VmHWM is per process; the other tests in the binary ran on parallel threads
and their allocations failed the 64 MiB bound intermittently.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A 16384x1 output stacks every grid row onto one pixel row, so a sane warp
visits ~100x its pixels; allow each triangle about two grid cells (CodeRabbit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An empty layers list cleared the composition's markers too (CodeRabbit);
the composition-marker menu uses comp.deleteAllMarkers instead.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Offset Paths' Amount, Line Join, Miter Limit, Copies and Copy Offset,
and Zig Zag's Size, Ridges per segment and Points reported their
internal ids as matchName, so property("ADBE Vector Offset Amount") and
the other documented names returned null. They now report and resolve
by ADBE Vector Offset Amount, Line Join, Miter Limit, Copies and Copy
Offset, and ADBE Vector Zigzag Size, Detail and Points. Lookups by
display name and internal id still work.
ConfinedFootage remembered the first answer for each footage path. A symbolic
link inside a read root that pointed inside when first read could be pointed
outside afterwards, and the next frame-cache miss read the outside file.
Each read now runs the confinement check again.
automation_roots_check_relinked_footage_again renders a link to a file inside
the root, points it outside, empties the memory cache and renders again: the
outside pixels came back before this change and don't now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The command checks look at the paths an agent gives, but a project it opens or
imports from a read root names its own footage, which was read wherever it is
when the project rendered: render_frame returned its pixels.
Backend::confine now wraps a headless session's footage and services:
ConfinedFootage reads frames, audio, aux channels, models and vector frames
only for footage inside the read roots (problem() says why the rest shows
nothing), and ConfinedServices checks read_file, exists, stamp, list_dir and
is_dir the same way. Writes are left to the command checks.
Test: automation_roots_confine_the_footage_a_project_names covers an absolute
path, `../`, a linked file and a linked folder, each opened and imported.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The ≡ menu of a panel in its own OS window ("Move Panel to New Window")
was a root-layer Area drawn by the main window's pass at window-local
coordinates, so it opened in the main window at the wrong place.
The open menu now remembers the viewport that owns the click and is
drawn by that viewport's pass (after the window's last panel body); the
main window draws only its own panels' menus. The menu takes at most its
window's width (a floated panel can be narrower than the menu), a press
in another window dismisses it, and a menu left behind by a window that
closed is dropped.
Ports the FilmCraft fix (filmcraft#3633).
Tested: new unit tests in panels.rs run passes of two viewports against
one context: the menu is drawn in the owner's pass and not the main
window's, a main-window menu stays there, a 160 pt window holds the
menu, a press in another window closes it, and a closed window's menu is
dropped. effectcraft-ui-egui lib tests (387), clippy -D warnings, fmt.
Brandon ThomasandClaude Opus 5.5
zitann2020andDeepSeek V4.1 Flash
dotCooCoo
adikshan11andClaude Opus 5.5
Dylan M. Taylor
aiongg
marshiyar