]> https://gitweb.dealii.org/ - dealii.git/log
dealii.git
8 years agoMerge pull request #1859 from luca-heltai/fix-diag-blop
Wolfgang Bangerth [Wed, 11 Nov 2015 17:40:10 +0000 (12:40 -0500)]
Merge pull request #1859 from luca-heltai/fix-diag-blop

Fixed bug in DiagonalBlockLinearOperator.

8 years agoMerge pull request #1860 from bangerth/add-output-object-to-get_data
Luca Heltai [Wed, 11 Nov 2015 15:52:53 +0000 (16:52 +0100)]
Merge pull request #1860 from bangerth/add-output-object-to-get_data

Add an output data object to FE::get_*_data().

8 years agoMerge pull request #1854 from davydden/slepc_unit_tests_initial_space
Luca Heltai [Wed, 11 Nov 2015 15:51:24 +0000 (16:51 +0100)]
Merge pull request #1854 from davydden/slepc_unit_tests_initial_space

Slepc unit tests and initial space

8 years agoFixed indentation. 1859/head
Luca Heltai [Wed, 11 Nov 2015 08:48:51 +0000 (09:48 +0100)]
Fixed indentation.

8 years agoAdded test for BlockDiagonalOperator bug.
Luca Heltai [Tue, 10 Nov 2015 23:22:26 +0000 (00:22 +0100)]
Added test for BlockDiagonalOperator bug.

8 years agoMerge pull request #1858 from luca-heltai/fix-header
Luca Heltai [Tue, 10 Nov 2015 22:57:13 +0000 (23:57 +0100)]
Merge pull request #1858 from luca-heltai/fix-header

Fixed header.

8 years agoAdded symmetric tensor to sacado_product_type. 1858/head
Luca Heltai [Tue, 10 Nov 2015 22:53:33 +0000 (23:53 +0100)]
Added symmetric tensor to sacado_product_type.

8 years agoAdd an output data object to FE::get_*_data(). 1860/head
Wolfgang Bangerth [Tue, 10 Nov 2015 21:31:21 +0000 (15:31 -0600)]
Add an output data object to FE::get_*_data().

This allows finite elements to already fill certain output data elements
at the time the get_data() function is called, rather than having to later
figure out whether this is the first time (or not) they get into
fill_fe_*_values().

In its current form, the patch only adds that possibility to passing the
same output object to FiniteElement::get_*_data() that will later be used
when FiniteElement::fill_fe_*_values() will be called. The patch does not
actually implement any of the elements using this -- I will leave this
to a later patch addressing #1824.

8 years agoExplicitly specify templates. Not specifying them would make the compiler confused...
Luca Heltai [Tue, 10 Nov 2015 19:33:31 +0000 (20:33 +0100)]
Explicitly specify templates. Not specifying them would make the compiler confused about what version of block_operator to use.

8 years agoFixed header.
Luca Heltai [Tue, 10 Nov 2015 18:48:57 +0000 (19:48 +0100)]
Fixed header.

8 years agominor renaming of data members and variables in SLEPc solver 1854/head
Denis Davydov [Tue, 10 Nov 2015 14:30:57 +0000 (15:30 +0100)]
minor renaming of data members and variables in SLEPc solver

8 years agoMerge pull request #1856 from tjhei/step-40-trilpetsc
Timo Heister [Tue, 10 Nov 2015 14:01:43 +0000 (09:01 -0500)]
Merge pull request #1856 from tjhei/step-40-trilpetsc

decide between Trilinos and PETSc in step-40

8 years agodecide between Trilinos and PETSc in step-40 1856/head
Timo Heister [Tue, 10 Nov 2015 13:42:46 +0000 (08:42 -0500)]
decide between Trilinos and PETSc in step-40

based on a patch sent in by Kevin Drzycimski on the mailing list

8 years agoMerge pull request #1855 from asartori86/symmetric_tensor_sacado
Luca Heltai [Tue, 10 Nov 2015 12:08:27 +0000 (13:08 +0100)]
Merge pull request #1855 from asartori86/symmetric_tensor_sacado

specialized scalar_product(SymmetricTensor,Tensor) with sacado

8 years agospecialized scalar_product symmetric*tensor with sacado 1855/head
alberto sartori [Tue, 10 Nov 2015 12:06:35 +0000 (13:06 +0100)]
specialized scalar_product symmetric*tensor with sacado

8 years agoa note in changes about SLEPc
Denis Davydov [Mon, 9 Nov 2015 16:37:03 +0000 (17:37 +0100)]
a note in changes about SLEPc

8 years agoGHEP and SHEP unit tests for SLEPc
Denis Davydov [Mon, 9 Nov 2015 14:41:02 +0000 (15:41 +0100)]
GHEP and SHEP unit tests for SLEPc

8 years agoextra options to Lanczos and GD SLEPc solvers
Denis Davydov [Mon, 9 Nov 2015 14:39:19 +0000 (15:39 +0100)]
extra options to Lanczos and GD SLEPc solvers

8 years agoadd set_initial_space() to SLEPc solvers
Denis Davydov [Mon, 9 Nov 2015 14:11:15 +0000 (15:11 +0100)]
add set_initial_space() to SLEPc solvers

8 years agoMerge pull request #1849 from bangerth/update-fill_fe_values-argument-list
Martin Kronbichler [Mon, 9 Nov 2015 14:10:39 +0000 (15:10 +0100)]
Merge pull request #1849 from bangerth/update-fill_fe_values-argument-list

Update fill fe values argument list

8 years agoMerge pull request #1847 from adamkosik/master
Wolfgang Bangerth [Mon, 9 Nov 2015 13:59:53 +0000 (22:59 +0900)]
Merge pull request #1847 from adamkosik/master

FETools::project_dg, initialize output to zero

8 years agoDocument FiniteElement::fill_fe_*_values() arguments. 1849/head
Wolfgang Bangerth [Mon, 9 Nov 2015 01:13:57 +0000 (19:13 -0600)]
Document FiniteElement::fill_fe_*_values() arguments.

8 years agoAdd entry in news/changes.h. 1847/head
Adam Kosik [Mon, 9 Nov 2015 08:18:16 +0000 (09:18 +0100)]
Add entry in news/changes.h.

8 years agoMerge pull request #1852 from tamiko/do_not_run_step45
Martin Kronbichler [Mon, 9 Nov 2015 08:14:56 +0000 (09:14 +0100)]
Merge pull request #1852 from tamiko/do_not_run_step45

Testsuite: Bugfix: Do not run step-45

8 years agoMerge pull request #1850 from bangerth/add-test
Martin Kronbichler [Mon, 9 Nov 2015 08:13:48 +0000 (09:13 +0100)]
Merge pull request #1850 from bangerth/add-test

Add test.

8 years agoTestsuite: Bugfix: Do not run step-45 1852/head
Matthias Maier [Mon, 9 Nov 2015 05:21:18 +0000 (23:21 -0600)]
Testsuite: Bugfix: Do not run step-45

step 45 is no longer configuration independent.

8 years agoAdd test. 1850/head
Wolfgang Bangerth [Mon, 9 Nov 2015 01:18:22 +0000 (19:18 -0600)]
Add test.

This is a simplified test of FESystems whose base elements are again FESystems.
The test helped me debug a couple of issues, and might be of help again
at a later time. It is a simpler version of fe/shapes_system.

8 years agoFETools::project_dg, initialize output to zero
adamkosik [Sun, 8 Nov 2015 22:10:57 +0000 (23:10 +0100)]
FETools::project_dg, initialize output to zero

8 years agoMerge pull request #1846 from gassmoeller/fix-real-to-unit-zero-discriminant-again
Denis Davydov [Sun, 8 Nov 2015 09:01:25 +0000 (10:01 +0100)]
Merge pull request #1846 from gassmoeller/fix-real-to-unit-zero-discriminant-again

Fix MappingQ1

8 years agoFix MappingQ1 1846/head
Rene Gassmoeller [Sun, 8 Nov 2015 06:06:47 +0000 (00:06 -0600)]
Fix MappingQ1

8 years agoReshuffle the argument list of FE::fill_fe_*_values().
Wolfgang Bangerth [Sat, 7 Nov 2015 20:59:23 +0000 (14:59 -0600)]
Reshuffle the argument list of FE::fill_fe_*_values().

This brings the argument list in line with the changes made to Mapping in #1204.
In essence, it is all about sorting the list in a halfway reasonable order. No
functional changes.

The patch changes all of the derived finite element classes as well. In some cases
it changes the name of arguments to match a uniform style.

8 years agoMerge pull request #1845 from bangerth/improve-loop-nesting
Wolfgang Bangerth [Sun, 8 Nov 2015 01:52:25 +0000 (10:52 +0900)]
Merge pull request #1845 from bangerth/improve-loop-nesting

Improve loop nesting.

8 years agoMake the same kind of changes also to FESystem. 1845/head
Wolfgang Bangerth [Sun, 8 Nov 2015 01:49:53 +0000 (19:49 -0600)]
Make the same kind of changes also to FESystem.

8 years agoImprove loop nesting.
Wolfgang Bangerth [Fri, 6 Nov 2015 22:13:43 +0000 (16:13 -0600)]
Improve loop nesting.

We had a number of places of the form
  for (...tight loop...)
    if (constant condition)
      data update;

The compiler almost certainly can hoist the condition out of the loop,
but why make it this complicated for the compiler. I also find the code
easier to read because the loop is really only over a counting index
and does not carry any particular meaning at a level higher than the
if-statement.

8 years agoMerge pull request #1841 from bangerth/undo-parallelization
Martin Kronbichler [Sat, 7 Nov 2015 09:52:29 +0000 (10:52 +0100)]
Merge pull request #1841 from bangerth/undo-parallelization

Undo parallelization

8 years agoMerge pull request #1844 from bangerth/eliminate-remote-NUMA-memory-accesses
Wolfgang Bangerth [Fri, 6 Nov 2015 22:51:08 +0000 (07:51 +0900)]
Merge pull request #1844 from bangerth/eliminate-remote-NUMA-memory-accesses

Address @kronbichler's comments for #1825.

8 years agoAddress @kronbichler's comments for #1825. 1844/head
Wolfgang Bangerth [Fri, 6 Nov 2015 22:47:49 +0000 (16:47 -0600)]
Address @kronbichler's comments for #1825.

8 years agoMerge pull request #1825 from bangerth/eliminate-remote-NUMA-memory-accesses
Wolfgang Bangerth [Fri, 6 Nov 2015 22:46:32 +0000 (07:46 +0900)]
Merge pull request #1825 from bangerth/eliminate-remote-NUMA-memory-accesses

Rather than creating a copy-data object in build_one_patch, build it locally and move it

8 years agoRe-inline a function previously broken out into its own context. 1841/head
Wolfgang Bangerth [Fri, 6 Nov 2015 16:02:01 +0000 (10:02 -0600)]
Re-inline a function previously broken out into its own context.

8 years agoUndo parallelization in FESystem.
Wolfgang Bangerth [Fri, 6 Nov 2015 07:49:36 +0000 (01:49 -0600)]
Undo parallelization in FESystem.

As explained in comments, parallelizing this sort of operation destroys
data locality in NUMA contexts.

8 years agoMerge pull request #1838 from bangerth/replace-update-once-each
Bruno Turcksin [Fri, 6 Nov 2015 17:59:51 +0000 (11:59 -0600)]
Merge pull request #1838 from bangerth/replace-update-once-each

Replace update_once/each

8 years agoReplace update_once/each by requires_update_flags() in FiniteElement. 1838/head
Wolfgang Bangerth [Tue, 3 Nov 2015 04:33:52 +0000 (22:33 -0600)]
Replace update_once/each by requires_update_flags() in FiniteElement.

The purpose of this patch is to align the finite element classes with the way the
mapping classes have already been converted. Specifically, there is no need for
any of the users of finite element classes to actually know whether a FE implementation
wants to treat a particular flags as update_once or update_each. This is an internal
decision. Rather, all we need to know is what flags they need overall. This is now
communicated by the new FiniteElement::requires_update_flags() function.

The update_once() and update_each() functions have been retained -- for now -- as
internal functions individual elements can implement, but they are no longer virtual.

8 years agoRather than creating a copy-data object in build_one_patch, build it locally and... 1825/head
Wolfgang Bangerth [Mon, 2 Nov 2015 13:53:24 +0000 (07:53 -0600)]
Rather than creating a copy-data object in build_one_patch, build it locally and move it.

This avoids a bunch of rather annoying remote-NUMA-region memory write contention issues
that @kronbichler identified.

8 years agoMerge pull request #1835 from bangerth/avoid-unnecessary-memory-allocation
Martin Kronbichler [Thu, 5 Nov 2015 19:36:59 +0000 (20:36 +0100)]
Merge pull request #1835 from bangerth/avoid-unnecessary-memory-allocation

Avoid unnecessary memory allocation.

8 years agoAvoid unnecessary memory allocation. 1835/head
Wolfgang Bangerth [Thu, 5 Nov 2015 05:59:59 +0000 (23:59 -0600)]
Avoid unnecessary memory allocation.

At least in cases where we know that the output vector of one function
has the same type as the result type of the current one, we can avoid
allocating a temporary vector.

8 years agoMerge pull request #1837 from tjhei/fix_examples
Martin Kronbichler [Thu, 5 Nov 2015 14:21:24 +0000 (15:21 +0100)]
Merge pull request #1837 from tjhei/fix_examples

Fix examples

8 years agostep-45 requires MPI and Trilinos 1837/head
Timo Heister [Thu, 5 Nov 2015 14:12:33 +0000 (09:12 -0500)]
step-45 requires MPI and Trilinos

8 years agoremove contract() deprecation warning in step-9
Timo Heister [Thu, 5 Nov 2015 14:12:22 +0000 (09:12 -0500)]
remove contract() deprecation warning in step-9

8 years agoMerge pull request #1836 from tjhei/fix_tests4
Martin Kronbichler [Thu, 5 Nov 2015 14:11:41 +0000 (15:11 +0100)]
Merge pull request #1836 from tjhei/fix_tests4

fix tests

8 years agofix tests 1836/head
Timo Heister [Thu, 5 Nov 2015 14:02:50 +0000 (09:02 -0500)]
fix tests

- use std_cxx11::bind instead of std::bind
- require Trilinos if we use it
- fe/bdm_*: do not set matrix entries <1e-14 to zero (different on
different machines) and always print 0 entries
- use check_solver_within_range where needed

8 years agoMerge pull request #1834 from gassmoeller/triangulation_weights_callback
Wolfgang Bangerth [Thu, 5 Nov 2015 05:18:40 +0000 (14:18 +0900)]
Merge pull request #1834 from gassmoeller/triangulation_weights_callback

Fix test output.

8 years agoFix test output. 1834/head
Rene Gassmoeller [Thu, 5 Nov 2015 03:54:27 +0000 (21:54 -0600)]
Fix test output.

Allow calling repartition() even when no_automatic_refinement flag is not set

8 years agoMerge pull request #1832 from luca-heltai/fix-matrix-creator
David Wells [Wed, 4 Nov 2015 22:07:31 +0000 (17:07 -0500)]
Merge pull request #1832 from luca-heltai/fix-matrix-creator

Fixed bug in one version of create_laplace_matrix.

8 years agoMerge pull request #1833 from drwells/add-latex-brackets
Denis Davydov [Wed, 4 Nov 2015 22:03:55 +0000 (23:03 +0100)]
Merge pull request #1833 from drwells/add-latex-brackets

Add brackets around a LaTeX subscript.

8 years agoFixed bug in matrix creator. 1832/head
Luca Heltai [Wed, 4 Nov 2015 21:45:25 +0000 (22:45 +0100)]
Fixed bug in matrix creator.

8 years agoAdd brackets around a LaTeX subscript. 1833/head
David Wells [Wed, 4 Nov 2015 21:46:57 +0000 (16:46 -0500)]
Add brackets around a LaTeX subscript.

8 years agoMerge pull request #1823 from bangerth/remove-first_cell-mechanism
Bruno Turcksin [Tue, 3 Nov 2015 19:29:34 +0000 (13:29 -0600)]
Merge pull request #1823 from bangerth/remove-first_cell-mechanism

Remove first_cell mechanism in FiniteElement

8 years agoMerge pull request #1777 from drwells/class-names-as-template-arguments
Bruno Turcksin [Tue, 3 Nov 2015 18:56:22 +0000 (12:56 -0600)]
Merge pull request #1777 from drwells/class-names-as-template-arguments

Class names as template arguments

8 years agoMerge pull request #1769 from jperryhouts/add_erfc
Martin Kronbichler [Tue, 3 Nov 2015 14:32:12 +0000 (15:32 +0100)]
Merge pull request #1769 from jperryhouts/add_erfc

Add erfc function to FunctionParser.

8 years agoMerge pull request #1828 from tjhei/fix_tests3
Bruno Turcksin [Tue, 3 Nov 2015 14:27:48 +0000 (08:27 -0600)]
Merge pull request #1828 from tjhei/fix_tests3

fix tests/multigrid/transfer_prebuilt_01

8 years agofix tests/multigrid/transfer_prebuilt_01 1828/head
Timo Heister [Tue, 3 Nov 2015 13:51:20 +0000 (08:51 -0500)]
fix tests/multigrid/transfer_prebuilt_01

I forgot to remove the .release.output file (output is now identical
between release and debug mode).

8 years agoMerge pull request #1827 from bangerth/add-include-guard
Timo Heister [Tue, 3 Nov 2015 13:11:53 +0000 (08:11 -0500)]
Merge pull request #1827 from bangerth/add-include-guard

Add an include guard to tests/base/functions.h.

8 years agoAdd an include guard to tests/base/functions.h. 1827/head
Wolfgang Bangerth [Mon, 2 Nov 2015 23:17:21 +0000 (17:17 -0600)]
Add an include guard to tests/base/functions.h.

Not important, but found while playing with collating testcases into
one.

8 years agoAdded muparser erfc test. 1769/head
Jonathan Perry-Houts [Mon, 2 Nov 2015 21:14:44 +0000 (13:14 -0800)]
Added muparser erfc test.

8 years agoAdd erfc function to FunctionParser.
Jonathan Perry-Houts [Tue, 20 Oct 2015 22:46:50 +0000 (15:46 -0700)]
Add erfc function to FunctionParser.

8 years agoMerge pull request #1707 from gassmoeller/triangulation_weights_callback
Timo Heister [Mon, 2 Nov 2015 17:50:41 +0000 (12:50 -0500)]
Merge pull request #1707 from gassmoeller/triangulation_weights_callback

Triangulation weights callback

8 years agoUse VectorType, not InVector, as a template type. 1777/head
David Wells [Mon, 19 Oct 2015 02:48:08 +0000 (22:48 -0400)]
Use VectorType, not InVector, as a template type.

Some functions use a template for the input and output types, but other
functions use InVector. This commit switches the second kind to use VectorType,
which is more consistent with the usage of both VectorType and InVector.

8 years agoUse FEValuesType, not FEValues, in templates.
David Wells [Mon, 19 Oct 2015 12:38:06 +0000 (08:38 -0400)]
Use FEValuesType, not FEValues, in templates.

This makes things less ambiguous because 'FEValues' is a valid class
name.

8 years agoPrefer QuadratureType to Quadrature in templates.
David Wells [Mon, 19 Oct 2015 02:36:43 +0000 (22:36 -0400)]
Prefer QuadratureType to Quadrature in templates.

Writing something like

template<typename Quadrature>

is ambiguous because Quadrature is also a class.

8 years agoUse typename BlockMatrixType in templates.
David Wells [Mon, 19 Oct 2015 02:27:59 +0000 (22:27 -0400)]
Use typename BlockMatrixType in templates.

This replaces the use of 'BlockMatrix', which, while it is not a class
name, looks a lot like one.

8 years agoUse SP, not SparsityPattern, as a template type.
David Wells [Mon, 19 Oct 2015 01:49:36 +0000 (21:49 -0400)]
Use SP, not SparsityPattern, as a template type.

This commit also exchanges 'SparsityPatternBase' for 'SP' for
consistency.

8 years agoRemove unnecessary dealii:: qualifiers.
David Wells [Thu, 22 Oct 2015 23:26:09 +0000 (19:26 -0400)]
Remove unnecessary dealii:: qualifiers.

8 years agoUse VectorType, not Vector, as a template type.
David Wells [Mon, 19 Oct 2015 01:29:18 +0000 (21:29 -0400)]
Use VectorType, not Vector, as a template type.

Writing something like

template<typename Vector>

is ambiguous because Vector is also a class.

8 years agoMerge pull request #1826 from drwells/remove-repeated-typedef
Denis Davydov [Mon, 2 Nov 2015 14:51:04 +0000 (15:51 +0100)]
Merge pull request #1826 from drwells/remove-repeated-typedef

Remove repeated typedef.

8 years agoRemove a repeated typedef. 1826/head
David Wells [Mon, 19 Oct 2015 02:48:53 +0000 (22:48 -0400)]
Remove a repeated typedef.

8 years agoUse a local typedef to shorten things.
David Wells [Mon, 19 Oct 2015 12:33:12 +0000 (08:33 -0400)]
Use a local typedef to shorten things.

8 years agoadd entry in changes.h 1707/head
Rene Gassmoeller [Mon, 2 Nov 2015 14:24:15 +0000 (08:24 -0600)]
add entry in changes.h

8 years agoIntroduce a signal for weighted repartitioning
Rene Gassmoeller [Wed, 30 Sep 2015 14:12:16 +0000 (16:12 +0200)]
Introduce a signal for weighted repartitioning

8 years agoMerge pull request #1817 from luca-heltai/fix-mac-pack
Matthias Maier [Mon, 2 Nov 2015 14:08:15 +0000 (08:08 -0600)]
Merge pull request #1817 from luca-heltai/fix-mac-pack

Reworded packaging options. We no longer need to export DYLD libraries.

8 years agoRemove the first_cell/is_first_cell/current_update_flags() mechanism from FiniteElement. 1823/head
Wolfgang Bangerth [Sun, 1 Nov 2015 15:56:00 +0000 (09:56 -0600)]
Remove the first_cell/is_first_cell/current_update_flags() mechanism from FiniteElement.

What this patch does is address the issue that
`FiniteElement::get_data()` (as was previously the case with Mappings
as well) sets a flag in the returned `InternalData` object that
indicates that we are visiting a cell for the first time. `FEValues`
later sets this flag back once we're done with the first cell.

The way this is used is that the various finite element
implementations query `current_update_flags` in
`FE::fill_fe_*_values()`, which equals `update_each|update_once` if we
are on the first cell, and `update_each` on later cells. That means
that on the first cell we update fields in the output object that we
need to initialize only once, e.g., the values of shape functions on
quadrature points, but we won't do this again on later cells.

I don't particularly like this approach. It is opaque and took even me
a long time to reconstruct when I saw it (see #1305 for a
discussion). My preference is to do things we only need to do once in
`FE::get_data()`, and do things we need to do every time in
`FE::fill_fe_*_values()`. This kind of change was already implemented
for mappings in the patches references from #1305, and the current
pull request goes into the same direction.

The current PR only does one half of the necessary changes in order to
keep the patch reasonably self contained and readable. In particular,
what it really does is simply *always* do everything, by storing
`update_once|update_each` in `FE::InternalDataBase::update_each`. In
other words, it is the equivalent to telling finite element
implementations that we are always on the first cell. This requires
more work than in the previous state, although copying shape values
every time is arguably of small expense compared to transforming
gradients.

I will of course fix this up in upcoming patches. It requires
significant shuffling around. I'll open an issue in a minute
discussing the details. I'd simply like to propose this here already
as an incremental step forward, without making patch review too
onerous. The patch is mostly mechanical once you understand what it
does.

8 years agoMerge pull request #1819 from davydden/shared_tria_artificial_bugfix
Denis Davydov [Mon, 2 Nov 2015 08:17:36 +0000 (09:17 +0100)]
Merge pull request #1819 from davydden/shared_tria_artificial_bugfix

Shared tria artificial bugfix

8 years agoadded extra Asserts to shared::Tria tests; do not output partition results 1819/head
Denis Davydov [Fri, 30 Oct 2015 22:48:53 +0000 (23:48 +0100)]
added extra Asserts to shared::Tria tests; do not output partition results

8 years agoMerge pull request #1822 from bangerth/doc-update-38
Denis Davydov [Mon, 2 Nov 2015 07:39:44 +0000 (08:39 +0100)]
Merge pull request #1822 from bangerth/doc-update-38

Augment documentation of FE::get_(sub)face_data.

8 years agoAugment documentation of FE::get_(sub)face_data. 1822/head
Wolfgang Bangerth [Mon, 2 Nov 2015 07:35:39 +0000 (01:35 -0600)]
Augment documentation of FE::get_(sub)face_data.

8 years agoMerge pull request #1820 from tamiko/fix_includes
Wolfgang Bangerth [Mon, 2 Nov 2015 00:13:58 +0000 (09:13 +0900)]
Merge pull request #1820 from tamiko/fix_includes

CMake: Bugfix: Also use DEAL_II_INCLUDE_DIRS for bundled boost

8 years agoAdd a news/changes.h entry 1820/head
Matthias Maier [Sun, 1 Nov 2015 23:21:30 +0000 (17:21 -0600)]
Add a news/changes.h entry

8 years agoCMake: Bugfix: Also use DEAL_II_INCLUDE_DIRS for bundled boost
Matthias Maier [Sun, 1 Nov 2015 23:17:24 +0000 (17:17 -0600)]
CMake: Bugfix: Also use DEAL_II_INCLUDE_DIRS for bundled boost

This fixes a compilation issue with a zlib library not in a default
location in combination with bundled boost. Many thanks to Lukas Korous for
pointing this out.

8 years agoMerge pull request #1808 from bangerth/further-FEValues-cleanups-37
Luca Heltai [Sat, 31 Oct 2015 16:44:27 +0000 (17:44 +0100)]
Merge pull request #1808 from bangerth/further-FEValues-cleanups-37

Break dependency of MappingQ from MappingQGeneric.

8 years agofixed a bug in shared::Tria where...
Denis Davydov [Fri, 30 Oct 2015 22:47:29 +0000 (23:47 +0100)]
fixed a bug in shared::Tria where...

true_subdomain_ids_of_cells was not set up for the case
without artificial cells.

8 years agoFixed wording in dealii.conf.in 1817/head
Luca Heltai [Fri, 30 Oct 2015 11:07:57 +0000 (12:07 +0100)]
Fixed wording in dealii.conf.in

8 years agoMerge pull request #1816 from tjhei/fix_expand_inst_ldgold
Wolfgang Bangerth [Fri, 30 Oct 2015 16:30:12 +0000 (11:30 -0500)]
Merge pull request #1816 from tjhei/fix_expand_inst_ldgold

fix cmake expand_instantiations ldgold workaround

8 years agofix cmake expand_instantiations ldgold workaround 1816/head
Timo Heister [Fri, 30 Oct 2015 16:12:26 +0000 (12:12 -0400)]
fix cmake expand_instantiations ldgold workaround

Avoid cmake error if DEAL_II_LINKER_FLAGS is empty.

8 years agoMerge pull request #1813 from bangerth/fix-1798
Matthias Maier [Fri, 30 Oct 2015 02:06:26 +0000 (21:06 -0500)]
Merge pull request #1813 from bangerth/fix-1798

Workaround linker bug when compiling expand_instantiations.

8 years agoWorkaround linker bug when compiling expand_instantiations. 1813/head
Wolfgang Bangerth [Fri, 30 Oct 2015 01:25:29 +0000 (20:25 -0500)]
Workaround linker bug when compiling expand_instantiations.

Do so by removing -fuse-ld=gold from the list of linker flags for this program.

8 years agoMerge pull request #1809 from tamiko/tensor_fixes
Matthias Maier [Thu, 29 Oct 2015 22:26:32 +0000 (17:26 -0500)]
Merge pull request #1809 from tamiko/tensor_fixes

Fix several Tensor<rank,dim,Number> related bugs:

8 years agoreindent 1809/head
Matthias Maier [Thu, 29 Oct 2015 22:24:16 +0000 (17:24 -0500)]
reindent

8 years agoMerge pull request #1812 from tjhei/fix_bdm_tests2
Wolfgang Bangerth [Thu, 29 Oct 2015 15:45:09 +0000 (10:45 -0500)]
Merge pull request #1812 from tjhei/fix_bdm_tests2

add missing changes to fix bdm tests

8 years agoadd missing changes to fix bdm tests 1812/head
Timo Heister [Thu, 29 Oct 2015 14:43:40 +0000 (10:43 -0400)]
add missing changes to fix bdm tests

the PR #1778 written to fix tests/fe/bdm_[89], also see #1770, did not
include the necessary changes to the .cc files.

8 years agoMerge pull request #1810 from davydden/sharedtria_fix_no_mpi
Denis Davydov [Thu, 29 Oct 2015 10:05:40 +0000 (11:05 +0100)]
Merge pull request #1810 from davydden/sharedtria_fix_no_mpi

fix shared::Tria for no-MPI case

8 years agofix shared::Tria for no-MPI case 1810/head
Denis Davydov [Thu, 29 Oct 2015 08:49:10 +0000 (09:49 +0100)]
fix shared::Tria for no-MPI case


In the beginning the Universe was created. This has made a lot of people very angry and has been widely regarded as a bad move.

Douglas Adams


Typeset in Trocchi and Trocchi Bold Sans Serif.