Improve slice eval

[lshaw8317]

Necessary bug fixes for integer slicing for reductions and refactoring slices_eval as suggested in #436.
Had to add a couple of modifications to ensure contiguity of numpy arrays in some special cases where an evaluation (and hence copy, to contiguous memory) did not occur (when not using numexpr and when the expression was simply "o0"). Consequently also had to improve the rebasing of operands which the lazyexpr machinery carries out (prior to this PR, the rebasing would not occur when the lazyexpr was just a blosc2 array, i.e. "a" would not be rebased to "o0").
Before merging have to do a benchmark comparing new slicing with non-unit steps for lazy expressions (which only uses numpy at chunk-level) with the old format (which filled the whole span of the slice before taking the steps at the very end). Consequently, the old version might load a very large result into memory, even if the slice was small - consider an array of shape (10**9, ) with slice [::10**8]; the old Blosc2 would fill an array of shape (9.9**9, ) and then slice to give a result of shape (10,). The new version would never have more than the chunksize in memory.

[lshaw8317]

pre-commit.ci autofix

[lshaw8317]

The proposed change in Blosc2 3.6.2 leads to reduced memory consumption, especially when the slice is small relative to the array size (i.e. the step is very large), relative to the previous implementation in Blosc2 3.6.1. See the benchmark in bench/ndarray/slice-expr-step.py