Make stricter checks on central longitude

[PaulWessel]

There are a few cases where GMT was inconsistent and wrong - we will label these as bugs:

1. When the user specified a specific central longitude, it is to be respected, and for a global map the w/e borders are set at 180 distance from that longitude.
2. If not a global map then if the distance from the specified central meridian to either border exceeds 180 then we yield a map region error. Currently, something slightly different may happen if -R's +r is used for rectangular areas.
3. If no central meridian is set then it continues to default to the mid longitude.

Tests pass and some of the odd and strange plots found in #7291 are now dealt with.

[seisman]

The following command no longer works after this PR:

gmt basemap -Bag -Bzag -JR270/10c -Jz5 -R10/70/-5/10/0/1 -p225/30 -pdf map                           
basemap [ERROR]: Western boundary is > 180 degrees from specified central meridian and thus your region is invalid
basemap [ERROR]: General map projection error

[seisman]

Before this PR was merged, the command produces the following figure. As you can see, the location of the z-axis doesn't match well with the map boundary. Is it intentional?

[PaulWessel]

Couple of issues I guess:

1. Since Winkel Tripel has separate west and east borders, they cannot be more than 180 degrees from the center, and the center longitude cannot be outside the range. This is a new check to prevent crazy situations.
2. Since no +r is used in setting the region, the sliver of Winkel map given w/e/s/n has variable height. The back wall mechanism has no knowledge of this and will not draw some curved back wall.

Probably best to improve the docs for -B on this.