Derrick Stolee <stolee@xxxxxxxxx> writes: > I think in this pruned form, the cache-tree is valid and serves all of > the roles it was designed for. We can quickly compute the root tree in > 'git commit'. We can navigate the cache-tree in traverse_by_cache_tree() > to accurately describe the trees in the index. These things are tested > in t1092 and the p2000 performance test. OK, then I was worried too much. Thanks for clarifying.
