Motion analysis review round 3: ui.start reads a study as motion.study does, keys off the timeline refused, pairs counted once, sub-assembly copies documented

This commit is contained in:
MateuszSychXpl committed 2026-10-16 17:52:39 +02:00
1 parent 3128762014
commit c34e607f5b
4 files changed
+26 -23

No files matched your search

+5 -4
View File
@@ -193,10 +193,11 @@ colliding pair with its `first_step`, `last_step`, the `steps` ranges where they
far into each other (`depth_mm`, the most at `deepest_step`) and their shared `volume_mm3`
there (null when it is not cheap to work out). It plays the study as `motion.play` does, but
contact sets don't stop it, and parts are tested at the samples only: a part that passes right
through another between two samples is missed. A check
takes at most 256 bodies; on a large design, list the `occurrences` to check (an arm, with
`against` the parts it must miss). When the work runs out it says `truncated` and the step it
`checked_to`.
through another between two samples is missed. Parts are told apart by occurrence, as contact
sets tell them apart: an occurrence inside a sub-assembly placed twice is one part, so its two
copies are not checked against each other. A check takes at most 256 bodies; on a large design,
list the `occurrences` to check (an arm, with `against` the parts it must miss). When the work
runs out it says `truncated` and the step it `checked_to`.
**Insert Design** (`component.insert_design {path, translate?}`, on the **ASSEMBLE** panel and
in the **File** menu) adds a component linked to another saved design, so each part can live in