Require openscad_cpp_evaluator 0.41.0

[revarbat]

Brings BelfrySCAD/openscad_cpp_evaluator#107 and #108.

children() takes a vector or range

children([3:1:5])    // children 3, 4, 5
children([3:-1:1])   // children 3, 2, 1 -- in that order
children([2,2,2])    // child 2, three times

Before, a vector or range was accepted syntactically and then silently ignored — every such form rendered child 0, with no warning of any kind.

Backwards ranges warn, but only for an implicit step

for (x=[3:1])      // WARNING: begin is greater than the end, but step is positive
for (x=[3:-1:1])   // silent
for (x=[1:3])      // silent
for (x=[1:1:3])    // silent

Spelling out the step is taken as deliberate. This is a divergence from OpenSCAD, which warns for the explicit forms as well — recorded in the wiki's Differences from OpenSCAD table, with the full rule on Language-Syntax → Backwards ranges.

The warning belongs to the range, not the loop, so this reports too:

r = [5:0];         // warns, even though nothing iterates r

That matches where the reference puts it, and is a case the first attempt at this feature missed entirely.

Release hygiene

The first uv lock after publishing caught only 5 of 7 files — the same PyPI simple-index propagation race that has bitten this process twice before. Re-locked after confirming the index listed all 7, then verified against the published wheel rather than a local build, under both engines:

VM=0 and VM=1, byte-identical:
WARNING: ... line 5     <- r = [5:0], never iterated
WARNING: ... line 1     <- for (x=[3:1])
ECHO: "c3" "c2" "c1"    <- children([3:-1:1])

Version bump only, 0.76.6 → 0.77.0.

🤖 Generated with Claude Code