Does the soil supporting a foundation form an isoceles trapezoid?

This is an interesting question. I don't know the answer, but it would be interesting to discuss. I'm not a structural engineer, but I recall seeing diagrams that show the region of soil that supports a foundation has a cross section that forms an isosceles trapezoid underneath the foundation. If this is true, then if the wall of the the trench intersects this trapezoid, the foundation would tend to push the wall of the trench and collapse it. By this theory, if the trench is far enough away so the wall of the trench does not intersect the trapezoid, the foundation would have no effect. I don't recall if the angle of the trapezoid depends on the soil type or not.

I believe you are missing the point. With answer to your question: yes, in case when eccentric load is applied to foundation then pressure at formation level has a trapezoidal shape, however it is not isoceles trapezoid. The pressure values can vary from positive to negative - albeit tension will not develop in soil. It is recommended that pressure below whole width of footing is positive i.e. compression; unless a temporary case is considere - like under windload.

The check of design is a simple comparison of allowable bearing pressure to the maxium applied, i.e. at the highest point of the trapezoid.

Howere, my question still remains unanswered: how close to the edge of unreinforced foundation can a wall be build before concrete cracks - is there an equesion somewhere in a code, or a good practice rule? For anchors it is allowable (depending on anchor type) to drill up to 50mm from concrete edge. Considering this it seems that the idealized load from wall could be applied as close as 50mm from concrete edge, if not closer to foundation face.