PostGIS ST_MakeEnvelope size direction change on geography

Question

This doesn't really effect me but I was just wondering if you find this a little bit strange or not. I think this issue is probably cause of this question also but I think the accepted answer does not apply to my question.

I am using ST_MakeEnvelope and it tax xmin,ymin,xmax,ymax values as arguments.

I realized that for example, if longitude distance is over 180 degrees, it seems to wrap around other way around (That is what I think is happening, I may be wrong?).

I thought it would always go from xmin to xmax. As in xmin=-81 to xmax=100 should go in direction -81,-80,-79...0,1...100 , yet it seems to make a boundary box going other way -81,-82,-83...(+/–)180,179...100

Am I imagining things and is this suppose to work like this? Any reason why this way is better than always assuming we should travel from xmin to xmin+ direction?

SELECT tbl1.column1, tbl2.column1, tbl1.column2 && tbl2.column2 AS overlaps FROM 
( VALUES
(1, 'POINT(0 0)'::geography), 
(2, 'POINT(-85 0)'::geography) ) AS tbl1,
( VALUES
(3, ST_MakeEnvelope(-80, -10, 100, 20, 4326)::geography), 
(4, ST_MakeEnvelope(-81, -10, 100, 20, 4326)::geography)) AS tbl2;

Results in:

 column1 | column1 | overlaps 
---------+---------+----------
       1 |       3 | t
       1 |       4 | f
       2 |       3 | f
       2 |       4 | t

Answer 1

By the geography advanced FAQ http://postgis.net/docs/using_postgis_dbmanagement.html#idm1391 what you experience is intentional

What is the longest arc you can process?

We use great circle arcs as the "interpolation line" between two points. That means any two points are actually joined up two ways, depending on which direction you travel along the great circle. All our code assumes that the points are joined by the shorter of the two paths along the great circle. As a consequence, shapes that have arcs of more than 180 degrees will not be correctly modelled.