https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121628
kargls at comcast dot net changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |kargls at comcast dot net
--- Comment #12 from kargls at comcast dot net ---
(In reply to Jerry DeLisle from comment #11)
> I see one testsuite failure here:
>
> Running /home/jerry/dev/trunk/gcc/testsuite/gfortran.dg/dg.exp ...
> FAIL: gfortran.dg/array_memcpy_2.f90 -O scan-tree-dump-times original
> "memcpy" 2
Well, technically that testcase is bogus. 'y' is unallocated
and is referenced on the RHS of 'z = y'.
If Christopher's patch is doing a deep copy, it likely
is trying to grab some undefined internal information
that an allocated 'y' would actually have defined.
