There are quite many T-junctions that could easily be removed by 'merging/moving' vertices. The handle could use ndis and chrd prims, even some more cyli.

If you want to fix these, do it.

I submitted this fix to clear the comment errors and I had to flip the BFC while I was at it. I have no time or interest in changing anything else.

OK, I will. Thanks for the permission. I guess the same goes for the cannon base.

Subparted half. Removed T-junctions, improved prims around handle.

Changed inner cyl to prim as suggested. Used standard 16-res, but it clashes a bit with the 24-fold symmetry at the top. Would 64-hires be overkill?
Removed history from subfile as instructed.

Note: all part fixes are fair game for anyone to edit.

Reverted to 24-fold inner cylinder to honor original authors design. I also find it pleasing my eye more than the 16-fold standard res cylinder.
Added missing cond lines to inner cyl and the small fin at the top. Removed more T-junctions. Added some box and rect prims.
Changed date as it is now over midnight.
I'm now done with the part (assuming I haven't introduced any direct errors). As Orion points out, anyone may edit.

cut to a quarter, used box prims for the ridges

does anyone have both the new (92926) and the old (2439) one to compare the inner diameter?
this has R17, 92926 has R16.25 -> should it be the same? which one?

I do. The new has clearly thicker walls even considering the change in patterning. The radiuses changes with depth, about 0.7 LDU wider at the top. It appears that the outer diameter has been kept the same to fit the same lids.

Does that mean 92926 has a step inside? At which heigth? Or is it sloped? What are the two diameters?

No step inside. The insides are slightly conical and not perfect cylinders. Presumably, this helps unmolding, but need not be modeled.

2439 - old part
Vertical grooves on the outside from the bottom all the way to the top.
Outer diameter approximately the same as the lid, i.e 39.4 LDU.
Inner radius at top about 17 LDU (the one I measured was a tiny bit (0.2 LDU) below 17 LDU. There may be some variation between batches). Radius at bottom about 0.7 LDU smaller.

92926 - new part
Vertical grooves from the bottom to about 9.8 LDU from the top.
The last 9.8 LDU from the top is a cylinder with rougly the same diameter as the lid.
The cylinder is about 39.4 LDU in diameter whereas the grooved section is 39.2 LDU in diameter.
Inner radius at the top about 16.36 LDU. Inner radius at bottom about 0.7 LDU smaller.

Visually, the apparent wall thickness of the new part is clearly larger than the old part. The difference accounts for a) the cylinder vs grooved section (0.2 LDU), groove depth (1 LDU), and the wall thickness itself.
The wall thickness for the old is between 2.7 and 3.15 LDU
The wall thickness for the new is between 3.34 and 3.75 LDU

In my opinion, the dimensions for both parts are ok since the differences between the model and the real part is below 1 LDU.

The wall thicknesses above are for the top. The wall is probably thicker at the bottom.