mirror of
https://github.com/Comfy-Org/ComfyUI_frontend.git
synced 2026-05-23 06:10:32 +00:00
9dd5bbb025a04d71aa8dafdbdcfc9fa8a9fb25c7
7999 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
dante01yoon
|
9dd5bbb025 |
fix(sharing): match regular load policy on shared workflow validation
Switch to raw workflow_json fallback when Zod validation fails, matching the policy at scripts/app.ts:1191-1198. Share service no longer validates directly — loadGraphData applies the same fallback under Comfy.Validation.Workflows, removing the cross-path policy inconsistency. Drops the tolerantArray combinator and per-field linearData.inputs relaxation added earlier in this PR. Zod schemas stay strict as the canonical spec; tolerance lives at the consumer boundary. Fixes the reported share=21e32125c692 load failure and any future extra.* shape drift, not just linearData.inputs. |
||
|
dante01yoon
|
db969dd50e |
refactor(validation): move linearData.inputs tolerance to tolerantArray combinator
Address review feedback: schema should represent the *specified* shape,
not "what will work and what won't". The previous in-schema transform
mixed canonical spec with tolerance, making the schema misleading to read.
- Restore `zLinearInput` to its strict union (2-tuple OR 3-tuple with
`{ height? }`) — this is the canonical write contract.
- Introduce `tolerantArray(itemSchema, label)` combinator that drops
invalid entries with a console warning, leaving valid ones intact.
The item schema stays strict; tolerance lives at the array container.
- Apply `tolerantArray` only to `extra.linearData.inputs` for now.
Other `extra.*` arrays remain untouched until proven necessary.
- Document that execution-critical arrays (`nodes`, `links`,
`widget_values`) must keep `z.array` so malformed entries fail loudly.
Behavior change vs the previous commit: legacy `[id, name, <number>]`
shapes are no longer coerced to `[id, name, { height: <number> }]` —
they're dropped. Without seeing the actual production data, coercing
guesses are riskier than dropping; a dedicated migration can be added
later if a recognizable legacy shape is confirmed.
|
||
|
dante01yoon
|
4c20d44e6b | docs(validation): convert zLinearInput comment to JSDoc | ||
|
dante01yoon
|
a3e8f8625d |
fix(validation): relax linearData.inputs schema to load legacy shares
Strict `z.union([3-tuple, 2-tuple])` rejected the entire workflow when
any single `extra.linearData.inputs` entry didn't match, producing
"Failed to load shared workflow: invalid workflow data" for published
shares whose entries were stored with legacy or unknown 3rd-element
shapes (raw height number, forward-compat fields, etc).
The cloud accepts and serves `workflow_json` as an opaque
`Record<string, unknown>` (see `packages/ingest-types/src/zod.gen.ts`),
so non-conforming shapes do get persisted and surface only on the
frontend's read path. Until the SOT gap is closed, validation here
must be tolerant of historical variants.
Replace the union with a permissive `tuple([nodeId, string]).rest(unknown)`
that normalizes the trailing element:
- `[id, name]` → unchanged
- `[id, name, { height? }]` → unchanged
- `[id, name, <number>]` → coerced to `[id, name, { height: <number> }]`
- `[id, name, <other>]` → trailing dropped, kept as 2-tuple
- `[id, name, _, _, ...]` → extras dropped
Only entries missing the `[nodeId, string]` prefix still reject.
|
||
|
Comfy Org PR Bot
|
a0150ffe17 |
1.45.6 (#12204)
Patch version increment to 1.45.6 **Base branch:** `main` ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12204-1-45-6-35f6d73d365081fc8539ca25a55aac74) by [Unito](https://www.unito.io) --------- Co-authored-by: christian-byrne <72887196+christian-byrne@users.noreply.github.com> Co-authored-by: github-actions <github-actions@github.com>v1.45.6 |
||
|
Christian Byrne
|
c92030b158 |
refactor: deduplicate Civitai hostname logic in getSourceName (#11822)
## Summary
- Extract `isCivitaiHost` private helper from `isCivitaiModelUrl` in
`formatUtil.ts` for DRY hostname checking
- Add `isCivitaiUrl` exported function for hostname-only Civitai URL
detection (distinct from `isCivitaiModelUrl` which also validates the
path format)
- Refactor `getSourceName` in `assetMetadataUtils.ts` to use the shared
`isCivitaiUrl` instead of inline duplicate hostname checks
- Add tests for `isCivitaiUrl` covering `.com`, `.red`, subdomain, and
invalid URL cases
## Changes
- `packages/shared-frontend-utils/src/formatUtil.ts` — add
`isCivitaiHost` private helper + export `isCivitaiUrl`; refactor
`isCivitaiModelUrl` to use helper
- `packages/shared-frontend-utils/src/formatUtil.test.ts` — add
`isCivitaiUrl` test suite
- `src/platform/assets/utils/assetMetadataUtils.ts` — import
`isCivitaiUrl` from `@/utils/formatUtil`; remove inline hostname logic
from `getSourceName`
## Testing
### Automated
- Added `isCivitaiUrl` test suite (6 cases: `.com`, `.red`, subdomains,
non-Civitai, invalid URL)
- All 71 existing `formatUtil` tests pass
- All 53 existing `assetMetadataUtils` tests pass (behavior preserved)
- TypeScript typecheck passes
### E2E Verification Steps
1. Run unit tests: `npx vitest run
packages/shared-frontend-utils/src/formatUtil.test.ts
src/platform/assets/utils/assetMetadataUtils.test.ts`
2. Expected: all tests pass
3. Verify `getSourceName('https://civitai.red/models/123')` returns
`'Civitai'`
4. Verify `isCivitaiUrl('https://civitai.com/models/any-path')` returns
`true`
5. Verify `isCivitaiModelUrl` still rejects non-API paths while
`isCivitaiUrl` accepts them
## Review Focus
`isCivitaiUrl` (new, hostname-only) vs `isCivitaiModelUrl` (existing,
hostname+path format): `getSourceName` needs to recognize ANY Civitai
URL as a source, so using `isCivitaiModelUrl` directly would incorrectly
reject valid browse URLs like `civitai.com/models/123`.
Closes #11357
┆Issue is synchronized with this [Notion
page](https://app.notion.com/p/PR-11822-refactor-deduplicate-Civitai-hostname-logic-in-getSourceName-3546d73d36508110974ccc3b7384d82b)
by [Unito](https://www.unito.io)
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
|
||
|
Dante
|
988d532467 |
fix(queue): contain JobDetailsPopover error message overflow (#12173)
## Summary Cap the Job Details "Error message" block at `max-h-96` (24rem / 384px) with an internal scroll, wrap long unbreakable tokens (filenames, JSON), and preserve newlines so the failed-job popover no longer grows unbounded. ## Changes - **What**: Added `max-h-96 overflow-y-auto whitespace-pre-wrap wrap-break-word` to the error message container in `JobDetailsPopover.vue`, plus a `FailedWithLongError` Storybook story covering the overflow case. ## Review Focus - 24rem cap was set per Alex's spec in the [Slack thread](https://comfy-organization.slack.com/archives/C0A4XMHANP3/p1778506109115989). - `wrap-break-word` (Tailwind 4 canonical of `break-words`) is needed because long underscore-joined filenames don't break naturally; `whitespace-pre-wrap` preserves any newlines in the raw error. - Not in scope: the popover z-index clipping issue Alex flagged later in the same thread — that's a separate follow-up. Fixes FE-660 ## Screenshots **Before** — error block grows unbounded with the panel:  **After** — error block capped at 384px and internally scrollable:  Reproduce locally via Storybook: `pnpm storybook` → Queue → JobDetailsPopover → **FailedWithLongError**. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12173-fix-queue-contain-JobDetailsPopover-error-message-overflow-35e6d73d3650812d9873e5d163cad0c6) by [Unito](https://www.unito.io) --------- Co-authored-by: Benjamin Lu <benjaminlu1107@gmail.com> |
||
|
AustinMroz
|
fe08ad2fcd |
Fix pre-commit linter skipping type checks (#12203)
Adds the `--type-aware` so that the typechecks performed by precommit
hooks have parity with the results output by a full `pnpm lint`
Most notably, unawaited promises would not be caught by the precommit
hooks prior to this PR.
```
× typescript-eslint(no-floating-promises): Promises must be awaited, add void operator to ignore.
╭─[browser_tests/fixtures/utils/vueNodeFixtures.ts:45:5]
44 │ async select() {
45 │ this.header.click()
· ───────────────────
46 │ }
╰────
```
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12203-Fix-pre-commit-linter-skipping-type-checks-35e6d73d365081a4adade833294df7ed)
by [Unito](https://www.unito.io)
|
||
|
Alexander Brown
|
93edf166d0 |
fix(website): link careers page to Ashby job description, not application form (#12200)
*PR Created by the Glary-Bot Agent*
---
## Summary
The careers page at comfy.org/careers was linking every role to its
Ashby application form (`.../{id}/application`) instead of the job
description page (`.../{id}/`). Users expect to first read the role
description, not land on the submit-resume page.
Ashby's job board API returns both `jobUrl` (description) and `applyUrl`
(application form). `toDomainRole` was preferring `applyUrl`; this PR
switches to `jobUrl` and renames the `Role` field accordingly so the
field name matches its meaning.
## Changes
- `apps/website/src/utils/ashby.ts`: use `job.jobUrl` directly instead
of `job.applyUrl ?? job.jobUrl`.
- `apps/website/src/data/roles.ts`: rename `Role.applyUrl` →
`Role.jobUrl`.
- `apps/website/src/components/careers/RolesSection.vue`: update the `<a
:href>` binding.
- `apps/website/src/data/ashby-roles.snapshot.json`: regenerated
fallback snapshot — URLs stripped of `/application`, `id`s recomputed
from the new URLs.
- Unit + E2E tests updated; new E2E assertion that links do not end in
`/application` prevents regressions.
The Ashby schema (`ashby.schema.ts`) still accepts `applyUrl` since the
API returns it — we just no longer consume it.
## Verification
- `pnpm test:unit` — 70/70 pass
- `pnpm typecheck` — 0 errors
- `pnpm build` — succeeds; inspected `dist/careers/index.html`, all 19
Ashby links now point to description URLs and zero contain
`/application`
- Oracle code review — 0 issues
Fixes user report in #hiring-ideas (Slack).
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12200-fix-website-link-careers-page-to-Ashby-job-description-not-application-form-35e6d73d3650815cbedadf974f7d3364)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
|
||
|
Dante
|
9fe19a2afb |
fix(settings): unify settings item heights and use 14px label text (#12180)
## Summary Body text in the settings dialog was still rendering at the inherited 16px (browser default) instead of the 14px design spec, and rows with different control types (toggle, slider, dropdown, radio) collapsed to different heights — making the list look uneven and cramped. ## Changes - **What**: `FormItem` label now uses `text-sm` (14px) and the row enforces `min-h-8` (32px) so toggle/slider/dropdown/radio rows align. `SettingGroup` bumps inter-item margin from `mb-2` to `mb-3` for breathing room between settings. ## Review Focus `FormItem` is also used by `ServerConfigPanel`, so the 14px/32px row also applies there — consistent with the same settings-dialog visual language, but worth a glance. Fixes #FE-525 ## Screenshots Lite Graph panel (1280×900 viewport) showing toggle/slider/dropdown/radio rows side-by-side: | Before (`origin/main`) | After (this PR) | | --- | --- | | <img src="https://raw.githubusercontent.com/Comfy-Org/ComfyUI_frontend/pr-12180-screenshots/before-litegraph.png" width="480"> | <img src="https://raw.githubusercontent.com/Comfy-Org/ComfyUI_frontend/pr-12180-screenshots/after-litegraph.png" width="480"> | Before: label text inherits 16px from `<body>`; toggle-only rows (e.g. "Always snap to grid", "Live selection") shrink to ~24px while dropdown/slider rows stay ~32px, so the list looks uneven and cramped. After: labels are 14px; every row is at least 32px tall so toggles/sliders/dropdowns/radios line up; `mb-3` adds 4px of breathing room between rows. ## References - Linear: https://linear.app/comfyorg/issue/FE-525/verify-settings-text-size-and-item-heights - Figma: https://www.figma.com/design/vALUV83vIdBzEsTJAhQgXq/Comfy-Design-System?node-id=6290-75412 - Origin thread: https://comfy-organization.slack.com/archives/C075ANWQ8KS/p1777657610484679?thread_ts=1776808927.654249 --------- Co-authored-by: github-actions <github-actions@github.com> Co-authored-by: Benjamin Lu <benjaminlu1107@gmail.com> |
||
|
Comfy Org PR Bot
|
6845d57a80 |
chore(website): refresh Ashby roles snapshot (#12191)
Automated refresh of `apps/website/src/data/ashby-roles.snapshot.json` from the Ashby job board API. **Flow:** 1. `Release: Website` workflow ran (manual trigger). 2. This PR opens with the regenerated snapshot. 3. `CI: Vercel Website Preview` deploys a preview for review. 4. Merging to `main` triggers the production Vercel deploy. The snapshot fallback in `apps/website/src/utils/ashby.ts` remains intact: builds without `WEBSITE_ASHBY_API_KEY` continue to use the committed snapshot. Triggered by workflow run `25746888214`. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12191-chore-website-refresh-Ashby-roles-snapshot-35e6d73d365081f4b2e1d802dd412a72) by [Unito](https://www.unito.io) Co-authored-by: Yourz <8287689+Yourz@users.noreply.github.com> |
||
|
Christian Byrne
|
469a5edf99 |
feat: cloud-nodes catalog at /cloud/supported-nodes (#11903)
*PR Created by the Glary-Bot Agent*
---
## Summary
Adds a comfy.org page that lists every custom-node pack supported on
Comfy Cloud, with per-pack detail subpages. Data is fetched at build
time from `cloud.comfy.org/api/object_info` (gated by
`WEBSITE_CLOUD_API_KEY`), sanitized of user content, joined with public
registry metadata from `api.comfy.org/nodes`, and falls back to a
committed snapshot — mirroring the existing Ashby careers integration
pattern.
- Index: `/cloud/supported-nodes` (en) and
`/zh-CN/cloud/supported-nodes` (zh-CN)
- Detail: `/cloud/supported-nodes/[pack]` and
`/zh-CN/cloud/supported-nodes/[pack]`, generated via `getStaticPaths()`
from the same fetcher as the index so the two routes can never diverge.
## What's new
**Shared package (extracted)**
- `@comfyorg/object-info-parser` — Zod schemas (`zComfyNodeDef`,
`validateComfyNodeDef`), node-source classifier (`getNodeSource`,
`isCustomNode`, `CORE_NODE_MODULES`), and helpers (`groupNodesByPack`,
`sanitizeUserContent`). `src/schemas/nodeDefSchema.ts` and
`src/types/nodeSource.ts` become 1-line re-export shims; existing
imports keep compiling.
**Build-time pipeline**
- `apps/website/src/utils/cloudNodes.ts` — Ashby-style fetcher:
retry/backoff `[1s, 2s, 4s]`, 10 s timeout via AbortController, Zod
envelope + per-node validation, snapshot fallback, memoized via
module-level `inflight` promise.
- `apps/website/src/utils/cloudNodes.registry.ts` — Public registry
enrichment (no auth, batches of 50, single retry, soft-fail).
- `apps/website/src/utils/cloudNodes.ci.ts` — GitHub Actions annotations
+ step summary mirroring the Ashby reporter.
- `apps/website/src/utils/cloudNodes.build.ts` — Single
`loadPacksForBuild()` consumed by both index and detail pages so they
share one source of truth.
- `apps/website/scripts/refresh-cloud-nodes-snapshot.ts` — atomic-rename
refresh CLI that walks pack/node string fields with a user-content
extension regex *before* renaming the snapshot into place.
- Mandatory user-content sanitization strips uploaded filenames from
combo lists (`LoadImage`, `LoadImageMask`, `LoadImageOutput`,
`LoadVideo`, `LoadAudio` zeroed; any combo value matching
`/\.(png|jpe?g|webp|gif|mp4|mov|webm|wav|mp3|flac|ogg|safetensors|ckpt|pt)$/i`
filtered).
**Page + components**
- `apps/website/src/pages/cloud/supported-nodes.astro` (en) + zh-CN
twin.
- `apps/website/src/pages/cloud/supported-nodes/[pack].astro` detail
(en) + zh-CN twin, async `getStaticPaths` driven by
`loadPacksForBuild()`.
-
`apps/website/src/components/cloud-nodes/{HeroSection,PackGridSection,PackCard,PackBanner,NodeList,PackDetail}.vue`
— Vue 3.5 destructured props, `cn()` from `@comfyorg/tailwind-utils`,
design-system tokens only, no PrimeVue.
- Pack card name links to its detail page; banner uses the shared
`fallback-gradient-avatar.svg` asset (copied into
`apps/website/public/assets/images/`) when `banner_url` and `icon` are
missing.
- 25 new `cloudNodes.*` i18n keys in `en` + `zh-CN`.
**Tests**
- 33 unit tests in `@comfyorg/object-info-parser` (schemas, classifier,
sanitizer, grouping).
- 19 new website unit tests covering fetcher (10), CI reporter (6),
registry enrichment (3) — Ashby patterns mirrored.
- E2E: index smoke + search + banner + detail click-through + direct
visit + zh-CN parity.
## Required maintainer follow-up
GitHub Apps cannot push `.github/workflows/*` changes (push was rejected
with `refusing to allow a GitHub App to create or update workflow …
without workflows permission`), so the workflow edits prepared in this
branch were reverted in commit `9be2abce8`. The intended diffs are
documented as copy-paste-ready snippets in `apps/website/README.md`
under the new "Cloud nodes integration → CI wiring" section.
A maintainer must:
1. Provision `WEBSITE_CLOUD_API_KEY` in the repo secrets and the Vercel
project env.
2. Apply the `ci-website-build.yaml` and
`ci-vercel-website-preview.yaml` diffs documented in the README directly
to `main` (or as a follow-up commit on this branch with a maintainer
account).
The committed snapshot lets builds succeed without the secret while the
maintainer step is pending — pages render from
`apps/website/src/data/cloud-nodes.snapshot.json`.
## Self-review (Oracle)
Two warnings caught and fixed in commits `deba5ab02` and `99dfc3381`:
- Index/detail pages now share a single source of truth
(`loadPacksForBuild`), so a fresh fetch can't expose packs whose detail
routes weren't generated.
- Refresh script validates parsed snapshot fields *before* the atomic
rename, instead of regex-scanning the serialized JSON after the file is
already in place.
## Quality gates (local)
```
pnpm --filter @comfyorg/object-info-parser test → 33 passed
pnpm --filter @comfyorg/website test:unit → 42 passed
pnpm --filter @comfyorg/website typecheck → 0 errors
pnpm --filter @comfyorg/website build → 47 pages built (incl. 6 cloud-nodes routes)
pnpm lint → 0 errors (1 pre-existing warning in unrelated test file)
pnpm knip → 0 errors (1 pre-existing tag hint in unrelated file)
```
E2E (`pnpm --filter @comfyorg/website test:e2e`) is intended to be run
by the Vercel/CI pipelines.
## Manual verification
Built `dist/`, served locally on port 4321, drove with Playwright:
- `/cloud/supported-nodes` renders both pack cards, search input, sort
dropdown
- `/cloud/supported-nodes/comfyui-impact-pack` renders the metadata grid
(publisher, downloads, stars, version, license, last updated) and 3
categorized node sections with 5 nodes total
- `/zh-CN/cloud/supported-nodes` localizes hero (`Comfy Cloud 上的自定义节点`),
label (`云端节点目录`), search placeholder (`搜索节点包或节点名称`), sort
- `/zh-CN/cloud/supported-nodes/comfyui-controlnet-aux` localizes every
metadata label (`查看仓库`, `发布者`, `下载量`, `GitHub 星标`, `最新版本`, `许可证`,
`最后更新`) and renders dates with `Intl.DateTimeFormat('zh-CN')`
(`2026年4月27日`)
- Search input narrows pack count from 2 to 1 when typing `impact`
(verified via DOM count)
Banners render the shared `fallback-gradient-avatar.svg` when the
snapshot's image URL doesn't resolve — expected in the local sandbox.
## Preview URL (after CI completes)
`https://comfy-website-preview-pr-{N}.vercel.app/cloud/supported-nodes`
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-11903-feat-cloud-nodes-catalog-at-cloud-supported-nodes-3566d73d36508194afdec5f389897585)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
Co-authored-by: GitHub Action <action@github.com>
|
||
|
Christian Byrne
|
35443e94f5 |
feat(website): SEO model pages — 207 models, FAQ JSON-LD, partner node support (#11892)
## Summary - Adds programmatic SEO model pages at `/p/supported-models/[slug]` for **207 models** auto-generated from `workflow_templates` (180 local + 27 partner nodes) - 3-file architecture: `generated-models.json` (auto-generated, checked in) + `model-metadata.ts` (editorial overrides) + `models.ts` (65-line merger) - Full JSON-LD per page: `SoftwareApplication` + `BreadcrumbList` + `FAQPage` (targeting AI Overviews / People Also Ask) - Partner node support: `directory: 'partner_nodes'` hides Download button, shows VIEW TUTORIAL - `generate-models.ts`: walks `workflow_templates` for local models + `API_PROVIDER_MAP` for 30+ partner integrations (Kling, Meshy, Luma, Runway, Stability AI, ByteDance, Google, etc.) - Weekly GH Actions workflow opens issue when new models appear in `workflow_templates` but not in `generated-models.json` - `add-model-page` Claude skill for Slack-driven model page PRs ## Files changed | File | Purpose | |------|---------| | `apps/website/src/config/generated-models.json` | Auto-generated, 207 models (27 partner + 180 local) | | `apps/website/src/config/model-metadata.ts` | Editorial overrides: docsUrl, blogUrl, featured (9 entries) | | `apps/website/src/config/models.ts` | 65-line merger — imports JSON + overrides, exports `models` + `getModelBySlug` | | `apps/website/scripts/generate-models.ts` | Build-time parser; run with `pnpm generate:models` | | `apps/website/src/i18n/translations.ts` | ~30 UI keys added (no per-model keys — displayName is plain string) | | `apps/website/src/pages/p/supported-models/[slug].astro` | Dynamic route with 3x JSON-LD schemas | | `apps/website/src/pages/p/supported-models/index.astro` | Model grid index page | | `apps/website/src/components/models/ModelHeroSection.vue` | Hero component | | `.github/workflows/model-page-discovery.yaml` | Weekly auto-discovery workflow | | `.claude/skills/add-model-page/SKILL.md` | Claude skill for adding/updating model pages | ## Test plan - [ ] `pnpm build` passes in `apps/website` - [ ] `/p/supported-models` index renders 207 model cards - [ ] `/p/supported-models/kling-ai` shows Partner Node eyebrow, no Download button, VIEW TUTORIAL CTA - [ ] `/p/supported-models/flux-1-dev` shows Diffusion Model eyebrow, Download + Tutorial buttons - [ ] `/p/supported-models/umt5-xxl-fp8-e4m3fn-scaled` redirects 301 to `umt5-xxl-fp16` (canonicalSlug) - [ ] Structured data validator shows FAQPage + SoftwareApplication + BreadcrumbList valid Fixes FE-421 --------- Co-authored-by: GitHub Action <action@github.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> |
||
|
Dante
|
e05e6cd2fb |
test: FE-230 e2e regression for asset-delete clearing Load Image preview (#12131)
## Summary Follow-up to #11493 — adds a `@cloud` Playwright spec that drives the asset-deletion flow end-to-end and asserts the four FE-230 outcomes wired up in `useMediaAssetActions.deleteAssets`. - Path: `browser_tests/tests/assetDeleteClearsLoadImage.spec.ts` - Flow: load `widgets/load_image_widget`, seed `widget.value` + `node.imgs`, reset changeTracker, open assets sidebar → Imported tab → right-click card → Delete → confirm dialog. - Asserts (auto-retrying): DELETE `/assets/:id` request issued, `widget.value === ''`, `node.imgs.length === 0`, `workflow.isModified === true`. - Tagged `@cloud` because input-asset deletion is gated on `isCloud` in `deleteAssetApi`. Addresses [this review thread](https://github.com/Comfy-Org/ComfyUI_frontend/pull/11493#pullrequestreview-4262129237) from #11493. ## Test plan - [ ] `pnpm build:cloud && pnpm exec playwright test --project=cloud browser_tests/tests/assetDeleteClearsLoadImage.spec.ts --reporter=list` passes locally with a running ComfyUI backend - [ ] CI `cloud` matrix passes ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12131-test-FE-230-e2e-regression-for-asset-delete-clearing-Load-Image-preview-35d6d73d36508192b4a2df7860f48c44) by [Unito](https://www.unito.io) |
||
|
pythongosssss
|
681915275e |
fix: remove failed to export toast when cancelling export workflow (#12134)
## Summary An incorrect error toast currently shows when cancelling the workflow export from an asset ## Changes - **What**: - skip toast on cancel - add e2e & unit tests - refactor asset tab open helper to wait by default & cleanup usage ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12134-fix-remove-failed-to-export-toast-when-cancelling-export-workflow-35d6d73d3650815b839ff26edd70a472) by [Unito](https://www.unito.io) |
||
|
pythongosssss
|
3d9c9ce327 |
fix: update color widget colors (#12133)
## Summary The widget was using a different fg/background compared to all other widgets, this updates to use the standard classes. ## Changes - **What**: - update styles ## Screenshots (if applicable) Before <img width="570" height="597" alt="Screen Shot 2026-05-11 at 07 19 12" src="https://github.com/user-attachments/assets/18a5330f-5e9a-4d16-b3f0-0acfab5d6f99" /> After <img width="570" height="597" alt="Screen Shot 2026-05-11 at 07 15 46" src="https://github.com/user-attachments/assets/81c9da58-fdda-4539-ae1e-98727f12b9ac" /> ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12133-fix-update-color-widget-colors-35d6d73d365081da8515cd48a8e8ecc2) by [Unito](https://www.unito.io) No e2e tests for this as they would be simple screenshot asserts on the color of nodes, which is not worthwhile. |
||
|
jaeone94
|
e765eb1bb2 |
fix: suppress missing media scan during uploads (#12111)
## Summary - Prevent missing media detection from scanning media loader nodes while a drag/drop, paste, or file-select upload is still in progress. - Align LoadAudio with the existing media upload lifecycle by setting `node.isUploading`, blocking concurrent uploads, and clearing the flag after upload completion. - Keep added-node model and missing-node scans on the original one-microtask path, while deferring added-node media scanning by one extra microtask so upload handlers can mark transient upload state before the scan reads widget values. ## Why Drag/drop and paste can create media loader nodes before the backing upload has settled. During that short window, the widget may contain a local filename that is not yet backend-resolvable, so missing media detection can surface a false missing asset. Refreshing works because the upload has completed by then. ## Follow-up - E2E coverage for this upload race will be handled in a follow-up PR together with E2E coverage for the annotated output-media path changes from #12069. ## Validation - `pnpm format` - `pnpm lint` - `pnpm typecheck` - `pnpm vitest run src/composables/graph/useErrorClearingHooks.test.ts src/platform/missingMedia/missingMediaScan.test.ts src/extensions/core/uploadAudio.test.ts src/composables/node/useNodeImageUpload.test.ts` - Re-ran `pnpm typecheck` after rebasing onto latest `main` - Pre-push `knip` hook passed Fixes FE-620 ## Screenshots Before https://github.com/user-attachments/assets/db7891de-a4b5-4cde-aa76-6340e6cdf7b2 After https://github.com/user-attachments/assets/9b99bb13-0d5b-4ff7-8f52-66eea6e417ec ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12111-fix-suppress-missing-media-scan-during-uploads-35b6d73d365081f3b54eed02874ccaa4) by [Unito](https://www.unito.io) |
||
|
Rizumu Ayaka
|
56434ae9ac |
perf: debounce template search input to keep typing responsive (#12183)
## Summary Route the templates dialog search through `FormSearchInput`'s debounced searcher so per-keystroke work no longer trips the heavy filter/render path. ## Changes - **What**: `WorkflowTemplateSelectorDialog` now uses `FormSearchInput` instead of `SearchInput`. The raw input is bound to a local ref; the actual `searchQuery` consumed by `useTemplateFiltering` is only written after the input debounce settles, so dependent computeds (notably `shouldUsePagination`, which used to flip on every keystroke and force a full grid rebuild) stay stable while typing. - **What**: `FormSearchInput` gains optional `debounceMs` (default `250`) and `debounceMaxWaitMs` (default `1000`) props. Existing callers are unchanged; the templates dialog passes `400` / `4000` to match the feel tuned in this PR. ## Review Focus - Reset path: `searchQuery` is still owned by `useTemplateFiltering` and cleared by `resetFilters`; a watch syncs the visible input back to empty when that happens. - `FormSearchInput` is currently under `src/renderer/...` but already imported by workbench-level components (rightSidePanel tabs). This PR follows that existing precedent rather than relocating the component. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12183-perf-debounce-template-search-input-to-keep-typing-responsive-35e6d73d365081b7a11ec4a84323095f) by [Unito](https://www.unito.io) |
||
|
Alexander Brown
|
25c2d828c0 |
test: enable vitest/consistent-each-for and migrate .each → .for (#12161)
*PR Created by the Glary-Bot Agent*
---
Enables the oxlint rule `vitest/consistent-each-for` (configured to
prefer `.for` for `test`, `it`, `describe`, and `suite`) and migrates
every `.each` parameterized test in the repo to `.for`. Using `.for`
avoids accidentally splatting tuple elements into separate callback
arguments and exposes `TestContext` as the second callback argument.
The first commit covers the 38 lint-detected files (88 callsites):
renames `.each` → `.for` and updates callback signatures to destructure
when the data is an array of tuples (objects/primitives already work
unchanged with `.for`).
The follow-up commit addresses code review feedback: oxlint's rule does
not recognize `test.each` on extended test bases
(`baseTest.extend(...)`) and skips files in `ignorePatterns`
(`src/extensions/core/*`). These were converted manually so the policy
is uniform across the codebase.
## Verification
- `node_modules/.bin/oxlint src` — 0 errors, 0 `consistent-each-for`
violations
- `pnpm typecheck` — passes
- `pnpm test:unit` — all modified test files pass; pre-existing
environmental flakes (`GraphView.test.ts`, `ColorWidget.test.ts`, etc.,
unchanged here and flaky on `main` in this sandbox) are unrelated
- `pnpm lint` / `pnpm knip` — clean
- Manual verification: 362 tests across 6 representative converted
suites re-run in an interactive shell — all passing
Manual UI verification (Playwright/screenshots) is not applicable:
changes are test-file-only refactors with no production runtime or UI
behavior change.
## Notes on `.for` semantics
- Array-of-tuples (`[[a, b], ...]`) passes the tuple as a single arg, so
callbacks were changed from `(a, b) => …` to `([a, b]) => …`.
- Array-of-objects (`[{a}, …]`) already used destructuring — unchanged.
- Array-of-primitives (`['a', …]`) — callback signature unchanged.
- A handful of complex cases use a small `type Case = [...]` alias plus
`it.for<Case>([...])` to preserve tuple inference where TS narrowed
unions otherwise broke parameter types.
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12161-test-enable-vitest-consistent-each-for-and-migrate-each-for-35e6d73d3650810c9417e07bdd9f27a2)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
|
||
|
Dante
|
ceb9936058 |
fix(i18n): clamp unsupported browser locales to a shipped tag (#11712)
## Summary
Sidebar buttons rendered literal i18n keys (e.g.
`sideToolbar.labels.assets`) on a fresh install when the user's
`navigator.language` base tag wasn't one of the 12 shipped locales —
German/Italian/Polish/Dutch/Brazilian-Portuguese users among others.
## Changes
- **What**: Add `resolveSupportedLocale()` that tries the full BCP-47
tag first (preserves `zh-TW`, `pt-BR`), then the base tag, then `'en'`.
Wire through both entry points (`createI18n`'s initial locale,
`Comfy.Locale`'s `defaultValue`) and clamp inside `loadLocale`,
propagating the resolved tag to `GraphView` so a stale stored
`Comfy.Locale='de'` from older builds also recovers.
- **Side benefit**: Brazilian Portuguese users were previously falling
through `pt-BR` → `pt` (unshipped) → broken. The full-tag-first lookup
now correctly lands them on the `pt-BR` bundle.
- **Breaking**: None.
- **Dependencies**: None.
## Root Cause
Three-link chain:
1. `Comfy.Locale`'s default was `() => navigator.language.split('-')[0]
|| 'en'`. German → `'de'` (unshipped).
2. `loadLocale('de')` silently `console.warn`'d and returned without
throwing.
3. `GraphView` then ran `i18n.global.locale.value = 'de'` anyway.
4. `st(key, fallback) = te(key) ? t(key) : fallback`. vue-i18n's `te()`
checks **only** the current locale and ignores `fallbackLocale` — every
key missed → `st()` returned the literal key string.
Two pathways reached the broken state (defaultValue path, and
unset-setting path through `createI18n`'s own `navigator.language`
snapshot); the new helper closes both.
## Review Focus
- `loadLocale` now returns `SupportedLocale` (was `void`). Old `void`
callers continue to compile; the only change is `GraphView` consuming
the return value.
- Unit-tested in `src/i18n.test.ts` (added `resolveSupportedLocale`
block + updated the `loadLocale` unsupported-locale case from "warn" to
"clamp to en").
- Self-reproduced via Playwright with `navigator.language='de-DE'` +
fresh-install state on both `main` (shows the bug) and this branch
(shows the fix). Spec saved at
`temp/scripts/issue-10563-locale-bug.spec.ts`.
Fixes #10563
FE-480 — https://linear.app/comfyorg/issue/FE-480
## Screenshots
**Before** (from #10563, on `main`):
<img width="258" height="399" alt="Sidebar with literal i18n keys"
src="https://github.com/user-attachments/assets/098d1d76-8e89-4237-813f-5f030b34e51e"
/>
**After** (this branch, same `navigator.language='de-DE'`):
<img width="367" height="793" alt="Screenshot 2026-04-28 at 2 07 38 PM"
src="https://github.com/user-attachments/assets/9d279de3-50a8-4774-999f-ab4c3018a9ef"
/>
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-11712-fix-i18n-clamp-unsupported-browser-locales-to-a-shipped-tag-3506d73d3650812f89d2f0fe3199de3a)
by [Unito](https://www.unito.io)
|
||
|
Deep Mehta
|
bb420fe2c7 |
feat: add model-to-node mappings for new model directories (#12151)
## Summary Add entries to `MODEL_NODE_MAPPINGS` so the model browser's "Use" button correctly creates a loader node for five model directories that currently have no mapping. ## Changes - **What**: 5 new entries in `src/platform/assets/mappings/modelNodeMappings.ts`: - `background_removal` → `LoadBackgroundRemovalModel` / `bg_removal_name` (ComfyUI v0.21+ core) - `frame_interpolation` → `FrameInterpolationModelLoader` / `model_name` (ComfyUI v0.21+ core) - `film` → `FILM VFI` / `ckpt_name` (ComfyUI-Frame-Interpolation) - `ultralytics/bbox` → `UltralyticsDetectorProvider` / `model_name` (ComfyUI-Impact-Pack) - `ultralytics/segm` → `UltralyticsDetectorProvider` / `model_name` (ComfyUI-Impact-Pack) - **Breaking**: none ## Review Focus - Node class names and input keys were cross-checked against the ComfyUI v0.21.0 source and the published Impact-Pack / Frame-Interpolation node definitions - Both `ultralytics/bbox` and `ultralytics/segm` map to the same node (`UltralyticsDetectorProvider`); its `model_name` combo accepts values from both subdirectories (`bbox/...` and `segm/...`) - `film` and `frame_interpolation` are separate directories serving different node packs — keeping them as distinct entries rather than collapsing under a parent ## Test plan - [ ] In the model browser, clicking "Use" on `birefnet.safetensors` creates a `LoadBackgroundRemovalModel` node with the model preselected - [ ] Same for one model in each of: `frame_interpolation`, `film`, `ultralytics/bbox`, `ultralytics/segm` ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12151-feat-add-model-to-node-mappings-for-new-model-directories-35d6d73d365081ff834bf6eb610da160) by [Unito](https://www.unito.io) |
||
|
pythongosssss
|
4504256f11 |
test: add test for custom node i18n (#12132)
## Summary Adds e2e test for custom node i18n ## Changes - **What**: - add e2e regression for previous fix with no e2e ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12132-test-add-test-for-custom-node-i18n-35d6d73d365081f7bed2db39af38f855) by [Unito](https://www.unito.io) |
||
|
Comfy Org PR Bot
|
1290bbd359 |
1.45.5 (#12152)
Patch version increment to 1.45.5 **Base branch:** `main` ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12152-1-45-5-35e6d73d365081c39aecddf42d8f7a2a) by [Unito](https://www.unito.io) --------- Co-authored-by: christian-byrne <72887196+christian-byrne@users.noreply.github.com> Co-authored-by: github-actions <github-actions@github.com>v1.45.5 |
||
|
Yourz
|
7ddf71d91b |
fix(website): center GitHubStarBadge text in Safari (#12138)
*PR Created by the Glary-Bot Agent* --- ## Summary The 10px star count text inside the desktop nav GitHub star badge rendered vertically off-center in Safari/WebKit. The text was visibly shifted upward inside the yellow badge body, while Chromium rendered it centered correctly. ## Root cause `NodeBadge.vue` centers its inner text span by setting `flex items-center justify-center` on the segment, then nudging the text with `translate-y-1` (or `translate-y-0` for the small variant). The text span itself is an `inline-block` with no explicit `line-height`, so it inherits the default `line-height: 1.5` (15px for a 10px font). Safari and Chromium distribute that extra leading differently for the `PP Formula Narrow` custom font: Safari pushes the glyph higher inside its 15px line box, while Chromium positions it near the middle. With only a 5px gap above/below the glyph to play with, that browser-specific divergence is enough to make the badge look misaligned in Safari. ## Fix Add `leading-none` to the star count text class so the inline-block's line box equals the font size (10px) and the parent flex container's `items-center` produces deterministic vertical centering across browsers. Verified at lg breakpoint (1440×900) in both WebKit and Chromium via Playwright; the badge now renders identically and is properly centered. ## Verification - `pnpm typecheck` (website) — clean - `pnpm build` (website) — 51 pages built successfully - Pre-commit hooks (stylelint, oxfmt, oxlint, eslint, typecheck, typecheck:website) — all passed - Visual inspection in WebKit and Chromium via Playwright Fixes FE-648 ## Screenshots    ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12138-fix-website-center-GitHubStarBadge-text-in-Safari-35d6d73d3650818aa0e8e0f341b60378) by [Unito](https://www.unito.io) --------- Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com> |
||
|
Christian Byrne
|
74caeb0b0b |
fix: detect V1/V2 draft storage keys in new-user check (#11728)
## Problem `checkIsNewUser()` in `useNewUserService` only checked legacy pre-V1 localStorage keys (`workflow`, `Comfy.PreviousWorkflow`) to determine if a user had prior workflow history. A returning user who had only ever used the V1 or V2 draft persistence system would have neither of those keys set, causing `isNewUser()` to return `true` and the getting-started tab to appear in the workflow templates dialog after a settings reset. ## Solution Extend the check to also cover: - **V1 draft store keys**: `Comfy.Workflow.Drafts`, `Comfy.Workflow.DraftOrder` - **V2 draft index key**: `Comfy.Workflow.DraftIndex.v2:personal` The `personal` scope is hardcoded for the V2 check because at the time `checkIsNewUser()` runs, the cloud workspace ID (stored in sessionStorage) may not be set yet. This is fine — any genuine new user will have no personal workspace index regardless. The original legacy keys are preserved for users who may still have them from older installs. ## Tests Added three new test cases covering V1 draft store keys, V1 draft order key, and V2 draft index key. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11728-fix-detect-V1-V2-draft-storage-keys-in-new-user-check-3506d73d3650819ca4cfc8e83d95c258) by [Unito](https://www.unito.io) --------- Co-authored-by: Connor Byrne <c.byrne@comfy.org> |
||
|
Alexander Brown
|
ced7c93e63 |
testing: Improve custom checks in .coderabbit.yaml (#12141)
Update coderabbit end-to-end check logic. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12141-testing-Improve-custom-checks-in-coderabbit-yaml-35d6d73d3650818e8be2f0b7d403683b) by [Unito](https://www.unito.io) --------- Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> |
||
|
Christian Byrne
|
759ed3d4e2 |
feat(website): add community-workflows demo page (#11942)
*PR Created by the Glary-Bot Agent* --- Adds a new interactive demo page at `comfy.org/demos/community-workflows` for the [Explore and Use a Community Workflow from the Hub](https://app.arcade.software/flows/mqZh17oWDuWIyhK0xwEV/view) Arcade walkthrough. Built on top of the demo infrastructure merged in #11436. ## Changes - `apps/website/src/config/demos.ts` — register the new demo - `apps/website/src/i18n/translations.ts` — add en + zh-CN strings (title, description, transcript) - `apps/website/public/images/demos/community-workflows-og.png` — 1200×630 OG image so email/social previews render correctly - `apps/website/public/images/demos/community-workflows-thumb.webp` — 1280×720 WebP thumbnail - `apps/website/e2e/demos.spec.ts` — refactored to iterate `demos` from config so every demo (current + future) is exercised in both en and zh-CN, and the iframe `src` is asserted to contain the correct Arcade ID Adding a new demo only requires editing `demos.ts` + `translations.ts` going forward; the e2e refactor is a one-time generalization that gives future demos coverage automatically. ## Verification - `pnpm typecheck:website`: 0 errors, 0 warnings, 0 hints - Pre-commit hook ran `pnpm typecheck`, `oxfmt`, `oxlint`, `eslint` — all clean on staged files - `npx astro build`: 53 pages built; `/demos/community-workflows/` and `/zh-CN/demos/community-workflows/` generated and present in `sitemap-0.xml` - Page rendered in Playwright preview: hero (title, GETTING STARTED, BEGINNER, ~2 min), Arcade embed loads, transcript section present, "What's Next" links to `image-to-video` - zh-CN page shows localized title (探索并使用社区工作流), description, badges, and "What's Next" heading - OG meta tag references the new 1200×630 PNG ## Screenshots   ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11942-feat-website-add-community-workflows-demo-page-3576d73d36508139b647c774b1d39323) by [Unito](https://www.unito.io) --------- Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com> |
||
|
pythongosssss
|
5d53e75d23 |
test: add tests for deprecated & api node badge visibility settings (#11681)
## Summary Adds coverage for untested node badges ## Changes - **What**: - add shared addNode helper - will follow up to standardize across tests - add deprecated & api node badge tests in LiteGraph & Vue nodes ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11681-test-add-tests-for-deprecated-api-node-badge-visibility-settings-34f6d73d365081569129ecffa608122e) by [Unito](https://www.unito.io) --------- Co-authored-by: github-actions <github-actions@github.com> |
||
|
Comfy Org PR Bot
|
d23e86d9a4 |
[chore] Update Ingest API types from cloud@0a03f3a (#12043)
## Automated Ingest API Type Update This PR updates the Ingest API TypeScript types and Zod schemas from the latest cloud OpenAPI specification. - Cloud commit: 0a03f3a - Generated using @hey-api/openapi-ts with Zod plugin These types cover cloud-only endpoints (workspaces, billing, secrets, assets, tasks, etc.). Overlapping endpoints shared with the local ComfyUI Python backend are excluded. --------- Co-authored-by: MillerMedia <7741082+MillerMedia@users.noreply.github.com> Co-authored-by: GitHub Action <action@github.com> |
||
|
Yourz
|
d901c63a0b |
feat: convert careers CategoryNav to scroll-spy locator (#12110)
*PR Created by the Glary-Bot Agent*
---
Converts the `CategoryNav` in the careers `RolesSection` from a
click-to-filter button into a scroll-spy section locator, matching the
pattern already used by `ContentSection.vue` (customer story details,
TOS, privacy policy).
## Changes
- **`apps/website/src/components/careers/RolesSection.vue`**
- Replaced category-based filtering with anchor navigation: clicking a
department in the sidebar smooth-scrolls (via existing Lenis/GSAP
`scrollTo` helper) to that department's section with a `-144px` header
offset.
- Removed the `ALL` button — every department is always rendered as its
own scroll target with `id="careers-dept-{key}"`.
- Added `useIntersectionObserver` (rootMargin `-20% 0px -60% 0px`) that
updates the active nav item as the user scrolls. An `isScrolling` guard
prevents the observer from fighting click-jumps mid-animation.
- Added a viewport-entry fade/slide-up animation on each department
section, gated by `motion-safe:` so users with `prefers-reduced-motion`
see content immediately. The reveal state is sticky (one-way) so
sections don't disappear once revealed.
- Active state is driven by raw department keys; both the nav model and
the observer's id-to-key mapping use a single consistent identifier.
- **`apps/website/e2e/careers.spec.ts`**
- Replaced the obsolete "ENGINEERING filter narrows the list" test with
one that validates locator behavior: clicking the department button
scrolls the section into the viewport, sets `aria-pressed="true"`, and
keeps the full role list rendered.
## Verification
- `pnpm --filter @comfyorg/website typecheck` — clean.
- `pnpm exec oxfmt` / `pnpm exec eslint` / `pnpm exec oxlint` — clean.
- Pre-commit lint-staged hooks (stylelint, oxfmt, oxlint, eslint,
typecheck) — passing.
- Manual smoke test via Playwright on `astro dev`: careers page renders
all departments stacked vertically, active department in the sidebar
highlights based on viewport position (DESIGN active on initial scroll),
nav items reflect each department instead of including an `ALL` button.
## Screenshots
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12110-feat-convert-careers-CategoryNav-to-scroll-spy-locator-35b6d73d3650818a9226e5dcb1244756)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Glary <glary@bot.local>
|
||
|
Yourz
|
5ca9f3e7e6 |
feat(website): remove left-edge fade-out from local hero illustration (#12137)
*PR Created by the Glary-Bot Agent* --- Removes the left-edge fade-out gradient overlay on the hexagonal hero illustration on the `/download` (local) page. The hex cluster now reads fully edge-to-edge instead of being blended into the page background on the left side. Tracked in [FE-650: Remove fade-out effect on local page hero illustration](https://linear.app/comfyorg/issue/FE-650/remove-fade-out-effect-on-local-page-hero-illustration). ## Change Drops the `<!-- Left-edge fade -->` `<rect>` and its `<defs><linearGradient id="localHeroFadeLeft">…</linearGradient></defs>` from `apps/website/src/components/product/local/HeroSection.vue`. Animation logic (panel expansion + hex ring rotation) is untouched. ## Verification - `pnpm nx typecheck website` — pass (0 errors) - `pnpm nx build website` — pass (51 pages built) - `pnpm exec eslint apps/website/src/components/product/local/HeroSection.vue` — clean - `pnpm format:check apps/website/src/components/product/local/HeroSection.vue` — clean - Manual: `pnpm dev` + visited `/download` at 390×844 (mobile), 1280×800, and 1440×900. Mobile screenshots clearly show the fade is gone; the leftmost hexagons are now fully visible. ## Screenshots Mobile (390×844), before — note the dark fade obscuring the left side of the hex cluster: ## Screenshots    ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12137-feat-website-remove-left-edge-fade-out-from-local-hero-illustration-35d6d73d365081cba690ed7d46a19882) by [Unito](https://www.unito.io) Co-authored-by: Glary Bot <bot@glary.dev> |
||
|
Yourz
|
6d5fa743b3 |
fix: seamless SocialProofBar marquee loop (#12139)
*PR Created by the Glary-Bot Agent* --- ## Summary The partner-logo marquee on the homepage `SocialProofBar` glitched on every loop restart — a visible jump where the strip snapped back to the start. ## Root cause The previous implementation rendered all logos as siblings of a single flex container and animated the track from `translateX(0)` to `translateX(-50%)`. Because the `gap` utility inserts spacing between every adjacent pair of items (including the seam between the two duplicated halves), `-50%` of the total width does not equal the distance from one half-start to the next half-start. The mismatch (`gap / 2`) is exactly what the eye sees as a jump. ## Fix - Wrap each duplicated half in its own flex group. - Place the two groups as siblings of an outer `flex w-max gap-X` track with a gap that matches the inner gap. - Animate each group by `translateX(calc(-100% - var(--marquee-gap)))`, where `--marquee-gap` is set inline to the same value as the Tailwind gap class. - Scope the `animation` declaration to `@media (prefers-reduced-motion: no-preference)` so reduced-motion users get a stable, non-animated client list instead of the global "snap to 0.01ms" jump. At `t = end`, the second group sits at `x = 0` — exactly where the first group started — so the next animation cycle is visually indistinguishable from the previous frame. The duplicate carries `aria-hidden="true"` so screen readers don't read the client list twice. ## Verification - `pnpm typecheck`, `pnpm format`, `npx eslint` on changed files: clean. - Geometry verified at runtime on desktop (1440×900) and mobile (390×844): copy widths match, second copy lands at `x = 0` at animation end. - New Playwright regression tests (`apps/website/e2e/responsive.spec.ts`) pause the CSS animation, sample bounding rects at `t=0` and `t≈duration`, and assert the seam invariant — covering desktop forward, mobile forward, and mobile reverse marquees. All 5 SocialProofBar tests pass on both `desktop` and `mobile` projects. - Reduced-motion behavior verified in the browser: `animationName: none`, `transform: none`, tracks at their natural positions. Fixes FE-649 ## Screenshots     ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12139-fix-seamless-SocialProofBar-marquee-loop-35d6d73d36508141b6ccf0167016b8c8) by [Unito](https://www.unito.io) Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com> |
||
|
Henry Lee
|
603dd3eb4e |
fix(assets): recognise m4v and mkv as video extensions (#12088)
## Summary `.m4v` and `.mkv` files render as a generic file icon in the assets sidebar instead of a video preview because `VIDEO_EXTENSIONS` doesn't list them, even though both formats are widely produced by ComfyUI custom nodes and are browser-playable when written with common codecs. ## Changes - **What**: Add `m4v` and `mkv` to `VIDEO_EXTENSIONS` in `packages/shared-frontend-utils/src/formatUtil.ts` and extend the existing test cases. Aligns with ComfyUI core's canonical video extension list (`tests-unit/folder_paths_test/filter_by_content_types_test.py:13`), which includes both. The frontend's format registry at `src/platform/workflow/core/types/formats.ts` also lists `.m4v` with mime `video/x-m4v` — `formatUtil.ts` was the inconsistent surface. - **Breaking**: None. - **Dependencies**: None. ## Review Focus `m4v` is Apple's MP4 container variant; `mkv` is the Matroska container. ComfyUI custom nodes most commonly produce both with H.264/VP9 codecs, which play natively in modern browsers via `<video>`. Adding the extensions routes those files through the existing `MediaVideoTop` component without any new rendering logic. If a user's `.mkv` happens to use an exotic codec the browser can't play (e.g. H.265/HEVC in Chrome), they get the same controllable failure mode as today's `.avi` entries — a `<video>` element with the browser's native unsupported-source UI. That is no worse than the current "show a generic file icon" behavior, and strictly better in the common case. ## Screenshots (if applicable) > **Note**: Screenshots taken from the OSS *input* assets sidebar with [#12086](https://github.com/Comfy-Org/ComfyUI_frontend/pull/12086) also applied locally. That PR fixes a separate regression where OSS input filenames carry a `[input]` annotation suffix that breaks all extension-based media detection — without it, `m4v`/`mkv` files (and every other file in that sidebar) still render as the generic icon. This PR alone is sufficient for cloud assets and OSS output history; the input-sidebar previews require both PRs. Before: <img width="1197" height="714" alt="2026-05-09-031123_hyprshot" src="https://github.com/user-attachments/assets/5c6ebc2d-aac2-411f-a2e4-51a111033184" /> After: <img width="1042" height="723" alt="2026-05-09-031005_hyprshot" src="https://github.com/user-attachments/assets/f0acc2cf-8571-4fd0-b0cd-2b8b87ff9b74" /> |
||
|
Yourz
|
d767a325a2 |
FE-604: fix(website): activate last section badge when scrolled to bottom (#12057)
*PR Created by the Glary-Bot Agent* --- ## Summary Fixes the bug where the last badge in `ContentSection`'s sticky sidebar nav stays unhighlighted when the user scrolls to the very bottom of the page on tall viewports (reported on a 14" MacBook M4 Pro at 3024×1964 / 2016×1310 logical, both Chrome and Safari). ## Root cause The scroll-spy uses an IntersectionObserver with `rootMargin: '-20% 0px -60% 0px'`, which makes only a 20%–40% horizontal band from the viewport top "active". When multiple intersecting entries are reported, the callback picks the one whose `boundingClientRect.top` is smallest (highest up on screen). On tall viewports, when the page is scrolled to the absolute bottom, the last *and* the second-to-last sections frequently both sit inside that 20%–40% band at the same time. The "smallest top" tiebreak then selects the second-to-last section, leaving the last badge inactive even though the user has reached the end of the page. ## Fix `apps/website/src/components/common/ContentSection.vue`: 1. Add `isAtBottom()` — true when the viewport bottom has reached the document bottom (within 4px to absorb sub-pixel rounding). 2. The IntersectionObserver callback bails out when `isAtBottom()` so it cannot overwrite the choice below. 3. A passive `scroll` listener (and a one-shot `onMounted` call) sets `activeSection` to the last section whenever the page is at the bottom — including when the component mounts already at the bottom (e.g. hash navigation to a trailing anchor, restored scroll position, or a page shorter than the viewport). 4. Both the scroll handler and the IO callback honor the existing `isScrolling` flag, so click-driven smooth scroll-to-section behavior is unchanged. ## Verification Reproduced the bug at viewport 2016×1310 (14" M4 Pro "More Space" mode) on `/privacy-policy`: - Before fix: at absolute bottom, IntersectionObserver picks `australian-privacy` (second-to-last) — bug confirmed via DOM inspection that showed multiple sections intersecting the active band, with the second-to-last winning the "smallest top" tiebreak. - After fix: - Scrolled to bottom → last badge `CONTACT` is active. - Scrolled to top → first badge `INTRO` is active. - Scrolled mid-page → correct mid-section is active. - Click on a badge → smooth scrolls and that badge becomes active. - Initial render at bottom (loaded `/privacy-policy#contact`, browser scrolls to the bottom on mount) → `CONTACT` active immediately. `pnpm typecheck` and `pnpm typecheck:website` pass; `pnpm lint` reports 0 errors; existing website unit tests pass. Note: The website app currently has no Vue component test setup (`vitest.config.ts` is configured for `node` env, no DOM). Adding component tests for this scroll-spy interaction would require setting up `happy-dom` and `@testing-library/vue` for the website app, which is out of scope for this bug fix. Fixes FE-604 ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12057-FE-604-fix-website-activate-last-section-badge-when-scrolled-to-bottom-3596d73d365081faa243f4dd8e6ee54a) by [Unito](https://www.unito.io) --------- Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com> |
||
|
Christian Byrne
|
39b2bb5eab |
[chore] Update Comfy Registry API types from comfy-api@dfcca37 (#12087)
## Automated API Type Update This PR updates the Comfy Registry API types from the latest comfy-api OpenAPI specification. - API commit: dfcca37 - Generated on: 2026-05-07T19:11:35Z These types are automatically generated using openapi-typescript. ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12087-chore-Update-Comfy-Registry-API-types-from-comfy-api-dfcca37-35a6d73d365081f49beec09f418d7c1a) by [Unito](https://www.unito.io) Co-authored-by: coderfromthenorth93 <213232275+coderfromthenorth93@users.noreply.github.com> |
||
|
pythongosssss
|
c643438601 |
fix: hide image buttons if load failed (#12136)
## Summary When an image fails to load in the image preview, the context buttons are still visible - clicking these does not work (Mask editor opens and closes, download does nothing) - this hides the buttons if load fails. ## Changes - **What**: - hide buttons if load failed - tests ## Review Focus <!-- Critical design decisions or edge cases that need attention --> <!-- If this PR fixes an issue, uncomment and update the line below --> <!-- Fixes #ISSUE_NUMBER --> ## Screenshots (if applicable) Current: <img width="622" height="857" alt="image" src="https://github.com/user-attachments/assets/26e391a0-5538-4c6c-ac8a-b6f2b6acabae" /> ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12136-fix-hide-image-buttons-if-load-failed-35d6d73d365081579c71f1849b9ab1bd) by [Unito](https://www.unito.io) |
||
|
Dante
|
02e1ba2968 |
fix: Load Image preview retains deleted asset (FE-230) (#11493)
## Summary After deleting an asset, the Load Image node kept displaying the deleted thumbnail — both in the node body and in the picker dropdown (All / Imported / Generated tabs), even after a workflow reload. - Fixes FE-230 - Source: Slack https://comfy-organization.slack.com/archives/C0A4XMHANP3/p1776715727656809 ## Root Cause Three distinct paths kept the deleted asset visible: 1. **Node-body preview cache** — `useMediaAssetActions.deleteAssets` never cleared `node.imgs` / `node.videoContainer` / the `nodeOutputStore` Vue ref, so the canvas renderer kept its cached frame. 2. **Live-delete dropdown gap** — the picker reads from `outputMediaAssets.media` (the asset list) and from `missingMediaStore.missingMediaCandidates` (verified-missing names). On live delete, neither was updated for the deleted asset, so the dropdown filter had nothing to drop. 3. **Synthetic "selected" placeholder** — `useWidgetSelectItems.missingValueItem` rebuilt any orphaned `modelValue` as a fake item with a `/api/view?filename=...` preview URL. Browsers had cached that URL pre-delete, so the deleted thumbnail still rendered with a blue checkmark even after the filter dropped the real asset entry. A subtler issue compounded #2/#3: candidate names stored in `missingMediaStore` are raw widget values (e.g. `sub/foo.png [output]`), but the dropdown computed comparison keys differently per source (asset list uses bare `asset.name`, widget option list uses bare filename). Names with a subfolder prefix slipped through the filter. ## Fix - **`clearNodePreviewCacheForFilenames`** (existing helper, refactored): exports `findNodesReferencingFilenames` + `extractFilenameFromWidgetValue`. Uses `nodeOutputStore.removeNodeOutputs` so the **reactive** Pinia ref updates, not just the legacy `app.nodeOutputs` mirror. Also clears `node.videoContainer` for Load Video. - **`markDeletedAssetsAsMissingMedia`** (new): on successful deletion, surfaces the affected widgets through `missingMediaStore` immediately so the dropdown filter has something to drop without waiting for verification. - **`useMissingMediaPreviewSync`** (new): watches `missingMediaStore` and clears `node.imgs` / `node.videoContainer` / Vue preview source for nodes referencing confirmed-missing media on workflow load — covers the post-reload case. - **`useWidgetSelectItems`**: normalizes both sides of the missing-media filter via `extractFilenameFromWidgetValue` (strips `[input|output|temp]` annotation + subfolder prefix), and suppresses `missingValueItem` when the value is in the missing-media store so the cached-thumbnail "selected" placeholder doesn't appear. ## Red-Green Verification | Commit | CI Status | Run | |--------|-----------|-----| | `test: FE-230 add failing test for Load Image preview cache clearing` | 🔴 Failure — test caught the bug | https://github.com/Comfy-Org/ComfyUI_frontend/actions/runs/24700188700 | | `fix: FE-230 clear Load Image preview cache when asset is deleted` | 🟢 Success | https://github.com/Comfy-Org/ComfyUI_frontend/actions/runs/24700265884 | ## Test Plan - [x] Unit coverage: 78 tests across 5 files (preview-cache helper, mark-deleted-as-missing, missing-media-preview-sync, widget-select-items missing-media filter incl. subfolder-prefix case, useMediaAssetActions integration) - [x] Live delete: Load Image node preview clears, dropdown drops the asset across All / Imported / Generated, no synthetic "selected" placeholder - [x] Post-reload: missing-media verification → `useMissingMediaPreviewSync` clears the preview, dropdown drops the asset - [x] Linear FE-230 auto-links via the Source line ## Scope note In-session and session-restore are both covered. If the backend/CDN continues serving the deleted `filename`/`asset_hash` after deletion, a cross-session reopen may still render stale bytes from cache — that's a backend/CDN concern tracked separately. ## demo ### before https://github.com/user-attachments/assets/e4d3a40e-0d46-43ad-985c-22ce7e0d3faf ### after https://github.com/user-attachments/assets/fcac9387-4c07-4be2-bcdd-d1a6192fe962 |
||
|
pythongosssss
|
15b8771cc2 |
fix: clear active job on reconnect if no longer in queue (#12067)
## Summary When a socket disconnects messages can be missed and lead to a stale UI state, this updates the state on reconnect and clears the active job if it is no longer running ## Changes - **What**: - add call to update queue on reconnect - clear active job if job not in queue response - tests ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12067-fix-clear-active-job-on-reconnect-if-no-longer-in-queue-3596d73d365081f79d42d73966420c50) by [Unito](https://www.unito.io) |
||
|
Comfy Org PR Bot
|
e68d50e677 |
1.45.4 (#12118)
Patch version increment to 1.45.4 **Base branch:** `main` ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12118-1-45-4-35d6d73d365081fcb5f5d06dec17bb59) by [Unito](https://www.unito.io) --------- Co-authored-by: christian-byrne <72887196+christian-byrne@users.noreply.github.com> Co-authored-by: github-actions <github-actions@github.com>v1.45.4 |
||
|
Comfy Org PR Bot
|
48b5e0165a |
1.45.3 (#12113)
Patch version increment to 1.45.3 **Base branch:** `main` ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12113-1-45-3-35c6d73d365081468180cefef02dca03) by [Unito](https://www.unito.io) --------- Co-authored-by: christian-byrne <72887196+christian-byrne@users.noreply.github.com> Co-authored-by: github-actions <github-actions@github.com>v1.45.3 |
||
|
Christian Byrne
|
fe1de3b254 |
refactor: remove dedup complexity from reportInactiveTrackerCall (#11833)
## Summary Remove the module-level `reportedInactiveCalls: Set<string>` and the early-return dedup check from `reportInactiveTrackerCall()` in `src/scripts/changeTracker.ts`. Every invocation now emits `console.warn` and (on Desktop) `Sentry.captureMessage` unconditionally. ## Why The dedup was added in #11328 but is unnecessary: - Every first-party call site already goes through the `activeWorkflow?.changeTracker` guard, so flooding from in-repo code is unlikely. - Repeated identical alerts may actually provide more diagnostic signal than the first-only approach suppresses. ## Changes - Drop `reportedInactiveCalls` Set - Drop the per-`(method, workflowPath)` early-return - Trim the JSDoc accordingly No behavior change for callers (`deactivate`, `captureCanvasState`); only the reporting frequency increases. ## Verification - `pnpm test:unit -- src/scripts/changeTracker.test.ts` — 16/16 passing - `pnpm typecheck` — clean - ESLint / oxfmt — clean - Fixes #11372 ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11833-refactor-remove-dedup-complexity-from-reportInactiveTrackerCall-3546d73d365081fabf57cbf1fa17051f) by [Unito](https://www.unito.io) |
||
|
Christian Byrne
|
1c2ae70343 |
chore(#11843): replace bare string NodeId typings in parameters tab components (#12014)
## Summary Replace `nodeId: string` with canonical `NodeId` type in right-side panel parameters tab components, eliminating redundant `String()` conversions at call sites. ## Changes - `TabNodes.vue`: `isSectionCollapsed` and `setSectionCollapsed` now accept `NodeId` instead of `string`; callers updated to pass `node.id` directly (removing `String()` wrapping) - `TabNormalInputs.vue`: same pattern ## Notes The other 6 files listed in the issue use `nodeId` parameters that carry execution IDs (`NodeExecutionId = string`), not graph node IDs (`NodeId = number | string`). Changing those to `NodeId` would be semantically incorrect. The two files changed here are the clear-cut cases where `node.id` (a `NodeId`) was being unnecessarily stringified before being passed. ## Testing ### Automated - `pnpm typecheck` — passes - `pnpm lint` — passes (0 warnings, 0 errors) - `pnpm format:check` — passes ### E2E Verification Steps 1. Open ComfyUI frontend 2. Load a workflow with multiple nodes 3. Open the right side panel (Parameters tab) 4. Verify node sections collapse/expand correctly per node 5. Verify "Collapse All" / "Expand All" toggle works correctly 6. Repeat with both TabNodes and TabNormalInputs views Fixes #11843 ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12014-chore-11843-replace-bare-string-NodeId-typings-in-parameters-tab-components-3586d73d365081ed84caf560277f0553) by [Unito](https://www.unito.io) |
||
|
jaeone94
|
8f68be5699 |
fix: handle annotated output media paths in missing media scan (#12069)
## Summary
This PR fixes missing-media false positives for annotated media widget
values such as:
```txt
photo.png [output]
clip.mp4 [input]
147257c95a3e957e0deee73a077cfec89da2d906dd086ca70a2b0c897a9591d6e.png [output]
clip.mp4[input] // Cloud compact form
```
The change is intentionally scoped to the missing-media detection
pipeline for:
- `LoadImage`
- `LoadImageMask`
- `LoadVideo`
- `LoadAudio`
It preserves the raw widget value on `MissingMediaCandidate.name` for UI
display, grouping, replacement, and user-facing missing-media rows.
Normalized values are used only as comparison keys during verification.
## Diff Size
`main...HEAD` line diff is currently:
- Production/runtime code: `+478 / -37` (`515` changed lines)
- Unit test code: `+960 / -47` (`1,007` changed lines)
- Total: `+1,438 / -84` (`1,522` changed lines)
The PR looks large mostly because it locks both Cloud and OSS/Core
runtime paths with unit coverage; the production/runtime change is about
one third of the total diff.
## What Changed
- Added missing-media-scoped annotation helpers for detection-only path
normalization.
- Core/OSS recognizes spaced suffixes like `file.png [output]`.
- Cloud also recognizes compact suffixes like `file.png[output]`.
- User-selectable trailing `input` and `output` annotations are
normalized for matching.
- Unknown annotations and middle-of-filename annotations are left
unchanged.
- Added shared file-path helpers in `formatUtil`:
- `joinFilePath(subfolder, filename)`
- `getFilePathSeparatorVariants(filepath)`
- Updated media verification to compare candidates against both raw and
normalized match keys.
- Kept input candidates and generated output candidates in separate
identifier sets so an input asset cannot accidentally satisfy an output
reference with the same name.
- Moved missing-media source loading into `missingMediaAssetResolver` so
`missingMediaScan` remains focused on scan/verification orchestration.
- Updated Cloud generated-media verification to use the Cloud assets API
instead of job history:
- Cloud input candidates use input/public assets.
- Cloud output candidates use `output` tagged assets.
- Kept OSS/Core generated-media verification history-based, matching the
current generated-picker/widget availability model.
## Runtime Verification Paths
### Cloud
Cloud stores generated outputs as asset records. For an annotated output
value, this PR verifies against the `output` asset tag rather than job
history.
```txt
Widget value
"147257...d6e.png [output]"
|
v
Detection keys
"147257...d6e.png [output]"
"147257...d6e.png"
|
v
Cloud asset sources
input candidates -> /api/assets?include_tags=input&include_public=true
output candidates -> /api/assets?include_tags=output&include_public=true
|
v
Match against
asset.name
asset.asset_hash
subfolder/asset.name
subfolder/asset.asset_hash
slash and backslash separator variants
```
Example:
```ts
candidate.name = 'abc123.png [output]'
asset.name = 'ComfyUI_00001_.png'
asset.asset_hash = 'abc123.png'
asset.tags = ['output']
// Result: not missing
```
### OSS / Core
Core widget options for the normal loader nodes are input-folder based.
Annotated output values are resolved by Core through
`folder_paths.get_annotated_filepath()`, but the current generated
picker path is history-backed. This PR keeps OSS generated verification
aligned with that widget availability model instead of treating the full
output folder as the source of truth.
```txt
Widget value
"subfolder/photo.png [output]"
|
v
Detection keys
"subfolder/photo.png [output]"
"subfolder/photo.png"
|
v
OSS generated source
fetchHistoryPage(...)
|
v
History preview_output
filename: "photo.png"
subfolder: "subfolder"
|
v
Generated match keys
"subfolder/photo.png"
"subfolder\\photo.png"
```
This means OSS/Core verification is about whether the generated media is
currently available through the same generated/history-backed path the
widget uses, not a full disk-level executability check across the entire
output directory.
## Why Not Consolidate All Annotated Path Parsers
There are existing annotated-path parsers in image widget, Load3D, and
path creation code. This PR does not replace them.
The helper added here is detection-only: it strips annotations to build
comparison keys for missing-media verification. Parser consolidation
across widget implementations is intentionally left out of scope to keep
this fix narrow.
## Known Follow-Ups / Out Of Scope
- FE-620 tracks the separate video drag-and-drop upload race between
upload completion and missing-media detection.
- Published/shared workflow assets are still not fully represented by
`/api/assets?include_public=true`; that remains a backend/API contract
issue.
- A future backend/API contract that answers “is this workflow media
executable?” would be preferable to stitching together runtime-specific
FE sources.
- OSS/Core full output-folder scanning via `/internal/files/output` was
considered, but that endpoint is internal, shallow (`os.scandir`), and
not the same source currently used by the generated picker flow.
## Validation
- `pnpm test:unit -- missingMediaAssetResolver missingMediaScan
mediaPathDetectionUtil formatUtil`
- touched files `oxfmt`
- touched files `oxlint --fix`
- touched files `eslint --cache --fix --no-warn-ignored`
- `pnpm typecheck`
- pre-commit `pnpm knip --cache`
- pre-push `pnpm knip --cache`
`knip` passes with the existing tag hint:
```txt
Unused tag in src/scripts/metadata/flac.ts: getFromFlacBuffer → @knipIgnoreUnusedButUsedByCustomNodes
```
## Screenshots
Before
https://github.com/user-attachments/assets/50eab565-3160-4a57-a758-87ec2c09071e
After
https://github.com/user-attachments/assets/08adcbbd-c3fc-43f9-b86c-327e4eb5abd8
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12069-fix-handle-annotated-output-media-paths-in-missing-media-scan-3596d73d365081f4afa3d4dd45cad3da)
by [Unito](https://www.unito.io)
|
||
|
Terry Jia
|
653ef1a4f0 |
Handle Load3D "none" model selection in frontend (#11178)
## Summary Load3D now supports panoramic images and HDRI loading, it can serve as a viewer for those without requiring a 3D model. Previously, the node required a model file to execute. Rather than making the input optional (which would break existing workflows that rely on it being required), a "none" option is added to the combo list, allowing users to run Load3D with no model loaded. BE change is https://github.com/Comfy-Org/ComfyUI/pull/13379 ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-11178-Handle-Load3D-none-model-selection-in-frontend-3416d73d365081e589b3d89bc67f75e7) by [Unito](https://www.unito.io) |
||
|
Christian Byrne
|
c16052e2e3 |
feat: sort right-click context menu categories alphabetically (#12039)
*PR Created by the Glary-Bot Agent*
---
## Summary
Sort the canvas right-click "Add Node" context menu by display name
(case-insensitive, natural numeric). Previously, both category submenus
and leaf nodes appeared in node-registration order, making the menu hard
to scan for users browsing for nodes.
This change is scoped specifically to the **smaller right-click
contextual menu**. It does NOT affect the double-click search menu or
the left-side Nodes panel.
## Changes
- `src/lib/litegraph/src/LGraphCanvas.ts` — In `onMenuAdd` →
`inner_onMenuAdded`, sort the deduplicated category submenu entries and
the leaf-node entries by `content` using `localeCompare` with `{
numeric: true, sensitivity: 'base' }`. Categories still appear before
leaf nodes within a level (preserves existing UX).
- `src/lib/litegraph/src/LGraphCanvas.onMenuAdd.test.ts` — New unit
tests that mock `LiteGraph.ContextMenu` and assert: case-insensitive
sort, natural numeric ordering (`Cat1`, `Cat2`, `Cat10`), leaf-node
sorting inside a category, and category-before-leaf placement.
## Verification
- `pnpm vitest run src/lib/litegraph/src/LGraphCanvas.onMenuAdd.test.ts`
— 4/4 pass
- `pnpm typecheck` — clean (ran via pre-commit hook on initial commit)
- `oxfmt` / `oxlint` / `eslint` — clean
- Oracle review against `main` returned 0 critical / 1 warning (test
coverage) / 1 suggestion (numeric sort) — both addressed in this PR.
## Notes
- The sort is applied at the menu-build site rather than inside
`LiteGraphGlobal.getNodeTypesCategories`/`getNodeTypesInCategory` to
keep the change scoped to the menu UX and avoid changing the iteration
order seen by extensions that consume those public methods.
- Per user request, this is opening as a draft PR for self-review +
CodeRabbit feedback in a single follow-up pass; manual browser
verification (right-click screenshots) was deferred to that pass.
- Slack thread context: user reported the contextual menu is "a mess"
for discovering native nodes; alphabetical sorting addresses the
discoverability problem without touching the search-oriented menus.
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12039-feat-sort-right-click-context-menu-categories-alphabetically-3596d73d36508107a87ffec1c353994e)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com>
Co-authored-by: Alexis Rolland <alexis@comfy.org>
|
||
|
Comfy Org PR Bot
|
3e94459340 |
1.45.2 (#12096)
Patch version increment to 1.45.2 **Base branch:** `main` ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12096-1-45-2-35b6d73d36508193be00c1c878d42c2a) by [Unito](https://www.unito.io) --------- Co-authored-by: christian-byrne <72887196+christian-byrne@users.noreply.github.com> Co-authored-by: github-actions <github-actions@github.com>v1.45.2 |
||
|
Henry Lee
|
ca54877f9d |
fix(assets): strip directory annotation from input filenames (#12086)
## Summary
Imported assets render as a generic check-check icon instead of a
thumbnail because the OSS `/internal/files/{type}` endpoint returns
annotated filenames (`photo.png [input]`) that the assets-sidebar mapper
passes through verbatim, which breaks extension-based media-type
detection.
## Changes
- **What**: Strip ComfyUI's trailing directory-type annotation (`
[input]`, ` [output]`, `[temp]`) in `mapInputFileToAssetItem` so `name`,
`id`, and the generated `/view?filename=…` URL all use the canonical
on-disk filename. Adds a focused unit test.
- **Breaking**: None.
- **Dependencies**: None.
## Review Focus
### Root cause
ComfyUI core PR
[comfyanonymous/ComfyUI#13078](https://github.com/comfyanonymous/ComfyUI/pull/13078)
(April 2026) changed `/internal/files/{type}` to append the directory
type to each entry:
```python
# api_server/routes/internal/internal_routes.py
return web.json_response(
[f"{entry.name} [{directory_type}]" for entry in sorted_files], status=200
)
```
The annotation is the wire format `LoadImage`-style widgets expect, so
the backend change is correct. The assets-sidebar mapper treated the
response strings as raw filenames. After
[#8914](https://github.com/Comfy-Org/ComfyUI_frontend/pull/8914) changed
`getMediaTypeFromFilename` to default unknown extensions to `'other'`,
every input asset now routes to `MediaOtherTop` and renders as
`icon-[lucide--check-check]`:
```
getMediaTypeFromFilename("photo.png [input]").split('.').pop() === "png [input]" → 'other'
```
The strip happens at data ingestion so every consumer of
`AssetItem.name` (sidebar grid, list, filter, gallery, drag-drop,
delete) gets the canonical filename automatically. OSS-only — Cloud
paths get clean names from the cloud API and are unaffected.
Reproduces locally on stock OSS ComfyUI on `main` of both repos; no
public issue tracker entry.
## Screenshots (if applicable)
Before:
<img width="1091" height="718" alt="image"
src="https://github.com/user-attachments/assets/ff1f070d-da39-4e5a-bc6d-99b7214f7da8"
/>
After:
<img width="1089" height="716" alt="image"
src="https://github.com/user-attachments/assets/7123d9bf-f7dd-4430-b6f7-f6702b70baaa"
/>
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-12086-fix-assets-strip-directory-annotation-from-input-filenames-35a6d73d365081e9b9eed7d8630d6f0b)
by [Unito](https://www.unito.io)
|
||
|
Simon Pinfold
|
a4faaa0159 |
fix: disable ultralytics asset-browser registration (#12075)
*PR Created by the Glary-Bot Agent* --- ## Summary Disable the `UltralyticsDetectorProvider` model-to-node mapping so the node falls back to the static combo populated from `/api/object_info`, restoring pre-#8468 behavior on cloud. ## Why PR #8468 opted `UltralyticsDetectorProvider` into the cloud asset-widget path, which exposed a latent mismatch in cloud asset metadata for nested-directory model folders. The bug has two independent halves, and a fix that addresses only one will still leave the workflow broken at execution time: - **Tag lookup mismatch.** Cloud stores tags as combined values like `ultralytics/bbox`, while the asset query asks for split tags (`models` + `ultralytics`) with exact-match filtering — so the dropdown returns no results. - **Submitted value mismatch.** Cloud stores filenames as basenames, but the node expects subdirectory-prefixed values (e.g. `bbox/face_yolov8m.pt`) that the static combo path normally produces. Both halves require cloud-side fixes (asset ingestion + metadata) before the asset-browser registration can be safely re-enabled. Until then, removing the registration restores the working static-combo behavior so users are unblocked. ## Changes - `src/platform/assets/mappings/modelNodeMappings.ts`: comment out the `['ultralytics', 'UltralyticsDetectorProvider', 'model_name']` entry with a note pointing at BE-689 and the re-enablement criteria. - `src/stores/modelToNodeStore.test.ts`: drop the now-stale ultralytics expectations from `EXPECTED_DEFAULT_TYPES`, `MOCK_NODE_NAMES`, and the hierarchical-fallback `it.each` cases. ## Verification Local quality gates: - `pnpm typecheck` — clean - `pnpm lint` — clean (3 pre-existing warnings, 0 errors) - `pnpm format:check` — clean - `pnpm knip` — clean (1 pre-existing warning unrelated to this change) - `pnpm test:unit -- src/stores/modelToNodeStore.test.ts` — 51/51 passing Manual runtime verification (dev server + Playwright against the live module): - `MODEL_NODE_MAPPINGS` no longer contains any entry where `[0] === 'ultralytics'` or `[1] === 'UltralyticsDetectorProvider'` (84 entries total, 0 ultralytics). - `useModelToNodeStore().getNodeProvider('ultralytics')` returns `null` after `registerDefaults()`, so the asset-widget path is no longer triggered for this node. - `getNodeProvider('ultralytics/bbox')` also returns `null`, confirming hierarchical fallback no longer resolves to the disabled mapping. - `getNodeProvider('checkpoints')` still resolves to `CheckpointLoaderSimple`, confirming unrelated mappings are intact. End-to-end cloud verification (actually exercising the asset-browser path against cloud-seeded ultralytics metadata) is not possible in the local sandbox, since the regression depends on the cloud's asset ingestion data shape. The change is a single mapping-table removal that reverts to the well-exercised static-combo path covered by the updated unit tests. Long-term cloud-side fix is tracked in BE-689. - Fixes BE-689 ## Screenshots  ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12075-fix-disable-ultralytics-asset-browser-registration-35a6d73d36508179b394f0915e69742e) by [Unito](https://www.unito.io) Co-authored-by: Glary-Bot <glary-bot@users.noreply.github.com> |
||
|
Dante
|
8108967d49 |
feat(dialog): migrate Prompt + Confirmation dialogs to Reka-UI (Phase 1) (#12041)
## Summary Phase 1 of the dialog migration kicked off in #11719. Migrates the two simplest production dialogs — `PromptDialogContent` and `ConfirmationDialogContent` — from PrimeVue `Dialog` onto the Reka-UI primitives landed in Phase 0. Public API of `useDialogService` / `dialogStore` is unchanged. Parent: [FE-571](https://linear.app/comfyorg/issue/FE-571/dialog-system-migration-primevue-reka-ui-parent) This phase: [FE-573](https://linear.app/comfyorg/issue/FE-573/phase-1-migrate-promptdialog-confirmationdialog-closes-11688) Predecessor: #11719 (merged at `0788e7139`) Refs #11688 (closed manually after Phase 0; the actual user-visible max-width fix ships in this PR) ## Changes ### `src/services/dialogService.ts` | Call site | Renderer | Size | Width override | | --- | --- | --- | --- | | `prompt()` | `'reka'` | `md` | — | | `confirm()` | `'reka'` | `md` | — | | `showBillingComingSoonDialog()` | `'reka'` | `sm` | `contentClass: 'max-w-[360px]'` | ### `src/components/dialog/content/ConfirmationDialogContent.vue` - Drops `import Message from 'primevue/message'` — the only PrimeVue dependency in the component - Replaces `<Message>` with a Tailwind `role="status"` alert keeping the `pi pi-info-circle` icon and muted-foreground severity ### `src/stores/dialogStore.ts` + `src/components/dialog/GlobalDialog.vue` - Adds `contentClass?: HTMLAttributes['class']` on `CustomDialogComponentProps` - Forwards it to `<DialogContent :class="...">` on the Reka branch (PrimeVue path keeps using `pt`) ## Why this scope 1. **Smallest content surface** — `PromptDialogContent` is 43 LOC; the only PrimeVue dependency in `ConfirmationDialogContent` is the `<Message>` info banner. 2. **Closes #11688 ergonomics** — Reka's `md` size = `max-w-xl` (576px / 36rem), exactly the max-width the issue reporter asked for. 3. **Three known callers** — all in `dialogService.ts`. No other callers needed to change. 4. **Renderer branch is already proven by Phase 0**; this PR just flips the flag. ## Visual proof Verified live in Storybook (`Components / Dialog / Dialog → Default` and `… → All Sizes`) at viewport `1920×1080`. DOM inspection confirms the rendered widths match the design intent: | Story | size | Rendered width | Computed `max-width` | | --- | --- | --- | --- | | `Default` | `md` | **576 px** | **576 px (= 36rem)** | | `All Sizes` (sm slot) | `sm` | 384 px | 384 px (= 24rem) | The `md` measurement directly answers the #11688 reporter screenshot (1558 px wide PrimeVue dialog → 576 px Reka dialog on the same display). Local screenshot artifacts (not committed): `temp/screenshots/phase1-md-576px-1920w.png`, `temp/screenshots/phase1-md-allsizes-1920w.png`, `temp/screenshots/phase1-sm-384px-1920w.png` — drag-drop into the PR body before marking ready for review. ## Quality gates - [x] `pnpm typecheck` — clean - [x] `pnpm lint` — clean - [x] `pnpm format` — applied (oxfmt) - [x] `pnpm test:unit` (touched files): **26/26 passed** - `ConfirmationDialogContent.test.ts` (9 tests, no longer needs PrimeVue plugin) - `PromptDialogContent.test.ts` (5 tests, unchanged) - `GlobalDialog.test.ts` (9 tests, Phase 0 coverage still passes after the contentClass forwarder addition) - `dialogService.renderer.test.ts` **new** — 3 tests asserting each call site sets `renderer: 'reka'` (regression net) - [ ] `pnpm test:browser:local --grep "@mobile confirm dialog"` — **could not run locally** (no ComfyUI Python backend on `localhost:8188` in this session); CI will gate the existing fixture, which is already renderer-agnostic (`getByRole('dialog')` + `getByRole('button', ...)` in `browser_tests/fixtures/components/ConfirmDialog.ts`). ## Public API impact None. `useDialogService().prompt(...)` / `confirm(...)` / `showBillingComingSoonDialog(...)` keep their existing signatures. Custom-node extensions calling `app.extensionManager.dialog.*` continue to work. ## Out of scope (later phases) - `ErrorDialogContent`, `NodeSearchBox`, `SecretFormDialog`, `VideoHelpDialog`, `CustomizationDialog` — Phase 2 (FE-574) - Settings dialog — Phase 3 (FE-575) - Manager dialog — Phase 4 (FE-576) - `ConfirmDialog` callers (`SecretsPanel`, `BaseWorkflowsSidebarTab`) — Phase 5 (FE-577) - Removing PrimeVue `Dialog` imports + `<style>` cleanup in `GlobalDialog.vue` — Phase 6 (FE-578) - Legacy `ComfyDialog` (`src/scripts/ui/dialog.ts`) - Deduplicating `Dialogue.vue` / `ImageLightbox.vue` ## Screenshot <img width="865" height="497" alt="Screenshot 2026-05-08 at 4 35 45 PM" src="https://github.com/user-attachments/assets/6aead2ad-2e0b-478a-9154-bb632a6bf3d1" /> <img width="1363" height="964" alt="Screenshot 2026-05-08 at 4 38 16 PM" src="https://github.com/user-attachments/assets/10647752-a063-4901-a206-842799cc5d7a" /> <img width="889" height="486" alt="Screenshot 2026-05-08 at 4 46 57 PM" src="https://github.com/user-attachments/assets/81899a81-205a-46f2-bddd-7639624607f6" /> ## Test plan - [x] Unit: 26/26 pass on touched files - [ ] CI: `@mobile confirm dialog` spec on the migrated path - [ ] Manual (post-CI on a real backend): open prompt and confirm dialogs on 1920×1080 viewport, verify ≤ 36rem max-width, ESC closes, backdrop click closes, Enter submits prompt, focus trap holds - [ ] Manual: open Billing Coming Soon dialog — verify it stays at the existing `max-w-[360px]` width |
||
|
Dante
|
0ef98de8eb |
fix: make credits help icon a tooltip button in cloud user popover (FE-617) (#12072)
## Summary The help icon (lucide circle-help) next to the credits balance in the cloud user popover was a bare `<i>` with `v-tooltip` and `cursor-help`. PrimeVue tooltip on a bare `<i>` did not fire reliably and the icon had no focus/keyboard semantics, so users saw "no hover action and not clickable". Wrap the icon in `<Button variant="muted-textonly" size="icon-sm">`, matching the existing pattern in `InfoButton.vue` and `MissingPackGroupRow.vue`. Same change applied to `CurrentUserPopoverLegacy.vue` and `CurrentUserPopoverWorkspace.vue`, which shared the broken pattern. - Fixes FE-617 - https://comfy-organization.slack.com/archives/C0A4XMHANP3/p1778191473621829 ## Red-Green CI Verification The branch was force-pushed back to the test-only commit so CI could run against it, then restored to the fix commit. | Commit | CI: Tests Unit | Outcome | |--------|---------------|---------| | `test:` ( |
