untranslated_report lists 167 status and error messages that no
complete-message catalog translates. zh-hans.tsv gains 114 @msg rows for
the ones no open pull request covers: 3D and Revolve limits, image export
and encoding errors, color profiles, color books and swatch exchange
files, binary DXF and DWG import, path operations, swatch, graphic style
and action set libraries, and zip, script and regular expression errors.
Placeholders, backticked names, command ids and listed values stay as the
English writes them.
The other 53 (composite fonts, SVG filters, read roots and graph design)
are left to #7211, #7378 and #7405. docs/localization-parity.md and
docs/target-app-parity.md give the new counts: 3,554 @msg rows, and 55
messages the report lists for zh-hans, all of them translated by open
pull requests.
A graph keeps paints for its series and labels, and character and
paragraph style names for its labels, which each redraw gives the
marks and labels. Copy did not follow them, so a graph pasted into
another document left those swatches and styles behind, and a paste
that renamed them left the graph pointing at the old names.
Copy now collects them and paste relinks them, as it does for type.
Paste without Formatting drops the labels' style names, so it still
leaves text styles behind.
On the core Wayland clipboard, a read waited for the app that copied to
close its pipe, and a send wrote until the app that pastes took all of
the data. An app that left its pipe open and stopped held the one read
worker, so later pastes failed at once, or held one of the eight send
threads until it closed its end.
Reads and sends now use the pipe nonblocking and wait for it with poll.
When no byte moves for 2 s (VectorCraft's own limit, the existing text
wait) the read fails without pasting partial data and the send stops;
both close their pipe. The wait starts again after each byte that moves,
so a large transfer that keeps going is not cut short.
Quitting can wait up to 27 s for the clipboard handles to let go before
eframe closes the windows. winit cannot hide a native Wayland window, so
the main window stayed on screen for that wait, and the X11 windows of
floating panels and documents were not hidden at all.
- A native Wayland main window is unmapped by attaching no buffer to its
surface. The window and its connection stay open for the Wayland
clipboard. An X11 main window, XWayland included, is hidden as before.
- On Linux the X11 windows of floating panels and documents are recorded
as eframe creates and drops them (viewport hooks), and quitting unmaps
them through a second X11 connection before the wait.
- The new module apps/vectorcraft/src/quit_windows.rs replaces
app::hide_before_teardown, with tests of which windows are hidden or
unmapped. portal_save's Wayland surface helper is shared with it.
The comment on the zigzag cases said the open zigzag grown by 1.5 leaves
stroker loops whose faces the distance test drops. With the turned branch's
distance test measuring points in the turned frame, or dropping no face at
all, the three zigzag cases still pass and only the plate case fails (201
contours instead of 101). The comment now says what the zigzag cases check:
one turn for the low path, none for the path stood on end, and the same
region from both. The plate comment says its narrow holes are dropped by
the distance test measured at points turned back into the plate's frame.
A drag of art across the document tabs ended whenever the source
document's revision changed. Since #7173, bringing a document with other
assigned color profiles to the front raises every open document's
revision so that each redraws in the new profiles, so holding art over
such a tab switched to it and then canceled the drag before the drop.
The drag now keeps the source document it left and ends only when that
document's content changes (DocState::same_content). A redraw keeps it,
and so does type measured again when the source comes back to the
front, which replaces the document without editing it.
A raster effect (Drop Shadow, Outer Glow, Gaussian Blur and the other
effects that paint pixels) on a symbol instance's topmost fill now draws
on each filled shape of the symbol's art, and one on its topmost stroke
draws on each stroked shape, after any geometry effects listed before
it. The same rules as for geometry effects apply: only a visible fill or
stroke with an opacity above 0 counts, a group's own fills and strokes
in the art take the effect, and an instance inside the art passes it on.
The canvas, Use Preview Bounds and the drawn bounds, SVG and PDF export
draw it, and Object > Expand Appearance is enabled for such an instance:
a shadow or glow becomes an image under the instance, and a blur,
feather or inner glow replaces it with the image, in one undo step.
Existing documents with a raster effect on a symbol instance's fill or
stroke draw differently: before, the effect drew nothing.
Character panel > Tsume, shown with Show East Asian Options, takes away a
percentage of the blank space on each side of a character's ink, inside
its own width, in horizontal type. The value is from 0 to 100, 0 by
default. 50 halves that space and 100 sets the ink against the
characters before and after it; the glyph keeps its shape. A character
without ink, such as a space, keeps its width, and a ligature or a
letter with its marks takes the value of its first character. The
font's pair kerning, tracking, manual kerning and Insert Aki are added
after it, and justification keeps it. Vertical type keeps its spacing.
- text.setFormat and text.setRangeStyle take `tsume`; a value that is
not a number or is outside 0 to 100 is refused, naming the key.
Character styles carry it, and a file holding a value outside the
range opens with the nearer end.
- The Character panel's Tsume field has 0, 25, 50, 75 and 100 in its
menu, and Reset Panel sets it to 0. Character Style Options shows it
in Advanced Character Formats.
- Live SVG text places each character Tsume moves, and type on a path
with Tsume is written as outlines.
Docs: docs/mcp.md (Tsume, Character Style Options fields),
docs/target-app-parity.md.
The centre-lines class gains a mixed-width body scene: one disc of radius
24 px with arms of 2, 3 and 5 px that start inside it, Stroke 6 px, and no
other ink. A test checks that the scene is one connected shape and that
both the anti-aliased and hard input trace as one fill and three strokes
that reach all three free ends.
docs/image-trace-quality.md's centre-lines section is measured again on
current main (24 scenes, two runs identical in every column but ms). The
new scene finds 100.0% of its arms with 99.3% / 98.3% of the disc filled;
mix now traces as 6+3 paths with 99.9% / 99.8% found and 7/7 ends, where
the earlier whole-blob rows showed 3+1, 57.3% and 6/7. All other rows
are unchanged except ms.
A paragraph whose characters differ in line metrics, in area type that is
not a rectangle or that wraps around an object, was broken line by line
with the Single-line Composer when the frame could not hold all of it.
The lines the frame holds are now composed with the whole paragraph: a
line the frame has no room for, and every line after it, is measured as
wide as the span of the last line set in the frame, less its indents.
Those lines overflow, or go on in the next threaded frame, which composes
them again at its own widths. A paragraph whose first line has no room,
and a search that passes its limits, still fall back to Single-line.
Docs: docs/mcp.md, docs/target-app-parity.md.
untranslated_report lists 167 messages under "messages, all 23 languages
checked". ru.tsv gains an @msg row for 114 of them, in one block at the end
of the file: image export and encoding errors, color profiles, color books
and swatch exchange files, binary DXF and DWG import, 3D and Revolve limits,
and script, zip, line and file errors. Placeholders, parameter names and
literal values stay as the English writes them.
The other 53 are left to the open pull requests that translate them
(composite fonts, SVG filters, read roots and graph design).
document.exportSelection run with no params or {} (a shortcut, or an
automation call) wrote the picked file straight away with each format's
defaults: a PDF used the default preset, so Save PDF's options, such as
writing long pages in points, could not be chosen. It now takes Export
As's route after the save panel: Save PDF without its Artboards
section, SVG Options without its artboards, the raster options, EPS
Options, DXF Options, Text Export Options with Selection Only on, or an
exporter plug-in's parameters. OK writes the selection alone, cropped
to it, with the chosen options. While a dialog is open the call returns
null, and OK returns what file.exportAs returns, without format or
bounds.
With a path or any option it still writes straight away.
- file.exportAs takes exportSelection: true and then writes the
document that document.exportSelection writes
(fileio::selection_doc), without artboard choices. The options
dialogs and the replace-file questions pass it on.
- Nothing selected stops before the save panel.
- SVG Options opened by Export Selection leaves the remembered Use
Artboards as it was.
- Exports through file.exportAs (Export As and the options dialogs)
apply Appearance of Black's rich black, as document.export does.
Docs: docs/mcp.md, docs/control-protocol.md, docs/target-app-parity.md.
A gradient stop that loses its link to a spot swatch kept the process
plates it printed with that ink, so it carried a DeviceN mix with no
ink. Paint::map_links, which deleting a swatch, recoloring and color
adjustments go through, now clears a stop's plates whenever it leaves
the stop without a link, as GradientStop::set_color(color, None)
already does. Library gradients and library character styles linked
to a swatch the document lacks unlink their stops the same way.
Docs: docs/target-app-parity.md.
A clipping path inside art cut to another clipping object is written as
its fills and its strokes. Its text wrap options went with the fills, so
when the outer region left nothing of the fills, the strokes were written
without them. The strokes now keep them unless a fill object was written.
A symbol definition that had no room left for its group under the object
limit returned without restoring the writer's drawn-art depth. Asked for
from art drawn for an object (a stained instance's art), the depth then
dropped below zero on the way out, which panics in debug builds. That
path now restores it as the others do.
The development guide and the progress doc now name all 23 languages in
COMPLETE_MESSAGES: Hindi was missing from the first, and Georgian, Hindi,
Turkish and Catalan from the second. The localization summary adds
Bosnian to the languages marked full that are not in the list.
The untranslated report listed 167 status and error messages that no
complete-message catalog translated. pt-br.tsv now has an @msg row for
the 114 of them that no open pull request already translates: messages
of the 3D and Revolve effects, the image encoders, color profiles, swatch
exchange files, color books and DXF import. Placeholders, parameter names
and literal values stay as the English writes them.
The other 53 (composite font, SVG filter, read root and graph design
messages) are translated by open #7211, #7378 and #7405; adding them here
too would make bundled_catalogs_are_consistent fail on a duplicate row
once those merge.
The continuous-tone check exempts a neutral or single-curve picture when
sixteen colors leave at most a fifth of the error of eight, but only
applied the exemption below the 0.1 ratio. Clean gray, sepia, single-hue
and duotone art of 10 to 16 evenly spaced levels lands above it, so it
was smoothed before tracing. The exemption now applies in both branches.
At 16 Colors, 12 sepia and 16 gray or sepia levels crossed by 1 px lines
keep every region the lines cut, and Automatic traces 16 clean sepia
levels in 9 colors instead of 4. Art in several unrelated colors is
unaffected.
path.setHandle checked for a no-op by running the handle edit on a copy
and comparing the whole anchor. On a smooth anchor the edit recomputes
the other handle, so when that handle was not exactly in line, or the
recomputed value rounded differently, a request for the handle's own
point realigned the other handle and recorded an undo step.
The check now compares the requested point with the handle itself, and
still lets independent turn a smooth anchor into a corner.
A clipping path with Roughen, Transform or another effect that reshapes
a path drew its own fill and stroke reshaped, but the canvas and raster
export still clipped the art by the plain path. SVG and PDF export bake
the effect into the path before writing the clip path, so the screen
and a saved SVG showed different art.
effects::clip_outline now applies the live geometry effects in the
clipping path before taking its region: the path's or compound path's
own effects, the effects of its fills and strokes (the region is then
the union of the paths they give, as an export bakes them), the effects
of a compound path's paths, of a group's visible members and of a
nested clip group's clipping path, shown or hidden, and a group's own
effects as an export evaluates them. Raster effects such as Gaussian
Blur leave the region as it is. The renderer, the engine commands that
measure a clip region (flatten, arrowheads, cutting) and the writers use
it; the writers receive the baked document, so their output for a path
with effects does not change.
The renderer measures a clip group by the region its clipping path's
effects give it, so a view or tile that shows only the moved part of
the clip still draws the clipped art, and it drops cached clip regions
that hold a plug-in effect when plug-ins change.
Two rules change in every export as well as on the canvas:
- When the clip group, or a group around it, has a geometry effect of
its own, the clipping path's effects apply first and its gradients
move with them. #7114 left the clipping path's effects live under the
group's effect, so they applied after it.
- Type used as a clipping path keeps its own effects when a group's
effect outlines it. #2073 and #2290 outlined it without them, to
match the plain clip the canvas drew.
When a recovery copy fails because a folder on the way to the recovery
folder is a file or is missing, the error names that folder, not the
recovery folder. Replies to a client confined to read roots hid only
the recovery folder itself, so the full path of that ancestor stayed
visible. A reply now also hides each ancestor of a recovery folder that
it writes as a whole path. The check compares text only and does not
look anything up on disk.
Follows #7287.
On main, the untranslated report lists 167 status and error messages
under "messages, all 23 languages checked": no complete-message catalog
has a row for them, so the Japanese interface shows them in English.
- ja.tsv gains 114 @msg rows for the listed messages: image export and
encoding errors, color profiles, color books and swatch exchange
files, binary DXF and DWG import, 3D and Revolve limits, and script,
zip, line and file errors. The 53 listed messages that #7211, #7378
and #7405 translate are left out.
- It gains 10 @msg rows for messages other catalogs translate and
Japanese lacked: the Banner Hems and Grommets errors, a parameter
range and the failed recovery save. The three script and RTF parsing
errors that #7314 adds are left out.
Placeholders, parameter names, command ids and literal values stay as
the English writes them, and the rows follow the catalog's Japanese
message conventions.
The script module docs, the script.run params text and the evidence line now
say that $.fileName in code that eval or Function compiled gives the file
being run, and the params text says that the empty name applies to code
compiled from a script given as text.
Two windows or processes using one Plug-ins folder could interleave their
writes: an install from one could leave its module with the other's grant,
a grant written for one plug-in could drop one written for another, and
plug-in settings and cache changes could lose each other's keys or pass the
cache caps together.
Each change to the folder now runs in one turn on disabled.lock, the file
the disabled list was already locked with:
- plugin.install takes the turn before it reads the stored grants and holds
it through the module, grant, put-back and disabled-list writes;
plugin.remove holds it through grant removal and storage deletion.
- Grant writes, the disabled list, loading the folder, and the plug-ins'
settings and cache changes read and write in a turn.
- When disabled.lock can't be locked, plugin.install and plugin.remove fail
and change nothing, plugin.setEnabled keeps its change in the process, and
loading the folder writes nothing to it.
Docs: docs/plugins.md, docs/mcp.md, the plugin.install and plugin.remove
params text, docs/target-app-parity.md.
The French row of docs/localization-parity.md gave the counts on 92594378a. After main was merged in, fr.tsv holds 9,166 rows, 3,436 of them @msg and 48 @msgplural, and the untranslated report lists 178 missing French messages: 11 for French alone and 167 that main added for every language.
An edge that two contours of an area type frame share lies inside the
frame, and lines already run across it. Inset Spacing still kept text
clear of it, so a frame of contours that meet along an edge left a gap
of twice the inset at the seam. The inset now applies only beside the
stretches of edges that bound the frame, counted by parity as the line
spans are. A negative text wrap offset no longer shrinks a wrap object
beside an edge two of its contours share.
The browser build stamped each snapshot of the Web Locks held with
Date.now() and judged its age against Date.now(). After the system
clock was set back, an old snapshot read as fresh, so a tab could
report no lock on an area another tab still held. Snapshots are now
stamped and aged with vectorcraft_ui_egui::now_ms(), and a snapshot
stamped later than now is not trusted.
craft-svg builds against an older vectorcraft-geom and vectorcraft-pathops by git, so the
graph holds two packages of each name and `cargo check -p vectorcraft-geom` is ambiguous.
The wasm step now passes each workspace crate's package id.
Expand as Wireframes reshaped the faces of a 3D effect with the geometry
effects after it, but the faces it built carried no stroke. An Outline
Stroke after the 3D effect then outlined a plain 1 pt stroke with miter
joins, while the canvas outlines the round-joined stroke each face is
drawn with: the seam stroke (half the facet's thickness, 0.05 to 0.75 pt;
0.75 pt on Revolve) or the 0.5 pt wire edge of the Classic Wireframe
surface. The wireframe's faces now carry that stroke, so the expanded
edges lie where the drawn outlines lie.
Linked images and placed documents in the art of the stored brushes (a
scatter or art brush's art, a pattern brush's tiles) are listed in the
Links panel after the other stored art, one row per linked or placed
object in each brush tile, with the brush as owner
(`{kind: brush, name, index, tile?}`). Check Links, the missing and
modified files reported on open, the link watch, Find in Folder, Relink
(including Relink All Instances), Update, Embed, Placement Options, Link
Info and Go To act on them, and their tooltip names the brush.
- vectorcraft-doc: `Document::mounted_art` holds art kept outside the
document's trees while a command looks it up by id; the stored-art
lookups, link visitors and `art_roots` include it, and
`number_mounted_art` gives its objects ids of their own.
- vectorcraft-brush: `stored_art::StoredArt` mounts the art of the
brushes that hold images or placed documents and writes back only the
entries whose art changed. Entries that do not read as a brush are left
as stored. `brush::store` keeps ids of their own in that art, so a row
keeps its id between commands.
- vectorcraft-engine: the links commands, opening a document, saving a
placed document's source, output of linked files and placed documents,
color masks after Update or Relink, and Package relinking go through
the mounted art.
Brush art rows follow the general Links rule as VectorCraft's own
application of it.
action.batch drops Open steps before it plays each file or data set, and
returned the per-file error of command.batch unchanged, so `batch step N`
counted from 0 within the filtered steps. A failed step is now named
``action step N (`id`) failed: reason``, N being its place in the
`commands` the caller sent, from 1, with skipped Open steps counted.
The batch's own open, data set choice and destination steps keep their
unnumbered errors, and command.batch messages are unchanged.
A prefs.import step, which command.batch refuses when it reaches that
step, rolling the batch back, is now refused before any file opens
instead of failing each file or data set the batch plays.
Docs: action.batch params text, docs/mcp.md, docs/target-app-parity.md.