The canvas keeps a hairline visible by raising a stroke's weight to half a device pixel. A width profile multiplies that weight, so a 1 pt stroke with a 6x width point swelled to twice its size at 25% zoom and shrank back on zooming in. The minimum now goes to the profile's widest point instead; plain strokes are unchanged.
Pattern tiles were drawn with the default render options, so an image
whose pixels cannot be decoded drew nothing inside a pattern fill even
when the render marks undecoded images, as the canvas does. Tiles now
follow the caller's mark_undecoded_images setting, and the tile cache
keys on it.
An opacity mask with such a pattern now treats the mark as part of the
mask, as it does for an undecodable image outside a pattern.
The renderer drew the root node of an opacity mask's art, of placed symbol
art, of an art brush piece, of an Image Trace source image and of a node
thumbnail without its own opacity mask. Make Opacity Mask uses the top
object as the mask art, so a mask made from an object that already had an
opacity mask ignored that inner mask. These roots now draw through the
same masked group as any other node, still drawn whatever their visibility
and bounds, and an incomplete inner mask still leaves out the outer object.
The renderer sized its worker pool from available_parallelism, which on
Windows counts every processor in the machine, also when the process's
affinity mask allows fewer. A process allowed one core started 8 workers,
and the renderer's flush spun on that core while the workers got no time,
so renders crawled or stalled.
The desktop app and vectorcraft-cli now read the affinity mask at startup
(src/windows_cores.rs, shared) and pass the count to the renderer with
vectorcraft_render::set_usable_cores. With one usable core the renderer
renders on the calling thread.
With undecoded-image marks off, an image in an opacity mask's art that
failed to draw left a hole in the mask. Without Clip, the area outside
the mask art counts as white, so the masked object showed in full.
The renderer now notes an image that fails to draw while mask art is
drawn and leaves the masked object out, as PDF export does since #4765.
This covers images in pattern tiles and in cached effect and shadow
rasters, which remember the failure and report it again when they are
reused. Nested masks pass the failure outward, and View Opacity Mask
shows no coverage for such a mask. An image still waiting to be decoded,
or drawn with the undecoded-image mark, keeps today's behavior.
A knockout element with one opaque Normal solid fill (a path or
compound, no mask, effects or isolation) is drawn once under
AntiAlias::None, including the internal hard-edge frame of Art
Optimized supersampled exports, instead of being erased and added back
through per-element backdrop layers. Analytical antialiasing keeps the
existing compositor and rounding.
Exact-pixel tests compare against the general knockout path for paths
and compound holes across precision, supersampling, translucent parents
and copied backdrops, and check that the backdrop rasters are skipped;
translucent fills keep their knockout compositing.
# Conflicts:
# docs/roadmap.md
# docs/target-app-parity.md
Highlight Alternate Glyphs compares each glyph with required shaping
alone, which turns liga and calt off. Older Arabic fonts (such as the
legacy Arial of the Microsoft core fonts) put the mandatory lam-alef
ligature in liga instead of rlig, so every lam-alef was highlighted.
A lam + alef cluster (any alef form, marks allowed) is now never an
alternate candidate, whatever feature the font uses.
Adds a test font built in code (lam-alef under liga or rlig) and
host-independent regressions. Also two equivalent one-line repairs
strict Clippy needs on main: a nonminimal_bool in the Path Precision
Tools inspect check and is_none() in a gradient panel test.
Effect › Color Effects › Duotone… (addon.color.duotone, also in Object ›
Plug-ins › Color Effects) prints the grey-equivalent tone of each colour
with one to four inks, each with its own tone curve. It is a colour
adjustment like the add-on's others, so it stacks with them and works as
a filter through addon.color.filter.
- One tone of a swatch the document has gives a tint of that swatch,
linked, so a spot tone separates onto its plate; several tones give
the CMYK mix of their tints (the inks overprint, so they multiply).
- Colour adjustments now map paints with their swatch links
(ColorMap::apply_link); the existing adjustments behave as before.
- Skip Objects with Spot Colors leaves spot-linked colours as they are.
- The dialog has a row per tone and the picked tone's curve (the Curves
graph, reused), its points and Reset to Linear / Inverse.
Print Ink Fixes (addon.color.printFixes, Object › Plug-ins › Color
Effects › Print Ink Fixes…) scales CMYK colours over a total ink limit
down to it, removes white overprint and sets black overprint (Overprint
Black's logic) in a CMYK document, as one undo step, and reports what
changed. RGB documents are refused.
The panel opens both under Inks.
Deferred: an opaque undercoat, tones for type only, extra gradient stops
for smoother tints, curve flip and all-100 % buttons.
Effect › Color Effects › Vector Halftone… (addon.color.halftone, also in
Object › Plug-ins › Color Effects) draws its object as the halftone dots
Object › Vector Halftone makes, live: mono (one dot colour, or each dot in
the art's colour under it) or CMYK screens, the five dot shapes,
frequency, angle, invert and clip to the art's outline.
- The screens, dots and caps move to vectorcraft_effects::addons::halftone,
shared by the command (its behaviour and tests unchanged) and the effect.
- The renderer samples the art (vectorcraft_render::addons::halftone) and
keeps the dots by the object's content and what it reads from the
document, so unchanged art isn't sampled again; past 100000 cells per
screen the object draws as it is.
- Expand Appearance turns the object into its dots, and the SVG, PDF, EPS
and AI exporters write them as paths.
- Its dialog reuses the command's fields, without Keep Original.
Deferred: stochastic and radial patterns, a dot-gain curve, per-screen
separation settings, type or symbols as dots.
A layered .ai or Illustrator EPS whose editing data (embedded rasters,
mostly) unpacks past 512 MiB opened as its PDF part or printed page, in
one layer, with "it decompresses to more than 512 MB" (#4628 on 0.17.0
hit the older 256 MiB limit the same way).
The limit is now an eighth of installed memory, from 512 MiB (what every
64-bit build read before) to 2 GiB: 1 GiB on an 8 GiB machine, 2 GiB on
16 GiB or more. Reading takes about as much memory again as the data
(1.1 GB of data with 24 embedded 4096 x 4096 images peaked 1.1 GB above
the start), so the ceiling keeps a small file that unpacks without end
from taking a large machine's memory. 32-bit and web builds keep 256 MiB.
vectorcraft_render::installed_memory exposes the reading the image cache
already sizes itself from.
The two nested_instances_..._culled_in_bounded_time tests in the
renderer's tests checked that a frame with 2^30 differently placed
symbol instances finishes within 30 s of wall-clock time. On a loaded
machine the frame took longer than that, and both tests failed. The
time bound also passed when the frame did far more work than it should:
with the 4,096 limit on measured placements raised to 65,536, the frame
measured symbol art 131,118 times and finished in under 2 s.
SymbolBoxes now counts how many times measure_art runs a measure on
symbol art (art_measured), and FrameStats::symbol_art holds the count
for the frame's symbol boxes. The tests assert that the frame measures
symbol art at most 20,000 times. It measures it 8,234 times, and 4,160
times when each instance has a Transform effect of its own. The wait
for the frame is 300 s, as a guard against a frame that never ends.
A new test in vectorcraft-doc checks that each symbol's art box is
measured once however many instances of it are measured.
On a machine busy with other builds, four tests each failed once and
passed when run again alone.
The two render tests that count the rasters a drag moves draw content
painted with a pattern or showing a symbol, whose raster key holds the
font generation. The fonts change even while a test holds
fonts_steady(): the first lookup by family name catalogs the installed
fonts, and resolving a cataloged family or a fallback character loads a
face. When that happened between the two frames of a drag, the count
was 0. The new with_fonts_unchanged runs a step up to 100 times, until
the generation is the same after a run as before it, and the two tests
take their frames from it. The second render test,
a_moved_group_draws_the_shadow_of_the_instance_art_in_it_while_interacting,
was not seen failing but has the same fault.
The screenshot test's allow_due moved the capture's due time but not
its 8 s wall-clock deadline, so on a slow run the reply was the timeout
error. allow_due now also sets the deadline to f64::MAX; the two timeout
tests set it in the past themselves.
The persistent control server test gave the server 2 s to wait for
each request, 120 ms to read a line and 150 ms to write a reply. It now
gives the line and the reply 1 s each and the wait 30 s. The handler
sleeps 1.25 s, longer than both budgets, so a deadline kept from the
first request still runs out.
The idle expiry test read until the server closed the connection after
its 150 ms idle budget, and the client gave up after 3 s, so a pause of
about 3 s in the server's thread failed it with TimedOut. Every test in
control_server/tests.rs that waits for the server to reply, to close a
connection or to return its slot now waits up to WAIT (20 s), and
line_timeouts() gives the wait for a request and the reply 30 s. The
budgets under test stay 120 and 150 ms (1 s in the persistent test).
Envelope Options stored Anti-Alias and Preserve Shape Using, but nothing that
draws an image read them: a distorted image was always smoothed and always
cut into clip groups.
- With Anti-Alias off, each piece of a distorted image is marked pixelated
(saved as `pixelated`). The renderer draws it with nearest-neighbor
sampling and without mips, SVG export writes `image-rendering: pixelated`,
and PDF export writes it without /Interpolate.
- With Preserve Shape Using Transparency, each piece is the image with a
white opacity mask in the piece's shape instead of a clip group. The
renderer paints an image whose mask is one plain white path straight into
that path, as it does for an image in a clip group, when the path lies
inside the image's frame.
Docs: docs/mcp.md, docs/target-app-parity.md.
On screen, placed documents (PDF, .ai and .vectorcraft) draw from
bitmaps made in the background. Each bitmap that became ready changed
a generation in the canvas key, so the canvas rendered the whole frame
again. The bitmaps share a 512 MiB budget, and making one dropped the
least recently drawn others, even those on screen: with more than that
on screen, every frame dropped a bitmap it drew and asked for it again,
without end. A bitmap that could not be made (the file can't be read
and there is no preview) also changed the generation and was asked for
again in every frame.
When a bitmap is ready, placed_document::made_between names its file,
and the canvas renders again only where the objects showing it are
(placed_area), as it does after an edit, when the frame shown was
rendered on the UI thread and that part is at most half the frame.
Art of a symbol, a graph design, a pattern or a brush that shows the
file still renders the whole frame.
Bitmaps drawn or made since a frame began are not dropped to make room
for the ones it asks for. They stay past the budget up to 2 GiB. A
bitmap that would take them past 2 GiB is not kept, and its file draws
exactly at that zoom. A bitmap that can't be made shows the
placeholder, does not change the generation, and is not asked for
again.
Docs: docs/target-app-parity.md.
A PNG, JPEG, GIF or WebP image that takes more than the decode
allowance as RGBA (2 GiB, 512 MiB in the web app), such as a 24000 ×
24000 JPEG, was placed with the warning "too large to draw" and then
drew as a crossed gray box on the canvas and was left out of raster
exports. A 16-bit PNG whose decoded samples were past the allowance
was shown the same way, with the same warning, although its RGBA fits.
The renderer now decodes such an image at the largest size, keeping its
proportions, that fits the allowance as RGBA. Each pixel of the result
is the mean of the image's pixels it covers. A PNG that isn't
interlaced is read a row at a time; other images are decoded whole when
their decoded samples fit the allowance. A 16-bit PNG that isn't
interlaced and whose RGBA fits is read a row at a time at its own size.
An image of more pixels than the allowance has bytes still draws as the
crossed box, and so does one whose decoded samples are past the
allowance, other than a PNG that isn't interlaced. The renderer
decodes an image at reduced size only while the result and the memory
reading it takes (the image decoded whole, or the PNG reader's row
buffers) fit beside the images being painted; otherwise the image is
not decoded on that render (on the canvas it draws as the crossed box)
and is tried again on the next render. The document keeps the whole
file for saving and export. A placed document's saved preview past the
allowance is decoded the same way, at the size it is drawn at or at the
reduced size when that is smaller.
Open, Place and the Links panel say "an image of W × H pixels is shown
at reduced size" for an image drawn smaller. They say nothing for an
image drawn at its own size, and they say "too large to draw" for one
that is not drawn at reduced size.
Docs: docs/mcp.md, docs/target-app-parity.md (the crossed box's line
now names an image over the allowance that can't be drawn at reduced
size).
A Drop Shadow, Gaussian Blur, Feather or Photoshop-style filter raster
too large to keep whole is cropped to the view. It was drawn again
only while a later frame's rect, panned back, lay inside the frame it
was made in, so every pan that showed new pixels made it again. For a
Drop Shadow on a 6000 x 4000 pt image in a 2560 x 1440 view at 100% to
400%, a pan took 60 to 84 ms; with Gaussian Blur or Feather, a pan took
158 to 217 ms.
When such a raster is made again at the zoom and rotation of the one
kept for the object, it now also covers a margin around the view: up
to half the view's longer side, while the raster stays within the
8.4 million pixel budget at full size (441 pixels on each side of a
2560 x 1440 view for a 20 pt blur at 100%). Later pans that keep the
view inside the margin draw that raster. The first frame at a zoom, a
zoom step and the frames of a drag make the raster without the margin,
as before. A shadow made over an earlier cropped shadow covers only
the rect the earlier one was made for.
A kept raster is drawn again only after a pan by whole device pixels,
but the wheel, a trackpad and the scroll bars scrolled the view by
amounts that are not whole pixels. The canvas now moves the view by
whole device pixels when it scrolls or is dragged. A scroll by the wheel, a trackpad, a touch screen or a scroll
bar is rounded, and the part of a pixel left over is added to the next
scroll of the same view; it is dropped after a zoom, a turn, Fit or a
document switch. A Hand tool, Space or middle-button drag and a scroll
bar thumb drag round the distance from where the drag began.
The silhouette of a shadow or glow raster is now drawn with art culled
against the raster's own area instead of the view, as filtered content
already is. Before, a group member outside the view cast no shadow
into a raster kept whole, so a pan that showed the member drew it
without its shadow. A cropped raster kept across the margin would have
done the same. The moved silhouette that a drag compares is drawn the
same way.
Docs: docs/target-app-parity.md.
The renderers share decoded images within a budget of an eighth of the
installed memory, at most 1 GiB. The installed memory was read only on
Linux, so on Windows, macOS and FreeBSD the budget was always 1 GiB,
also on machines with less than 8 GiB.
The installed memory is now read with GlobalMemoryStatusEx on Windows
(through winsafe, which vectorcraft-cli already uses), sysctl
hw.memsize on macOS and sysctl hw.physmem on FreeBSD. The budget rule
is unchanged: an eighth of the installed memory, at most 1 GiB, 1 GiB
where the memory can't be read, and 256 MiB in the web build.
An image whose pixels can't be decoded, such as a PNG cut short by an
interrupted download, drew nothing on the canvas and the status bar
said nothing. Every render decoded it again, so each frame that drew it
paid for a decode that fails.
The renderer now notes such an image by its content and does not decode
it again. Each render still counts it in decode_failures, so an export
reports it as before. An image refused only because the images being
drawn leave it no room within twice the decoding allowance is not
noted, and the next render tries it again.
On the canvas the image draws as a light gray box crossed from corner
to corner, and the status bar says once for each image that it couldn't
be decoded. Exports still draw nothing there.
Docs: docs/target-app-parity.md.
The brush cache dropped brush art that can show a symbol or a pattern
only when the art of the document's symbols or patterns changed. Renaming
one, for example swapping the names of two symbols, left their art as it
was, so a brush whose art holds an instance of one of those names kept
drawing the old symbol. The cache now compares the names along with the
art.
Replace glob imports (`use super::*`, `use crate::*`, `use sig::*`) with
the names each file uses, in scriptcheck, pdf, render, text and ui-egui.
No code changes.
Highlight Substituted Fonts is now a preference in Preferences › Type,
next to Highlight Alternate Glyphs, on by default and saved with the
other preferences (`prefs.set highlightSubstitutedFonts`). The canvas
reads it through `RenderOptions::highlight_substituted_fonts` for every
open document; a document's own setup no longer decides it. File ›
Document Setup no longer shows the checkbox and keeps Highlight
Substituted Glyphs.
A file saved with Document Setup's Highlight Substituted Fonts off now
shows the pink highlight when it opens, while the preference is on.
The file still opens with the stored value, and saving it again keeps
that value.
`document.setup` no longer reports `highlightSubstitutedFonts`. It
ignores the key and adds a `warnings` entry that names `prefs.set` to
its reply, so an older reply passed back still applies its other keys.
The raster cache of type under a raster effect now also redraws when
Highlight Alternate Glyphs changes.
The test checked only that each outline had a dark pixel, which a line
two or more pixels wide also has. It now counts the dark pixels across
the path, a stem of the type and the guide, and expects one, with and
without supersampling.
After imageTrace.setView, screenshot and render_preview still drew the
tracing result, because ui.render renders with the export options and only
the canvas turns on trace views. Only the app window showed the view.
ui.render now takes traceViews. With traceViews: true, Image Trace objects
draw in the view set with imageTrace.setView, as on the canvas, both
headless and in the running app. screenshot and render_preview take
traceViews and pass it on; without it they draw the tracing result, as
exports do. The trace-and-style prompt names the argument.
A supersampled render drew Outline view's lines, type outlines and dotted
guides as little as a quarter of an output pixel wide, so they came out
pale gray. The renderer now scales those widths and the guides' dashes by
the supersampling factor in the larger frame, so the lines stay one output
pixel wide.
Docs: docs/mcp.md (render_preview, screenshot), docs/control-protocol.md
(ui.render).
`tests_alternate`'s Arabic tests look up installed fonts by name
(`FontDb::face`), which loads each one into the global database and so
changes its generation, without `tests_fontchange::adding_fonts()`.
When that lands in the first frame of
`tests_charstroke::character_stroke_outlines_are_kept_between_frames`,
the second frame lays the type out again and expands the stroke again,
and the test fails (3 runs in 30 on macOS, where Arial and Geeza Pro
are installed; where no Arabic font is, those tests return early and
nothing changes). They now hold the lock, and its documentation says
looking up an installed font counts as changing the fonts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The outlines of a run's character stroke are kept with the glyph geometry the renderer keeps for the type, in a byte-limited cache swept by frame, so type that changes in every frame (an animation) frees them within a few frames.
The Image Trace panel had no way to compare a trace with its source
image except changing View and changing it back.
The panel now has an eye button to the right of View, enabled with an
Image Trace object selected. While it is held down, the canvas draws the
selected Image Trace objects with their source image, whatever their
View, and draws their View again on release. The document, its View and
the undo history are unchanged. The renderer takes the objects in
RenderOptions::trace_source, which is also part of the canvas frame key
and the raster-effect cache key.
To check what it kept for a live object, or for effect art made from
live objects, type or symbols, the renderer walked the object in every
frame to hash the symbols, patterns and placed files it shows, and it
looked up each symbol instance it found by name in the list of every
symbol. At 1000 x 1000 pixels, the walk took most of the frame on a
page of 2,000 envelopes around instances of the last 2,000 of 10,000
symbols: 83 ms a frame when every frame walked them.
The renderer now keeps what the walk found with the node, which it
holds weakly, and with the document's symbols and patterns as the walk
read them. A later frame makes the key from what was kept while the
node and the document's symbols and patterns stay the same, and walks
the node again when any of them changes. A walk that finds a placed
file is not kept. An evaluation that the live cache does not keep, such
as a thumbnail's copy, is no longer walked. The same page now takes
30 ms a frame, with the same pixels.
Docs: docs/target-app-parity.md
Brush art leaves out the symbol instances and pattern paints whose art
draws a brushed stroke. The brush cache now records whether a brush's art
or the stroke's paint can show a symbol or a pattern, and drops those
entries when the art of the document's symbols or patterns changes, so an
edit that makes a symbol or pattern start or stop drawing a brushed stroke
draws as a new renderer does.
The parity line gives the median and fastest-frame figures of the
benchmark for both documents it names.
The renderer kept brush art per node id and stroke. The copies drawn
for symbol instances, inline graphics and repeat copies share their
art's ids, so they replaced each other's brush art and it was made
again in every frame. The pieces were drawn as plain nodes, so brush
art that is a live object, such as an envelope that a scatter brush
places, was evaluated again in every frame, for a path on the page too.
The brush cache now also keys brush art by the stroked path, so each
copy keeps its own. Kept pieces are drawn through the renderer's
caches: their stroke outlines, live-effect geometry and live object
evaluations are kept between frames. Each frame draws what a new
renderer draws. On one thread at 1000 x 1000 pixels, 2,000 paths on
the page whose scatter brush art is an envelope took 400 ms a frame on
main and take 94 ms, and the fastest frames of 2,000 instances of a
symbol whose art is a brushed path went from 16 ms to 12 ms, with the
same pixels.
Docs: docs/target-app-parity.md
CodeRabbit findings #32, #38, #41, #44, #102:
- A frame (and a PDF export) draws symbol instances holding at most 16M objects together
(`vectorcraft_doc::MAX_SYMBOL_NODES`): nested instances multiplied past the depth limits.
- A dotted guide whose control polygon would make more than 100,000 dashes is drawn solid.
- Groups blending with their backdrop, drawn twice each, are drawn isolated past 4 nested.
- An image is decoded only while it and the images paints hold fit twice the decode allowance.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* CI: line tables only for workspace crates, no debug info for dependencies
* CI: one integration-test binary per crate
Each crate's tests/*.rs files become modules of tests/it/main.rs. Files that change
process-wide state (Color Settings, the font database, installed plug-ins, the working
directory), the fuzz sweeps and the test targets the Affinity corpus workflow names stay
their own binaries.
* Docs: name the integration tests' new paths
* CI: xtask test with nextest and doctests, one wasm check, clippy --keep-going
cargo xtask ci runs the tests with cargo nextest (cargo test when it is not installed),
builds them with --lib --bins --tests, runs the doctests as their own step, and checks
the wasm crates in one cargo check --keep-going call.
Taha-Mahmoodi