[Patch, fortran] PR42736 - [4.3/4.4/4.5 Regression] Wrong-code with allocatable or pointer components in elemental functions

[Paul Richard Thomas]

Dera Tobias,

> Thanks! I had also some thoughts about examples, but they never called
> gfc_check_dependency; I will also try to find an example the evening -
> and be it just to convince myself that the current algorithm is mostly
> sufficient.

It turns out to be trivial to find a failing example.  Make the
component 'Value' of UnitValue a pointer, provide CENTIMETER%Value
with a target 'forty_two' and, Bob's your uncle, the segfault is back.

I am on the search for a fix to this.

Paul