egui places a popup by its size the frame before, so on the frame a late-measured name widens the
menu the filter text can stay where it was while the menu's rect already reaches past the window.
The helper now waits until both the filter and the menu's rect stay put, which fixes the
intermittent failure of a_menu_wider_than_its_field_stays_in_the_window (the menu ending 0.7 pt
past the window's edge).
Type on a path sits at fractions of its path's length. Scaling the path
non-uniformly changes where a fraction lands (CodeRabbit's case: M0 0 L100 0
L100 100 under diag(2, 1) moves the start bracket from (200, 0) to (150, 0)).
Path type now bakes the vertical scale into the character sizes and scales
its path uniformly by it, which keeps every fraction where it was; the
horizontal stretch stays in the transform.
A Swatch Options rename of the spot color base keeps it the base, also
as a command.batch step, and so do an Undo and a Redo of the rename.
Deleting the swatch or its color group, merging it into another, making
it a process color or closing its document ends the link, so a later
spot color of the same name does not become the base until it is chosen
again. The base keeps the color it showed.
A batch's Activate Document, New, Open, Close and Move Document steps
are followed, so steps in another document leave the link alone. When
the swatch goes in a way the panel does not follow, the link waits for
an Undo or Redo, and the next other command ends it.
Baking an object's scale into the character attributes removed that scale
from the type's transform, but the area frame (and a path type's path), which
is drawn through that transform too, kept its unscaled size. A 120 x 80 frame
scaled 200% shrank back to 120 x 80 on the page, so the 24 pt type broke into
two lines and the rest became overset.
The frame and the path now take the scale the transform loses, so they, the
line breaks and every glyph stay where they were.
Fixes#8297
It still said Illustrator's editing data isn't written, while the other
two File formats rows give what a saved `.ai`'s editing copy leaves out.
It now has their words. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The area rows and the evidence list name the change and its tests, as
AGENTS.md asks of a change that lands. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Type with dashes aligned to corners, and strokes past the 256 checks an
import makes, keep their drawn look; the status line said only that such
strokes come in live where they draw as their look.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It still said Illustrator's editing data isn't written, while the other
two File formats rows give what a saved `.ai`'s editing copy leaves out.
It now has their words. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The status line, the M4 Files row and the revision history name the
change, as AGENTS.md asks of a change that lands. Estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The check that a stack's patterns are each its own took each style item's
pattern once, but not that the drawn look places it one way:
`placed_paints` keeps a pattern's first placement, so a look filling with
one pattern at two placements, for a style that names it once, came in
with the first placement for both. `placed_paints` now also notes each
pattern the look places otherwise than its first placement, and a stack
with such a pattern keeps its drawn look. The same placement again, as a
compound path's look has for each subpath, is still one.
Test `a_look_placing_a_stacks_pattern_twice_keeps_the_drawn_look` (two
placements keep the look; the same one twice comes in as the path).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A path or compound path whose art style fills it with a pattern (alone,
or among other fills and strokes) came in as its drawn look: the style's
stack (`ArtStyle::stacked`) took solid colors and gradients only, though
the reader already places a style's pattern fill, as type takes it
(#6856).
A style's pattern fill now counts in the stack. As for the stack's
gradients, the reader takes it only where the drawn look paints with that
pattern (fills only), once (`placed_paints` keeps the first placement of
each name), and places it as the look does; otherwise the drawn look
stays. Type keeps taking its style's pattern from wherever its look paints
with it, as before.
Under effects on the object, a stack with a pattern fill keeps its drawn
look: a conservative rule, as in one garment file a Scribble over such a
fill drew unlike the look though the tiles were placed as the look's are,
for a reason not established.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The status line, the M4 Files row and the revision history name the
change, as AGENTS.md asks of a change that lands. Estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Besides the reader of its layers, a file whose layers come from its PDF
part is read again over all of its data for its artboards (once more for
the page scale), its variable library, its swatches, brushes, graphic
styles and symbols. Each of those readers started with every envelope
check (`MAX_CHECKED`) and every aligned-dash check (`MAX_DASH_CHECKS`)
an import has, so such a file could draw several times the strokes the
limit allows.
- The artboard and variable-library readers start with no checks
(`reader_aside`): neither depends on how a stroke or an envelope draws.
- The swatch, brush, style and symbol readers, which need only what
comes before the first layer, make no more checks once they reach it
(`Reader::defs_only`). What they read is unchanged in files that write
their definitions before their layers, as saved files do (symbols after
the first layer aren't read at all). A pattern tile or brush art defined
after the first layer is still read, without checks: it draws the same,
and keeps its drawn art where it could have come in live.
Definitions read before the layers (pattern tiles, brush art, symbols)
are still checked by each of those readers within the limit.
Test `the_reads_apart_from_the_layers_check_no_aligned_dashes`, which
counts the checks made (`DASH_CHECKS_MADE`, in tests only).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The row said an envelope of type that paints nothing opens live, which
reads as if the type painting nothing were enough. It also takes drawn
art that paints nothing, as the other docs say.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A stroke in an art style whose dictionary turns on aligning dashes to
corners and path ends (`BSDashAdj`) made the reader keep the object's
drawn look, the dashes fitted to the path as outlines that can't be
edited. VectorCraft draws such dashes itself (`Dash::align_corners`).
The style now reads the stroke with its dashes aligned. The reason the
look was kept (#1439: the style's dash array doesn't hold the fitted
pattern) still holds, as the fitting is VectorCraft's own, so the reader
takes the stroke only where the object with it draws as its drawn look,
through the check images with a Warp already use (now `look_alike`).
Aligned dashes have checks of their own (`MAX_DASH_CHECKS`) and don't
weigh the page, so the envelopes' checks and share of the page check stay
theirs. Elsewhere, past those checks and on type, the look stays.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The status line, the M4 Files row and the revision history name the
change, as AGENTS.md asks of a change that lands. Estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
When nothing paints the shapes aren't compared, so a fill added to the
type later may bend it unlike the unpainted outlines the file saved.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It still said Illustrator's editing data isn't written, while the other
two File formats rows give what a saved `.ai`'s editing copy leaves out.
It now has their words. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The area rows and the evidence list name the change and its tests, as
AGENTS.md asks of a change that lands. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An envelope of type whose type has no fill and no stroke, its drawn art
unpainted too, inks no pixel either way. The drawn-art check took that
as a difference and kept the drawn art, with its type hidden.
`alike_within` now first asks whether either side has any paint at all
(`paints`). When neither has any, they draw alike (nothing) and add
nothing to the page check's share. Art that paints, however faintly, is
compared as before. The shapes aren't compared when nothing paints: the
first way tried (the envelope as saved) is taken.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The status line, the M4 Files row and the revision history name the
change, as AGENTS.md asks of a change that lands. Estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It still said Illustrator's editing data isn't written, while the other
two File formats rows give what a saved `.ai`'s editing copy leaves out.
It now has their words. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The area rows and the evidence list name the change and its tests, as
AGENTS.md asks of a change that lands. Scores and estimates unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Illustrator writes a clipping mask of type from the text document as a
clipping group (`q` ... `Q`) with no clipping path (`W`), the text object
(`/AI11Text`) last. The reader took only `W` paths (and legacy type in
rendering modes 4 to 7) as clipping paths, so it read such a group unclipped:
its art painted past the letters, and where the layers then differed too much
from the page the file opened as its page.
- `Reader::end_container` records each clipping group with no clipping path
whose last object is a text slot (not the type a look draws) in
`Structure::type_clips`, by the group's id. The group is still read as it is:
the slot is filled later.
- `clip_by_type` (crates/eps/src/import/ai/read.rs) moves the type to the
front as the group's clipping path once its slot became one text object that
is still the group's last object. A slot left empty or filled with several
objects (type and the strokes the page draws around it) leaves the group as
read, unclipped, as before.
- The layers apply it after `fill_slots`; such type is never weighed as type
the page doesn't draw (`MADE` is taken off), since the page shows it by what
it clips. Envelopes of type apply it to their contents after `put_type`, so
one opens live where its drawn art shows the clip (`WarpedType::type_clips`).
Measured on 224 garment files we own (used locally, never committed): 14 renders
change; 12 come closer to the file's own page (for example 15.7 % → 0.07 %,
26.5 % → 0.28 %, 58.5 % → 1.4 % of the page differing), one file has no PDF
content to compare with, and in one the texture inside a word changes along
the letters' edges (0.35 % → 0.56 %). Envelopes of type that open live stay at
185; no file opens another way and no note changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A PDF/X-4 export wrote xmp:ModifyDate and xmp:CreateDate in its XMP
metadata but no xmp:MetadataDate. The PDF/X-4 metadata edit now adds
xmp:MetadataDate next to xmp:ModifyDate, with the same value, and grows
the stream's /Length to match. When the export dates the file as modified
later, post::modification_date already rewrites every xmp:MetadataDate,
so the two dates stay equal.
PDF/X-1a, PDF/X-3, PDF/A and plain PDF exports are unchanged.
A fill and a stroke at the same opacity below 1 were joined into one
type object whose opacity then applies to both together, losing the
darker overlap the file composites. stroke_filled now takes a stroke
only when the fill and the stroke are opaque.
An upright glyph filled and stroked was kept as an outline for the
reshaped-outline fallback while upright glyphs filled only were not;
it is no longer kept either, so such a line can't come in as the
outlines of its stroked glyphs alone.
Docs: the #8015 entries say what is left (translucent type, and a TJ
that starts a new line part way through).
When one Add to Swatches call added a pattern swatch and then a global or
spot color its art links to, the pattern brought the color in and the color
was added a second time as a numbered copy. The color the pattern brought
now takes the library swatch's place and color group instead.