While the models load, an empty picture area says the run is going instead
of asking for a run. With steps on, the last step stays up until the saved
picture has loaded, so it no longer blinks out. Between a tiled upscale's
split and stitch (and for Ultimate SD Upscale) the step pictures are tiles:
the whole picture stays and the chip says "upscaling in tiles".
Image Crop + Rotate + Pad, Load Image + Pad and Load Video refused a
picture named the way ComfyUI names a generated file
("photo.png [output]", "[temp]") with "The selected source is outside
ComfyUI's input folder." Core Load Image takes those names.
resolve_input_path now reads the name the same way the video nodes'
Outputs list already does (resolve_server_file): the tag picks the
input, output or temp folder, and the rest is a name below it. A whole
path, "~", "." and "..", and a network path are refused from the text
alone; a link that leads out of the folder is refused once resolved.
Both functions share that code. A whole path inside the input folder,
accepted before, is now refused too, as the security rule asks.
Load Video's player and the shared upstream preview stripped the tag
and asked /view for the input folder; they now keep the folder.
Tests cover input, [input], [output] and [temp] names through the
resolver and through ComfyUI's own validation next to core Load Image,
and the refused escapes. Help pages, SECURITY.md and CHANGELOG say so.
A list of pictures runs these nodes once per picture, and the panels
only showed one of them (#204). Arrows (2 / 5) now appear on the node
when a run brings more than one picture: Compare steps through its
pairs, A and B together, and Erase saves the pair on screen. The shared
preview panel gets the same arrows, so Save Image, Mask Refine, Mask by
Name, LaMa Inpaint and Select Frame step through their results. A node
too narrow for the arrows widens to fit while they show.
Switching a preview off and on again (Mask Refine, LaMa Inpaint, Select
Frame, Save Image) left an empty box until the next run, and the node came
back shorter than it was.
Now the node remembers its height when the preview goes off (saved with
the workflow) and returns to it, with the last result in the box. If the
node ran while its preview was off, the box says "Run the workflow again
to see it here" rather than showing an older picture. With nothing to show
yet it says "Run the workflow to see it here".
Mask Refine's AUTO button now says what it does in its tooltip.
In Nodes 2.0 a node's saved height is only a minimum and its content can
make it taller. The preview panel's picture was in normal flow, so a
portrait result drawn at full width grew Save Image from 730 to 1276 px in
the Krea 2 Outpaint example, over the Run Timer under it. Nodes 2.0 also
drew its own copy of the result under the panel, because the panel never
set hideOutputImages.
- The stage's picture is out of flow and centred, so the stage gets its
size from the node and the picture scales to fit, in both renderers.
- suppressCoreImagePreview sets hideOutputImages; classic never reads it.
- In Nodes 2.0 that copy was the only thing filling node.imgs, so the
panel falls back to the node's stored output (app.nodeOutputs) and
re-reads it when a workflow tab restores its outputs. node.imgs still
wins where it exists, so classic keeps its old path.
Select Frame, Mask Refine and LaMa Inpaint share the panel and grew the
same way. Core Save Image and Preview Image keep their size in Nodes 2.0.
Classic is unchanged.
Tests: the stage-picture guard in panel_guards, hideOutputImages in
core_preview and the stored-output lookup in input_preview fail on main
and pass here.
Five things surfaced by release testing.
Node search: ComfyUI's search dialog cuts at 64 results and ranks by
[match tier, global popularity, name length], where tier 1 needs the name
to START with what was typed. Popularity is a fixed global table in which
every node from a new pack scores zero, so tier is the only lever. As
"Refine Mask" this node came 121st for "mask" and was never on screen.
Renamed to Mask Refine, which is 15th, and kept "refine mask" as an alias.
The mapping key is untouched, so saved workflows are unaffected. Aliases
alone cannot fix this: the frontend indexes them but drops them from the
tier scoring because they are a list rather than a string, so they decide
whether a node is findable, not where it lands.
Previews: Mask Refine, LaMa Inpaint and Select Frame showed whatever fed
their input, which is nothing when a segmentation node is upstream - Mask
Refine's panel was blank however the mask turned out. Each now returns its
result as a preview and the panel shows that, falling back to the input
before the first run. A node does not need to be an OUTPUT_NODE for this;
returning a ui payload is enough. ComfyUI's own preview for these nodes is
stood down so the picture appears once, in the panel, rather than also
underneath the node. That also fixes LaMa finishing a single image with an
empty panel: frame previews were only attached for batches over one frame,
on the assumption that one image would preview through the output path,
which a non-output node does not have.
Mask Refine opens on expand and blur with the other five behind MORE.
Hidden widgets keep their values and their saved positions. AUTO sets both
from the mask's size, scaled off the 8/4 hand-tuned for the watermark
workflow at 576px, because a feather is a fraction of the picture rather
than a fixed pixel count.
Save Video gains nvenc h264/h265, webm av1, ProRes mov, lossless ffv1 mkv,
gif and webp, plus pingpong and save_metadata. Every format is encoded and
read back in the tests rather than trusted from its table row. fps steps in
whole frames but stays FLOAT: an INT would refuse Load Video's fps link and
round 29.97 to 30, drifting picture off audio. mov files now carry the
embedded workflow, which they silently did not. The reload button is gone -
it only re-fetched what was already on screen.
Not done: Select Frame's preview updates on a run, not as the number
changes. Seeking to a frame needs the source fps, which the browser never
learns, so live scrubbing needs a server route - the surface Frame Chooser
was just demoted for.
ausboss