This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: gfortran 4.4 regressions
- From: "Paul Richard Thomas" <paul dot richard dot thomas at gmail dot com>
- To: "Tobias Burnus" <burnus at net-b dot de>, "Mikael Morin" <mikael dot morin at tele2 dot fr>, "Daniel Kraft" <d at domob dot eu>
- Cc: "fortran at gcc dot gnu dot org" <fortran at gcc dot gnu dot org>
- Date: Mon, 19 Jan 2009 07:06:19 +0100
- Subject: Re: gfortran 4.4 regressions
- References: <4973AFC3.6040405@net-b.de>
Dear Tobias, Mikael and Daniel,
> 38859 wrong-code [4.3/4.4 Regression] ubound and lbound treat
> structure co...
> (a one-line patch by Mikael exists, which may nor may not be enough)
As it happens, I have been looking at PR38852, which is strongly
related but on the runtime. Mikael's patch parallels code in
trans-intrinsic.c(gfc_conv_intrinsic_bound) and will be enough IMHO.
>
> 38887 wrong-code [4.4 Regression] run-time abort for MVBITS with
> run-time ...
> 38883 [4.4 Regression] ICE for MVBITS with derived type argumen...
> (For these two, something goes wrong with copying/array (un)packing
> the TO= argument)
If Daniel is available, he should certainly look at these two, since
he has most recently laid hands on this part of the compiler.
>
> 38907 ice-on-valid-code [4.3/4.4 Regression ] ICE when contained
> function has sam...
This has my grubby fingers all over it - I'll look tonight
>
> 37754 -- [4.4 Regression] READ I/O Performance regression from 4.3...
> (If I understood correctly, Jerry might have a patch which should
> still get in 4.4. I think fixing this can be deferred to 4.5.)
Agreed - I think that a performance fix is too risky for this stage.
Paul
--
The knack of flying is learning how to throw yourself at the ground and miss.
--Hitchhikers Guide to the Galaxy