Hi PA,
Paul-Antoine Arras wrote:
This is now fixed and the testcase included (map-subarray-15.f90).
...
Multiple levels of nested allocatable components are now supported as
well and the relevant testcase included (map-subarray-14.f90).
There is still one specific case that is not handled properly (no
regression though) as demonstrated by another test
(map-subarray-16.f90): mapping both a subcomponent and the parent
struct on the same target directive.
This revised patch also incorporates a few formatting fixes.
Thanks!
Can you update PR120505 after the commits to state what remains to be done? -
Or,
alternatively and possibly better, create a new PR – and add a refer to it in
PR120505. If you file the PR first, you could update the should-fail TODO of
map-subarray-16.f90 to the new PR before committing.
(But I would refer to the new PR only after 1/2 and 2/2 have been committed, i.e
as last comment in PR120505 - such that it stands out at the end.)
Regarding map-subarray-16.f90: That not only "Demonstrate[s] a bogus error"
(preexisting) but it is also a check that this code doesn't ICE (which it
did with in an older version of this patch.)
* * *
Commit log text:
This patch ensures that all such intervening allocatables in a reference chain
is properly mapped. For the above example, the frontend has to create the
'are properly …' (now that you changed it to 'all allocatables').
Otherwise: LGTM. Thanks for the patch and sorry for the slow review!
Tobias
PS: Likewise for this patch, I'd prefer if the AuthorDate of the commit could
be newer than Dec 9, 2025, esp. given that later changes happened. (I know that
CommitterDate also exists, but it is not shown by default and I find old
AuthorDates confusing, esp. if it does not match the latest change. But this
is not a showstopper.)
