wrenchroom 0.5.4

What's Changed

• Open 0.5.4.dev0 by @kmoneil in #191
• A report says the positions its step came to, and check()'s docstring what no sidecar means by @kmoneil in #193
• wrenchroom 0.5.4 by @kmoneil in #194

Full Changelog: v0.5.3...v0.5.4

wrenchroom 0.5.3

What's Changed

• Open 0.5.3.dev0 by @kmoneil in #183
• A swing's step is from 0.1 to 30 degrees, and every report says the one it ran at by @kmoneil in #186
• A swing's positions are spaced evenly round the circle, whatever its step by @kmoneil in #187
• The sidecar says the step every way of running the check swings at by @kmoneil in #189
• wrenchroom 0.5.3 by @kmoneil in #190

Full Changelog: v0.5.2...v0.5.3

wrenchroom 0.5.2

What's Changed

• Open 0.5.2.dev0 by @kmoneil in #177
• A tube pushed into a part is no threaded insert, though it is named one by @kmoneil in #180
• Two parts drawn in one place, one a solid the boolean fails on, are not asked by @kmoneil in #181
• wrenchroom 0.5.2 by @kmoneil in #182

Full Changelog: v0.5.1...v0.5.2

wrenchroom 0.5.1

What's Changed

• Open 0.5.1.dev0 by @kmoneil in #155
• The clash list measures a fastener as its verdict does (closes #156) by @kmoneil in #159
• An insert set into its part is no clash, and a nut in its trap is said so (closes #158) by @kmoneil in #160
• A plug is found by its name wherever "plug" stands in it (closes #157) by @kmoneil in #161
• A tapping screw is sized as a tapping screw: ISO 1478's ST sizes (closes #142) by @kmoneil in #162
• A boolean's answer is held to what the two meshes share (closes #163) by @kmoneil in #170
• A screw that forms a machine thread keeps its machine size (closes #164) by @kmoneil in #172
• A tool's contact under --exact is held to the meshes too (closes #169) by @kmoneil in #171
• A screw terminal's plug is a plug: "screw" is about the terminal (closes #166) by @kmoneil in #173
• A No. 5, and a tapping screw's No. 7 and No. 14, are sizes a name gives (closes #165) by @kmoneil in #174
• Inch heat-set inserts sit in their makers' holes as metric ones do (closes #167) by @kmoneil in #175
• wrenchroom 0.5.1 by @kmoneil in #176

Full Changelog: v0.5.0...v0.5.1

wrenchroom 0.5.0

What's Changed

• Open 0.5.0.dev0 by @kmoneil in #145
• Every boolean runs non-destructively, so measuring one part never changes another (closes #146) by @kmoneil in #147
• Clashes: wrenchroom clashes, and checks: {clashes: true} inside check (M8) by @kmoneil in #148
• The 3D view draws each clash: its two parts, and the overlap between them in red (M8) by @kmoneil in #149
• Build order: a build: list, each fastener checked in the step that adds it (M8) by @kmoneil in #150
• Connectors: each plug pulled off its receptacle, and with hand room, gripped (M9) by @kmoneil in #151
• The 3D view draws each plug: where it comes to, and the fingers tried (M9) by @kmoneil in #152
• A ring comes down over whatever its nut is on, and its handle with it (closes #143) by @kmoneil in #153
• wrenchroom 0.5.0 by @kmoneil in #154

Full Changelog: v0.4.0...v0.5.0

wrenchroom 0.4.0

What's Changed

• Open 0.4.0.dev0 by @kmoneil in #132
• The tools a model needs: wrenchroom tools --used, and a tools list in every report by @kmoneil in #133
• CI runs macOS on pull requests only, on Python 3.13 by @kmoneil in #137
• A bolt's hex head in a hex pocket is held by it, and a joint whose two halves are both held fails (closes #134) by @kmoneil in #138
• A screw whose shank sits on no size is sized by its Torx or cross recess, or its shank is said (closes #135) by @kmoneil in #139
• The tools list names the tools blocked fasteners would need once reached (closes #136) by @kmoneil in #140
• A weekly run of the unit suite on macOS and Python 3.14 by @kmoneil in #141
• wrenchroom 0.4.0 by @kmoneil in #144

Full Changelog: v0.3.1...v0.4.0

wrenchroom 0.3.1

What's Changed

• Open 0.3.1.dev0 by @kmoneil in #112
• The Chrome view test says why a run delivered no page, and tries it once more, alone (closes #113) by @kmoneil in #114
• A name's thread size the solid is drawn as another of is noted and taken as drawn, and a leadscrew's nut is passed over by @kmoneil in #118
• A clash on a screw's head is measured on all of its head, a dome, a crown or a countersunk cone, its shank left out by @kmoneil in #119
• Cross recesses are read as makers draw them, and cross recessed, Type I and ASME names are read by @kmoneil in #120
• A ring gets on along the axis, over the end of the bolt through it, and a part in its way leaves it off (closes #121) by @kmoneil in #126
• A set screw's thread is no clash, and its core is: one stopped by the part tapped for it is blocked by it (closes #122) by @kmoneil in #127
• A nut's clash runs on to its end on the tool's side: a cap nut's dome is measured (closes #123) by @kmoneil in #128
• A Torx recess is read from its lobes, and a recess of round walls that is no Torx isn't guessed a keyed head (closes #124) by @kmoneil in #129
• A cross recess's drawn span picks its Phillips number, and one drawn for another number than its thread's is noted (closes #125) by @kmoneil in #130
• wrenchroom 0.3.1 by @kmoneil in #131

Full Changelog: v0.3.0...v0.3.1

wrenchroom 0.3.0

What's Changed

• Open 0.2.1.dev0 by @kmoneil in #92
• A thread and length alone is a screw, taken on its drive; a part shaped like a fastener but not named as one is said (closes #95) by @kmoneil in #97
• Washers named after their screws aren't screws, set screws take their own keys from their socket's end, and thumb screws are checked for fingers (closes #96) by @kmoneil in #98
• A nut in a trap, a hex pocket or a slot its own width, is held by it, and its screw must turn (closes #93) by @kmoneil in #99
• A screw drawn into what its way out meets is a clash, one drawn twice is said once, and a name's length the solid disagrees with is noted (closes #94) by @kmoneil in #100
• M18 and M22 socket heads' keys are DIN 912's, said so, and the drive tables' catalogue rows name the standards that confirm them (closes #102) by @kmoneil in #105
• An overlap is measured near the origin, so a graze far out can't read as the whole tool (closes #107) by @kmoneil in #108
• The 5.5 mm spanner is Elora's 123, as the 5 mm is, not Hazet's 105 (closes #103) by @kmoneil in #106
• M8 and M10 Phillips screws take PH4, as every ISO cross-recess standard says, and the full kit has a PH4 driver (closes #101) by @kmoneil in #109
• A part drawn as a surface is read for what it bounds: a closed shell is its solid, and an open shell or loose faces are left out and said (closes #104) by @kmoneil in #110
• wrenchroom 0.3.0 by @kmoneil in #111

Full Changelog: v0.2.0...v0.3.0

wrenchroom 0.2.0

What's Changed

• Open 0.1.1.dev0, and write down the pypi environment's rules by @kmoneil in #76
• A rule's tool: must fit its fastener, or it is not covered, saying why (closes #72) by @kmoneil in #77
• A state's model lacking a fastener says so and fails the run (closes #74) by @kmoneil in #78
• A state's model: is found beside the sidecar, and a missing one says so (closes #73) by @kmoneil in #79
• Small things: key arms checked, unused custom tools noted, part-noun names quiet, alike notes once (closes #75) by @kmoneil in #80
• Makers' solids: a cone is countersunk only as the head, the head is the wide end, the name's keyed head stands (closes #81) by @kmoneil in #86
• A real export's names: Hexnut is a nut, a bare insert needs a bore, a fixed thread takes its screw's size and says it must turn (closes #84) by @kmoneil in #87
• Below M3: M1.6 to M2.5 and #0 to #3 in every table, the small tools in the full kit, and a missing head names its standard (closes #83) by @kmoneil in #88
• A socket drawn at its standard's size takes its key, a loose one says so, and a Torx recess picks its size (closes #82) by @kmoneil in #89
• Parts that won't mesh: mended first, referred only at their surface, judged by their own faces' strays, and said (closes #85) by @kmoneil in #90
• wrenchroom 0.2.0 by @kmoneil in #91

Full Changelog: v0.1.0...v0.2.0

wrenchroom 0.1.0

What's Changed

• ci: bump astral-sh/setup-uv from 10.0.1 to 10.2.0 in the actions group by @dependabot[bot] in #1
• M3: the mesh engine is the default; --exact keeps the referee by @kmoneil in #3
• Names from a model can't steer the terminal; SECURITY.md by @kmoneil in #4
• Lock Pillow 12.3.0 past threejs-materials' pin by @kmoneil in #5
• M4: read fasteners from part names by @kmoneil in #6
• M4: read what a named fastener's solid shows by @kmoneil in #7
• M4: check finds fasteners by name and solid; a sidecar is optional by @kmoneil in #8
• M4: wrenchroom detect writes the sidecar to correct and keep by @kmoneil in #9
• M5: the report as Markdown (--md) by @kmoneil in #10
• M5: the HTML view (--html), one self-contained file with the verdicts in 3D by @kmoneil in #11
• M6: kits are data, and a check tries only the kit's tools by @kmoneil in #12
• M6: the pytest plugin, and wr.check that stays a function by @kmoneil in #13
• M6: room for the hand on each handle, off until tuned by @kmoneil in #14
• M6: the imperial kit, and inch tools named with their unit by @kmoneil in #15
• M6: Torx keys, T10 to T40, in the full kit by @kmoneil in #16
• M6: the open end of the combination spanner by @kmoneil in #17
• M6: ball-end keys, leant up to 25 degrees off the axis by @kmoneil in #18
• M6: nut drivers, the last of the full kit's tools by @kmoneil in #22
• An --only glob that matches nothing fails the run (closes #20) by @kmoneil in #23
• A report's FILE may be -, for stdout (closes #21) by @kmoneil in #24
• A drive word anywhere in a name says the head (closes #19) by @kmoneil in #34
• Sidecar mates take globs, and one matching nothing fails the run (closes #32) by @kmoneil in #35
• An in-tolerance hex takes its standard's spanner (closes #27) by @kmoneil in #36
• A nut with both ends clear is turned from the end with more room (closes #26) by @kmoneil in #37
• A fastener noun with words after it is a candidate, never silent (closes #30) by @kmoneil in #38
• A flat-topped head is held to the standards' outlines (closes #31) by @kmoneil in #39
• Fixed threads hold themselves, and a round nut gets no spanner (closes #29) by @kmoneil in #41
• A fastener drawn as several solids is one fastener (closes #28) by @kmoneil in #42
• Big spanners, a gland's label, and the CAD kernel kept quiet (closes #33) by @kmoneil in #43
• The engines agree on grazes, and a verdict decided by one says so (closes #25) by @kmoneil in #44
• Shoulder screws take the key their thread sets (closes #40) by @kmoneil in #45
• Every tool on a hex's flats tests the hex's corner sweep, and a socket its wall by @kmoneil in #53
• explain and --only keep a fastener's pair (closes #46) by @kmoneil in #54
• A hex drawn a little small takes its own thread's spanner, saying so (closes #50) by @kmoneil in #55
• A nut bored at its thread's minor diameter has a free face its bolt doesn't cover by @kmoneil in #56
• A plain shoulder-screw head is ISO 7379's, and a guessed head says so (closes #48) by @kmoneil in #57
• A gland's dome is no bore, its hex is its band, and a ring gets on over it or not (closes #47) by @kmoneil in #58
• Spanners as long as makers make them, and stubbies only where sold (closes #49) by @kmoneil in #59
• A pass only a ball end reaches says so, and shallow sockets get none (closes #51) by @kmoneil in #60
• blocked_by names the parts that decided it, and each attempt its best arc (closes #52) by @kmoneil in #61
• Read a part's faces once per frame, and an assembly's boxes once per check by @kmoneil in #62
• Custom tools in the sidecar, in the shape of the built-in tables (spec 5.3) by @kmoneil in #65
• A FAIL line names each part once, and a few at most (closes #64) by @kmoneil in #66
• A hex its own corners stop says so once, and a fastener drawn into a part is a clash (closes #63) by @kmoneil in #67
• A README that shows what wrenchroom is for and how to use it, with an example the tests run by @kmoneil in #68
• The view harness ends a script as a browser does (CodeQL js/bad-tag-filter) by @kmoneil in #69
• Release to PyPI from a version tag, through a workflow PyPI trusts by @kmoneil in #70
• wrenchroom 0.1.0 by @kmoneil in #71

New Contributors

• @dependabot[bot] made their first contribution in #1
• @kmoneil made their first contribution in #3

Full Changelog: https://github.com/kmoneil/wrenchroom/commits/v0.1.0