This meta-bug tries to list all rejects-valid and ice-on-valid-code that can be triggered with standard Fortran 95 conforming code, are not arch specific, and can not reasonably be called a limit of the compiler. In full agreement with pault, I think fixing these should be a priority.
only 1 open since 2005, 2 open since 2006, others are 2007.
Should we open another PR for wrong-code errors?
(In reply to comment #2)
> Should we open another PR for wrong-code errors?
no, I overlooked that keyword, and they belong to this category (though I'll again ignore arch specific ones). I'll add them to the list shortly.
from the 19 wrong code bugs I've only retained 8 that I judged as user visible, F95, and triggered without additional options.