https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87043
Dominique d'Humieres <dominiq at lps dot ens.fr> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |WAITING
Last reconfirmed| |2018-10-25
Ever confirmed|0 |1
--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> Not sure if
>
> allocate(error_messages(world_size), stat=allocstat)
> max_err_len = maxval(len_trim(error_messages))
>
> is well-defined fortran iff allocation fails? I suppose not, so
> warning is correct, testcase is broken?
There is no warning if I replace
if (world_rank == 0) then
with
if (world_rank == 0 .and. allocstat == 0) then
> Oh, I see your point. But then still the question is why this warning
> is not shown when compiling with -Og.
Why do you expect middle-end warnings to be independent of the optimization
level?
In any case, if this PR is not closed, its component should be moved to
middle-end.
