Libmesh has been distributing and running with the new ExodusII library for
over a month now and I haven't seen any issues on the mailing list regarding
the new version so I'm preparing the next patch in this set.
IMPORTANT:
This change could potentially break users doing regression testing with the
Exodus output format as the output meshes will contain a block, sideset and
nodeset with an id of zero (previously these entities were all set to 255),
which may differ with a mesh differencing tool. On the plus side this will
make the Exodus output more consistent with the other output formats in Libmesh
and will allow us to move forward with increasing the size of subdomain_id
without further effecting Exodus output.
Let me know if there is any reason not to commit this patch, otherwise I'll get
it applied in the next day or two.
Cody
Begin forwarded message:
> From: Cody Permann <[email protected]>
> Date: July 12, 2011 8:25:30 AM MDT
> To: Roy Stogner <[email protected]>
> Cc: libmesh-devel <[email protected]>
> Subject: Re: [Libmesh-devel] Exodus II Update
>
> Sounds good - here is my proposed plan
> - New ExodusII library -> COMMITED
> - Exodus/Nemesis patches to write 0 IDs instead of
> std::numeric_limits<T>::max() for certain operations
> - Configurable subdomainID size patch (defaults to 2 bytes)
>
> Vacation time
>
> - NetCDF patch
>
> Cody
>
> On Jul 12, 2011, at 12:04 AM, Roy Stogner wrote:
>
>>
>>
>> On Mon, 11 Jul 2011, Cody Permann wrote:
>>
>>> This update looks good from my perspective, however maybe we need to
>>> update netCDF too? Opinions?
>>
>> My vote would be to update netCDF as long as we're at it, but if
>> you've got things working without such an update there's no reason to
>> bundle everything into a single commit; start with exodus-alone.
>>
>>> I'm happy to commit my changes but I'd like to hear back from the
>>> developers before doing so.
>>
>> I'd say go ahead and commit so everyone else can start testing too.
>>
>> Besides, the svn head is already in a state of shame at the moment.
>> My new NodeConstraints feature goes into an infinite loop with
>> ParallelMesh, so there's four "--enable-everything --enable-parmesh"
>> regression test failures that I'm still trying to correct. Even after
>> that, we aren't planning a new official release soon. That nasty
>> PetscVector::localize() issue ought to get a real fix instead of just
>> a workaround first.
>> ---
>> Roy
>
------------------------------------------------------------------------------
BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA
The must-attend event for mobile developers. Connect with experts.
Get tools for creating Super Apps. See the latest technologies.
Sessions, hands-on labs, demos & much more. Register early & save!
http://p.sf.net/sfu/rim-blackberry-1
_______________________________________________
Libmesh-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
