MuonReadoutGeometry - Correct the input width

Hi everybody,

in !56646 (merged), large parts of the MMReadoutGeometry were rewritten to accommodate for a better phi constraint from the fit. The inputWidth parameter is defined as the shortest distance between two neighbouring strips. I.e. the length of the line AB or AE in the drawing below for the eta and stereo strips, respectively.

In the parameter description, the input width is calculated from the input pitch, which is commonly expressed as the length of the line AB for both eta and stereo strips. To obtain the actual length of AE, one has to multiply the line with the cos of the stereo angle instead of dividing it. Otherwise, the strip would intersect point E. The inputWidth is only used in the micromega digitization to determine the evolution of the avalanche electrons. Eventhough the effect is small for a realistic stereo angle of 1.56 deg, we should fix it.

UPDATE

We checked the latest data run which is reconstructed with 22.0.100 and the latest geometry. The plot below shows the residuals of the stereo layers hot of the press. They're just horrible.

We traced down the issue to a confusion of the actual definition of the stereo angle. It's defined from the positive y-axis in a clockwise manner.

Adding the labels urgent and frozen-tier0-violating

Also done in this MR. Remove an overseen correction of the boundary parameters in the MS track fit.

Tagging for information: @pscholer, @sangelid, @tlari, @mbandier, @nstyles, @elmsheus, @christos, @jmaurer, @pberta, @iodice, @rosati, @pfleisch, @jcatmore, @mhodgkin