encode_values() accepted strings containing "\n" or "\r". ezdxf writes
1000-tag values verbatim, so the separator split the tag pair: the saved
DXF failed to reopen (DXFStructureError) and, because every undo and
transaction snapshot is also a DXF save, transaction_rollback broke after
a single such write. Both engines route through encode_values, so the
refusal (ValueError naming values[i]) now happens there, before either
backend touches the document. Only LF and CR split a DXF tag; tabs and
the other Unicode separators round-trip and stay allowed.
Tests: pure encoder refusal for \n, \r, \r\n; other control characters
still typed; ezdxf refuses-then-rollback-and-reopen round trip; COM fake
sees no SetXData / RegisteredApplications call.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
`text`, ATTDEF `prompt` and `default` were only isinstance(str)-checked, so
'A\nB' and '10^' passed validation and ezdxf's DXFAttr fixer
(`fix_one_line_text`) rewrote them on `add_text` / `add_attdef` at DEBUG
log level — the declared ATTRIB value and the one the P&ID reader gets back
differed silently, and COM (raw string) diverged from ezdxf. Same hole the
tag check closed in 4158575, left open on its three siblings.
A `_one_line(value, where)` helper now refuses '\n', '\r', any other C0
control, and a trailing '^' (ezdxf's own `is_valid_one_line_text` rule),
naming `entities[i]: 'text'` / `attdefs[i]: 'prompt'|'default'`. Tests cover
the refusals, the strings ezdxf keeps verbatim ('10^2', 'Ø20 H7', '%%C20'),
and an ezdxf round-trip proving accepted values are stored unchanged.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Three ways `engineering/pid/lines.py` could emit a silent wrong result or
an opaque crash instead of an honest refusal:
- `format_line_number` only recognised a token that was exactly `{field}`;
`P-{seq:04d}`, `{size}"-{service}` or `{bogus}x` were copied verbatim,
braces and all, into the line number. Tokens are now parsed with
`string.Formatter`: one named known field per token, a format spec is
honoured (`{seq:04d}` -> `0007`), literal decoration around the field
survives with it and drops out with it, and conversions, positional
fields, multi-field tokens, unbalanced braces and specs the value cannot
take are all refused by token name.
- `route` never checked finiteness: a NaN `end` collapsed the U candidate
to a valid two-point "straight" path ending at the stub point. Every
coordinate of start/end/waypoints and the stub are now checked and the
refusal names the offending `end.x` / `waypoints[1].y` / stub.
- A non-axis port direction (45.0, 90.5, nan) raised `KeyError` from the
AXIS lookup. `axis_direction()` now normalises with a 1e-6 deg tolerance
for engine drift and otherwise raises ValueError naming `start_dir` /
`end_dir` and how to resolve it (rotate to a multiple of 90 or pass
waypoints).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Every grammar error is prefixed with the tag's own letters, so the old
single-letter fragments (FCI/"I", PSVX/"X", ...) matched whatever the
grammar said: a mutation that dropped the letter and position from every
error left 9 of 11 invalid rows green. The fixture now quotes the error
text the spec demands ("unexpected letter I at position 3", "... after
valve V", "trailing O (open/closed) ...", "needs at least one function
letter after F"), and the test pins it to the single first error.
ABCDEF-106 was mispinned as a grammar rejection of "D": six letters
exceed the regex cap, so it never reached the grammar. It is now pinned
as a format rejection, and ABCDE-106 carries the "letter D at position 4"
case.
A fixture lint rejects any fragment that is a substring of its own tag
and requires grammar rows to name a letter, position or trailing rule,
so a future row cannot reintroduce the hole.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Two ways a malformed block_define request slipped past the pre-write
validator and reached the engine:
* `ATTDEF_TAG_RE.match` with a `$` anchor accepts 'TAG\n' because `$`
matches before a trailing newline. ezdxf strips the newline on write, so
'TAG\n' + 'TAG' landed as two ATTDEFs with one tag, defeating the
duplicate-tag refusal, and the tag-keyed attribute readers then dropped
one silently. Now `fullmatch`.
* `_layer` only required a non-empty string, so a DXF-illegal name such as
'BAD<LAYER' passed and ezdxf raised DXFValueError mid-loop, after the
block existed and earlier primitives were written (with overwrite=True,
after the old definition was already wiped). The gate now refuses on the
same forbidden-character rule `sanitize_symbol_name` uses, exposed as
`security.illegal_symbol_name_chars`, with a TypeError naming
`entities[i]` and 'layer'.
Tests cover both repros and fail against the previous code.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
`add_blockref` never checks the name and `add_auto_attribs` skips autofill
when `Insert.block()` resolves to None, so a typo'd block name returned a
handle whose attribute values were silently dropped and whose dangling
INSERT `doc.audit()` later deletes. Check `name in self._doc.blocks` and
raise `ValueError` on both the attribute and no-attribute paths, matching
the COM engine's `InsertBlock` refusal for an unknown name.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Plan covers the prerequisite block_insert attribute fix (ezdxf wrapped the
INSERT in an anonymous *U block and lost its ATTRIBs), the two generic
contracts, the 43-symbol catalogue, routing, the ISA-5.1 grammar, the
graph reader, deliverables, six critique focuses, pid_from_spec, TOOL_PACKS,
resources, benchmark v4, and the live COM smoke.
Spec reconciled with the plan: 43 base symbols, FUNC/LOOP at +-2.0,
solid = DXF SOLID entity, shape/direction options, example spec uses an
actuated FCV-101 (a hand valve has no signal port), refine hints in
detail["hint"], preflight LAYER_SET_INTENT_MISMATCH, radial port matching
on the circle.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FKFStjEFzpGyLq5bMpWxTd
Research snapshot of the AutoCAD MCP field (20 READMEs, 2026-09-14), the
seven 1.6 tracks with measurable acceptance criteria, their order
(A -> E -> B -> G -> F -> H -> D), and the tool-pack decision.
Track A: symbol catalogue authored in code (ISO 10628-2 + ISA-5.1), two
generic contracts (block_define, entity_get/set_xdata), nine pid_* tools,
routing, the ISA-5.1 tag grammar, the pid_graph reader, six critique
focuses, TOOL_PACKS, and the test/evidence plan.
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>
**COM discarded areas it had already read.** `entity_measure` fell back to
rebuilding geometry from `.Coordinates` when `.Area` *or* `.Length` was
missing, coupling two independent property reads. HATCH, REGION and 3DSOLID
expose `.Area` and neither of the others, so the fallback fired, found no
vertices, and raised "AutoCAD reported neither an area nor usable vertices" --
while holding the area. Those are the exact types the README sells the COM
backend on. Now only a missing `.Area` triggers the fallback; an absent
perimeter is reported as `None` with `perimeter_exact: False` instead of being
invented, which also removes a `round(None, 6)` TypeError waiting for the first
type that answers `.Area` and nothing else.
**A plain square hatch reported itself self-intersecting.** `ezdxf.path
.flattening()` repeats the start point, and `is_self_intersecting` excluded
adjacent edges *by index*, so with the duplicate present the first and last
real edges were cross-tested and registered as crossing. Measured: a 40x40
SOLID hatch came back `area: 1600.0, self_intersecting: True`. A closed loop
now answers the same whether or not it carries its closing vertex, and a real
bowtie is still caught (`area: 0.0, self_intersecting: True`).
**The bowtie warning never fired on the live backend.** `self_intersecting` was
only computed inside the COM fallback, so an LWPOLYLINE bowtie -- which has
both `.Area` and `.Length` and never reaches it -- returned `area: 0.0,
self_intersecting: null`. The confident zero this field exists to flag, going
out unflagged. It is now computed on the normal path too.
Also corrects a docstring written earlier today in `contracts/analysis.py`
claiming the live backend "asks AutoCAD for `.Area`, which has already applied
the [hatch island] style". Whether ActiveX's hatch area applies the style is
not verified against a live seat, and the method could not measure a hatch at
all. It now says so.
The ActiveX member profiles behind the COM tests are assumptions, not facts,
and the tests say so; a live seat settles them. `.Circumference` for CIRCLE and
`.Perimeter` for REGION are deliberately NOT added here -- writing more members
from the reference is what produced four wrong beliefs in v1.4/v1.5.
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>
`_c131_cpu.py`, `_c131_micro.py` and `_c131_probe.py` were throwaway probes
written by a subagent while checking a README claim, and a7c8a7a swept them
in with `git add -A`. They are not part of the project and nothing imports
them.
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>
Leaderboard restyled to the LLM-benchmark look: dark background, rank-numbered labels, per-rank color palette, 'NN / 100 - grade' annotations. The live-run, task-matrix and performance charts share the same dark theme for a consistent README identity. Charts regenerated deterministically from their data files.
Co-authored-by: Cursor <cursoragent@cursor.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>
- TOOL_PROFILE=lean|core|full applied in the lifespan via server-level enable/disable; system_about reports the active profile; opt-in solid_* tools are hidden while ENABLE_3D=false
- layouts: layout_list/create/set_current + viewport_create (scaled model viewports) on both backends; drawing_export_pdf(layout=...); ezdxf uses TILEMODE + active-layout binding, COM switches ActiveLayout and restores it
- ISO 286: engineering/fits.py authored IT4-11 + fundamental-deviation tables (1-500 mm; shafts d..p/h/js, holes D..H/JS) with clear out-of-scope errors; dimension_linear/radius/diameter gained fit= resolved from the measured nominal
- tolerances: DIMTM now signed in deviation/limit modes so double-positive fits (p6 +0.035/+0.022) render correctly; legacy positive-magnitude contract preserved
- solids: solid_box/cylinder/extrude/revolve/boolean native on COM (AddBox/AddCylinder/AddRegion+AddExtrudedSolid/AddRevolvedSolid/Boolean); ezdxf reports the honest solid_3d capability boundary
- capability maps: paper_space native, new viewport_render feature, ENABLE_3D-gated solid_3d
- 60 new tests (profiles, layouts, viewports, layout PDF, fits, gating); version pins in tests now read version.__version__
Co-authored-by: Cursor <cursoragent@cursor.com>
ruff format was never run repo-wide; CI now enforces 'ruff format --check', so normalize every module once. No behavior changes.
Co-authored-by: Cursor <cursoragent@cursor.com>
