This pull request refines the loading and error handling logic for both
the `VideoPreview.vue` and `ImagePreview.vue` components. The main
improvements include making the loading skeleton more accurate and
visually consistent, updating how loading and error states are managed
when URLs change, and ensuring that the main media elements are hidden
while loading. These changes enhance the user experience by providing
clearer feedback during media load operations.
**Loading and error state improvements:**
* The loading skeleton in both `VideoPreview.vue` and `ImagePreview.vue`
now only appears when loading and no error is present, with updated
styling and fixed dimensions for better consistency. (`VideoPreview.vue`
[[1]](diffhunk://#diff-17b5c19b4628f22e45570b66a85ed1fc16e931dd368fe420584d487e522ab8aaL29-R41)
`ImagePreview.vue`
[[2]](diffhunk://#diff-0c0b17c5c68464e0284398ad42b823509d414c9cf297f3bc2aa2b00e0f9c2015L29-R48)
* The main video and image elements are now hidden (using the
`invisible` class) while loading, preventing display glitches before the
media is ready. (`VideoPreview.vue`
[[1]](diffhunk://#diff-17b5c19b4628f22e45570b66a85ed1fc16e931dd368fe420584d487e522ab8aaL29-R41)
`ImagePreview.vue`
[[2]](diffhunk://#diff-0c0b17c5c68464e0284398ad42b823509d414c9cf297f3bc2aa2b00e0f9c2015L29-R48)
* The loading state (`isLoading`) is now set to `true` whenever new URLs
are provided, and reset appropriately when navigating between media
items, ensuring accurate feedback to the user. (`VideoPreview.vue`
[[1]](diffhunk://#diff-17b5c19b4628f22e45570b66a85ed1fc16e931dd368fe420584d487e522ab8aaL145-R152)
`ImagePreview.vue`
[[2]](diffhunk://#diff-0c0b17c5c68464e0284398ad42b823509d414c9cf297f3bc2aa2b00e0f9c2015L164-R176)
[[3]](diffhunk://#diff-0c0b17c5c68464e0284398ad42b823509d414c9cf297f3bc2aa2b00e0f9c2015L224-R236)
**Code consistency and maintainability:**
* Both components now import and use the shared `cn` utility for
conditional class names, improving code consistency and maintainability.
(`VideoPreview.vue`
[[1]](diffhunk://#diff-17b5c19b4628f22e45570b66a85ed1fc16e931dd368fe420584d487e522ab8aaR115)
`ImagePreview.vue`
[[2]](diffhunk://#diff-0c0b17c5c68464e0284398ad42b823509d414c9cf297f3bc2aa2b00e0f9c2015R132)
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-7094-Fix-skeleton-loaders-for-Image-Video-Previews-2bd6d73d3650817989e1f4d597094016)
by [Unito](https://www.unito.io)
Increases the `/features` endpoint polling interval (runtime config
polling) from the current 30 seconds to 10 minutes.
## Background
The polling was originally added for two main purposes:
1. Server alert badges
2. Extra assurance that states are synchronized between frontend and
backend
However, both of these use cases are not critical:
- Server alert badges are unlikely to be needed in practice. WebSocket
(WS) can be used eventually as an alternative
- For synchronization, the system should be redesigned to be explicit
about marking client state as stale, or use WebSocket/Server-Sent Events
(WS/SSE) on a per-data basis rather than polling for the entire feature
flag set
## Motivation
The reason to reduce polling frequency is that per-user feature flags
are being added, which will make `/features` endpoint handling
significantly heavier. The endpoint will no longer just return static
JSON with high cache age, making frequent polling more costly.
## Future Considerations
- Eventually migrate server alert badges to use WebSocket
- Design a system that explicitly marks client state as stale when
needed
- Consider using WebSocket or Server-Sent Events for targeted data
updates instead of polling entire feature flag set
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-7100-cloud-increase-feature-flag-polling-interval-to-10min-from-30s-2bd6d73d365081dfa8abeec901d0d975)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Alexander Brown <drjkl@comfy.org>
## Summary
Fixes the `ReferenceError: __DISTRIBUTION__ is not defined` error when
running i18n collection tests.
## Problem
PR #6879 added conditional menu commands based on distribution (hiding
memory unload commands in cloud). This introduced a dependency on
`isCloud` which uses the `__DISTRIBUTION__` Vite define variable in
`coreMenuCommands.ts`.
When Playwright's test runner imports this file during i18n collection,
it fails because Vite define variables are only replaced during Vite's
build/dev process, not during Playwright's TypeScript compilation.
## Solution
Created a simple shim (`scripts/vite-define-shim.ts`) that:
1. Defines all Vite define variables as global constants with default
values
2. Provides a minimal `window` shim for Node environment
3. Is imported at the top of `collect-i18n-general.ts` before any code
that uses these variables
This approach is simpler than:
- Creating a custom Babel plugin (attempted in this PR, see commit
history)
- Using `ctViteConfig` (only works for component testing, not regular
Playwright tests)
- Post-build regex replacement (fragile and error-prone)
## Test Plan
Run `pnpm collect-i18n` and verify:
- ✅ No more `ReferenceError: __DISTRIBUTION__ is not defined`
- ✅ No more `ReferenceError: window is not defined`
- ⏱️ Tests may timeout if dev server is not running on port 5173, but
that's a separate issue
## Related
- Fixes issue introduced by PR #6879
- Related to Notion task:
https://www.notion.so/comfy-org/Implement-Babel-plugin-for-Vite-define-replacements-in-Playwright-2b56d73d365081d5bb63e02712912b17🤖 Generated with [Claude Code](https://claude.com/claude-code)
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-6906-bugfix-Add-vite-define-shim-for-Playwright-i18n-collection-2b66d73d36508182b4d6d69b88ae2771)
by [Unito](https://www.unito.io)
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: github-actions <github-actions@github.com>
Co-authored-by: Alexander Brown <drjkl@comfy.org>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
## Summary
Makes the area a bit to the left and right of the dot also clickable.
Addresses complaints about it being tricky to connect nodes in Nodes
2.0.
---------
Co-authored-by: github-actions <github-actions@github.com>
## Summary
Temporarily simplifies the resize logic to only work on bottom right
corner and eliminates edge cases where corner resizing caused position
drift issues.
- Remove multi-corner resize handles in favor of bottom-right only
- Delete resizeMath.ts and its tests (no longer needed)
- Simplify useNodeResize to only handle bottom-right resize
- Remove position tracking from resize callback
---------
Co-authored-by: github-actions <github-actions@github.com>
Follow-up adjustments to the weekly ComfyUI release automation workflow.
## Changes
1. **Rename workflow to follow conventions**
- File: `weekly-comfyui-release.yaml` → `release-weekly-comfyui.yaml`
- Name: "Weekly ComfyUI Release" → "Release: Weekly ComfyUI"
- Matches pattern of other `release-*` workflows
2. **Sync fork with upstream before creating PR**
- Fetches latest upstream/master before making changes
- Ensures PR only shows requirements.txt diff, not stale fork commits
- Does not modify fork's master branch (only pushes automation branch)
## Testing
After merge, can test via manual workflow dispatch in Actions tab.
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-7060-fix-Weekly-release-workflow-adjustments-2bb6d73d365081008436d1b9e5f7dd65)
by [Unito](https://www.unito.io)
Before node v2, you could use getInputPos, getOutputPos. In node v2,
that is not possible.
## Summary
Want to get the position of the output / inputs on the graph like before
node v2.
## Changes
- **What**: <!-- Core functionality added/modified --> Added
LGraphNode.getSlotPosition
## Review Focus
## Screenshots (if applicable)
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-7042-Expose-LGraphNode-getSlotPosition-2ba6d73d36508195a0fed2d1fe2b64f6)
by [Unito](https://www.unito.io)
Adds scheduled workflow to bump ComfyUI frontend RC releases every
Monday at noon PST.
## Implementation
- **Resolver script** (`scripts/cicd/resolve-comfyui-release.ts`):
Checks ComfyUI `requirements.txt` and determines next minor version,
compares branch commits to latest patch tag
- **Workflow** (`.github/workflows/weekly-comfyui-release.yaml`):
- Scheduled for Monday 20:00 UTC (noon PST)
- Manual dispatch supported for testing/off-cycle runs
- Three jobs: resolve version → trigger release if needed → create
ComfyUI PR
- Reuses existing `release-version-bump.yaml` workflow
- Creates draft PR from fork to `comfyanonymous/ComfyUI` with updated
`requirements.txt`
- Includes error handling and validation for all steps
- Force pushes to same branch weekly to maintain single open PR
## Testing
- Resolver script tested locally: correctly identified v1.28.8 → v1.29.4
with release needed
- yamllint passes
- knip passes with `gh` binary added to ignore list
## Follow-up
Can test via workflow_dispatch once merged, or in this PR if we enable
Actions on branch.
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-6877-feat-Add-weekly-ComfyUI-release-automation-2b46d73d36508154aa05c783c6942d9a)
by [Unito](https://www.unito.io)
There's a warning toast shown if the frontend is considered out-of-date
(relative to the version in the requirements.txt of
comfyanonymous/ComfyUI). As a result, e2e tests run on older release
branches (e.g., when backporting or hotfixing) can sometimes trigger the
warning which obviously causes visual regression tests to fail. This PR
adds a hidden setting to disable the warning and sets it to `true` in
the e2e test fixtures.
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-7004-test-add-setting-to-ignore-version-compatibility-toast-warnings-in-e2e-tests-2b86d73d3650812d9e07f54a0c86b996)
by [Unito](https://www.unito.io)
explicitly prevents subgraphs from making an api call since they don't
have docs, this was previously reliant on a non-ok resolution
also doesn't try returning anything that has contenttype of text/html to
prevent the markdown renderer from crashing
## Summary
- short-circuit blueprint/subgraph nodes in help: skip doc fetch and
return the node description, avoiding SPA fallback responses
- guard node help fetch against HTML/SPA fallbacks using content-type
checks; treat them as missing and trigger the existing description
fallback
- keep base URL logic unchanged for non-blueprint nodes
## Testing
- pnpm typecheck
- pnpm lint:fix
- pnpm test:unit
## Summary
In cloud distribution, completed jobs now show "Finished in Xh Ym Zs" as
the primary text instead of the filename.
- Uses `formatDuration` to display time as `1h 30m 45s`, `30m 45s`, or
`45s`
- Gated with `isCloud` - non-cloud continues to show filename
- Added i18n key `queue.completedIn` for localization
Filename is not fetchable right now in cloud. This is what design wanted
as the alternative.
<img width="679" height="1097" alt="image"
src="https://github.com/user-attachments/assets/291deb42-77d8-4de9-b4f8-ee65f3c25011"
/>
Co-authored-by: Claude <noreply@anthropic.com>
The original video was encoded with AV1 codec which Safari does not
support, causing "Plug-in handled load" errors. Re-encoded to VP9
which is supported across all major browsers including Safari 14.1+.
Summary
- Re-encode cloud-subscription.webm from AV1 to VP9 codec for Safari
browser compatibility
Problem
Safari does not support AV1 codec in webm containers, causing the cloud
subscription video to
fail with "Plug-in handled load" error.
Solution
Re-encoded the video using VP9 codec (libvpx-vp9) which is supported by
all major browsers
including Safari 14.1+.
Test plan
- Verify video plays correctly in Safari
- Verify video plays correctly in Chrome/Firefox/Edge
[cloud-subscription.webm](https://github.com/user-attachments/assets/ab2dcaa1-f3b3-47c9-84d0-9fc8ea8c6f17)
┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-7006-fix-Re-encode-cloud-subscription-video-to-VP9-for-Safari-compatibility-2b86d73d365081ad9262f97fc1ebb5ee)
by [Unito](https://www.unito.io)
