Simplex tree ctors with data

[VincentRouvreau]

rec_copy can copy or not simplex_data.
Some documentation and tests to explain when data are copied or not.
Fix #1127

[hschreiber]

Is it wanted that the non assigned data seems to take a random value ? For unit test "simplex_data", the output is:

(2 1 0 ) filtration=0 - data='4'
(1 0 ) filtration=0 - data='5'
(2 0 ) filtration=0 - data='43808336'
(0 ) filtration=0 - data='2'
(2 1 ) filtration=0 - data='43808304'
(1 ) filtration=0 - data='43808304'
(2 ) filtration=0 - data='43808304'

Data is of type int and only 4 and 5 were assigned.

[VincentRouvreau]

Is it wanted that the non assigned data seems to take a random value ? For unit test "simplex_data", the output is:

(2 1 0 ) filtration=0 - data='4'
(1 0 ) filtration=0 - data='5'
(2 0 ) filtration=0 - data='43808336'
(0 ) filtration=0 - data='2'
(2 1 ) filtration=0 - data='43808304'
(1 ) filtration=0 - data='43808304'
(2 ) filtration=0 - data='43808304'

Data is of type int and only 4 and 5 were assigned.

It is the behaviour we want. We do not say it, but data are not initialized with a default value, that would take time. It is up to the user to deal with it.

[mglisse]

We do not say it, but data are not initialized with a default value

If it was enough to disturb @hschreiber , maybe we should be explicit in the doc that this gets a default-initialization, and remind people that for scalars this may mean an indeterminate value.

Or we could change to value-initialize, I guess the main case where this could make a difference is if Simplex_data is a class that contains a large array of scalars, I don't know how likely that is.

(I haven't looked at the code)

[VincentRouvreau]

If it was enough to disturb @hschreiber , maybe we should be explicit in the doc that this gets a default-initialization, and remind people that for scalars this may mean an indeterminate value.

I did so on 961919a

[mglisse]

Suggested change

	// Specific for optionnal data
	// Specific for optional data

(same elsewhere)

[VincentRouvreau]

Done on 72bbb88 for the 2 typos. Codespell pass on the rest of the code