A dimension's block took a recolored copy named COMPARE_<tag>_D1, and
the DXF writer redraws a dimension whose block is not a *D block, so
saving the snapshot replaced geometry read from a file. The copy now
gets a free *D name.
A named block both drawings define was reused when its images and
data links had the same relative paths, even when they were relative
to different folders. These paths are now taken from each drawing's
folder before the blocks are compared.
A multiline attribute is drawn from its own multiline text, so an
inline color code in it overrode the comparison color set on the
attribute: an unchanged attribute could stay red, or a removed one
green. The export now takes the inline colors out of the multiline
text of a reference's attributes and of attribute definitions in the
recolored blocks, as it does for MTEXT and multileader text.
The comparison of a named block both drawings define counted each
object's geometry, layer, color, linetype, lineweight and linetype
scale. A block whose line was hidden, more transparent, thicker or
extruded another way in the other drawing counted as the same, so the
other drawing's references drew with this drawing's definition. These
properties now count as well.
A named block such as PART was compared with the other drawing's by the
blocks it draws, including a dimension's block. That block's generated
name (`*D1` here, `*D3` there) was part of its signature, so the same
PART counted as different and was copied as PART_2 with a copy of the
dimension block. The comparison now leaves out an anonymous block's
name and compares its contents.
* Pin craft-libs at 420d21a
* Pin craft-libs at 16fb3d4
* Use craft-ui's shared docking workspace and ui.dock commands
Replace the local Workspace, its validation, apply, visibility and the
ui.dock parser with craft_ui::docking::Workspace and DockOperation.
CADCraft keeps its panels, the drawing view's rules, the default layout
and the workspace file. Add tests for loading a v1 workspace file and for
dragging a panel group by its bar.
* Pin craft-libs at 6e7f285
* Docking: reveal panels through craft-ui's set_visible, which reopens groups and places new panels itself
* Docking: palette switches regroup a panel whose recorded anchor is hidden instead of opening it alone
DimStyle::apply_fields rounded a value for a whole-number field such as
DIMDEC or DIMTMOVE and clamped it to 0..255, so DIMDEC -1 became 0 and
DIMALTD "300" became 255. DIMSTYLE, DIMOVERRIDE and the SETVAR overrides
that new dimensions take all accepted these values.
A value below 0 or one that rounds above 255 is now refused, as an
out-of-range value is. DIMADEC still takes -1. A lineweight given as
-1, -2 or -3 (ByLayer, ByBlock, Default) is no longer read as 0 when
the style's lineweight is positive, and a lineweight above 211 given as
text is refused instead of read as 211.
SETVAR wrote a DIM* variable to the drawing header without checking it,
so values such as DIMTMOVE 3, DIMATFIT 7 or DIMLWD 500 were accepted.
New dimensions then ignored them, because the style override goes
through DimStyle::apply_fields, which refuses them.
SETVAR now runs the value through apply_fields on a copy of the current
dimension style, as DIMSTYLE and DIMOVERRIDE do, and reports an invalid
value instead of writing it.
DIMTFILL and DIMTFILLCLR were not modeled, so a dimension's text never had a
background and lines under it stayed visible. Dimension styles now carry both
variables. DIMTFILL 1 hides what is drawn under the text in the drawing's
background color, and DIMTFILL 2 fills the box in DIMTFILLCLR. The box is the
text box widened by DIMGAP on every side. The arc length symbol over the text
and the DIMINSPECT frame are drawn over the background.
The variables work with SETVAR, DIMOVERRIDE and the style fields. DXF stores
them as DIMSTYLE groups 69 and 70 from R2007 and in the style's CADCRAFT
xdata in older files; the reader skips the record's flags, which also use 70.
For DIMTFILL 2, the dimension's *D block and EXPLODE add a SOLID in
DIMTFILLCLR under the text. A dimension drawn from the *D block it was read
with gets a background around each of the block's texts, before the block's
first text or first segment of the text's marks, except for a text that a
SOLID of the block already covers. The background stays after the drawing is
saved and reopened.
Past the limit on blocks copied, a block of the other drawing that this
drawing also defines was skipped without a note, and the red group drew
this drawing's definition even when the two differed. Such a block now
counts as left out unless both drawings' blocks draw the same thing, so
the export says that some blocks were too many to copy.
Tools > Inquiry > Radius had no prompts, so with nothing selected it
stopped with "`handle` is required". It now asks "Select arc or circle"
and reports the radius and diameter of the arc or circle picked. A
selected arc or circle is still measured at once.
MEASUREGEOM's Radius option and Tools > Inquiry > Radius also measure
the arc segment of a polyline under the pick. radius.measure takes an
optional `point`, and for a polyline it measures the arc segment nearest
that point.
A named block that holds a dimension was kept as this drawing's
definition when only the block the dimension draws differed between the
drawings, so the red group drew this drawing's dimension graphics. The
export's block comparison now covers a dimension's block, as it covers a
multileader's.
AREA and Volume measured the selected object when `points` was given but
was not a list of points; they now report an error, and only an absent
or null `points` measures the selection. A volume too large for f64
(an area times a height) is an error in the JSON form. At the typed
prompt, a volume or an Add/Subtract volume running total too large for
f64 makes the prompt say "The volume is too large." and ask for the
height again, where it reported "Volume = 0" or "Total volume = 0".
MEASUREGEOM typed at the command line offered Distance, Radius and ARea,
and Tools > Inquiry > Angle ran without a prompt, so it failed asking for
{vertex,p1,p2} or {h1,h2}. MEASUREGEOM now has an Angle option, and it
and Tools > Inquiry > Angle ask "Select arc, circle, line <specify
vertex>". An arc reports its included angle. A circle takes the pick as
the first endpoint, uses the center as the vertex, and asks for the
second endpoint. Two lines or polyline line segments report the angle
where the lines meet, extended. Enter asks for a vertex and two
endpoints. The JSON form of angle.measure is unchanged.
cutbase and copybase now fail with "`base` ([x, y]) is required" when
`base` is missing or not a point, as their parameters say, instead of
taking the lower-left corner of the objects. cutbase checks it before
erasing anything.
Copy, Copy with Base Point, Cut and Cut with Base Point put an object on
the clipboard once however often `handles` lists it, so the counts they
return match the objects copied or cut.
A block reference of the other drawing whose named block exists in this
drawing with other contents drew this drawing's definition in the red
group. The export now keeps this drawing's block only when both
drawings' blocks draw the same thing: the same objects, base point,
units, annotative and overlay flags and xref file, in the block and in
the blocks its references and multileaders use. Otherwise the other
drawing's block is copied under a free name (PART_2, PART_3, ...) and
its references point at the copy.
MEASUREGEOM offered Distance, Radius and ARea, and Tools > Inquiry had
no Volume item. The Volume option measures a polygon of three or more
points, or a closed 2D object as AREA measures it, and asks for a
height. It reports the area times the absolute value of the height.
Add volume and Subtract volume keep a running total, as AREA's Add
area and Subtract area do. Tools > Inquiry > Volume (volume.measure)
starts the same prompts, and the JSON form takes points or a handle
and a height.
Engineering and architectural units show the volume as a decimal
number. The de, fr and ru catalogs get a "Volume" row, and
docs/parity-checklist.md is regenerated with cargo xtask parity.
ATTDISP had no menu entry. View > Attribute Display now has Normal, On
and Off, which set ATTMODE to 1, 2 and 0 as ATTDISP's options do. ATTDISP
can also be typed as 'ATTDISP while another command runs, and that
command resumes afterward.
The eight catalogs without a "Normal" row get one, and
docs/parity-checklist.md is regenerated with cargo xtask parity.
BLOCKREPLACE (and -BLOCKREPLACE) turns every reference of one block, in
model space, the layouts and other blocks, into a reference of another
block at the same insertion point, scale and rotation. Each new
reference carries the replacement's attributes, with the values of the
tags both blocks define, and is annotative when the replacement is.
The command then asks whether to purge the replaced definition, and
removes it only when no dimension style names it and no block
reference, dimension or multileader in the drawing or in another block
definition does.
The command refuses a replacement that references the block it
replaces, a dynamic block as the block to replace, and blocks that
come from an xref. It does not run while the Block Editor or REFEDIT
is open. The ATTSYNC helpers in blocks/attribs.rs walk the references
and build the new attributes.
The self-reference check that BLOCK, INSERT of a file, REFCLOSE and
Lisp block creation share now also follows dimension blocks and
multileader block content, so BLOCK no longer accepts a redefinition
of A from a multileader whose content is A.
docs/parity-checklist.md is regenerated.
* Depend on craft-pdf (krilla writer, reader for tests) at craft-libs 9f75d3dc
The root patches pin the patched krilla and hayro-syntax/hayro-jbig2 to
the same revision, as craft-pdf requires.
* Plot PDFs with craft-pdf's krilla writer
The hand-written object, xref, stream and font writer goes. Each sheet's
display list is built, written as its page and dropped before the next.
Strokes with the same colour, width and transparency stay one path, and
each filled primitive is one path of its triangles. Text in embeddable
TrueType fonts is drawn as glyphs; krilla subsets the fonts and writes
their ToUnicode maps, so the hand-written subsetter goes too and
cadcraft-fonts keeps only the fsType embedding check. Primitives with a
point that is not finite or beyond 1e7 pt are left out. A font krilla
cannot embed after all makes the export retry with text as outlines.
Optional content (PDF layers) is not written until craft-libs' writer
supports it: the layer transitions go through one empty function, layers
that are off are not plotted, and the layer tests are ignored.
The tests read the PDF back with craft-pdf's reader.
* Pin craft-pdf and its krilla and hayro patches to craft-libs cd160e73
cd160e73 adds optional content groups to the patched krilla.
* Write drawing layers as PDF optional content groups
Each drawing layer a page draws on gets one group, named after the layer
and shared by all pages, created before the first page that uses it. A
group starts hidden when its layer is off in the drawing, so `offLayers`
again plots layers that are off on hidden PDF layers. At most 4096
groups per file; objects on further layers print unmarked. A single
sheet with groups opens with the viewer's layer panel (/PageMode /UseOC).
Layer changes end and start the marked content, and the last mark ends
before the page finishes.
The layer and text tests that check the groups run again, and a test
covers the group cap.
* Plot a zero-length line as a round dot in PDFs
krilla leaves out a path without extent, so a lone zero-length line drew
nothing. It is filled as the dot its round-capped stroke would make, as
wide as the stroke, with 0.25 pt for a hairline.
* PDF: retry as outlines when the writer fails inside a font; cap layer names; restore two test checks
* Tests: read plotted PDFs back through the shared reader in plot_layout_scale and plot_styles
* Pin craft-pdf, krilla, hayro-syntax and hayro-jbig2 to craft-libs main (7895f5e)
* PDF tests: read path coordinates in page space; cover the font retry and out-of-range points
- pdf_read maps path operands through the CTM (`page_ops`, used by `drawn` and
`clip_rect`) and checks that the plot area is drawn in page space.
- The 1:1 layout test finds the viewport corner at (72, 72) in page space, so a
mirrored page fails it.
- A font with a short hmtx passes the writer's font checks but fails subsetting;
the plot retries and draws its text as outlines.
- A primitive with a point past 1e7 pt is left out, and the rest still plots.
* PDF: clip a primitive with a far vertex to the plot area instead of leaving it out
A polyline or fill with a point past the writer's f32 range now keeps the
parts inside the plot area (polylines with the area widened by the stroke
width), each piece its own subpath. Primitives in range are written as before.
The test reader requires a decodable content stream.
New lines, arcs, circles, points and lightweight polylines (for example
PLINE, POLYGON, DONUT and BOUNDARY) now take the THICKNESS system
variable as their 3D thickness. THICKNESS could already be set with
SETVAR and with the Properties palette when nothing is selected, but new
objects were always created with a thickness of 0, apart from RECTANG's
own Thickness option. RECTANG keeps that option, and single-line text
does not take THICKNESS, as in AutoCAD.
The new THICKNESS command is on Format > Thickness. Typed or picked from
the menu, it asks "Enter new value for THICKNESS <0.0000>:"; Enter keeps
the current value, and a negative value is accepted.
docs/parity-checklist.md is regenerated.
Edit > Cut with Base Point asks for a base point, then for the objects,
copies them to the clipboard with that base point and erases them, as
COPYBASE does without the erase. A pickfirst selection skips the object
prompt. Objects on locked layers stay, as with CUTCLIP, which now shares
the cut helper. The JSON form takes {handles?, base: [x,y]}.
Catalog rows for the new menu label are added in all 15 languages, and
docs/parity-checklist.md is regenerated with cargo xtask parity.
* refactor: replace wildcard imports in two test modules (S2208)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* style: fmt
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
COMPAREEXPORT compares the two drawings again and opens the result as a
new untitled drawing, compare_<this> vs <other>.dwg. Model space holds
the other drawing's removed and changed objects as group
COMPARE_COMPARED and the compared tab's objects as group
COMPARE_CURRENT, in the overlay's colors: green for objects only in
this drawing or changed in it, red for the other drawing's, gray for
objects in both.
Block references, dimensions and multileaders refer to recolored copies
of their blocks, tables and tolerances use recolored copies of their
styles, and inline MTEXT and cell colors, dimension style colors,
multiline element colors and gradients follow the object. The new
drawing takes the layers and styles of the other drawing that this
drawing lacks, and the blocks its exported objects refer to; an
anonymous block whose name is taken is copied under a free name. The
comparison bar gets an Export button.
* usage: a read-only cross-reference of layers, blocks, styles and linetypes
USAGE reports, for every layer, block, text style and linetype, how many
objects use it in model space and the layouts, how many inside block
definitions, whether it is used only inside block definitions, and whether
it is empty or unused. Its unused lists are PURGE's own scan (candidates and
block reachability), shared read-only; PURGE's behaviour is unchanged.
The JSON form takes {type?, limit?, offset?}, returns schema 1 with a paged
items list, per-kind summary and PURGE's unused names in table order.
Fixes#1439
* docs: document the functions touched by #1439
* usage: mark a scan cut at its limit as incomplete (#1439)
USAGE's one pass over model space, the layouts and the block definitions was
capped at 5,000,000 objects with `take`, so a larger drawing silently omitted
later references while the counts, statuses, summary and message looked
complete — an unused layer or block could be reported as used, or vice versa.
The limit is now `MAX_SCAN`, and `Scan::new` takes it as a parameter: it counts
the drawing's objects first and records an `Incomplete { scanned, total }` when
the limit cuts the scan short. The report carries `incomplete: true` and an
`incompleteReason` naming both counts, and the message ends with a note, so a
partial report can never be mistaken for a complete one. A drawing holding
exactly the limit is still complete.
PURGE's scan limit and candidate behaviour are untouched; USAGE only reads
PURGE's unused lists as before.
Tests: unit tests cut a small drawing at a small limit (a 5,000,000-object
drawing is impractical) and check the boundary and the miscount the signal
prevents; the integration test checks a complete drawing says so. Also adds the
missing Ukrainian translation for the command label, which the i18n gate
requires.
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Two tests that start in the same millisecond got the same instance_id(), so this test's
shutdown() removed the other test's lock from the process-wide SAVEFILEPATH (after #1868).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Found while testing crates/engine/src/units.rs:176-179 (parse_mixed)
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Found while testing crates/color/src/lib.rs:115 and 126
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
aiongg