Merge branch 'feature/bump-to-2.8' into 'master'

See merge request [!64]

  [!64]: Nonestaging/dune-typetree/merge_requests/64

Merge branch 'feature/improve-treecontainer' into 'master'

ref:staging/dune-typetree Add alias TreeContainer using a meta function to
determine the per node value

-   Add support for TreeContainer<LeafToValue,Tree> where LeafToValue<Node> is
    the value associated to a node of type Node.
-   Add a manual lambda implementing the LeafToValue<Node> transformation by
    default constructing the values. This is needed because you cannot use a
    lambda expresion in a type alias.
-   Rename the existing TreeContainer<Value,Tree> alias to
    UniformTreeContainer<Value,Tree>. This is an **interface change** but:
    -   This is needed because you cannot have to templates of the same name.
    -   The new name is better.
    -   The old interface was not published in a release so far.

See merge request [!61]

  [!61]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/61

Merge branch 'feature/mailing-list-as-maintainer' into 'master'

ref:staging/dune-typetree As Steffen steps down as the maintainer of
dune-typetree and nobody claims this his or her project, I propose to set the
Dune developer mailing list as the dune.module email address. Further, there
will be no single maintainer to be asked for approval. Typetree will be like
the core module, were everybody does as he thinks it is the best. We help each
other by discussions and comments. I assume that especially users of
dune-functions and PDELab will follow the development and contribute further
improvements or features.

@smuething @dominic @oliver.sander @carsten.graeser @christi Would that be ok?

See merge request [!62]

  [!62]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/62

* Add support for `TreeContainer<LeafToValue,Tree>` where
  `LeafToValue<Node>` is the value associated to a node of
  type `Node`.
* Add a manual lambda implementing the `LeafToValue<Node>` transformation
  by default constructing the values. This is needed because you cannot
  use a lambda expresion in a type alias.
* Rename the existing `TreeContainer<Value,Tree>` alias to
  `UniformTreeContainer<Value,Tree>`.
  * This is needed because you cannot have to templates of the same name.
  * The new name is better.
  * The interface was not published in a release so far.

This simply calles the default constructor of the stored
container class.

Merge branch 'bugfix/deterministic-child-transformation' into 'master'

ref:staging/dune-typetree As a side effect of using parameter packs in the
implementation of the transformation behaviour of composite nodes, the
traversal order on the tree was not defined. The commit reimplements the
transformation such that evaluation order of children is left to right.

See merge request [!59]

  [!59]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/59

Merge branch 'bugfix/deterministic-child-transformation-cpp14' into 'bugfix/deterministic-child-transformation'

ref:staging/dune-typetree This should make [!59] compatible with C++14.

See merge request [!60]

  [!59]: gitlab.dune-project.org/NoneNone/merge_requests/59
  [!60]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/60

Merge branch 'feature/remove-vintage-treepaths' into 'master'

ref:staging/dune-typetree This MR (based on [!18]) removes DynamicTreePath,
MutableDynamicTreePath, MakeableDynamicTreePath, TreePathType::mostlyStatic,
and TreePathFactory. None of them is used any more in dune-typetree and
dune-pdelab when applying the modernization of the tree traversal algorithms
in [!18].

The key is, that the modernized tree traversal only uses HybridTreePath which
nicely does the job. The only other remaining tree path implementation is the
fully static TypeTree::TreePath. Getting rid of this one is harder, because
TreePath<i...> is used heavily as template parameter in dune-pdelab.

Notice that I'm not sure if removing all those implementations without
deprecation is a good idea. Hence this MR mostly serves for demonstrating how
this could be cleaned up. However, if the removed stuff is considered
internal, one may consider merging directly (after [!18] of course).

See merge request [!21]

  [!18]: gitlab.dune-project.org/NoneNone/merge_requests/18
  [!21]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/21

As a side effect of using parameter packs in the implementation of the
transformation behaviour of composite nodes, the traversal order on the
tree was not defined. The commit reimplements the transformation such
that evaluation order of children is left to right.

Merge branch 'feature/update-author-list' into 'master'

See merge request [!58]

  [!58]: Nonestaging/dune-typetree/merge_requests/58

Merge branch 'feature/remove-deprecated-and-unused-cmake-modules' into 'master'

ref:staging/dune-typetree They were deprecated in Dune-Typetree 2.6. They are
unused since half a year.

See merge request [!57]

  [!57]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/57

The modules were deprecated in Dune-Typetree 2.6.
The modules are unused since half a year.
Drop whole cmake/modules which became empty.

Merge branch 'bugfix/add-childextraction-include' into 'master'

ref:staging/dune-typetree Thanks to Simon Praetorius for pointing out the
issue. This is a fixup of [!53].

See merge request [!54]

  [!53]: gitlab.dune-project.org/NoneNone/merge_requests/53
  [!54]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/54

Thanks to Simon Praetorius for pointing out the issue.

Merge branch 'feature/leaftreepathtuple' into 'master'

ref:staging/dune-typetree This generates a tuple with all tree paths referring
to the leaf nodes of a tree.

See merge request [!53]

  [!53]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/53

Merge branch 'feature/constexpr-hybridtreepath' into 'master'

ref:staging/dune-typetree While the member functions of HybridTreePath are
were already constexpr this was not the case for all utility functions for
HybridTreePath. This avoided constexpr successive construction.

See merge request [!52]

  [!52]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/52

This generates a tuple with all tree paths refering to the
leaves of a tree.

While the member functions of `HybridTreePath` are were
already `constexpr` this was not the case for all utility
functions for `HybridTreePath`. This avoided `constexpr`
successive construction.

Merge branch 'doc/update-version' into 'master'

ref:staging/dune-typetree The version was still stuck at 2.6.0-rc1.

See merge request [!50]

  [!50]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/50

Merge branch 'feature/make-child-template-invalid-if-child-not-found' into 'master'

ref:staging/dune-typetree The type alias Child<Node,...> used to simply be
void if the child type could not be extracted. That is rather unhelpful, as it
makes it impossible to use in SFINAE.

This patch changes the behavior by making it impossible to instantiate the
type alias if there is no valid child.

See merge request [!49]

  [!49]: gitlab.dune-project.org/staging/dune-typetree/merge_requests/49

The type alias `Child<Node,...>` used to simply be `void` if the child type
could not be extracted. That is rather unhelpful, as it makes it impossible to
use in SFINAE.

This patch changes the behavior by making it impossible to instantiate the type
alias if there is no valid child.

Merge branch 'ci/centralize-config' into 'master'

See merge request [!48]

  [!48]: Nonestaging/dune-typetree/merge_requests/48

Closes #8

See merge request !47

See merge request !46

Implement helper makeTreeContainer() to construct container matching the tree structure

See merge request !37

Revert "Merge branch 'cherry-pick-9cf4ae48' into 'releases/2.6'"

See merge request !43

This reverts merge request !42

[LeafNode] Add empty child() method for consistency

See merge request !41