This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Patch, fortran] Derived type extension


BTW I forgot to mention that there is a separate patch included with
the derived type extension, which answers the remark about temporaries
being produced in assignments to pointer arrays.  Clearly, the
possibility of aliassing means that, in general, the appearance of
pointers on both sides of the assignment means that a temporary must
be produced.  However, in the specific case that the same pointer
symbol appears on the lhs and the rhs, aliassing has no effect and the
normal dependency analysis is good.  I was going to commit it as
"obvious":-)

Paul

On Mon, Jul 28, 2008 at 5:46 PM, Salvatore Filippone
<salvatore.filippone@uniroma2.it> wrote:
> Il giorno lun, 28/07/2008 alle 17.41 +0200, Paul Richard Thomas ha
> scritto:
>> Dear All,
>>
>> > I also applied his patch this morning to a clean tree with the same Rev.
>> > 138204 and I got no patch failures. (I think I had some fuzzy merges.) The
>> > trunk also successfully bootstrapped and so far (char_eoshift_1.f90)
>> > check-gfortran also shows no problems.
>> >
>> > Thus, I do not know why you had problems and I had none.
>>
>> Nor do I - the patch is based on last night's trunk.  Could we be
>> partially tupled?
>>
>> Paul
>
> Never mind, the patch text got corrupted (somehow) while transferring to
> my hard disk.
> Refreshed  both source tree and patch file (more carefully this time),
> and now it has bootstrapped.
>
> Thanks
> Salvatore
>
>



-- 
The knack of flying is learning how to throw yourself at the ground and miss.
 --Hitchhikers Guide to the Galaxy


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]