- std::initializer_list has its own size() method, so no need to use
std::distance.
STYLE/BUG: use separate iterator de-reference and increment in List
- avoids unnecessary copying of iterators, and avoids any potentially
odd behaviour with the combination with incrementing.
ENH: support construct from iterator pair for DynamicList, SortableList
- inherit from std::iterator to obtain the full STL typedefs, meaning
that std::distance works and the following is now possible:
labelRange range(100, 1500);
scalarList list(range.begin(), range.end());
--
Note that this does not work (mismatched data-types):
scalarList list = identity(12345);
But this does, since the *iter promotes label to scalar:
labelList ident = identity(12345);
scalarList list(ident.begin(), ident.end());
It is however more than slightly wasteful to create a labelList
just for initializing a scalarList. An alternative could be a
a labelRange for the same purpose.
labelRange ident = labelRange::identity(12345);
scalarList list(ident.begin(), ident.end());
Or this
scalarList list
(
labelRange::null.begin(),
labelRange::identity(12345).end()
);
- provides const/non-const access to the underlying list, but the
iterator access itself is const.
- provide linked-list iterator 'found()' method for symmetry with
hash-table iterators. Use nullptr for more clarity.
- lookup(): with a default value (const access)
For example,
Map<label> something;
value = something.lookup(key, -1);
being equivalent to the following:
Map<label> something;
value = -1; // bad value
if (something.found(key))
{
value = something[key];
}
except that lookup also makes it convenient to handle const references.
Eg,
const labelList& ids = someHash.lookup(key, labelList());
- For consistency, provide a two parameter HashTable '()' operator.
The lookup() method is, however, normally preferable when
const-only access is to be ensured.
- retain(): the counterpart to erase(), it only retains entries
corresponding to the listed keys.
For example,
HashTable<someType> largeCache;
wordHashSet preserve = ...;
largeCache.retain(preserve);
being roughly equivalent to the following two-stage process,
but with reduced overhead and typing, and fewer potential mistakes.
HashTable<someType> largeCache;
wordHashSet preserve = ...;
{
wordHashSet cull(largeCache.toc()); // all keys
cull.erase(preserve); // except those to preserve
largeCache.erase(cull); //
}
The HashSet &= operator and retain() are functionally equivalent,
but retain() also works with dissimilar value types.
- less clutter and typing to use the default template parameter when
the key is 'word' anyhow.
- use EdgeMap instead of the longhand HashTable version where
appropriate
- ensure that each have found() and which() methods
- add faceTraits for handling compile-time differences between
'normal' and tri-faces
- provide line::unitVec method (complimentary to edge::unitVec)
- provide key_iterator/const_key_iterator for all hashes,
reuse directly for HashSet as iterator/const_iterator, respectively.
- additional keys() method for HashTable that returns a wrapped to
a pair of begin/end const_iterators with additional size/empty
information that allows these to be used directly by anything else
expecting things with begin/end/size. Unfortunately does not yet
work with std::distance().
Example,
for (auto& k : labelHashTable.keys())
{
...
}
- add increment/decrement, repositioning. Simplify const_iterator.
- this makes is much easier to use labelRange for constructing ranges of
sub-lists. For symmetry with setSize() it has a setStart() instead of
simply assigning to start() directly. This would also provide the
future possibility to imbue the labelRange with a particular policy
(eg, no negative starts, max size etc) and ensure that they are
enforced.
A simple use case:
// initialize each to zero...
List<labelRange> subListRanges = ...;
// scan and categorize
if (condition)
subListRanges[categoryI]++; // increment size for that category
// finally, set the starting points
start = 0;
for (labelRange& range : subListRanges)
{
range.setStart(start);
start += range.size();
}
- previously had a mismash of const/non-const attributes on iterators
that were confused with the attributes of the object being accessed.
- use the iterator keys() and object() methods consistently for all
internal access of the HashTable iterators. This makes the intention
clearer, the code easier to maintain, and protects against any
possible changes in the definition of the operators.
- 'operator*': The standard form expected by STL libraries.
However, for the std::map, this dereferences to a <key,value> pair,
whereas OpenFOAM dereferences simply to <value>.
- 'operator()': OpenFOAM treats this like the 'operator*'
- adjusted the values of end() and cend() to reinterpret from nullObject
instead of returning a static iteratorEnd() object.
This means that C++ templates can now correctly deduce and match
the return types from begin() and end() consistently.
So that range-based now works.
Eg,
HashTable<label> table1 = ...;
for (auto i : table1)
{
Info<< i << endl;
}
Since the 'operator*' returns hash table values, this prints all the
values in the table.
This uses a concept similar to what std::valarray and std::slice do.
A labelRange provides a convenient container for holding start/size
and lends itself to addressing 'sliced' views of lists.
For safety, the operations and constructors restricts the given input range
to a valid addressible region of the underlying list, while the labelRange
itself precludes negative sizes.
The SubList version is useful for patches or other things that have a
SubList as its parameter. Otherwise the UList [] operator will be the
more natural solution. The slices can be done with a labelRange, or
a {start,size} pair.
Examples,
labelList list1 = identity(20);
list1[labelRange(18,10)] = -1;
list1[{-20,25}] = -2;
list1[{1000,5}] = -3;
const labelList list2 = identity(20);
list2[{5,10}] = -3; // ERROR: cannot assign to const!
- optimize erasure using different HashTable based on its size.
Eg, hashtable.erase(other);
If 'other' is smaller than the hashtable, it is more efficient to
use the keys from other to remove from the hashtable.
Otherwise simply iterate over the hashtable and remove it if
that key was found in other.
- some functionality similar to what the standary library <iterator>
provides.
* stdFoam::begin() and stdFoam::end() do type deduction,
which means that many cases it is possible to manage these types
of changes.
For example, when managing a number of indices:
Map<labelHashSet> lookup;
1) Longhand:
for
(
Map<labelHashSet>::const_iterator iter = lookup.begin();
iter != lookup.end();
++iter
)
{ .... }
1b) The same, but wrapped via a macro:
forAllConstIter(Map<labelHashSet>, lookup, iter)
{ .... }
2) Using stdFoam begin/end templates directly
for
(
auto iter = stdFoam::begin(lookup);
iter != stdFoam::end(lookup);
++iter
)
{ .... }
2b) The same, but wrapped via a macro:
forAllConstIters(lookup, iter)
{ .... }
Note that in many cases it is possible to simply use a range-based for.
Eg,
labelList myList;
for (auto val : myList)
{ ... }
for (const auto& val : myList)
{ ... }
These however will not work with any of the OpenFOAM hash-tables,
since the standard C++ concept of an iterator would return a key,value
pair when deferencing the *iter.
The deduction methods also exhibits some slightly odd behaviour with
some PtrLists (needs some more investigation).
- make construct from UList explicit and provide corresponding
assignment operator.
- add construct,insert,set,assignment from FixedList.
This is convenient when dealing with things like edges or triFaces.
- explicitly mention the value-initialized status for the operator().
This means that the following code will properly use an initialized
zero.
HashTable<label> regionCount;
if (...)
regionCount("region1")++;
... and also this;
if (regionCount("something") > 0)
{
...
}
Note that the OpenFOAM HashTable uses operator[] to provide read and
write access to *existing* entries and will provoke a FatalError if
the entry does not exist.
The operator() provides write access to *existing* entries or will
create the new entry as required.
The STL hashes use operator[] for this purpose.
- more hash-like methods.
Eg, insert/erase via lists, clear(), empty(),...
- minVertex(), maxVertex() to return the smallest/largest label used
- improved documentation, more clarification about where/how negative
point labels are treated.
