Hello,
This implements something I have been thinking about doing for
a little while. The idea is have the code for each testcase be
in its own subdirectory. For instance, this change moves the code
for null_record.exp to the null_record/ subdir.
The reason for this is to avoid name clashes. It is much more common
in Ada to split the code between several compilation units. The model
used in C where you usually have one .c file per testcase doesn't
apply well to Ada.
The framework I'm trying to implement is that we would have one
.exp file named <testcase>.exp, and have the sources located in
subdir <testcase>/.