With the Flare tool, an Alt-click (Option-click on macOS) now draws a flare
centered at the click at once, with the tool's current options, instead of
opening the Flare Tool Options. Alt counts when it is held as the button goes
down; a drag with Alt draws the flare in two stages as before. While the
perspective grid shows with an active plane, the flare is drawn on that plane
from the click, as the dialog's OK draws it, and a click the plane cannot take
reports the error. A plain click still opens the dialog.
A trim checked and changed the selection once for every trimmed path
and every piece, and each check or removal went through all selected
objects, so a trim of thousands of selected paths spent time
quadratic in their number on the selection alone. The command now
reads the selection through a set and removes and adds objects once
at the end. The resulting selection is the same as before, including
object order, the key object and key anchor, the direct-selected
anchors and segments, and the fields that a change to the object
selection clears.
The crossing search took a crossing only where the straight pieces it
splits two curves into actually cross. A line trimmed at a curve ends
where that search found the crossing, up to about 0.0001 pt off the
curve. Once the curve was cut as well, its segments were split
differently, the line's end could fall just short of them, and the end
no longer counted as a place on the curve. Trim at Crossings then
removed the curve past that line, to the next crossing or the whole
path.
An open end of a path that lies within 0.001 pt of another path now
meets it at the nearest point there, at any angle. Corners of a path
and closed paths keep the exact test, and paths that lie along each
other still don't cross. Trim at Crossings and its hover, Add Points
at Crossings (with and without cut) and the Join tool's trim and join
of crossing paths use this search. The params text of Trim at
Crossings and Add Points at Crossings and docs/mcp.md say so.
A Cmd-click with the Eyedropper lends Direct Selection or the Lasso for
the gesture. The lent tool started with no mesh points picked, and the
Eyedropper came back with its old picks, so a point picked that way was
not sampled into and a Shift-click could not add to the picks.
The lent tool now starts from the Eyedropper's picked points, and the
points it leaves picked go back to the Eyedropper, none when it unpicked
them all. A Cmd-click lent to Group Selection, which picks no mesh
points, leaves the Eyedropper's picks as they were.
Mesh points picked with the Mesh tool, Direct Selection or the Lasso
were dropped when switching to the Eyedropper, so a sample from an
image went to the whole mesh object instead of the picked points.
The picks now carry over to the Eyedropper, which shows them filled,
and go back to the Mesh tool, Direct Selection or a Lasso chosen next.
While a selected, editable mesh has points picked, a click or
Shift-click on an image or other art samples the color under the
pointer into those points alone (one undo step, none when the color is
unchanged); other selected objects keep their paint. A click on the
mesh itself, Alt-click and Shift+Alt-click are unchanged.
Package copied a placed EPS file as it is, but did not read the native
document a VectorCraft EPS file carries, so the images that document
links stayed out of the package. Placing the EPS reads that document
and looks for its linked files, so once the original folder was gone
the packaged copy showed only the previews.
Package now reads the carried document of a placed EPS or PostScript
file the way it reads a .ai file's editing data, and copies the files
it links beside the EPS file's copy, which finds them there by name.
Cut Path at Selected Anchor Points cleared the selection and then added
each cut path in turn, and each add searched the paths already added, so
selecting the cut ends took time that grows with the square of the number
of cut paths. The command now selects them with one Selection::set call,
which skips repeats through a hash set and keeps the order in which each
path first appears, and then adds each cut end's anchor. The selection is
the same in every field; a new test compares it with the one-by-one build
on open, closed and multi-subpath cuts, long lists and an empty list.
The gain shows on large documents. With the 8 middle anchors of each of
5,000 open paths of 10 anchors selected, building the selection of 40,000
cut ends takes 9.7 ms instead of 232 ms of CPU time, and with 2,000 paths
3.2 ms instead of 39 ms (medians of 7 runs of the new
cut_path_select_bench example, both ways in one process).
The Print dialog's General section (and the print preset editor) gets a
link button after Scale W and H. While it is on, typing either percentage
sets the other, and turning it on sets H to W. It is on for new settings,
dimmed with the fields unless Scaling is Custom or a tiling choice,
toggled with a click, Space or Enter, and screen readers hear its name
and on or off state. W and H now take 1 to 1000 %.
The link is saved as scale.linked with the other print settings, in the
document and in print presets. A scale saved without it reads as unlinked,
so earlier settings print as they did. print.setup, print.preview,
file.print and print.presets.save apply the same rules: one percentage
sets the other while linked, a different width and height without a
linked key turn the link off, and with linked: true (or linked: null over
a linked scale) they are an error.
Under Include Bleed, a PDF row that sets no bleed now keeps the bleed values a saved PDF preset writes out and turns on useDocument, instead of replacing the whole bleed object. Its file key then matches a row of the same preset that asks for the document's bleed, so the two rows write one PDF.
links.unembed now reads `ids` of one image as `id`, and `id` with
`folder` as `ids` of that image, as the other commands that name both
keys do. `id` with `ids`, and `path` with `folder`, are refused with the
shared "give `a` or `b`, not both" error, in a replayed step too, and the
catalog rows for the old combined message are removed.
The export walks for type and fonts and the artboard crop shared one
SymbolBoxes for every object. A SymbolBoxes also holds measuring budgets
(effect reaches, 9-slice placements), so in a large document objects late
in a walk could be measured by estimates where main measured them exactly.
SymbolBoxes::unmeasured makes new boxes that share the name index, and each
object is measured in boxes of its own again.
- The live-object walk (render/src/live.rs) finds symbols through the
frame's SymbolBoxes.
- SVG export's filter regions measure each object in unmeasured boxes, and
the region boxes share the writer's index.
- SVG import keeps each imported symbol's index, so a symbol whose art
holds instances no longer scans or indexes the list again.
- Object › Rasterize's clipping mask is left as it is on main, with
drawn_outline_with removed.
- The symbol-read-count feature of vectorcraft-doc gives other crates'
tests vectorcraft_doc::symbols_read. New tests check that a frame, an SVG
export, a PDF export, the type walk and the artboard crop read the symbol
list once among 2,000 symbols and 300 instances with drop shadows.