apply_map on sparse vectors returns vectors of smaller degree

[tfeulner]

Calling apply_map on a sparse vector with zeros at the last positions returns a vector of wrong degree.

sage: F.<a> = GF(9)
sage: v = vector([a, 0,0,0], sparse=True)
sage: f = F.hom([a**3])
sage: v.apply_map(f)
(2*a + 1)

[rbeezer]

It looks like you have duplicated some code for finding the common ring. I think that ideally this step should remain in the vector constructor.

Before returning, you have the mapped elements, the degree, and the sparseness. Did you try calling vector() with all the arguments supplied named, ie

arg0, arg1, arg2, sparse

much as calling format number 4, but with sparseness supplied? Or maybe calling format 4 needs to be sure to handle the sparse keyword properly.

I could be wrong about this, since I have not fully tested everything, but is there a way to just do the mapping and pass the degree in the sparse and non-sparse cases? Let the constructor determine the new ring for the mapped entries?

[tfeulner]

Replying to rbeezer:

It looks like you have duplicated some code for finding the common ring. I think that ideally this step should remain in the vector constructor.

I agree with you, this should stay in the constructor.

The updated patch now just explicitly adds, in the cases where it is necessary, one zero entry to the dictionary.

[rbeezer]

Sorry to have lost track of this one.

Fix looks good, behaves as expected, passes tests, and docs look good. Positive review.

Thanks for sticking with this one.

Rob