Misleading comments about "counter" in NMT cacher

[rootulp]

Context

celestia-app/pkg/inclusion/nmt_caching.go

Lines 76 to 85 in bce5e5f

	// EDSSubTreeRootCacher caches the inner nodes for each row so that we can
	// traverse it later to check for blob inclusion. NOTE: Currently this has to
	// use a leaky abstraction (see docs on counter field below), and is not
	// threadsafe, but with a future refactor, we could simply read from rsmt2d and
	// not use the tree constructor which would fix both of these issues.
	type EDSSubTreeRootCacher struct {
	mut *sync.RWMutex
	caches map[uint]*subTreeRootCacher
	squareSize uint64
	}

celestia-app/pkg/inclusion/nmt_caching.go

Lines 95 to 113 in bce5e5f

	// Constructor fulfills the rsmt2d.TreeCreatorFn by keeping a pointer to the
	// cache and embedding it as a nmt.NodeVisitor into a new wrapped nmt.
	func (stc *EDSSubTreeRootCacher) Constructor(axis rsmt2d.Axis, axisIndex uint) rsmt2d.Tree {
	// see docs of counter field for more
	// info. if the counter is even or == 0, then we make the assumption that we
	// are creating a tree for a row
	var newTree wrapper.ErasuredNamespacedMerkleTree
	switch axis {
	case rsmt2d.Row:
	strc := newSubTreeRootCacher()
	stc.mut.Lock()
	stc.caches[axisIndex] = strc
	stc.mut.Unlock()
	newTree = wrapper.NewErasuredNamespacedMerkleTree(stc.squareSize, axisIndex, nmt.NodeVisitor(strc.Visit))
	default:
	newTree = wrapper.NewErasuredNamespacedMerkleTree(stc.squareSize, axisIndex)
	}
	return &newTree
	}

Problem

Both of the code blocks above refer to a counter field that doesn't exist 😕

Proposal

Investigate if the counter field was present in a previous implementation. If it has been removed, consider deleting these comments too.

[rootulp]

Counter was introduced in #549 but it was removed in #926. The second PR didn't remove all the comments related to counter.