libMesh: A C++ Finite Element Library

On 11/13/2012 11:38 PM, Roy Stogner wrote:
>
> On Tue, 13 Nov 2012, David Knezevic wrote:
>
>> I'd like to add a new reduced basis example that models an acoustic 
>> horn (attached). The nice thing about this one is that it tests out 
>> the RB code in the --enable-complex case.
>>
>> However, I didn't want to check it in right away since it breaks some 
>> tests with 0.8.0 around the corner. I was wondering if you might run 
>> it through buildbot etc, and if it looks good at your end then I'd 
>> like to check it in before 0.8.0 is finalized if possible.
>
> I'm afraid that right now BuildBot only does its full suite of
> regression tests on our svn trunk, and our build server is too
> hammered right now (mostly with other PECOS stuff, but libMesh is
> probably the most expensive single component) to add a separate
> libmesh/testing branch to it.
>
> If others agree, I'd say go ahead and check it in to trunk now but be
> prepared to revert it and save it for 0.9.0 if there are any problems
> requiring more than trivial fixes.


OK, it's in. It's working fine for me, though as I mentioned it might be 
good to change the solver settings in the Makefile this case since GMRES 
seems to be slow... I've been using direct solvers for this problem, 
which work very well.

David

On Tue, 13 Nov 2012, David Knezevic wrote:

> However, I didn't want to check it in right away since it breaks some tests 
> with 0.8.0 around the corner. I was wondering if you might run it through 
> buildbot etc, and if it looks good at your end then I'd like to check it in 
> before 0.8.0 is finalized if possible.

It's giving compilation failures with our BuildBot configurations at
the moment - try it on a libMesh configured with
--disable-legacy-using-namespace and see if you can fix those?
---
Roy

On 11/14/2012 04:10 PM, Roy Stogner wrote:
>
> On Tue, 13 Nov 2012, David Knezevic wrote:
>
>> However, I didn't want to check it in right away since it breaks some 
>> tests with 0.8.0 around the corner. I was wondering if you might run 
>> it through buildbot etc, and if it looks good at your end then I'd 
>> like to check it in before 0.8.0 is finalized if possible.
>
> It's giving compilation failures with our BuildBot configurations at
> the moment - try it on a libMesh configured with
> --disable-legacy-using-namespace and see if you can fix those?


hmm, I'm not getting any compilation errors with

./configure --enable-everything --disable-slepc --enable-complex 
--disable-legacy-using-namespace

or

./configure --enable-everything --disable-legacy-using-namespace

What is an example configuration that is failing?

David

On Nov 14, 2012, at 7:43 PM, David Knezevic <dknezevic@...> wrote:

> 
> On 11/14/2012 04:10 PM, Roy Stogner wrote:
>> 
>> On Tue, 13 Nov 2012, David Knezevic wrote:
>> 
>>> However, I didn't want to check it in right away since it breaks some 
>>> tests with 0.8.0 around the corner. I was wondering if you might run 
>>> it through buildbot etc, and if it looks good at your end then I'd 
>>> like to check it in before 0.8.0 is finalized if possible.
>> 
>> It's giving compilation failures with our BuildBot configurations at
>> the moment - try it on a libMesh configured with
>> --disable-legacy-using-namespace and see if you can fix those?
> 
> 
> hmm, I'm not getting any compilation errors with
> 
> ./configure --enable-everything --disable-slepc --enable-complex 
> --disable-legacy-using-namespace
> 
> or
> 
> ./configure --enable-everything --disable-legacy-using-namespace
> 
> What is an example configuration that is failing?

Roy, what svn revision was failing?

I fixed an issue with r6366, it may be OK now David.

-Ben

On 11/14/2012 08:46 PM, Kirk, Benjamin (JSC-EG311) wrote:
> On Nov 14, 2012, at 7:43 PM, David Knezevic <dknezevic@...> wrote:
>
>> On 11/14/2012 04:10 PM, Roy Stogner wrote:
>>> On Tue, 13 Nov 2012, David Knezevic wrote:
>>>
>>>> However, I didn't want to check it in right away since it breaks some
>>>> tests with 0.8.0 around the corner. I was wondering if you might run
>>>> it through buildbot etc, and if it looks good at your end then I'd
>>>> like to check it in before 0.8.0 is finalized if possible.
>>> It's giving compilation failures with our BuildBot configurations at
>>> the moment - try it on a libMesh configured with
>>> --disable-legacy-using-namespace and see if you can fix those?
>>
>> hmm, I'm not getting any compilation errors with
>>
>> ./configure --enable-everything --disable-slepc --enable-complex
>> --disable-legacy-using-namespace
>>
>> or
>>
>> ./configure --enable-everything --disable-legacy-using-namespace
>>
>> What is an example configuration that is failing?
> Roy, what svn revision was failing?
>
> I fixed an issue with r6366, it may be OK now David.
>
> -Ben

Ah OK, thanks.