[Dune-devel] idset.subId() only for codim=0?
Martin Nolte
nolte at mathematik.uni-freiburg.de
Mon Feb 22 14:21:18 CET 2016
Hi Aleksejs, hi all,
as I was the original proposer of this interface change, I should maybe say
some words about it.
First of all, the IndexSet interface class forwards the subIndex methods for
all codimension to the implementation (since 2010, as entioned by Carsten) and
all grid implementation in dune-grid support it.
For some reason, the IdSets seem to have been overlooked. But as the
information is available for index sets, I would be very surprised if it
cannot be provided for IdSets, too. Maybe we should add this as an open issue.
Unless you are working on an adaptive code or require an absolute orientation
(in contrast to a relative one), you should be able to use the index sets instead.
Best,
Martin
On 02/22/2016 01:56 PM, Carsten Gräser wrote:
> Hi,
>
> Am 22.02.2016 um 13:20 schrieb Steffen Müthing:
>> Hi,
>>
>>> Am 22.02.2016 um 12:27 schrieb Jö Fahlke <jorrit at jorrit.de>:
>>>
>>> Am Mon, 22. Feb 2016, 11:54:42 +0100 schrieb Aleksejs Fomins:
>>>> Date: Mon, 22 Feb 2016 11:54:42 +0100
>>>> From: Aleksejs Fomins <aleksejs.fomins at lspr.ch>
>>>> To: "dune-devel at dune-project.org" <dune-devel at dune-project.org>
>>>> Subject: [Dune-devel] idset.subId() only for codim=0?
>>>
>>>> Can somebody please explain what is the motivation behind idset.subId(Entity, index, codim) in dune-grid being restricted to codim=0 entities?
>>>>
>>>> I was trying to find the Ids of the vertices of a face in 3D.
>>>
>>> Well, I think you are the first one to request that feature. And I think that
>>> if you require that functionality in general for all grids, most grids would
>>> have to store *a lot* more connectivity information than they are doing
>>> already to implement it. So that may be the reason why this has never been
>>> part of the interface.
>>
>> yes, AFAIK that’s basically the reason - for the same reason, there is only an
>> index_set.subIndex() and an entity.subEntity() for codim 0 entities.
> that's not correct: On the 2010 developer meeting we decided that
> the subIndex() method should also be added for other codimensions:
>
> http://users.dune-project.org/projects/dune-dev-meeting-201/wiki/Protokoll#subIndex-interface
>
> The reason for this is that otherwise there's no way to get the
> embedding of subentities without doing coordinate comparisons.
> See also the corresponding flyspray task:
>
> https://gitlab.dune-project.org/flyspray/FS/issues/818
>
> Unfortunately, while trying to fix this one
>
> https://gitlab.dune-project.org/flyspray/FS/issues/1709
>
> for 2.4, it turned out, that this method seems to be broken
> (i.e. compilation fails). Since noone ever noticed this it
> was postponed to 3.0.
>
> Best,
> Carsten
>
>
>
>
> _______________________________________________
> Dune-devel mailing list
> Dune-devel at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune-devel
>
--
Dr. Martin Nolte <nolte at mathematik.uni-freiburg.de>
Universität Freiburg phone: +49-761-203-5630
Abteilung für angewandte Mathematik fax: +49-761-203-5632
Hermann-Herder-Straße 10
79104 Freiburg, Germany
More information about the Dune-devel
mailing list