fix: seven defects an adversarial audit found, and the gates that missed them
Release / test (full, windows-latest, 3.11) (push) Canceled after 0s
Release / test (pdf, ubuntu-latest, 3.11) (push) Canceled after 0s
Release / test (pdf, ubuntu-latest, 3.12) (push) Canceled after 0s
Release / build (push) Canceled after 0s
Release / publish-pypi (push) Canceled after 0s
Release / github-release (push) Canceled after 0s

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>
This commit is contained in:
Umut CAN
2026-08-06 23:07:50 +03:00
co-authored by Claude Opus 5
parent 6b9860e68e
commit ca1c79fd01
28 changed files with 3376 additions and 148 deletions
+79 -14
View File
@@ -198,6 +198,11 @@ class CapabilityRefusalMiddleware(Middleware):
only channel that carries both a machine-readable payload *and* the error
flag; rebasing the exception on ``FastMCPError`` carries neither, which is
why no tool function changes here.
R32 added a sibling kind. A quarantined document is *not* a capability
boundary — the engine can draw a circle perfectly well, the document it would
draw into is untrusted — so it gets ``kind: "quarantined"`` and carries the
abandoned-call record instead of a capability key.
"""
async def on_call_tool(self, context: MiddlewareContext, call_next):
@@ -205,19 +210,25 @@ class CapabilityRefusalMiddleware(Middleware):
return await call_next(context)
except Exception as exc:
capability = _batch_capability_of(exc)
if capability is None:
quarantine = None if capability else _batch_quarantine_of(exc)
if capability is None and quarantine is None:
raise # not ours — leave it byte-for-byte as it was
message = _refusal_message(exc)
payload = {
"ok": False,
"error": message,
"tool": context.message.name,
"backend": _backend_name_or_none(context),
}
if capability is not None:
payload["kind"] = "unsupported"
payload["capability"] = capability
else:
payload["kind"] = "quarantined"
payload["quarantine"] = quarantine
return ToolResult(
content=message, # prose-only clients still get a full sentence
structured_content={
"ok": False,
"kind": "unsupported",
"capability": capability,
"error": message,
"tool": context.message.name,
"backend": _backend_name_or_none(context),
},
structured_content=payload,
is_error=True,
)
@@ -3514,6 +3525,43 @@ def _batch_capability_of(exc: BaseException) -> str | None:
return None
_quarantine_error_type: type | None = None
def _quarantine_refusal_type() -> type | None:
"""``DocumentQuarantineError``, imported lazily like the capability types.
It lives in ``backends.quarantine``, which deliberately imports no ezdxf, so
this resolves even on a box where a backend's dependencies are missing.
"""
global _quarantine_error_type
if _quarantine_error_type is None:
from backends.quarantine import DocumentQuarantineError
_quarantine_error_type = DocumentQuarantineError
return _quarantine_error_type
def _batch_quarantine_of(exc: BaseException) -> dict | None:
"""The abandoned-call record ``exc`` refuses on behalf of, or None.
Recognised by type, not structurally: ``DocumentQuarantineError``
deliberately has no ``capability`` attribute, because
``_batch_capability_of`` would then report it to the client as
``kind: "unsupported"`` — telling them the ezdxf engine cannot draw a circle
when the engine is fine and the document is not.
"""
try:
quarantine_type = _quarantine_refusal_type()
except Exception as exc_import: # pragma: no cover - core module
log.debug("quarantine refusal type unavailable: %s", exc_import)
return None
for error in _batch_cause_chain(exc):
if isinstance(error, quarantine_type):
return getattr(error, "quarantine", None) or {}
return None
def _refusal_message(exc: BaseException) -> str:
"""The most specific message on the cause chain, in full.
@@ -3560,6 +3608,16 @@ def _classify_batch_error(exc: BaseException) -> dict:
"capability": capability,
"message": _batch_message(exc),
}
quarantine = _batch_quarantine_of(exc)
if quarantine is not None:
# R32: its own kind. Filing it under "unsupported" would tell the batch
# reader the engine lacks a feature; filing it under "failed" would hide
# that every remaining step will refuse for the same reason.
return {
"kind": "quarantined",
"quarantine": quarantine,
"message": _batch_message(exc),
}
chain = _batch_cause_chain(exc)
if any(isinstance(error, NotFoundError) for error in chain):
return {"kind": "unknown_tool", "message": _batch_message(exc)}
@@ -4849,8 +4907,10 @@ async def drawing_settings(
description=(
"Omit to READ every setting; pass a dict to CHANGE them. Friendly keys: "
"units (mm/cm/m/inch/feet), linear_precision, angular_precision, ltscale, "
"dimscale, text_size, point_mode, point_size, osmode, fillet_radius. "
'Example: {"units": "mm", "dimscale": 1.0, "linear_precision": 2}.'
"dimscale, dim_text_height, dim_arrow_size, dim_decimals, "
'decimal_separator ("." or ","), zero_suppression, text_size, point_mode, '
"point_size, osmode, fillet_radius. "
'Example: {"units": "mm", "dimscale": 1.0, "dim_text_height": 3.5}.'
),
),
] = None,
@@ -4859,9 +4919,14 @@ async def drawing_settings(
"""Read or change common AutoCAD drawing settings by friendly name.
A convenience facade over the system variables (INSUNITS, LUPREC, LTSCALE,
DIMSCALE, TEXTSIZE, OSMODE, …) so the user can say "set units to mm and
dimension scale to 1" without memorising sysvar names. Call with no argument
to get a full snapshot of the current settings.
DIMSCALE, DIMTXT, DIMASZ, DIMDEC, DIMDSEP, DIMZIN, TEXTSIZE, OSMODE, …) so
the user can say "set units to mm and dimension text to 3.5" without
memorising sysvar names. Call with no argument to get a full snapshot of the
current settings.
`dim_text_height` / `dim_arrow_size` / `dim_decimals` / `decimal_separator`
/ `zero_suppression` shape the *dimension* — `text_size` is TEXTSIZE, the
height of a standalone TEXT entity, and does not touch dimensions.
"""
if settings:
await ctx.info(f"Applying drawing settings: {', '.join(settings)}")