Reestablish CPT hinge for master tables

[PaulWessel]

We (I) somehow lost a line or two in gmt_support.c. Added a test as well so that we are warned if this tests fails again. The CPT polar has a soft hinge at 0 and that is only activated via the +h[hinge] modifier at the end of the selected CPT. Output is below. Note that

1. top plot sets no hinge so the white center is at the mid-point -1368.5.
2. Middle panel uses a hinge of 2000 and creates a new cpt via makecpt. Now the white point is at z = 2000.
3. Bottom panel activates the soft-hinge at 0 so that is where the white point falls.

All tests pass (well, that use to pass).

[Esteban82]

Closes #8102