On 9/23/25 1:21 AM, Paul Richard Thomas wrote:
Hi All,
This one was incorrectly labelled as a PDT PR by yours truly :-) That
said, it seemed so peculiar that I investigated.
The fix is somewhat minimalist, on the assumption that resolve.cc will
pick up any problems while identifying the specific procedure. The
check that I have implemented is that there is at least one proc
pointer component that is a subroutine or function. It could have been
taken further by delving into the derived type default initializer
component to check that there is indeed a generic with the
'interface_name'. However, this is exactly what is done further
downstream.
Regtested on FC42/x86_64 - OK for mainline?
Paul
Yes OK. Tested here as well.
Regards,
Jerry
