[Bug fortran/55427] [OOP] ICE with class-array-pointer result

["dominiq at lps dot ens.fr" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55427

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|dominiq at lps dot ens.fr          |

--- Comment #4 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> This is fixed on trunk and GNU Fortran (GCC) 5.1.1 20150618 (Red Hat 5.1.1-4).
> Your example even runs correctly!

The ICE is gone with at least 4.9.3, but this leads to wrong code

 F

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x1026dd7b2
#1  0x1026ddf6e
#2  0x7fff8f042529
#3  0x1026d7b38
#4  0x1026d7cd0
Segmentation fault

> I have added Dominique to the CC list in the hope that he can pin down
> better when it was fixed. He seems to maintain the world's largest collection
> of archaic gcc's :-)

Well this fall in one my blind spots!-( 

The ICE disappeared between revisions r201266 (2013-07-26, ICE) and r201631
(2013-08-09, wrong code). In this range, I see r201284 (pr57285), r201286
(pr57991), r201294 (pr58009), r201328 and r201329 (pr57530), r201521 (pr57306),
and r201526 (pr57987). I did not look in detail at these PRs, but I think it is
related to/duplicate of pr57530.

The wrong code has been fixed between revisions r219797 (2015-01-17, wrong
code) and r219823 (2015-01-18, expected results). In this range, I see r219798
(pr60334), r219801 (pr61275, pr60357), r219802 (pr64578), r219814 (pr55901),
and r219818 (pr57959). I let you pick the candidate.

> This can be closed once we pin down approximately when it was fixed.