GitLab

STYLE: use 'forAll(' NOT 'forAll (', as per coding guide

An empty list (m_) has no pointer so cannot be referenced. Instead we
now return a null pointer (which is no problem for UList)

- the size_ was not being adjusted in two assignment operators:
      PackedList<nBits>::operator=(const PackedList<nBits>&)
      PackedList<nBits>::operator=(const UList<label>&)

- provides a convenient (and lazy) means of removing entries

- resize with factor 2 as per DynamicList

Old insertion speed:
    1000000 in 0.61 s
    2000000 in 2.24 s
    3000000 in 3.97 s
    4000000 in 5.76 s
    5000000 in 7.54 s
    6000000 in 9.41 s
    7000000 in 11.5 s

New insertion speed:
    1000000 in 0.01 s
    2000000 in 0.02 s
    3000000 in 0.01 s
    4000000 in 0.02 s
    5000000 in 0.01 s
    6000000 in 0.01 s
    7000000 in 0.01 s

- now that I re-examined the code, the note in commit 51fd6327
  can be mostly ignored

  PackedList isMaster(nPoints, 1u);
  is not really inefficient at all, since the '1u' is packed into
  32/64-bits before the subsequent assignment and doesn't involve
  shifts/masking for each index

  The same misinformation applies to the PackedList(size, 0u) form.
  It isn't much slower at all.

  Nonetheless, add bool specialization so that it is a simple assign.

- this probably should have been in commit c7c9a6e0 (2009-11-04)
  but got overseen in the HashTbl -> HashTable switch

- allow insert() and set() from a UList of Key
  This complements the existing erase(const UList<Key>&) method

- this builds on Mattijs' commit 968f0bbd but with a first()
  as well.

- Added both to FixedList, IndirectList, UIndirectList and *PtrList and
  since they can certainly be useful there. Did not add to BiIndirectList,
  since I'm not sure what it should mean there. Did not add to PackedList,
  since it's not clear how useful they'd be yet in these contexts (and I'm
  not sure how it would interact with the iterator proxy implementation).

- Note: STL defines front() and back() for these methods.

- previous draft version was HashTbl
- accidentally still had canonicalSize in templated code

- don't let automatic resize into integer overflow
- avoid possible overflow in increment()

StaticHashTable - fix erase()/++ as per HashTbl

- might still be worth dropping it, but at least it'll be in better
  condition when it gets chucked!

- the end iterators and canonicalSize are independent of the templated
  parameters, so move them to HashTblCore

- this avoid allocating separate end iterator at all.

  end(), cend() return a iteratorEnd.
  iterator == iteratorEnd just checks for non-NULL pointer on iterator.
  The compiler should optimize most of it away.