9.3 | [Release](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/tag/v3.9.3) | [ZIP](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.3/ComfyUI-H3-Continuum-v3.9.3.zip) | [SHA256SUMS.txt](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.3/SHA256SUMS.txt) | | 3.9.2 | [Release](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/tag/v3.9.2) | [ZIP](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.2/ComfyUI-H3-Continuum-v3.9.2.zip) | [SHA256SUMS.txt](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/releases/download/v3.9.2/SHA256SUMS.txt) | The original [3.9.2 GitHub source archive](https://github.com/ukr8b3g-cmyk/ComfyUI-H3-Continuum/archive/refs/tags/v3.9.2.zip) remains available from its unchanged tag. This custom-node package ZIP is distinct from the official workflow ZIPs. 3.9.4 changes package selection and version metadata only. The Registry publication workflow checks the actual ZIP file set, CRC and all declared hashes before uploading. The 3.9.3 Registry ZIP omitted docs/REVIEW_RESTORATION_REPAIR.md and included three undeclared prompt-authoring files; GitHub 3.9.3 assets are unaffected. Historical validation counts below retain their original version and scope. No new GPU, long-form performance or physical SSD-write validation is claimed. Scope - Server-owned scratch path for Finalize and Decode Cache. Blank inherits existing defaults; selected storage never silently falls back to OS temp. - Settings panel: locations, logical bytes, free capacity, explicit refresh and conservative unused-file cleanup. No periodic filesystem polling. - Protect live Tensor/storage aliases, cache mappings, active processes, unknown owners, unregistered files and directory/link substitutions. - Atomic settings/revision checks and queue/operation coordination. New files use the selected destination; existing outputs retain their paths. - Manual RAM and Disk-backed preflight; discoverable cgroup memory limits. Auto's 4 GiB threshold and reserve, dtype and generation contracts remain. Candidate validation - Real ComfyUI Core 0.38.0 CPU regression: 1740 passed / 2 skipped / 2 expected failures; zero unexpected failures and CUDA initializations. Expected failures still track the independent Inspector reconfigure bug. Skips: Windows private-commit metric and absent historical V3.5 workflow. - New storage regression gate: 37 cases, including real small Tensor/view mappings, lost/changed roots, failed writes/deletes, ownerless/link protection, detached helper leases, abrupt subprocess exit, two-process settings conflict, HTTP queue/read-only guards, and frontend lifecycle. - tools/validate.ps1: compileall and Core runtime verification passed using a CPU-only Python adapter; full pytest was run separately as above. Native PackedLayout, Fixed 3x5s planning and twelve-node registration pass. - Actual Core 0.38.0 / frontend 1.53.6 / Chromium 151 smoke passed: real settings navigation, invalid-path rejection, validation/save/cleanup, ten modal reopens without polling/duplicate requests, page reload with persisted settings, native synthetic graph serialization/reload, and Japanese controls at a narrow viewport. No page errors or prompt submits. The separate production-JS fixture covers 100 reopen cycles. - All 25 tracked JSON/ZIP assets, including official workflows, retain their pre-change bytes. Source and Registry payload hashes are verified. - Windows-focused storage/file-backed/cache CI is configured; its hosted result is a separate gate from these local Linux checks. The browser gate is headless desktop Chromium, not mobile hardware or GPU generation. 3.9.3 frontend follow-up - Empty HTTP 404 reproduced the reported parser error and empty selector. User startup logs identified the failed updated process binding to a port already held by an older process. Starting the updated backend on an available port restored status and controls in the user's environment. - Improve HTTP/invalid-response diagnostics, initialized candidate labels, full-width layout and manual recovery. Preserve Core URL/auth/profile handling; no automatic retries or replay of mutating requests. - Targeted CPU/JavaScript/HTTP regression: 59 passed / 2 existing expected Inspector failures. Full Core 0.39.0 CPU regression: 1740 passed / 2 skipped / 2 existing Inspector expected failures; CUDA attempts zero. - Core 0.39.0 / frontend 1.53.10 static assets / Chromium smoke passed on an alternate loopback host and port: initial empty-404 reproduction, explicit HTTP 404/403/empty-200 errors, initialized full-width selector, Refresh recovery without page reload, and actual settings save/cleanup. No page errors, prompt submissions or automatic mutation replay. This used Linux CPU dependencies; it is separate from Windows/GPU quality validation. Core registration/planning and compileall also passed. - Source/Registry hashes and unchanged workflow bytes are checked again. Limits and update No GPU generation, long-form throughput, physical SSD write measurement or mobile hardware validation was performed for this storage change. Memory and capacity checks do not reserve resources or prevent all OOM/I/O faults. External drives must remain connected while outputs refer to them. Unknown legacy cache files remain protected. Multi-user UI mutations are disabled. The existing Inspector reconfigure expected failures remain independent. Update custom nodes, restart ComfyUI normally, then hard-refresh the browser. Changing the saved scratch setting afterwards requires no restart. Source publication is separate from a GitHub Release/tag or Registry publication. ====================================================================== Historical 3.9.2 validation (counts and scope below belong to that version) ====================================================================== ComfyUI-H3-Continuum 3.9.2 / Issue #30 frontend hotfix - 2026-10-13 Package metadata: 3.9.2. Public registration: twelve nodes. Continue using V3.9 nodes; V3.8X2 remains a separate workflow. Repair scope - Skip upstream Review settings snapshots only without a comparison baseline. - Notify each connected Sampler once per Reference selection, preserving synchronous native values and the original widget callback. - Skip setSize when the complete Reference node target dimensions equal its actual dimensions. Height-style same-value assignment remains. - Retain Review monitoring with a baseline and fresh Take/Queue checks. - Keep Sampling Contract v6 / Graph Contract v4, Run Storage v3, public IDs, widget/socket order and official workflow bytes. Validation scopes (do not add counts from different environments) - Frontend repair with real Core, CPU: 1703 passed / 2 skipped / 2 xfailed. The two explicit expected failures track one independent Inspector bug: stale verified text after 2 or 100 Sampler configure passes. - Skips: Windows private-commit metric and absent historical V3.5 workflow. - Existing Review Queue JavaScript: 62/62 passed; external INT: 22/22 passed. - Offline native PackedLayout, Fixed 3x5s planning and twelve-node registration passed. CUDA initialization attempts: zero. - Isolated desktop Chromium, actual frontend JS/DOM with graph/API fixtures: 600 baseline-free draw callbacks: snapshots 600 -> 0. With a baseline, snapshots remain 600. Thirty connected checks: notifications 60 -> 30, unchanged-dimension setSize calls 30 -> 0. Table rebuilding remains. - Release-candidate CPU/metadata and real-ComfyUI browser results are recorded below after final preparation. Standalone GitHub Actions is a separate gate. Remaining limits Mobile freeze resolution is unconfirmed. Review snapshots with a baseline, Reference table rebuilding and the independent Inspector display bug remain. No new GPU generation was run for this frontend hotfix. Earlier 3.9.1 GPU results are historical evidence, not 3.9.2 GPU validation. Full official workflow generation and all accelerators remain outside this repair gate. Updating and publication Hard-refresh the browser after updating JavaScript. Restart ComfyUI normally if the displayed package version also needs refreshing. 3.9.1 Runs/Takes require no migration for this UI hotfix. Users upgrading from pre-v6 code must still follow the older v5 history instructions below. GitHub Release/tag and Registry publication are distinct; old tags are kept. Current publication status is listed in GitHub Releases. Registry publication is not part of this hotfix task. Final 3.9.2 candidate validation - 2026-10-13 - Real-Core CPU regression: 1703 passed / 2 skipped / 2 xfailed; no unexpected failures; zero CUDA initialization attempts. The independent Inspector reconfigure bug remains explicitly tracked. - Actual ComfyUI 0.38.0 / frontend 1.53.6 / Chromium 151.0.7922.173 smoke: physical checkbox interactions, Prompt/link/Bypass edit detection and restoration, Chunks shrink/restore, pan/zoom, native assignment graph serialization and app.loadGraphData reload passed. - Before/after frontend JS responses used the actual Core/frontend/LiteGraph with additive instrumentation. Baseline-free 600 draw callbacks: 600 -> 0 snapshots. With a baseline: 600 -> 600. One checkbox: 2 -> 1 Sampler notifications; same-dimension resize: 1 -> 0. - Scope: desktop headless browser, synthetic connected image/string nodes, model-free CPU server and graph reload. UI file-save, real stored Take adoption, mobile hardware and full official workflow acceptance remain separate. No POST /prompt, generation or GPU execution occurred. - GitHub Actions status must be read from the published candidate commit; the real-Core count above is not a claim that Core cases ran in standalone CI. ====================================================================== Historical 3.9.1 stability validation - 2026-10-04 The original record below describes that package, its tests and publication. ====================================================================== ComfyUI-H3-Continuum 3.9.1 / V3.9 stability validation - 2026-10-04 Current package metadata: 3.9.1. Public registration: twelve nodes. The V3.9 Sampler and Reference Images helper retain their public IDs and official workflows; V3.8X2 remains a separate workflow. LoRA Plan / V4.0 is not included. Run Storage v3 and State/Session formats are unchanged. Repair scope - Preserve audio-only/mixed Second Pass keyframes; spatially adapt video only. - Match partial/middle-chunk Review conditioning and References to the actual output physical groups, retaining Terminal Merge groups as atomic units. - Use Sampling Contract v6 / Graph Contract v4 for Take compatibility, including sampler closures and MODEL CFG/wrapper settings. - Crop Review Driving Audio once from the original physical-group time range. - Prefer Core audio resampling. Direct weight mutation inside the same VAE object requires explicit Reference Encode Cache clearing or VAE reload. Recorded validation (separate environments; do not add these counts) - Full CPU with real ComfyUI Core: 1,696 passed / 1 skipped / 0 failed. CUDA initialization attempts: zero. - Release preparation with real Core: 66 passed / 0 skipped / 0 failed. - Standalone dependency check: 48 passed / 18 skipped / 0 failed. - Release document/manifest checks: 9 passed / 0 failed. - GitHub Actions at release commit 94eaf70d674a58244e1cb139f021290d63fb2e6c: 1,676 passed / 21 skipped / 0 failed. Eighteen cases require real Core and are explicitly skipped without it; three retain existing environment or fixture limits. This is not 1,696 real-Core cases executed in CI. - Offline native PackedLayout, Fixed 3x5-second planning and twelve-node registration passed. - Targeted GPU partial-prefix, single-middle-chunk and complete-sequence Review Second Pass checks passed at 512x512 / 24 fps, including conditioning inheritance, source-PCM alignment and saved Take/raw preservation. - Integrated generation and Take-reuse GPU checks passed for the repaired runtime. These are targeted results, not universal quality/accelerator gates. Current manifests declare 429 development files and 144 Registry payload files. Hash integrity does not establish semantic freshness of every document. The official V3.9 workflow retains 32 nodes / 39 links, the original JSON bytes and matching workflow ZIP. Models/media are not included. Publication and remaining gates - GitHub Release v3.9.1 is published. Its tag remains fixed at the release commit above; later documentation corrections are follow-up main updates. - Registry publication is separate and has not been performed for this patch. - Browser save/reload acceptance and full acceptance of the unchanged default workflow remain separate gates; do not infer them from CPU or targeted GPU results. Record the tested workflow and runtime when completing them. - Optional latent-upscale doubled outlines remain separate in Issue #27. Updating: restart ComfyUI and refresh the browser. Old v5 Takes/history remain readable and are not deleted or rewritten, but cannot supply prefixes to new v6 generation. Start a new Full Video or Start again from Chunk 1. Continue using the V3.9 nodes. See README.md, README_JA.md and docs/VALIDATION_JA.md for current instructions. ====================================================================== Historical V3.9.0 main-source record - 2026-09-29 The original record below is retained as history. Its metadata, test counts, manifest counts and publication status do not describe package 3.9.1. ====================================================================== ComfyUI-H3-Continuum 3.9.0 / V3.9 main-source validation — 2026-09-29 Status: MAIN SOURCE PUBLICATION. The ComfyUI_W worktree has 12 public node IDs: the existing V3.8 and V3.9 Samplers plus the new V3.9 Reference Images helper. Package metadata is 3.9.0. A GitHub main commit is not a new GitHub Release or Registry publish; earlier Releases are retained. The user-selected official V3.9 JSON is copied byte-for-byte and has 32 nodes/39 links. Its matching one-JSON ZIP and gallery copy preserve that graph. The standard V38X2 ZIP now includes both unchanged V38X2 and V3.9 JSONs for one-download setup; the V38X2 JSON itself is unchanged. V3.9 retains the V38X2-based model/LoRA, Decode Cache, Finalize and Save chain; Image loaders 1–6 are wired to the new helper but bypassed by default, while 7–9 require user-added loaders. Its Sequence Prompt is blank. Its saved size controls are First Image + Draft 0.30 MP, with the First Image loader enabled; users must select their own image before Queue. For Reference-only, turn it off and choose Manual. Saved 480x640 applies only to Manual mode. Other saved controls include two 10-second chunks, Full Run mode, Spectrum/Turbo LoRA off, and a separate Project ID. V3.9 connects fixed Reference Images 1–9 to one helper, then connects its one typed output to the V3.9 Sampler. The helper's in-node table selects All chunks or per-chunk assignments. The backend retains sparse fixed IDs, uses the existing effective plan/Run Storage contracts, and rejects incompatible payloads. The V3.8 Sampler and official V38X2 workflow JSON/ZIP are unchanged. Older V3.9 prototype workflows need manual rewiring; saved Runs/Takes are not automatically migrated. The latest recorded Windows CPU run is 1536 passed, 1 skipped, 0 failed. A separate Linux/Python 3.13.5 CPU audit reported 1534 passed, 3 skipped, 0 failed; that external count was not rerun during this documentation edit. The selected workflow JSON has 32/39 nodes/links. The development Manifest contains 406 entries and the Registry Manifest contains 138 entries. Both were resynchronized and hash-checked after this documentation reconciliation. The earlier RR-R6B browser result predates this helper UI, and configured Reference-only GPU/API evidence does not establish browser save/reload or GPU acceptance of the current default official template. That acceptance remains HOLD. This source publication makes no generation-code changes relative to the validated local V3.9 candidate and does not create a tag, GitHub Release, Registry package, or new GPU run. Historical V3.8X2 / package 3.8.3 validation record follows. Its test counts, node count and workflow SHA describe that earlier preparation, not the current V3.9 candidate: ComfyUI-H3-Continuum V3.8X2 / package 3.8.3 — package validation record Status: local commit/push preparation PASS. No commit, push, tag, GitHub Release, or Registry publication is authorized or performed by this record. Public surface: exactly nine nodes - H3 Continuum Sampler V3.8 - H3 Continuum Reference Audios - H3 Continuum Finalize - H3 Continuum Load Image - H3 Continuum Load Audio - H3 Continuum Load Video - H3 Continuum Second Pass - H3 Continuum Reference Images - Decode Cache Helper Official V3.8X2 workflows - MiniMax_H3_Continuum_V38X2.json / .zip - MiniMax_H3_Continuum_V38X2+Decode_Cache_Helper.json / .zip - Both names contain the same graph. The current standard V38X2 ZIP also includes V3.9; the helper-explicit ZIP remains single-JSON. The historical V38X2 JSON SHA-256 is 6c445d62979632bc1a801b1c28dbd639c23b51cc156ebae18837baaa757e592b. Reference Image 1–9 resource gate - Tested system: NVIDIA GeForce RTX 5060 Ti 16 GB / RAM 64 GB. - Tested run: 512 x 608, 3 x 5 seconds, Full Run, Spectrum Off, LoRA Off, Sage Attention, Balanced 22-frame continuity. - Nine test references were reduced to about 0.30 MP each. - Sampling used about 15.0–15.5 GiB of the available 16,311 MiB VRAM. - This is a high-resource configuration with little 16 GB headroom. Pre-resize source-image copies to about 0.30 MP before loading all nine. Larger inputs, output sizes, wrappers, or concurrent workloads may require more VRAM/RAM. Decode Cache Helper boundary - It can reuse Decode only when latent contents, shape, dtype, decode metadata, VAE identity, and a process-local cache entry all match. - It does not accelerate Sampling. A first MISS can be slower. - The measured 20%+ reductions are configuration-specific observations, not an unconditional speed guarantee. Preparation verification - Full CPU suite: 1,367 passed / 1 skipped / 0 failed. - Frontend Review Queue: 46/46 PASS. - Workflow JSON structure: 32 nodes / 38 links / zero broken saved links. - Workflow aliases: byte-identical JSON; ZIP payload parity required. - Development and Registry manifests: regenerated after documentation and workflow bytes were finalized; final hash verification is part of the commit-ready Gate. Registry package policy - Include production runtime modules and only the explicitly re-included tools, documents, and four V3.8X2 workflow files from .comfyignore. - Exclude tests, internal AGENTS/PROJECT_STATE/WORKLOG files, development MANIFEST.sha256, historical workflow files, caches, and unrelated ZIPs from the Registry payload. - MANIFEST.sha256 covers the prepared source tree except itself and the three internal project-state files. REGISTRY_MANIFEST.sha256 covers the actual Registry payload except itself.