Multi profile extrusion performance (#2064)

Thanks @Koschtetnix.

tests/robustness/goldens.json

@@ -98,6 +98,7 @@
   "94": {"area":3480.70333,"bodies":1,"edges":61,"error":null,"faces":26,"outcome":"ok","rejected":1,"shells":2,"steps":["revolved section","cut cylinder","fillet","hole","rectangular pattern","cut box","cut box"],"unsupported":0,"volume":5456.446856},
   "95": {"area":2375.340299,"bodies":1,"edges":33,"error":null,"faces":12,"outcome":"ok","rejected":2,"shells":1,"steps":["revolved section","hole","cut cylinder","cut sphere","circular pattern","cut sphere"],"unsupported":1,"volume":6704.220611}
  },
+ "platforms": {"windows":{"seeds":{"1":{"area":2642.450837,"bodies":1,"edges":32,"error":null,"faces":13,"outcome":"ok","rejected":0,"shells":1,"steps":["revolved section","cut cylinder","join cylinder"],"unsupported":0,"volume":6991.569339},"23":{"area":1795.696636,"bodies":1,"edges":15,"error":null,"faces":7,"outcome":"ok","rejected":0,"shells":1,"steps":["revolved section","cut sphere"],"unsupported":0,"volume":3198.694618},"34":{"area":1086.303608,"bodies":1,"edges":30,"error":null,"faces":14,"outcome":"ok","rejected":0,"shells":1,"steps":["revolved section","cut sphere","hole"],"unsupported":0,"volume":2134.3653},"67":{"area":7943.46222,"bodies":2,"edges":86,"error":null,"faces":36,"outcome":"ok","rejected":0,"shells":2,"steps":["extruded profile","cut cylinder","hole","join cylinder","join box","hole","chamfer","rectangular pattern"],"unsupported":2,"volume":27607.962186},"94":{"area":3470.837042,"bodies":1,"edges":63,"error":null,"faces":27,"outcome":"ok","rejected":1,"shells":2,"steps":["revolved section","cut cylinder","fillet","hole","rectangular pattern","cut box","cut box"],"unsupported":0,"volume":5456.438772}}}},
  "notes": [
   {"reason":"First record: what each seed builds on main at 1a0100cd","seeds":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95]},
   {"reason":"seed 0 meshes differently since display tessellation plans (#805), which landed alongside these goldens (#826)","seeds":[0]},
@@ -135,6 +136,8 @@
   {"reason":"After integrating #1718, replay the regenerated revolved-section recipes on the combined branch: seed 13 has 40 faces/101 edges (one face/two edges fewer), seed 94 has 27 faces/63 edges (one face/two edges more). Both complete fixed command sequences pass validity and volume-direction checks after every step, and two independent golden runs per seed agree exactly. Steps, outcomes, body and shell counts are unchanged; area and volume match within existing 1e-6 relative tolerances (seed 13 volume differs by 0.000002 mm3). Only topology counts are updated. The finer unshifted conical-cut route and its geometry guards remain in place.","seeds":[13,94]},
   {"reason":"#1670 and #1753 (mesh export: twisted ruled faces divided to the chord tolerance, and the faces of an edge sharing its points exactly) change the meshes the kernel's validity and volume checks measure: seed 55's body (a filleted box cut by spheres and patterned) now passes the validity check it failed, and its volume moves by 1.3e-5 relative. Pull requests ran only every 8th seed, so main's full run found it. Recorded from main's CI output (Linux).","seeds":[55]},
   {"reason":"#853 recorded seeds 13, 71 and 94 as its own branch built them (13: 40 faces/101 edges, 71: 19/50, 94: 27/63), but with main's kernel they build as before (41/103, 20/52, 26/61), so every full run on main failed on them after #853 landed. This puts back the counts main builds; #853's conical-cut refinement (seed 20) is unchanged. Steps, outcomes, volumes and areas are the same.","seeds":[13,71,94]},
-  {"reason":"#1853 (a part revolve of a profile with an edge on the axis is swept a third of a turn at most per face, so the boolean can intersect its faces): seeds 28 and 95 start with a rectangle on the axis turned half a turn, now two quarter turns. Seed 28 (joined by a cylinder) has more faces for the same shape (5902.1214 -> 5902.1117 mm3, the new faces' meshes). Seed 95's hole into it failed with 'boolean Cut: no result' on the half-turn faces and now builds, so the generator's later picks move. Values from Linux CI.","seeds":[28,95]}
+  {"reason":"#1853 (a part revolve of a profile with an edge on the axis is swept a third of a turn at most per face, so the boolean can intersect its faces): seeds 28 and 95 start with a rectangle on the axis turned half a turn, now two quarter turns. Seed 28 (joined by a cylinder) has more faces for the same shape (5902.1214 -> 5902.1117 mm3, the new faces' meshes). Seed 95's hole into it failed with 'boolean Cut: no result' on the half-turn faces and now builds, so the generator's later picks move. Values from Linux CI.","seeds":[28,95]},
+  {"reason":"Independently reproduced on unmodified main 1e984c8e on Windows; the extrusion branch produces the same values. Shared references and tolerances remain unchanged.","platform":"windows","seeds":[1,23,34,67]},
+  {"reason":"Seed 94 independently reproduced on unmodified main c5992386 on Windows after the extrusion rebase: the same steps, outcome and geometry values as the rebased branch. Shared references and tolerances remain unchanged.","platform":"windows","seeds":[94]}
  ]
 }