engineering/mech/draw.py is the only module in the package that touches a
backend: it turns the pure engine's primitive descriptions into entity_create_*
and dimension_* calls, places them in WCS, and writes the part model back onto
the drawing as ACADMCP_MECH XDATA on an anchor POINT. It is not a contract
member and neither are the tools - exactly like gear_draw_* and
titleblock_apply_iso_a3, it composes the generic members, so it is written once
and works on both engines by construction and no capability key is needed. No
new ActiveX call was written, so there was nothing to probe on the live seat;
measured instead that every one of the 23 backend members draw.py calls exists
on ComBackend and that ComBackend has no abstract methods left.
mech_part_draw takes one model - a turned profile or a plate outline plus typed
features - and draws its views in first-angle positions (third on request),
dimensions them and records the placement. mech_view_add reads the model back
out of the drawing and adds a side, top, section (full/half/offset/revolved) or
detail view months later without the caller re-describing the part.
mech_dimension_part lays out ISO 129 dimensions with no overlapping text,
resolves ISO 286 fits against the measured nominal, and moves the holes into a
table past a threshold. mech_hole_pattern exposes the same expansion for a
drawing that was never built from the model - with its own diameter guard,
because Hole.validate needs a part and a standalone pattern has none.
mech_part_inspect reads the payload back and never modifies the drawing.
mech_part_from_spec draws a whole sheet inside one transaction. Every part is
validated before the transaction opens, so a refusal leaves the drawing
untouched and names the path (parts[1]: segments[0].d_outer: ...), and the
rollback is shielded with the shape engineering/pid/spec.py documents: anyio
cancellation is level-triggered, so a bare await in the handler would be
cancelled before the rollback body ran and leave half a sheet inside an open
transaction.
fits goes through dimension_intents(fits=...), the API Task 6 shipped, rather
than being stamped on at draw time: that is the path that validates a key by
name and refuses a fit on a reference dimension or on a hole that moved into
the table. It is scoped per view first - a fit on segment[0] names nothing in
the end view, and refusing there would make a correct three-view sheet
undrawable - and any key no view claimed is handed back to dimension_intents so
the refusal is written in its own vocabulary.
SECTION 21 registers the six tools in the new mech pack, with alias records
(SECTIONPLANE, ARRAY plus shop-floor synonyms in English and Turkish - mil,
kademeli mil, kama, pah, kesit, gorunus), a regenerated tool inventory, the
README release snapshot at 210 tools in 24 groups, and the frozen group-size
snapshot gaining mech: 6. Lean is deliberately untouched here: spec section 11
puts three of these on the lean surface together with group G's two sheet tools,
and Task 15 makes that edit once across all four branches.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
engineering/mech/stdparts.py turns the Task 8 tables into real blocks. catalogue()
searches 83 transcribed parts by designation, standard, size or family;
block_spec() emits the typed primitives and DESIG/STD/SIZE/MAT ATTDEFs
block_define takes, in side/top for fasteners and ISO 8826-1 simplified /
ISO 8826-2 detailed for bearings; insert_std_part() defines the block once,
inserts it and writes the std_part ACADMCP_MECH payload the skeleton pins, so
the parts list is a read of the drawing rather than a naming convention.
Threads follow ISO 6410 - minor at 0.8 x major, a thread-length limit line to
the major, a three-quarter minor arc in the end view.
std_feature_draw() adds a thread, a DIN 509 undercut, a DIN 471/472 groove, a
DIN 332 centre hole or an ISO 3601-2 O-ring groove to geometry that already
exists. Its dimensions come from the shared standards registry, never from this
module: an absent table, a size outside coverage or a row missing a dimension is
refused by name, and the DIN 509 relief is drawn from the profile its table
publishes rather than reconstructed here. DIN 332 form R is refused outright.
An unknown designation is refused with the nearest catalogue entries; nothing is
drawn approximately.
SECTION 22 registers the three tools in the new mech pack, with alias records
(INSERT plus shop-floor synonyms in English and Turkish), a regenerated tool
inventory, the README release snapshot at 207 tools in 24 groups, and the frozen
group-size snapshot gaining mech: 3.
Two plan corrections, both verified by running the code:
* the plan's @cad_tool summaries for std_part_insert (118 chars) and
std_feature_draw (130) overran discovery.serialize.MAX_SUMMARY_CHARS = 110, so
tests/test_tool_registry.py::test_no_authored_summary_is_long_enough_to_be_truncated
failed - they would have shipped rendered as "...". Both are shortened to 99
and 104 characters with the same content.
* the plan's file list omitted tests/test_tool_packs.py, which pins the pack
registry, the tool_packs report and the group count. It now carries MECH_TOOLS
and distinguishes ALL_PACKS (declaration order, what "available" reports) from
ALL_PACKS_ENABLED (sorted, what "enabled" reports) - those two lists stopped
being the same list when "mech" was appended after "settings".
No new ActiveX member is called: the module drives block_list / block_define /
block_insert / entity_set_xdata / entity_create_* only. The COM path was still
executed live once on AutoCAD 2026 in a scratch document created and closed by
the probe: the bolt block defined with layers CENTER+GEOMETRY, attributes read
back as DESIG/STD/SIZE/MAT = ISO 4014 - M12x60 / ISO 4014 / M12x60 / 8.8, the
ACADMCP_MECH std_part payload round-tripped through entity_get_xdata,
STD_ISO_15_6205_SIMPLIFIED inserted, and the M20 thread's three primitives
landing on GEOMETRY.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
centre_marks draws ISO 128-23 centre marks or full centre lines and reads
each circle's real centre and radius back out of the drawing through
entity_get instead of being told them - a radius restated from memory is
the silent wrong number this repository keeps refusing to ship. A handle
that is not a CIRCLE or ARC, and a circle in a plane tilted out of WCS XY
(which has no radius in this frame), are both refused by handle before
anything is written. The dash proportions are not transcribed here: the
prims are plain lines on CENTER, whose linetype already carries them.
section_line draws the ISO 128-40 cutting-plane line - wide segments at
the ends on GEOMETRY, the thin long-dash-dot middle on CENTER, an arrow at
each end whose head sits on the line end and points the way the section is
viewed, the same letter beyond both arrows - and returns the five-key
plane dict {p1, p2, label, direction, style} that mech_view_add(kind=
"section") consumes, so the label on the plan and the label on the view
are the same string by construction. A viewing direction parallel to the
plane is refused rather than drawn.
hatch_material hatches through standards.materials.hatch_for, so the
ISO 128-50 map and its refusal stay in one place; scale multiplies the
material's own pattern scale and angle replaces its angle. An unknown
material is refused with the list before any entity is written.
DEVIATION - one defect in the plan's own draw_material_hatch, fixed here
with two tests that fail against the plan's code. It called
entity_create_hatch with ROLE_LAYER["hatch"] without creating that layer,
the same omission review found in Task 10's draw_annotation_prims.
MEASURED live on AutoCAD 2026 (probe: GetActiveObject, Documents.Add(),
doc.Close(False) in a finally): on a scratch document whose layer table is
['0'], msp.AddHatch(0, "ANSI31", False) + AppendOuterLoop + Evaluate()
leaves the AcDbHatch in model space, and the very next statement
hatch.Layer = "HATCH" - which is the one line _apply_entity_attrs runs -
raises com_error (-2147352567, ..., 'Key not found', ..., -2145386476);
model space then still reports [('AcDbHatch', '0')] and the layer table is
still ['0'], so the refusal leaves an orphan hatch on layer 0. After
doc.Layers.Add("HATCH") the same assignment succeeds. drawing_open runs no
bootstrap, which is exactly when a cut face has no HATCH layer to land on.
draw_material_hatch now calls ensure_annotation_layers before creating the
hatch and reports layer / layers_created, as draw_annotation_prims does.
The whole task re-measured live afterwards through ComBackend on a fresh
scratch document (layer table ['0']): hatch_material -> {'ok': True,
'handle': '82', 'pattern': 'ANSI31', 'layer': 'HATCH', 'layers_created':
['HATCH']} with the entity read back as HATCH on HATCH; centre_marks on a
live CIRCLE r=12.5 -> sources [{'handle': '84', 'center': [80.0, 0.0],
'radius': 12.5}]; section_line -> 9 entities and the five-key plane; layer
table after ['0', 'CENTER', 'GEOMETRY', 'HATCH', 'TEXT']; document closed.
tests/test_mech_marks.py: 17 passed. Gates: ruff check clean, inventory
regenerated to 209 tools, test_tool_registry (engineering 12 -> 15, total
206 -> 209), test_discovery_aliases, test_tool_inventory,
test_capability_contract and test_release_consistency all pass. Full
suite: 3343 passed, the 6 pre-existing failures verified identical with
this change stashed, and
test_release_consistency.py::test_document_numbers_match_the_tree
deselected as Global Constraints instructs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
ISO 1302:2002's seven transcribed symbol-proportion rows (text height ->
line width d', short-leg height H1, minimum height H2) drive the 60 deg
vee, the machining-required bar at H1, the machining-prohibited circle
inscribed in the vee, and the all-around circle at the kink; the Ra/Rz
values, the process, the lay symbol and the machining allowance land
around the extension line, which grows with the widest of them. A text
height outside the seven rows is refused by name rather than
interpolated, and ISO 1302 is accepted as an alias designation for
drawings still issued under it - the graphical symbol is the same one.
ISO 2553:2019's reference line, dashed identification line (omitted for a
symmetrical weld, as the standard permits), size written before the
symbol, length and pitch written after it, field-weld flag, all-around
circle and process tail. Six elementary symbols are transcribed - square,
single-V, single-bevel, single-U, single-J, fillet. The rest of the
elementary table is deliberately not shipped: a weld kind outside the six
is refused with the list of the six, because a guessed weld symbol is a
wrong instruction to a welder.
Both tools place the symbol on a plain leader - a line plus an arrowhead
with the proportions the backend's own mleader already draws - rather
than through leader_create_mleader, which refuses empty text and would
hang an MTEXT label off every symbol.
Proportions neither standard dimensions (the all-around circles, the
flag, the identification line's dashes, the text pitch and the character
width estimate) are named constants documented as this implementation's
declared choices, never as transcribed values.
No new ActiveX member is called: both tools compose LINE / ARC / CIRCLE /
LWPOLYLINE / TEXT through the entity-creation contracts both engines
already implement, so no live COM probe was needed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
CLAUDE.md gains items 18-20 (styles, page setup & templates, environment),
the grown drawing_settings facade on item 11, lean 55 / 52 and
TOOL_PACKS=all|core,pid,settings, and three conventions: layer states are
ours (ACADMCP_LAYERSTATES XRECORD, mode "xrecord"), page-setup evidence is
the PDF's own /MediaBox, nothing interprets coordinates in UCS.
README token table re-measured on this tree with benchmarks/token_suite.py
(ratio/v1 estimates): full 199 tools / 59,315 tokens, core,settings 190 /
55,593, core 168 / 48,825, lean 55 / 16,360, search 2 / 356; corpus
re-derived from discovery.aliases: 186 AutoCAD command names, 933 synonym
phrases, 204 tools. Styles & standards row, page setup & templates row,
final environment wording, and "Known limitations of the settings track".
CHANGELOG Unreleased: the track E intro, Added (styles, page setup &
templates, environment, drawing_settings growth, benchmark evidence),
Fixed (the Track A hardening list), Changed (three groups, packs, measured
idle cost, release snapshot 204 · 8 · 5 · 3284). Spec status: implemented.
Where the plan's text and the tree disagreed, the tree wins:
mleaderstyle_create is implemented on both engines through the
ACAD_MLEADERSTYLE dictionary and no `mleaderstyle` capability key exists
(spec §4, capability maps), so the docs say so instead of calling it
ezdxf-only; the Task 7 fix (tolerant ACADMCP_PID payloads, crossings
tested against the arc) joins the hardening list the plan left it out of.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
A fresh Layout1 is already ISO A3 landscape (ezdxf R2018 default: 420 x 297,
plot_rotation 0), so an A3-only /MediaBox read-back passed with
page_setup_apply stubbed to a no-op: the A/B gate row and the v4 matrix task
could not fail on the setter. Both now apply ANSI B (432 x 279, a size no
fresh document has; spec section 9 names it) and read that PDF first, then
apply ISO A3, read its PDF, and require the setter's changed.size_mm to have
moved from B to A3. The stub fails at the first PDF.
The moved-sheet comparison is +/-0.5 mm, the spec's tolerance, not exact:
measured live on AutoCAD 2026 (scratch Documents.Add(), closed unsaved), the
COM engine reports ANSI B as 431.8 x 279.4 and the DWG To PDF.pc3 file's
MediaBox as 431.8 x 279.4 then 420.158 x 297.039 for A3; the task passes
there with changed.size_mm [[431.8, 279.4], [420.0, 297.0]]. Headless it is
[[432.0, 279.0], [420.0, 297.0]].
tests/test_benchmark_v4.py gains the mutation test the finding describes
(setter stubbed -> both rows must fail; both passed on the old code) and a
run of the gate row itself; the headless task test asserts the B stage.
README rows and the tasks_v4 docstring say what the task now does; the README
collected-tests figure follows the two new tests (3282 -> 3284). The
regenerated A/B report against v1.5.1 is byte-identical (26 -> 32/32).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Task 24 (merge sanity) on main after the five track E groups merged.
- TOOL_PACK_NAMES = ("core", "pid", "settings"); PACK_TOOL_NAMES["settings"]
is exactly the 22 SECTION 20 tools. Styles and page setup stay in core.
- LEAN_TOOL_NAMES += page_setup_apply, batch_plot (53 -> 55).
- _GROUP_TAG_PRIORITY / _GROUP_TAG_LABELS: "style" moved up next to "pid",
"pagesetup" -> page_setup and "environment" -> environment added; the 28
SECTION 19/20 decorators gain their section tag, so _tool_groups() reports
styles 10 / page_setup 6 / environment 22 (23 groups, 204 tools) and
pack == group == section for the settings pack.
- SECTION headers, backends/base.py composition, contracts/__init__.py
exports and the layer_states feature entry were already correct after the
merge (read and confirmed, nothing edited).
- Pack text: config.py, .env.example, README config row; README <sub> line
204 tools in 23 groups; release snapshot 3259 collected tests (measured).
- docs/tool-inventory.json regenerated; registry snapshot re-derived.
- tests: test_tool_packs.py replaced (11 tests), test_capability_contract.py
+3 (composition, the agreed capability flags, layer_states honesty),
golden set +10 tuning cases (81 = 66 tuning + 15 holdout); all ten route
in the top 3 without alias edits.
Deviation from the plan text: TRACK_E_KEYS carries five keys, not six. The
plan's "mleaderstyle" (ezdxf True / com False) contradicts spec §8.1, which
retracted that key because the ActiveX route exists (§4.2); measured on the
tree: capability_keys() == {boundary_trace, chspace, interactive_prompt,
live_application, preferences, revcloud, wipeout}, and
ComBackend.mleaderstyle_create is a native implementation
(backends/com_backend.py:6752, is_capability_default -> False). The test now
also asserts the key is absent.
No new ActiveX call was written in this task, so no live probe was needed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Live AutoCAD writes a real .dwt through SaveAs(path, ac2018_Template = 66) and
puts the description into the summary Comments; headlessly a .dxf template is
written through drawing_save_as and a .dwt request is refused with capability
dwt_write, naming the DXF route and the COM route. drawing_save_as's "dwt"
entry was 5 (acR13_dxf): a DWT request wrote an R13 DXF under a .dwt name.
Probed live on AutoCAD 2026 (typelib: ac2018_Template = 66, acR13_dxf = 5,
acTemplateDwg absent; the saved file starts AC1032, the document is rebound,
Comments takes). SECTION 19 is now 6 tools; SAVEAS joins SHARED_ACAD_COMMANDS;
templates/README.md documents the two files per template.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
SECTION 19 grows to five tools with drawing_template_list; drawing_new takes a
bundled name (iso_a3_mech, iso_a1_arch, iso_a3_pid, ansi_b_mech, ansi_d_arch)
or a .dwt/.dxf path and reports template: {name|path, path, source}.
scripts/build_templates.py builds templates/*.dxf through EzdxfBackend
(layer set, header DIM variables until group S's styles land, drawing_settings,
sheet tab, ISO 5457 frame on A3, page_setup_apply) and --check pins them
reproducible. One drift the plan's same-process measurement could not see is
fixed at the source rather than masked: ezdxf orders CLASSES by set iteration,
so two processes disagreed on LAYOUT / ACDBPLACEHOLDER; the section is re-keyed
(required list first, then by name) before the save.
The root .gitignore ignores *.dxf as generated output; templates/*.dxf is
un-ignored there, since the bundled set is the one DXF that is committed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
user_pick_point / user_select answer {cancelled: true} on ESC (DISP_E_EXCEPTION)
and {timed_out: true} past COM_CALL_TIMEOUT; preferences_set refuses read-only,
unknown and out-of-range keys before any ActiveX call.
The read-only key is Files.PrinterStyleSheetPath, not the spec's
PrintStyleSheetPath: IAcadPreferencesFiles has no member of that name
(measured live on AutoCAD 2026 / R25.1), and a misspelt key would make
preferences_get() raise on every live call.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Summary fields are refused headlessly with capability dwgprops before any
write; custom properties are $CUSTOMPROPERTYTAG/$CUSTOMPROPERTY pairs and
survive save + reopen. Validation is shared, by name, and never str()-coerces.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
90 authored rows (type, range/enum, default, meaning, saved-in, per-engine
support, friendly key); every ezdxf claim is checked against the real
backend and the registry-saved set is pinned to what the backend refuses.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Layer states are ACADMCP_LAYERSTATES XRECORD chunks (<= 255 ASCII chars each),
in the file and not in AutoCAD's Layer States Manager. Headless view restore
writes the *Active VPORT (ezdxf's header has no $VIEWCTR; measured). UCS is
stored and made current; tool coordinates stay WCS.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
EnvironmentContract carries every member of the track E interface; the two
drawing_properties_* members are @capability('dwgprops') in this worktree only
and become @abstractmethod when group C's implementations merge.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
scripts/smoke_pid_com.py runs the whole track against the live AutoCAD in
a new document. ComBackend.connect() is lazy by design, so the script reads
reachability from system_status()["connected"] rather than from connect()
as the plan sketched, and closes its document without saving unless --keep
is passed. On this machine (AutoCAD 2026, "25.1s (LMS Tech)") it exits 0:
5 nodes, 4 edges, 0 dangling, confidence_min 1.0, 0 critique issues,
2 XDATA chunks — identical to the headless engine on the same spec. The
first attempt exited 2 on a stale win32com gen_py cache (a package folder
holding only __pycache__); an environment fault, no backend change needed.
A/B against v1.5.1 published as ab-v1.5.1-vs-v1.6.0-dev.json: 26/29 ->
29/29, three miss -> pass, zero regressed. check_doc_numbers.py now names
the report and the two row labels in one place (AB_REPORT, AB_*_LABEL),
and README.md / benchmarks/README.md carry the new rows; the v1.5.0 and
v1.4.0 tables stay as history.
README: P&ID row, the showcase and catalogue images, the token table
re-measured (161 / 44,930 by default, 152 / 41,420 with TOOL_PACKS=core,
lean 50 / 14,007), every stale count reconciled (154 -> 166 registered,
149 -> 161 advertised, corpus 158/672 -> 162/731 by the same count, 19 ->
20 groups, 6 -> 8 resources, six-task matrix with pid_roundtrip), a
"Known limitations of the P&ID track" subsection, and the snapshot line at
this machine's 2423 collected tests (CI has not run on this tree; noted in
the changelog with the pre-existing 1369 vs 1363 drift).
CLAUDE.md: section 17, sections 5 and 7 counts, TOOL_PACKS, the two 1.6
contracts and the reader rules. Spec status -> implemented. The plan and
spec markdown were reformatted by ruff (fenced Python only) because
`ruff format --check .` was already red on them at HEAD and the gate
requires the whole tree clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
tasks_v4.py extends v3 with one task the reader can fail on its own: the
example spec is drawn through pid_from_spec's code and read back by the
graph builder, which never sees the spec. The runner, the chart and the
doc-number gate now default to v4; --matrix v3 / v2 reproduce the earlier
sets, and tests/test_benchmark_v3.py labels its full-matrix run "v3"
explicitly now that the default has moved.
The plan's adapter check expected the two unnumbered lines to come back
with no line number; draw_line builds one from the default number_format
(bare sequence when size/service/spec/insulation are empty), so the task
asserts the two authored numbers on the two process lines and sequence-
built XDATA numbers on the other two instead.
The published A/B report still covers 26 checks; Task 21 re-runs it
against v1.5.1, which is what clears the remaining doc-number drift.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
TOOL_PACKS=all (default) | comma list of core,pid. core is everything not
claimed by another pack and can never be dropped; unknown entries are
ignored with a warning. Pack hiding composes with TOOL_PROFILE=lean (the
lean set now carries pid_symbol_insert / pid_line_draw / pid_graph, 50
tools) and ENABLE_3D. system_about reports {enabled, available, ignored};
cad_batch's unknown_tool rejection names the TOOL_PACKS gate.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
420 -> 288 lines. The source-review leaderboard, the live-run score table and
the task-matrix heatmap all ranked this project against other people's
servers; they are gone. What is left is what this repo can prove about
itself: the correctness A/B against its own previous tags, the five benchmark
tasks that can fail, the performance lane with its regression stated, and the
limits section.
Client wiring, the environment table and the showcase tool-call sequence move
into <details> so the page scans in one screen without losing them.
Star history now charts U-Pool alongside this repo.
Every gated number still matches the tree (16/16) and the release-consistency
suite is green.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every one of these was reported ok/clean to the caller while being wrong.
Silently wrong numbers
- selection_filter(min_area) read `entity.dxf.get("area", 0.0)` for SOLID,
TRACE and HATCH, none of which carry that attribute, so every hatch
measured 0.0. A 1600 mm^2 hatch was excluded by min_area=1.0 while
`filtered_by` asserted the filter had run, and COM (which reads ActiveX
.Area) returned a different selection for the same call. It now answers
from the measurement `entity_measure` uses, so the two cannot disagree.
- Every headless dimension rendered through ezdxf's bare `Standard` style:
1.0 mm text with a comma decimal marker, against ISO 3098's 2.5 mm minimum
and ISO 129's point -- while the document header itself said $DIMTXT 2.5.
`ezdxf.new(setup=True)` is not the fix; measured, it renders a 100 mm
dimension as "10000". The attributes are now stored explicitly.
- Saving writes `dimrnd = 0.0` onto the Standard style, and ezdxf applies
`xround(value, dimrnd)` whenever the attribute is present at all --
`xround(33.333, 0.0)` is 33. So on any reopened file, including one this
server saved, every dimension was rounded to a whole unit before DIMDEC
formatted it: a 12.75 mm feature dimensioned 13. AutoCAD means DIMRND 0 as
*no* rounding; the sentinel is now discarded at every load site, and a real
0.5 mm rounding is left alone.
- The DIM* header fold gave out-of-range values teeth they did not have when
they died in the header: dim_decimals=-1 returned ok and then killed the
next dimension with `ValueError: Format specifier missing precision` from
inside ezdxf, and 40 printed forty digits of float noise onto the sheet.
Ranges are now AutoCAD's own (DIMDEC 0-8, DIMZIN 0-15, lengths positive).
Security
- security.py recorded the policy as "sanitize_command and sanitize_lisp each
have exactly ONE caller ... the policy is scoped to the CHANNEL". The
premise was false. `_ensure_linetype_loaded` and `ComBackend.linetype_load`
build `_-LINETYPE _LOAD {name} {file}` and hand it to SendCommand, which
treats a newline and ';' as segment separators -- so a `linetype=` argument
carrying either injected further AutoCAD commands, and the 36-verb denylist
never saw that channel. Guarded with the DXF symbol-name rule (the
injection characters fall out of it rather than being enumerated) plus
validate_path for the file argument, which was also the one path-taking
tool reaching the filesystem unchecked. The policy comment is corrected.
Gates that could not see
- drawing_critique returned zero issues on a quarantined document: all seven
checks wrapped their backend call in a bare `except Exception`. CLAUDE.md
rule 7 makes that call the gate an agent is told to trust. Only the
quarantine now escapes; an ordinary broken check still degrades.
`_check_dimension_legibility` also read `_doc` directly from the event loop
thread, missing the gate and performing exactly the unsynchronised access
the quarantine exists to prevent. `construction_clear` answered
"deleted: 0" -- there was nothing to clear -- for the same reason.
- The iso128 focus checked layer lineweights only, so a sheet lettered at
1.0 mm critiqued clean. It now reads the effective dimension height and
separator, header-then-style, in the order the renderer resolves.
- check_doc_numbers.py derived five values it never checked, and those were
the ones that had drifted. The collected-test count (README said 1274, the
tree had 1369) and the A/B pass-rate and Fixed columns are now gated.
- benchmarks/README.md's v1.4.0 table was impossible on its face -- 21
passing plus 5 fixed is 26, not 24 -- and contradicted the report it
linked, because `ab-v1.4.0-vs-v1.5.0.json` was regenerated from a tree that
already carried the dimension fixes and so never measured the v1.5.0 tag.
Replaced with a real run: ab-v1.4.0-vs-v1.5.1.json, 21/26 -> 26/26.
Release gate
- release.yml ran one ubuntu/3.11 lane, so a Windows-only defect could reach
PyPI with no Windows job having executed. That is not hypothetical: v1.5.0
shipped `except pywintypes.com_error` raising NameError wherever pywin32 is
absent. It now runs the same three lanes as ci.yml, Windows included.
1368 passed, 1 xfailed. ruff clean. Tool inventory and all 16 gated document
numbers match the tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The triage over the 43 unaddressed verdicts returned 4 code and 35 doc items.
The code ones shipped in e24b46d; these are the documents.
Twenty-one applied mechanically after a validation pass confirmed each locator
still matched exactly once and each replacement was supported by the tree. Ten
were held back because a re-verify pass found them mutually exclusive -- three
patches rewriting the same benchmarks/README block, three rewriting the same
OCS paragraph -- and applying them in sequence would have corrupted the files.
Those were resolved by hand into one coherent edit each. One was rejected.
The corrections worth naming:
- **The A/B table contradicted itself in three files.** benchmarks/README.md
and CLAUDE.md said 24 checks / 21-24 / 87.5% / "fixed 3" while the suite has
26 and the published JSON says 21/26, 80.8%, five fixed. All three now agree,
and the five are split into the three new-capability `miss -> pass` and the
two repaired-defect `fail -> pass`, which is the distinction the lane exists
to make.
- **"Every number below is produced by a script"** was false twice over:
section 5's nine scores are literals a human assigned, and section 4 now
quotes hand-collected timings. The claim is narrowed to the charts, which
really are all rendered from committed data, and both exceptions are named.
- **The OCS paragraph over-quantified.** TEXT rotation was given as the single
exception; z components and HATCH boundaries are also entity-frame, and the
live backend records `2d_polyline_coordinates_not_verified_against_live_autocad`
in the same reason string. It now points at `ocs_normalized` as the live list.
- **Transactions were shown as bare parity.** The server itself grades them
differently -- `FeatureCapability(True, "undo_mark")` on COM against
`(True, "snapshot")` headless -- and COM rollback sends `_UNDO B` without
confirming it landed. The row says so.
- **"Raw commands: Opt-in"** named a switch that does not exist. Both tools are
registered unconditionally; `DANGEROUS_COMMANDS_ENABLED` only disables the
sanitizer.
- **"Wave A shipped 13 of a planned 39"** hardened an approximation: the only
source writes `~39`.
- **CAD_PROGID** said nothing beyond the connection was tested against
BricsCAD/ZWCAD/GstarCAD, implying the connection was. It was not.
Perf lane reworked against a fresh paired measurement. The old text claimed
"28% slower than v1.4.0, median of three runs" and that
`EZDXF_CALL_TIMEOUT=0` restored v1.4.0's numbers "exactly". Three problems:
`perf_suite` runs each workload once per invocation and cannot produce a
median, "exactly" was refuted by the figures in its own sentence, and the
number moved when the timeout wrapper was rebuilt. It now reports the guard's
cost -- +20.8% and +19.9%, three paired runs, median taken by hand -- says the
figures have no published artifact, and says the wrapper rework recovered about
a third of the v1.5.0 overhead rather than closing it.
perf-ezdxf.json and the perf chart regenerated against the shipped code, and
the README table rebuilt from that artifact so the two cannot drift.
1267 passed, 1 xfailed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Three defects found after 1.5.0 reached PyPI, two of them by the adversarial
pass over the rewritten README.
The HTTP bind guard failed open on the launch path it was written for.
`_guarded_run_async` read the bind address from the `host` kwarg and treated an
absent one as loopback. It is not: `fastmcp run` forwards that kwarg only when
`--host` was passed, and fastmcp otherwise resolves the host from
`fastmcp.settings.host`, which is environment-backed. So
FASTMCP_HOST=0.0.0.0 fastmcp run server.py:mcp --transport http
bound every interface -- 149 tools including file open/save and AutoCAD command
execution, no token -- while the guard inspected the literal "127.0.0.1" and
waved it through. v1.4.0's changelog claimed this guard "fires on every launch
path"; it fired, and looked at the wrong host. It now resolves the effective
host, and warns rather than staying silent if it cannot.
`[com]` and `[full]` could not be installed off Windows. Both listed
`pywin32>=306` with no environment marker, and pywin32 publishes Windows wheels
and no sdist, so `pip install autocad-mcp-pro[full]` died at resolution on
Linux and macOS -- and `[full]` is the extra the README recommended to
contributors. CI never caught it because every Linux lane installs `[pdf]` or
syncs from uv.lock, whose resolution markers hide the gap. Both entries carry
`; sys_platform == "win32"` now, with a release-consistency test to keep them.
`drawing_settings({"dimscale": ...})` reported success and changed nothing
headlessly. It writes `$DIMSCALE`, which is what AutoCAD reads when it draws a
dimension, so the call worked on COM; ezdxf renders from the dimstyle table
entry and never consults the header. One call, one reported success, two
different drawings. The header value now folds into the per-dimension override
on all five dimension tools -- including `dimension_aligned` and
`dimension_angular`, which do not share the tolerance-override path and the
obvious fix would have missed. DIMSCALE 1.0 writes nothing.
README: every engine-conditional family in the feature table now says which
engine (boundary tracing, WIPEOUT/REVCLOUD/MTEXT masks, CHSPACE, typed hatch
edge paths, headless undo), and the checkout install line stopped recommending
an extra that could not resolve on the reader's platform.
1257 passed, 1 xfailed. 1.5.0 is to be yanked: its diameter and radius callouts
put wrong numbers on drawings at default settings.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
`view_screenshot` renders at 100 DPI with a tight bounding box, so the
showcase PNG is intrinsically 698 x 500. The README displayed it at
width="820" -- a ~17% browser upscale of a raster, on a page whose subject
is drafting precision, and next to four charts that are resolution-
independent SVG. Native width instead; sharp beats big.
The rest of the hero caption was re-audited against the DXF it describes
and every item holds: 7 engineering layers, involute gear outline, DIN 6885
keyed bore, hatched section A-A, and two ISO 129 dimensions now measuring
144.000 and 40.000 with the H7 callout -- the two that were wrong before the
diameter fix in 8f405bd.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
`add_diameter_dim` / `add_radius_dim` measure to the point they are handed;
`leader_length` is a text placement. The ezdxf backend passed
`centre + (radius + leader)` as that point, so the leader became part of the
measurement: a true 40 mm bore was dimensioned 60 at the default
`leader_length=10`, and 108 at 34. Radius callouts were `leader_length` too
large the same way.
This is the worst class of defect for this project -- a wrong number, at
default settings, on the one part of a drawing that exists to carry numbers --
and it shipped in v1.4.0 and earlier. The size now comes from `radius` and
`location` only places the text.
The live COM backend was never affected: ActiveX's AddDimDiametric takes the
two chord points and the leader length as separate arguments. The two engines
had been disagreeing by 2 x leader_length on the same call.
Found while building the drawing for the new README, whose own bore callout was
wrong. Two checks added to correctness_suite.py; both report fail -> pass
against v1.4.0, not miss -> pass, which is what separates a repaired defect
from new capability in that lane. A/B is now 21/26 -> 26/26, five fixed, zero
regressed.
README rewritten from scratch
- Hero is a real ISO A3 sheet the server drew: m6 z24 involute spur gear, DIN
6885 keyed bore, section A-A, ISO 129 dimensions, ISO 286 H7 fit, ISO 7200
title block, 56 entities, drawing_critique clean. New
scripts/render_readme_showcase.py rebuilds it, and the tool-call sequence
that produces it is printed in the page.
- Star history chart, badge row, capability matrix, five evidence lanes, and a
"what this release is bad at" section carrying the 28% throughput regression.
- Every image and repo link is now an absolute URL: README.md is the PyPI long
description, where relative paths 404.
Corrections this work forced, all measured
- The df=0 example was wrong. FILLET is already in entity_fillet's own text;
the commands that genuinely scored zero are WBLOCK, PEDIT, OVERKILL,
LAYTRANS and MATCHPROP.
- 40,305 tokens is the tools/list wire payload; ~25,143 is the part forwarded
to the model. The old wording said "of schema", which is neither.
- `--tokenizer anthropic` needs a package and credentials this repo does not
lock, and refuses rather than falling back. Said so.
- Discovery ranks over the 149-tool advertised catalog, not the 154 registered.
- Five middlewares, not four; the fifth is what makes refusals machine-readable.
- Five of the 27 capability keys are environment-dependent, so the matrix
carries footnotes instead of bare ticks.
- "Every coordinate is WCS" has one deliberate exception: TEXT rotation.
- render_live_chart's axis still said "fixed 10-task matrix" while our bar came
from v3's fifteen. Each bar now carries its own matrix.
- .env.example said MAX_UNDO_STACK=50 (code: 5) and omitted four real
variables, two of them the remote-HTTP controls.
- server.json described an empty ALLOWED_PATHS as blocking access. It disables
the allowlist.
- server.py called the lean profile ~46 tools; it is 47.
- backends/capability.py cited a test file that does not exist.
1249 passed, 1 xfailed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Python evaluates an `except` clause on the way out of every failed call,
not only the ones it is meant to catch. `pywintypes` is bound only when
`sys.platform == "win32"` and the import succeeded, so on any host
without pywin32 -- all of Linux, and Windows installed without the
`[com]` extra -- an ordinary exception leaving `ComBackend._run` hit that
clause and became `NameError: name 'pywintypes' is not defined`, burying
the real error.
Every Linux CI leg failed on exactly this, in
test_ordinary_errors_are_unchanged. The timeout tests survived only
because they matched an earlier `except`; nothing else in the suite sends
a non-timeout exception through `_run`.
Fixed with a module-level `_COM_ERROR` tuple: `(pywintypes.com_error,)`
where the import worked, `()` where it did not. An empty tuple catches
nothing, which is correct when no com_error type exists to be raised -- a
placeholder Exception subclass would be worse than the bug, silently
swallowing unrelated errors.
Reproduced on Windows before fixing by deleting the module attribute,
which resolves the name the same way the absent import does; that is now
a test, alongside one pinning `_COM_ERROR` as a tuple. Verified through
the locked environment CI uses: 1244 passed, 1 xfailed.
README test-count snapshot 1243 -> 1245.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The v1.5.0 lock was generated before the version bump, so its own
[[package]] entry said autocad-mcp-pro 1.4.0 while pyproject said 1.5.0.
No dependency differed -- the transitive graph is byte-identical -- but
`uv sync --locked` will not re-resolve, so lint, test-linux 3.11,
test-linux 3.12 and test-windows all failed at install before running a
single test.
`uv lock` fixes it in one line. The gate that should have caught it did
not exist: test_release_consistency chains pyproject -> version.py ->
CHANGELOG -> README -> server.json, and the lockfile was the one release
artifact carrying a version string nothing checked. Added
test_uv_lock_records_the_pyproject_version, verified RED against the
broken lock and green against the fixed one.
Reproduced the CI lanes locally into a throwaway environment before
pushing this time: `uv sync --locked --only-group dev`, then
`--locked --extra pdf`, then ruff and pytest under `uv run --no-sync`.
1242 passed, 1 xfailed.
README test-count snapshot 1242 -> 1243 for the new gate.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Tool count 131 -> 154; 474 -> 1242 collected tests.
The release exists for two measured reasons. A 149-tool catalog costs a client
~40,300 tokens before it has asked for anything, and a drafter searching the way
drafters think -- FILLET, BPOLY, QSELECT, WBLOCK -- scored df=0 against the stock
index, because those words were simply not in the data.
Discovery and batching
- DISCOVERY_MODE=search replaces the catalog with search_tools + call_tool,
ranked over an AutoCAD command and synonym corpus (discovery/aliases.py) on
top of fastmcp's BM25 index. Measured: 40,305 -> 356 tokens advertised.
- cad_batch runs a step list in one round trip; fields= projects result-heavy
responses. The `core` profile is gone -- discovery replaced it.
Honesty debt (Tier 0)
- OCS/ECS: `extrusion` appeared nowhere in this codebase, so mirrored and
rotated geometry reported coordinates that were quietly wrong. backends/ocs.py
translates at the backend boundary; every tool coordinate is WCS on both
engines, and a tilted plane refuses rather than answering with a number that
looks fine.
- drawing_save_as(fmt="dwg") wrote DXF bytes to a .dwg path. drawing_audit never
repaired. analysis_measure_area shoelaced points the caller typed in.
- ezdxf calls gained a per-call timeout, so one hung call can no longer wedge a
server whose document lock is a single asyncio.Lock.
Architecture
- The 1854-line backend ABC is 19 contract modules; @capability(key, reason=)
makes optional methods declarative and a gate fails on undeclared holes.
- 8 layout/viewport tools, CHSPACE, CAD_PROGID escape hatch.
- Wave A: selection filters, boundary tracing, hatch depth, annotation objects,
measurement -- 13 tools of a planned 39. The REGION and 2D-boolean family was
cut on evidence: add_region() yields zero ACIS bytes and the greiner_hormann
substitute loses 28.2% of the area on a square with one semicircular edge.
Verified against a live AutoCAD 2026, which found four COM defects no amount of
reading would have caught -- AcadPViewport has no ViewCenter, and the line that
set it had been swallowed by a bare except since v1.4.
Benchmark v3 (15 tasks) and what it caught
- v2 scored this server 10/10, which carried no information: every task in it
exercised something the server was built around. Five tasks were added because
they can fail, and three did while being written -- a hatch could not be
measured though the refusal listed HATCH as measurable, the boundary tools and
the measure tool disagreed about the same shape (139.2177 vs 139.2699), and a
circle stored as two bulged vertices measured 0.0.
- Competitor reports are NOT re-scored. They are pinned v1.4 runs against v2 and
were never asked these five questions; the chart shows those rows `not run`.
- Correctness A/B: 21/24 -> 24/24, zero regressions.
Performance, stated plainly
The published v1.4 perf report was recorded on CPython 3.14 and the new one on
3.11.15, so the apparent 2-6x speedup is the interpreter, not this release.
Re-measured on one interpreter, v1.5.0 is 28% slower creating entities and 22%
slower on the 10k roundtrip. Fully attributed: EZDXF_CALL_TIMEOUT=0 restores
v1.4.0's exact numbers. That is the deadlock fix being paid for, the knob is
documented, and winning it back is on the 1.6 roadmap.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
source_review.json extended with three projects reviewed 2026-07-23 (LokmenoWer/best-cad-mcp B/74, daobataotie/CAD-MCP B/35, zh19980811/Easy-MCP-AutoCad C/30) and our entry updated to the released v1.4.0 (95: CI matrix, PyPI packaging, A/B gate, paper space, ISO 286, opt-in solids). Chart regenerated via render_chart; README benchmark section now opens with the leaderboard + 9-row table, source-review lane merged into it.
Co-authored-by: Cursor <cursoragent@cursor.com>
Ran benchmarks/compare_versions.py against the v1.3.0 tag: 21/21 on both versions, 0 fixed / 0 regressed - the v1.4.0 feature+infra work introduced no correctness regressions. Main README gains lane 4 (version A/B, with the historical v1.0.0 -> v1.1.0 contrast); benchmarks/README records the release-gate result and drops the stale 'performance coming' stub.
Co-authored-by: Cursor <cursoragent@cursor.com>
