Where
The Sketch chapter, "constraint-based sketching" section (doc/sketch.rst, rendered at https://cadquery.readthedocs.io/en/stable/sketch.html). Observed on the stable build (v2.8.0, commit fb4c6d4).
The inconsistency
The worked example constrains with "Fixed":
result = (
cq.Sketch()
.segment((0, 0), (0, 3.0), "s1")
.arc((0.0, 3.0), (1.5, 1.5), (0.0, 0.0), "a1")
.constrain("s1", "Fixed", None)
.constrain("s1", "a1", "Coincident", None)
.constrain("a1", "s1", "Coincident", None)
.constrain("s1", "a1", "Angle", 45)
.solve()
.assemble()
)Two paragraphs below, the "Following constraints are implemented" table lists FixedPoint, Coincident, Angle, Length, Distance, Radius, Orientation, ArcAngle — Fixed is not in the table, and FixedPoint's documented argument (None for arc center or a 0..1 float) is not what the example passes.
A reader can't tell whether Fixed is an undocumented alias, a distinct constraint the table omits, or something deprecated that the example should stop using.
Measured on 2.8.0 (this is how the inconsistency was found)
Running the example verbatim on cadquery 2.8.0 / Python 3.12: it executes successfully — solve() converges and assemble() returns a usable face (extrudes to a valid solid). So the example's "Fixed" form is real, working API that the table doesn't document.
Suggestion
Either add a Fixed row to the constraint table (arity 1, argument None, "entity is fully fixed"), or — if FixedPoint is the intended public spelling — update the example to use it with its documented argument form. Happy to PR whichever direction maintainers prefer.
Where
The Sketch chapter, "constraint-based sketching" section (
doc/sketch.rst, rendered at https://cadquery.readthedocs.io/en/stable/sketch.html). Observed on thestablebuild (v2.8.0, commit fb4c6d4).The inconsistency
The worked example constrains with
"Fixed":Two paragraphs below, the "Following constraints are implemented" table lists
FixedPoint,Coincident,Angle,Length,Distance,Radius,Orientation,ArcAngle—Fixedis not in the table, andFixedPoint's documented argument (Nonefor arc center or a 0..1 float) is not what the example passes.A reader can't tell whether
Fixedis an undocumented alias, a distinct constraint the table omits, or something deprecated that the example should stop using.Measured on 2.8.0 (this is how the inconsistency was found)
Running the example verbatim on cadquery 2.8.0 / Python 3.12: it executes successfully —
solve()converges andassemble()returns a usable face (extrudes to a valid solid). So the example's"Fixed"form is real, working API that the table doesn't document.Suggestion
Either add a
Fixedrow to the constraint table (arity 1, argumentNone, "entity is fully fixed"), or — ifFixedPointis the intended public spelling — update the example to use it with its documented argument form. Happy to PR whichever direction maintainers prefer.