Fixes #2482

And also add proj_context_errno_string()

Revise gie 'expect failure errno XXXX' strings

Gone are pj_malloc, pj_calloc, pj_dalloc and pj_dealloc. Their primary
function as API memory functions in proj_api.h is no longer there and
the other use as a workaround for old errno problems is no longer valid
either.

Replaced with malloc and free across the codebase.

- Move the generic method initiated from adams_ws2 to a
  pj_generic_inverse_2d() method
- Use it in adams_ws2
- Use it in wink2

Fixes https://github.com/qgis/QGIS/issues/35512

Otherwise it returns junk (negative latitudes are mapped to
the same location as positive latitudes)

I'm a bit confused by PROJ peirce_q implementation. Looking at
the projection of the world, it looks like this matches the diamond
formulation of the right map shown at
https://desktop.arcgis.com/en/arcmap/latest/map/projections/peirce-quincuncial.htm,
but limited to the inner square of this diamond (which corresponds
to the northern hemisphere). We lack the 4 triangles on left, top,
right and bottom for the southern hemisphere. Furthermore, this
formulation of peirce_q does not seem to have the quincuncial
property of the square formulation (left images of the above ESRI
doc), or the one at https://en.wikipedia.org/wiki/Peirce_quincuncial_projection ...

This commit adds five new projections to PROJ:

    adams_hemi: Adams Hemisphere in a Square
    adams_wsI:  Adams World in a Square I
    adams_wsII: Adams World in a Square II
    guyou:      Guyou
    peirce_q:   Pierce Quincuncial

The code originates from Gerry Evendens libproject and has been adapted
to work with modern PROJ. To ensure that the modified code works as
intended extensive test data has been created using libproject and
sproj so that no errors occured when porting from libproject to PROJ.
The test data is wrapped in a gie files. All test cases reproduce
results from libproject at the mm level.