open_menu_with waited once for the samples rendering and then drew its
last frame. A frame starts at most four samples, and none while eight
are rendering, so that last frame could start the samples of the rows
it showed instead of drawing them, and
show_samples_on_white_draws_dark_type_on_white_in_every_brightness
failed with "dark: samples are drawn" when the workers were slow.
wait_for_samples now returns whether any sample was rendering, and the
helper draws and waits until a frame leaves none rendering. Only test
code changes: wait_for_samples is cfg(test) and open_menu_with is a test
helper.
Diagnosis and approach by Taha-Mahmoodi in #7093.
Spot Color Objects listed every spot color linked from a gradient's stops
once any stop was above opacity 0. A spot stop at opacity 0 between other
clear stops, at the end of the ramp after a clear stop, or meeting an
opaque stop only at one offset was listed though it prints no ink, and so
was an opaque stop out of order whose pairs have no width.
A stop now counts when it bounds a stretch of some width that the
gradient sampler mixes and either stop of that stretch is above opacity
0, or when it is the first or last stop and above opacity 0, since those
color the offsets before and after the others. Stops in any order are
judged by the stretches the sampler uses. A freeform gradient keeps the
old rule, since it draws from its points. Type runs follow the same rule.
The document.info params text, docs/mcp.md and the progress line
describe it.
object.distort warped the anchors of paths, groups, compound paths and
clip groups but left their bounding-box angle as it was, so a rotated
path kept its old box after a warp and an upright one stayed square to
the page however far the warp tilted it. Text, images and symbol
instances already took the angle of the warp's affine approximation.
- Each path, group, compound path and clip group now takes the angle
the warp turns its x axis to at its own center before the warp, the
members of a warped group and the art of a linked opacity mask each
at their own. An unlinked mask's art keeps its angle. Mirroring keeps
the nearer angle and a collapsing warp keeps the old one. Reset
Bounding Box squares the boxes again.
- Each subtree's bounds are measured once while it is warped.
- Corners that equal the corners of the box being warped change
nothing and add no undo step.
Docs: object.distort params text, docs/mcp.md, docs/target-app-parity.md.
Spreading two keys along a spine whose length overflows splits nothing, so
it returned the end anchors, and a closed spine was then opened at its start.
spread_anchors now rejects a length that is not finite before it makes any
target, so two keys get no anchors and the path stays as it was, as for
three or more keys.
Several threaded-text paths in threads.rs searched a list once for each
frame, so their work grew with the product of the frame count and the
list length:
- leave_threads (Create Outlines, Expand Appearance, Rasterize, Flatten
Transparency, symbol replacement, opening with text as outlines)
searched the threads for each frame that stayed and ran continued over
that frame's whole thread again, about n^2 frame visits for a thread of
n frames. It now marks each changed thread with one continued call,
and keeps the frames that left in a map instead of a list.
- has_thread_selection (the availability of Release Selection and Remove
Threading), with_threads, keep_shown_text and cut_to_shown, release and
remove test ids against a hash set instead of a list.
- create finds the thread of each selected frame from a table built
once, and each frame still takes in the first thread not yet merged
that lists it.
- release writes each thread back at its own index instead of searching
the thread list for an equal thread.
Marks, thread membership, frame order, text and overflow are unchanged.
A 256-frame thread that loses one frame visited 65,025 frames before
this change and at most 256 now. Six tests on many threads count the set
lookups: with a list in place of the set they make 67,650 to 258,226.
Another test compares the marks with a scan of the frames before each
one.
With repeated names numbered when a brush library is read, a stroke that
names a repeated brush draws the first brush of that name, and a later
brush of the name is read as "Name 2". Three tests built a library with a
repeated name and expected the later brush:
- Document Info lists a name two stored brushes share with the first
brush's type (art), not the last one's.
- The two render memory tests add their own `Dots` scatter brush beside
the default library's `Dots`. They now remove the default `Dots` first,
so the stroke draws their brush.
- The .ai writer test that leaves out a Pattern brush "X" gets the
Calligraphic brush after it back as "X 2", the name it had in the
document.
Giving the keys of a blend whose spine came from Replace Spine anchors of
their own flattened and measured the whole spine again for every key,
then walked its segments one by one. 1,000 keys on a spine of 4,096
segments flattened it 999 times.
The spine is now flattened once into one buffer. A split flattens only
its two halves and measures the segments after it again, in order, only
as far as the next key, so the arc lengths are summed as before and the
keys get the same anchors. A binary search finds each key's segment.
A spine so large that its length overflows no longer gets an anchor
split at a parameter that is not a number: spreading keys along it
returns nothing and leaves it as it was.
With the selection as its scope, a fix that removes or unwraps objects
(Needless Groups, Stray Points, Empty Text and the others) kept the list
of selected objects current by searching it and splicing it once for each
object it changed. Removing many selected objects above a long run of
others took time that grew with the square of their number.
The list is now rebuilt in one walk after each pass: each changed object
gives way to what replaces it, in its place, and the others keep their
order. The objects found, the counts, the document and the undo step are
the same as before.
Empty Artboards deleted each empty artboard with its own Vec::remove and
walked every ruler guide again for each one. It now drops the empty
artboards in one walk and deletes their guides in one more. The same
artboards and guides are deleted as before.
Tests cover the order after unwrapping nested and empty groups, Fix All
reaching an unwrapped member on a later round, the rebuild itself, the
number of objects looked at for 150 and 300 removals, and the number of
walks over the artboards and guides when 40 empty artboards are deleted.
`span_at` read the end of a closed subpath's closing segment as past
every stretch, so a click that landed there could leave the path
unchanged. It now reads that place as the first anchor.
BFlat