From an outside audit (same change as 32a3719 on feature/audio-engines):
- The scan wizard's offline mode still sent the greyed-out rename and
overwrite options, and scraper.py fetched Civitai model pages and preview
media for models whose .info held a Civitai id. Offline now implies no
media and no model page; the folder scan and scan_all refuse a missing or
non-object body instead of defaulting to an online scan.
- The Prompt Workshop's snap dock referenced leftSearch from the source deck
module (ReferenceError on click); the deck now exposes focusSearch().
- translate, delete_model, delete_gallery_image and the image-material
routes answer malformed bodies with 400 instead of 500.
- Removed six unused imports in ui_doctor.js and 16 duplicate locale keys
per language (the later, effective entry is kept, so nothing changes on
screen).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sending a Prompt Note with a UNet main model chained its LoRAs through
LoraLoader, whose clip input nothing could feed. UNet LoRAs now use
LoraLoaderModelOnly and the editor says to connect a text encoder.
Models without a baseModel in their metadata matched no base model, so
they never appeared; base_models now reports them and
compatible_models serves them for "__unlabeled__". Both routes share
one walker and run it off the event loop.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The grid used one fixed row height, so the extra apply button squeezed the
badges, date and node types on top of each other. Cards in apply mode now get
a taller row (the list is marked is-apply-mode when it renders), the apply
button sits at the bottom, and only the tag row may be clipped when space
runs out. The normal row is a little taller so a tag row fits as well.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeQEub5
(cherry picked from commit 9e761d0705)
In apply mode each card rendered a second apply button (a clone meant for
the list layout's hover actions) that showed as an extra row in the grid,
and both labels prefixed another icon to a locale string that already has
one. Cards now render a single button whose text comes straight from the
locale (textContent), in grid and list layouts alike.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeQEub5
(cherry picked from commit 01fc925bd6)
`body:has(#…-modal[style*="display: flex"]) #…-tooltip-bubble` made the
browser re-check the whole page on every inline-style change anywhere.
ComfyUI moves every DOM widget (text boxes) on each pan/zoom frame, so every
frame paid for it, with the browser closed too, and much more with plugins
that add DOM per widget (ComfyUI Prompt Assistant).
Measured in ComfyUI (frontend 1.53.6, Chromium) panning a 300-node
workflow for 120 frames, style recalculation in total:
- no plugins 60 ms; AMB 690 ms -> 63 ms
- AMB + Prompt Assistant 2300 ms (~19 ms/frame) -> 86 ms (PA alone 80 ms)
The rule was redundant: both popups hide the tooltip when they open and
showSharedTooltip refuses while either is open (checked in the browser).
Cache-busting versions bumped so browsers load the new stylesheet.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012RMuwbsWZHnx2eCWpxYqAB
The footer version sat right under the tool dock and invited misclicks.
The "!" update guide now opens with the current version and a
"Check for updates / switch version" button that closes the guide and
opens the version panel.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeQEub5
In the header it pushed the dock and close buttons out of view when the
browser is docked or the display is scaled. It now sits under the tool
dock and hides with the collapsed sidebar.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeQEub5
- api/version_manager.py: installed version (local only), published
releases on request (GitHub API, git tag fallback), switch to a
published tag, return to the default branch (fast-forward only), undo
the last switch; refuses to overwrite modified tracked files
- header version chip and "Version & updates" panel; nothing goes online
until "Check for updates" is clicked; restart via ComfyUI Manager
- zh/en strings, styles, architecture notes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeQEub5
Splitting the query on spaces hid how it was read and made phrases like
"cyberpunk city" impossible to search.
- ui_search_chips.js: Enter or a comma (also "，" / "、" inserted by an IME)
turns the typed text into a block; blocks keep their spaces and match as a
phrase; hash-like and numeric blocks are labelled; × removes a block,
Backspace in an empty box removes the last one, one button clears all; text
still being typed joins the search after a short pause; IME composition
never commits a block
- gallery_images accepts repeated `term` params (q still works)
Verified: tests/test_image_search.py (phrase blocks vs split query, 107 Python
tests in total) and a jsdom walkthrough of blocks, IME commas, backspace,
remove, clear and the request parameters.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeqEub5
- api/image_search.py reads only the PNG text chunks before IDAT (ComfyUI
prompt/workflow, A1111 parameters), builds one small record per image and
caches it by mtime; all space-separated terms must match
- hex terms of 8+ characters also match SHA256 values recorded in the
workflow (extra.anomalous_hashes) and local model files with that hash, so
renamed models are still found; A1111 "Model hash" text matches as text
- gallery_images accepts q; the search box above the gallery debounces input,
shows the match count, keeps paging/infinite scroll and the detail viewer on
the filtered list, and lets a newer query supersede one still loading
On the local output (774 PNG) the first search reads ~10 MB of text chunks;
later searches use the cache.
Verified: tests/test_image_search.py (prompt, seed, AND terms, A1111, file
names, recorded and local-index hashes, cache, chunk reader), full Python suite
(106), jsdom walkthrough of the gallery search box.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W411Z69FC3em9CzpeqEub5
