]> https://gitweb.dealii.org/ - dealii.git/log
dealii.git
7 years agoFix compilation of documentation. 3695/head
David Wells [Thu, 22 Dec 2016 18:15:47 +0000 (13:15 -0500)]
Fix compilation of documentation.

Since c60382365dc removed changes.h, we should also update our CMake scripts to
not look for build rules that use this (nonexistant) file.

7 years agoMerge pull request #3684 from bangerth/doc-update-80
Bruno Turcksin [Wed, 21 Dec 2016 21:43:12 +0000 (16:43 -0500)]
Merge pull request #3684 from bangerth/doc-update-80

Update CellData and SubCellData documentation.

7 years agoMerge pull request #3690 from bangerth/remove-changes.h
Bruno Turcksin [Wed, 21 Dec 2016 21:36:31 +0000 (16:36 -0500)]
Merge pull request #3690 from bangerth/remove-changes.h

Remove the changes.h file.

7 years agoRemove the changes.h file. 3690/head
Wolfgang Bangerth [Wed, 21 Dec 2016 18:24:29 +0000 (11:24 -0700)]
Remove the changes.h file.

This file is now automatically generated. Let's not leave a stale file sitting around.

7 years agoMerge pull request #3688 from drwells/clean-up-constraintmatrix-headers
David Wells [Wed, 21 Dec 2016 17:23:50 +0000 (12:23 -0500)]
Merge pull request #3688 from drwells/clean-up-constraintmatrix-headers

Move some headers into the .templates.h file.

7 years agoMove some headers into the .templates.h file. 3688/head
David Wells [Wed, 21 Dec 2016 16:33:35 +0000 (11:33 -0500)]
Move some headers into the .templates.h file.

7 years agoMerge pull request #3685 from davydden/doc/astyle_version
Denis Davydov [Tue, 20 Dec 2016 21:00:17 +0000 (22:00 +0100)]
Merge pull request #3685 from davydden/doc/astyle_version

explicitly say we use astyle 2.04

7 years agoMerge pull request #3687 from bangerth/make-IndexSet-threadsafe
Daniel Arndt [Tue, 20 Dec 2016 16:51:57 +0000 (17:51 +0100)]
Merge pull request #3687 from bangerth/make-IndexSet-threadsafe

Add a mutex to IndexSet and use it to make IndexSet::do_compress() thread-safe.

7 years agoexplicitly say we use astyle 2.04 3685/head
Denis Davydov [Tue, 20 Dec 2016 12:32:36 +0000 (13:32 +0100)]
explicitly say we use astyle 2.04

7 years agoAdd a mutex to IndexSet and use it to make IndexSet::do_compress() thread-safe. 3687/head
Wolfgang Bangerth [Tue, 20 Dec 2016 16:00:01 +0000 (09:00 -0700)]
Add a mutex to IndexSet and use it to make IndexSet::do_compress() thread-safe.

IndexSet has a number of 'mutable' member variables. The only function that modifies them is
'do_compress', which is called by 'compress', which is in turn called by all of the
'const' member functions. To make all of these 'const' member functions thread-safe, we
need to use a mutex in 'do_compress'. This patch does this.

I have verified that no other 'const' function actually modifies any of the 'mutable'
member variables, so only guarding 'do_compress' by the mutex is sufficient.

7 years agoUpdate CellData and SubCellData documentation. 3684/head
Wolfgang Bangerth [Tue, 20 Dec 2016 07:21:28 +0000 (00:21 -0700)]
Update CellData and SubCellData documentation.

7 years agoMerge pull request #3683 from bangerth/add-op-eq
Denis Davydov [Tue, 20 Dec 2016 05:55:12 +0000 (06:55 +0100)]
Merge pull request #3683 from bangerth/add-op-eq

Add a copy operator to Mutex so that it can be used in class with an operator=.

7 years agoMerge pull request #3682 from bangerth/fix-IndexSet-nth_index_within_set
Timo Heister [Tue, 20 Dec 2016 01:18:36 +0000 (17:18 -0800)]
Merge pull request #3682 from bangerth/fix-IndexSet-nth_index_within_set

Let IndexSet::nth_index_within_set() call compress().

7 years agoAdd a copy operator to Mutex so that it can be used in class with an operator=. 3683/head
Wolfgang Bangerth [Mon, 19 Dec 2016 23:07:01 +0000 (16:07 -0700)]
Add a copy operator to Mutex so that it can be used in class with an operator=.

7 years agoAdd test. 3682/head
Wolfgang Bangerth [Mon, 19 Dec 2016 23:05:43 +0000 (16:05 -0700)]
Add test.

7 years agoLet IndexSet::nth_index_within_set() call compress().
Wolfgang Bangerth [Mon, 19 Dec 2016 22:50:20 +0000 (15:50 -0700)]
Let IndexSet::nth_index_within_set() call compress().

This follows the pattern of all other const functions of the class.

7 years agoMerge pull request #3676 from bangerth/fix-DataOutFaces
Bruno Turcksin [Mon, 19 Dec 2016 13:30:40 +0000 (08:30 -0500)]
Merge pull request #3676 from bangerth/fix-DataOutFaces

Also check for the invalid FaceDescriptor that first_face() may return.

7 years agoMerge pull request #3678 from jppelteret/fix_symm_tensor_constructor_01
Jean-Paul Pelteret [Sat, 17 Dec 2016 11:22:38 +0000 (12:22 +0100)]
Merge pull request #3678 from jppelteret/fix_symm_tensor_constructor_01

Fix invalid loop range in SymmetricTensor class constructor; add test

7 years agoFix invalid loop range in SymmetricTensor class constructor; add test 3678/head
Jean-Paul Pelteret [Fri, 16 Dec 2016 08:26:44 +0000 (09:26 +0100)]
Fix invalid loop range in SymmetricTensor class constructor; add test

Fixes #3674

7 years agoAdd changelog. 3676/head
Wolfgang Bangerth [Fri, 16 Dec 2016 05:38:23 +0000 (22:38 -0700)]
Add changelog.

7 years agoAdd test.
Justin Kauffman [Fri, 16 Dec 2016 05:38:13 +0000 (22:38 -0700)]
Add test.

7 years agoAlso check for the invalid FaceDescriptor that first_face() may return.
Wolfgang Bangerth [Fri, 16 Dec 2016 05:37:32 +0000 (22:37 -0700)]
Also check for the invalid FaceDescriptor that first_face() may return.

This happens if a process has no faces to deal with. The bug was reported by Justin Kauffman <jak5378@gmail.com>.

7 years agoMerge pull request #3679 from masterleinad/split_changes_h
Matthias Maier [Fri, 16 Dec 2016 17:44:50 +0000 (11:44 -0600)]
Merge pull request #3679 from masterleinad/split_changes_h

Split changes.h

7 years agoSplit changes.h 3679/head
Daniel Arndt [Fri, 16 Dec 2016 12:14:31 +0000 (13:14 +0100)]
Split changes.h

7 years agoFix split_summary to allow hyphens in authors' names
Daniel Arndt [Fri, 16 Dec 2016 13:15:33 +0000 (14:15 +0100)]
Fix split_summary to allow hyphens in authors' names

7 years agoMerge pull request #3617 from masterleinad/recreate_changes_h
Daniel Arndt [Fri, 16 Dec 2016 11:23:03 +0000 (12:23 +0100)]
Merge pull request #3617 from masterleinad/recreate_changes_h

Recreate changes.h automatically

7 years agoReplace bash script by CMake script and adjust files to be installed 3617/head
Daniel Arndt [Tue, 13 Dec 2016 16:23:59 +0000 (17:23 +0100)]
Replace bash script by CMake script and adjust files to be installed

7 years agoMerge pull request #3677 from bangerth/fix-quicktest
Denis Davydov [Fri, 16 Dec 2016 07:56:54 +0000 (08:56 +0100)]
Merge pull request #3677 from bangerth/fix-quicktest

Fix a quicktest.

7 years agoFix a quicktest. 3677/head
Wolfgang Bangerth [Fri, 16 Dec 2016 05:43:56 +0000 (22:43 -0700)]
Fix a quicktest.

The test assigned MPI return codes to a variable, but then did not do anything with this
variable. We may as well test for errors.

7 years agoMerge pull request #3675 from davydden/diag_matrix_comp_warning
Wolfgang Bangerth [Thu, 15 Dec 2016 16:58:39 +0000 (09:58 -0700)]
Merge pull request #3675 from davydden/diag_matrix_comp_warning

fix compiler warning in DiagonalMatrix

7 years agofix compiler warning in DiagonalMatrix 3675/head
Denis Davydov [Thu, 15 Dec 2016 16:20:13 +0000 (17:20 +0100)]
fix compiler warning in DiagonalMatrix

7 years agoMerge pull request #7 from bangerth/masterleinad-recreate_changes_h
Daniel Arndt [Mon, 12 Dec 2016 09:01:56 +0000 (10:01 +0100)]
Merge pull request #7 from bangerth/masterleinad-recreate_changes_h

Make sure changes.h is created as input for doxygen

7 years agoMake sure the 'changes.h' file is automatically generated.
Wolfgang Bangerth [Sat, 10 Dec 2016 00:20:45 +0000 (17:20 -0700)]
Make sure the 'changes.h' file is automatically generated.

In particular, list it as an input to the doxygen run.

7 years agoOutput results to the console instead.
Wolfgang Bangerth [Fri, 9 Dec 2016 23:45:13 +0000 (16:45 -0700)]
Output results to the console instead.

7 years agoMerge pull request #3672 from drwells/remove-extra-semicolons
Denis Davydov [Sun, 11 Dec 2016 19:31:09 +0000 (20:31 +0100)]
Merge pull request #3672 from drwells/remove-extra-semicolons

Remove some extra semicolons.

7 years agoRemove some extra semicolons. 3672/head
David Wells [Sun, 11 Dec 2016 18:37:46 +0000 (13:37 -0500)]
Remove some extra semicolons.

7 years agoMerge pull request #3671 from drwells/fix-doc-typos-6
Wolfgang Bangerth [Sun, 11 Dec 2016 00:57:08 +0000 (17:57 -0700)]
Merge pull request #3671 from drwells/fix-doc-typos-6

Fix some typos in the documentation headers.

7 years agoFix some typos in the documentation headers. 3671/head
David Wells [Sun, 11 Dec 2016 00:53:52 +0000 (19:53 -0500)]
Fix some typos in the documentation headers.

7 years agoMerge pull request #3664 from bangerth/rename-functions
David Wells [Sat, 10 Dec 2016 15:55:12 +0000 (10:55 -0500)]
Merge pull request #3664 from bangerth/rename-functions

Rename DataPostprocessor functions.

7 years agoMerge pull request #3670 from jppelteret/doc_fix_03
Wolfgang Bangerth [Fri, 9 Dec 2016 21:39:04 +0000 (14:39 -0700)]
Merge pull request #3670 from jppelteret/doc_fix_03

Fix some minor documentation errors from #3034

7 years agoFix some minor documentation errors from #3034 3670/head
Jean-Paul Pelteret [Fri, 9 Dec 2016 20:25:48 +0000 (21:25 +0100)]
Fix some minor documentation errors from #3034

Predictably some minor errors were only spotted after a merge! The
biggest point is ensuring that definitions of static members of the
Elasticity::StandardTensors class don't get inlined in their description
as "Initial values".

7 years agoMerge pull request #3034 from jppelteret/standard_tensors
Wolfgang Bangerth [Fri, 9 Dec 2016 19:32:01 +0000 (12:32 -0700)]
Merge pull request #3034 from jppelteret/standard_tensors

Implemented standard tensors for elasticity.

7 years agoStep-44 now uses the newly implemented physics classes and functions. 3034/head
Jean-Paul Pelteret [Mon, 5 Dec 2016 09:26:31 +0000 (10:26 +0100)]
Step-44 now uses the newly implemented physics classes and functions.

7 years agoStandard transformations, and kinematics and tensors for elasticity.
Jean-Paul Pelteret [Mon, 5 Dec 2016 07:36:09 +0000 (08:36 +0100)]
Standard transformations, and kinematics and tensors for elasticity.

This commit adds a set of functions to perform various push forward /
pull back operations that regularly feature in finite deformation
mechanics. It also adds some standard tensor definitions for elasticity,
as well as functions that compute kinematic quantities typically used in
finite strain elasticity.

Refers to discussion in dealii/code-gallery/#7.

7 years agoAdd adjugate and cofactor functions to tensor class
Jean-Paul Pelteret [Sun, 25 Sep 2016 20:24:54 +0000 (22:24 +0200)]
Add adjugate and cofactor functions to tensor class

7 years agoMerge pull request #3667 from bangerth/default-constructor-for-ArrayView
Daniel Arndt [Thu, 8 Dec 2016 23:16:27 +0000 (00:16 +0100)]
Merge pull request #3667 from bangerth/default-constructor-for-ArrayView

Add a default constructor for class ArrayView.

7 years agoAdd changelog entry. 3667/head
Wolfgang Bangerth [Thu, 8 Dec 2016 17:30:39 +0000 (10:30 -0700)]
Add changelog entry.

7 years agoAdd testcase.
Wolfgang Bangerth [Thu, 8 Dec 2016 17:30:39 +0000 (10:30 -0700)]
Add testcase.

7 years agoAdd a default constructor for class ArrayView.
Wolfgang Bangerth [Thu, 8 Dec 2016 17:30:04 +0000 (10:30 -0700)]
Add a default constructor for class ArrayView.

Right now, one can only create such an object pointing to a specific
array. It is not possible to default construct it, and consequently
you can't put them into collections, for example.

7 years agoMerge pull request #3666 from bangerth/fix-ArrayView
Jean-Paul Pelteret [Thu, 8 Dec 2016 15:12:43 +0000 (16:12 +0100)]
Merge pull request #3666 from bangerth/fix-ArrayView

Allow copying of ArrayViews of length zero.

7 years agoAdd changelog entry. 3666/head
Wolfgang Bangerth [Thu, 8 Dec 2016 04:29:54 +0000 (21:29 -0700)]
Add changelog entry.

7 years agoAdd test.
Wolfgang Bangerth [Thu, 8 Dec 2016 04:29:53 +0000 (21:29 -0700)]
Add test.

7 years agoAllow copying of ArrayViews of length zero.
Wolfgang Bangerth [Thu, 8 Dec 2016 04:29:17 +0000 (21:29 -0700)]
Allow copying of ArrayViews of length zero.

7 years agoMerge pull request #3665 from gassmoeller/add_pop_front_index_set
Wolfgang Bangerth [Thu, 8 Dec 2016 03:35:40 +0000 (20:35 -0700)]
Merge pull request #3665 from gassmoeller/add_pop_front_index_set

Add convenience functions to IndexSet

7 years agoAddress comments 3665/head
Rene Gassmoeller [Thu, 8 Dec 2016 02:23:26 +0000 (19:23 -0700)]
Address comments

7 years agoAdd IndexSet::pop_front, pop_back, and is_empty
Rene Gassmoeller [Fri, 4 Nov 2016 00:13:08 +0000 (18:13 -0600)]
Add IndexSet::pop_front, pop_back, and is_empty

7 years agoMerge pull request #3663 from drwells/distribute-dofs-linear-time
David Wells [Wed, 7 Dec 2016 20:17:36 +0000 (15:17 -0500)]
Merge pull request #3663 from drwells/distribute-dofs-linear-time

Distribute dofs linear time

7 years agoTest that distributing hp dofs runs in linear time. 3663/head
David Wells [Wed, 7 Dec 2016 15:09:57 +0000 (10:09 -0500)]
Test that distributing hp dofs runs in linear time.

7 years agoCache end_active() when assigning hp dofs.
David Wells [Wed, 7 Dec 2016 15:01:57 +0000 (10:01 -0500)]
Cache end_active() when assigning hp dofs.

This fixes a performance bug where distributing dofs on some grids would take an
amount of time quadratic in the number of dofs when it should be linear.

7 years agoRename DataPostprocessor functions. 3664/head
Wolfgang Bangerth [Wed, 7 Dec 2016 17:14:19 +0000 (10:14 -0700)]
Rename DataPostprocessor functions.

In #3333, I added virtual functions to DataPostprocessor with the same name as the existing
functions. This leads to issues where we get a warning in every derived class that only
overloads one of these functions, because that hides the other function. This is, well,
suboptimal.

This patch is therefore a redo of my earlier attempt in which I continue to deprecate
the old functions, but the new functions have a different name. I think they also
have a better name (for a discussion of the naming, see
https://github.com/geodynamics/aspect/issues/1284 ). The different names avoid the
problem of getting the warning and should therefore lead to less discontent. They
also avoid the need to try to work around the warnings using 'using' declarations, like
in #3528.

7 years agoMerge pull request #3662 from jppelteret/tensor-silence_warning
Wolfgang Bangerth [Wed, 7 Dec 2016 16:06:49 +0000 (09:06 -0700)]
Merge pull request #3662 from jppelteret/tensor-silence_warning

Tensor class: Silence compiler warning

7 years agoImplemented the inverse of rank-2 SymmetricTensors. 3662/head
Jean-Paul Pelteret [Wed, 7 Dec 2016 10:27:00 +0000 (11:27 +0100)]
Implemented the inverse of rank-2 SymmetricTensors.

Includes a test case, gratis.

7 years agoRefactor definition of rank-2 Tensor inverse function; add test case
Jean-Paul Pelteret [Wed, 7 Dec 2016 09:05:24 +0000 (10:05 +0100)]
Refactor definition of rank-2 Tensor inverse function; add test case

This silences warnings produced by `-Warray-bounds`.

7 years agoMerge pull request #3658 from bangerth/move-functions
Wolfgang Bangerth [Tue, 6 Dec 2016 21:20:09 +0000 (14:20 -0700)]
Merge pull request #3658 from bangerth/move-functions

Move DataOutInterface::write_pvd/visit_record() to namespace DataOutBase.

7 years agoAlso extract the functionality of the write_pvtu_record(). 3658/head
Wolfgang Bangerth [Tue, 6 Dec 2016 02:19:41 +0000 (19:19 -0700)]
Also extract the functionality of the write_pvtu_record().

Move it to the DataOutBase namespace, and let the original function call the one
into which it has been refactored.

7 years agoMerge pull request #3660 from jppelteret/step-44_bugfix_01
Wolfgang Bangerth [Mon, 5 Dec 2016 16:49:14 +0000 (09:49 -0700)]
Merge pull request #3660 from jppelteret/step-44_bugfix_01

Fixed condition in assert checking for convergence in nonlinear solver

7 years agoFixed condition in assert checking for convergence in nonlinear solver 3660/head
Jean-Paul Pelteret [Mon, 5 Dec 2016 16:42:30 +0000 (17:42 +0100)]
Fixed condition in assert checking for convergence in nonlinear solver

Fixes #3604

7 years agoMerge pull request #3657 from drwells/vectorspacevector-virtual-dtor
Wolfgang Bangerth [Sat, 3 Dec 2016 23:28:57 +0000 (16:28 -0700)]
Merge pull request #3657 from drwells/vectorspacevector-virtual-dtor

Make VectorSpaceVector's destructor virtual.

7 years agoAlso adjust testsuite and tutorial programs.
Wolfgang Bangerth [Sat, 3 Dec 2016 23:18:47 +0000 (16:18 -0700)]
Also adjust testsuite and tutorial programs.

7 years agoMove DataOutInterface::write_pvd/visit_record() to namespace DataOutBase.
Wolfgang Bangerth [Sat, 3 Dec 2016 23:03:08 +0000 (16:03 -0700)]
Move DataOutInterface::write_pvd/visit_record() to namespace DataOutBase.

These functions did not depend on the state of the DataOutInterface object and could therefore
be made 'static'. On the other hand, we have traditionally kept such functions in namespace
DataOutBase. Move them there, and deprecate the old versions of these functions.

7 years agoMake VectorSpaceVector's destructor virtual. 3657/head
David Wells [Sat, 3 Dec 2016 22:24:28 +0000 (17:24 -0500)]
Make VectorSpaceVector's destructor virtual.

Since the classes inheriting from VectorSpaceVector all implement their own
memory management strategies, the destructor of the base class should always be
virtual.

7 years agoMerge pull request #3655 from bangerth/limit-TimerOutput-data
Martin Kronbichler [Sat, 3 Dec 2016 20:34:41 +0000 (21:34 +0100)]
Merge pull request #3655 from bangerth/limit-TimerOutput-data

Only show TimerOutput percentages down to 0.1%.

7 years agoOnly show TimerOutput percentages down to 0.1%. 3655/head
Wolfgang Bangerth [Sat, 3 Dec 2016 13:57:03 +0000 (06:57 -0700)]
Only show TimerOutput percentages down to 0.1%.

Currently, if a TimerOutput section takes a very small fraction of the overall
time, we happily output in the last column that it took 6.5e-4%. This is not
only visually awkward compared to all of the other numbers that are not using
the e-04 notation, but also pretty pointless: nobody cares whether something took
that small a fraction of the overall time, and if they happen to do anyway: the
absolute amount of time is printed in the previous column.

Consequently, show everything that takes less than 0.1% of time as 0%.

7 years agoMerge pull request #3654 from gassmoeller/optimize_spherical_manifold_new_point
Wolfgang Bangerth [Sat, 3 Dec 2016 02:49:58 +0000 (19:49 -0700)]
Merge pull request #3654 from gassmoeller/optimize_spherical_manifold_new_point

Optimize SphericalManifold::get_new_point()

7 years agoOptimize SphericalManifold::get_new_point() 3654/head
Rene Gassmoeller [Sat, 3 Dec 2016 00:44:29 +0000 (17:44 -0700)]
Optimize SphericalManifold::get_new_point()

7 years agoMerge pull request #3651 from tcclevenger/fix_MemoryStats_doc
Wolfgang Bangerth [Fri, 2 Dec 2016 22:01:12 +0000 (15:01 -0700)]
Merge pull request #3651 from tcclevenger/fix_MemoryStats_doc

Fixed documentation for Utilities::System::MemoryStats Struct Reference

7 years agoMerge pull request #3329 from kronbichler/extend_step-37
Denis Davydov [Fri, 2 Dec 2016 13:29:02 +0000 (14:29 +0100)]
Merge pull request #3329 from kronbichler/extend_step-37

Extend step 37 to parallel vectors

7 years agoFixed documentation for Utilities::System::MemoryStats Struct Reference 3651/head
tcclevenger [Thu, 1 Dec 2016 04:12:14 +0000 (23:12 -0500)]
Fixed documentation for Utilities::System::MemoryStats Struct Reference

Comments in deal.II/base/utilities.h were behind variable declarations
leading to wrong placement in online documentation

7 years agoMerge pull request #3653 from kronbichler/master
Wolfgang Bangerth [Thu, 1 Dec 2016 22:28:14 +0000 (15:28 -0700)]
Merge pull request #3653 from kronbichler/master

Fix DEAL_II_COMPILER_USE_VECTOR_ARITHMETICS with AVX512 and ICC

7 years agoFix DEAL_II_COMPILER_USE_VECTOR_ARITHMETICS with AVX512 and Intel. 3653/head
Martin Kronbichler [Thu, 1 Dec 2016 17:49:02 +0000 (18:49 +0100)]
Fix DEAL_II_COMPILER_USE_VECTOR_ARITHMETICS with AVX512 and Intel.

7 years agoMerge pull request #3650 from drwells/add-zlib-enum-docs
Wolfgang Bangerth [Wed, 30 Nov 2016 22:45:40 +0000 (15:45 -0700)]
Merge pull request #3650 from drwells/add-zlib-enum-docs

Add documentation to the zlib compression levels.

7 years agoAdd documentation to the zlib compression levels. 3650/head
David Wells [Wed, 30 Nov 2016 19:05:13 +0000 (14:05 -0500)]
Add documentation to the zlib compression levels.

Enumerations are not listed by doxygen unless they have documentation
strings, so add them.

7 years agoMerge pull request #3649 from jppelteret/collect_periodic_faces_update_01
David Wells [Wed, 30 Nov 2016 19:09:54 +0000 (14:09 -0500)]
Merge pull request #3649 from jppelteret/collect_periodic_faces_update_01

Added warning for unsuccessful case in GridTools::collect_periodic_faces

7 years agoAdded warning for unsuccessful case in GridTools::collect_periodic_faces 3649/head
Jean-Paul Pelteret [Wed, 30 Nov 2016 16:10:14 +0000 (17:10 +0100)]
Added warning for unsuccessful case in GridTools::collect_periodic_faces

A note and assert has been added to provide a warning for the case where
collect_periodic_faces doesn't find any matched faces. This may happen
if there is no colouring on the coarsest level mesh (i.e. grid
refinement before boundary id definition) or where they've accidentally
called the function with the wrong boundary id's (or direction?).

Fixes #3280

7 years agoMerge pull request #3645 from bangerth/update-laplace_transform
Bruno Turcksin [Tue, 29 Nov 2016 20:56:51 +0000 (15:56 -0500)]
Merge pull request #3645 from bangerth/update-laplace_transform

Update GridTools::laplace_transform().

7 years agoMerge pull request #3644 from bangerth/update-step-49
Bruno Turcksin [Tue, 29 Nov 2016 20:53:45 +0000 (15:53 -0500)]
Merge pull request #3644 from bangerth/update-step-49

Be consistent about how we name the Triangulation object.

7 years agoMerge pull request #3648 from bangerth/remove-stray-typenames
Bruno Turcksin [Tue, 29 Nov 2016 20:53:02 +0000 (15:53 -0500)]
Merge pull request #3648 from bangerth/remove-stray-typenames

Remove stray 'typename's.

7 years agoRemove stray 'typename's. 3648/head
Wolfgang Bangerth [Tue, 29 Nov 2016 18:17:30 +0000 (11:17 -0700)]
Remove stray 'typename's.

7 years agoMerge pull request #3647 from tjhei/doc_fixes
Wolfgang Bangerth [Tue, 29 Nov 2016 18:01:27 +0000 (11:01 -0700)]
Merge pull request #3647 from tjhei/doc_fixes

doxygen fixes

7 years agodoxygen fixes 3647/head
Timo Heister [Tue, 29 Nov 2016 17:55:55 +0000 (12:55 -0500)]
doxygen fixes

7 years agoUpdate GridTools::laplace_transform(). 3645/head
Wolfgang Bangerth [Tue, 29 Nov 2016 00:01:02 +0000 (17:01 -0700)]
Update GridTools::laplace_transform().

Incomprehensively, the function looped over all faces of a cell, and then
all vertices of the face. This visited each vertex exactly 'dim' times.
This can be done easier.

While there, also fix a number of issues where we confuse 'unsigned int'
with 'types::global_dof_index'. This doesn't matter here because the
function only works in sequential settings anyway, but we should be
consistent.

Finally, instead of setting vertex locations from each adjacent cell, only
do it once by keeping track when we touch the vertex the first time.

7 years agoBe consistent about how we name the Triangulation object. 3644/head
Wolfgang Bangerth [Mon, 28 Nov 2016 22:16:41 +0000 (15:16 -0700)]
Be consistent about how we name the Triangulation object.

Most of the functions in step-49 call the triangulation 'triangulation', but some
called it 'tria'. Be consistent about it.

While there, also be consistent about spaces between function name and opening
parenthesis.

7 years agoMerge pull request #3643 from tjhei/doc_nonzero_flux
Wolfgang Bangerth [Mon, 28 Nov 2016 16:55:54 +0000 (09:55 -0700)]
Merge pull request #3643 from tjhei/doc_nonzero_flux

clarify doc for compute_nonzero_normal_flux_constraints

7 years agoclarify doc for compute_nonzero_normal_flux_constraints 3643/head
Timo Heister [Mon, 28 Nov 2016 16:38:42 +0000 (11:38 -0500)]
clarify doc for compute_nonzero_normal_flux_constraints

7 years agoMerge pull request #3641 from bangerth/revise-3639
David Wells [Mon, 28 Nov 2016 16:05:45 +0000 (11:05 -0500)]
Merge pull request #3641 from bangerth/revise-3639

Find a better way to be dimension independent.

7 years agoMerge pull request #3622 from kalj/static-asserts
Wolfgang Bangerth [Mon, 28 Nov 2016 13:47:57 +0000 (06:47 -0700)]
Merge pull request #3622 from kalj/static-asserts

Add more static asserts for sensible template parameters

7 years agoAdd more static asserts for sensible template parameters 3622/head
Karl Ljungkvist [Tue, 22 Nov 2016 15:06:14 +0000 (16:06 +0100)]
Add more static asserts for sensible template parameters

7 years agoAdd funding information 3329/head
Martin Kronbichler [Thu, 24 Nov 2016 17:35:44 +0000 (18:35 +0100)]
Add funding information

7 years agoFind a better way to be dimension independent. 3641/head
Wolfgang Bangerth [Mon, 28 Nov 2016 00:15:13 +0000 (17:15 -0700)]
Find a better way to be dimension independent.

This patch revises commit 9386b1e30c in #3639. There, I tried to avoid instantiating types
that have a template signature of the kind <3,spacedim> if spacedim is less than three,
since that leads to invalid types. I did so by using <3,max(3,spacedim)>, which works
because we can only get to these places if dim==3, and consequently spacedim>=3, but the
compiler may not always recognize that if spacedim<3, then we must be in dead code.

This patch goes the other way around, which I find conceptually clearer: in code where we
have
    if (dim == 3)
      {
        typename Triangulation<3,spacedim>::cell_iterator cell = ...;
we clearly only got into the 'if' block because dim==3. Consequently, we might as well
have written
    if (dim == 3)
      {
        typename Triangulation<dim,spacedim>::cell_iterator cell = ...;
to the same effect. On the plus side, however, whenever dim<3 (i.e. in cases where
this is dead code anyway), we still do not instantiate invalid types.

This patch therefore undoes the changes made in #3639 and replaces them with the second
option above. There are a number of places where we then have to replace calls to
    GeometryInfo<3>::...
by
    GeometryInfo<dim>::...
as well as one place where we have to do a cast from
    RefinementCase<3>
to
    RefinementCase<dim>
or the other way around; the latter is the identity cast if we do get into this block.

7 years agoMerge pull request #3639 from bangerth/do-not-instantiate-invalid-classes
Luca Heltai [Sun, 27 Nov 2016 09:55:01 +0000 (10:55 +0100)]
Merge pull request #3639 from bangerth/do-not-instantiate-invalid-classes

Do not instantiate invalid classes.


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.