CVE-2020-26269: TensorFlow vulnerable to heap out of bounds read in filesystem glob matching

The general implementation for matching filesystem paths to globbing pattern is vulnerable to an access out of bounds of the array holding the directories:

if (!fs->Match(child_path, dirs[dir_index])) { ... }

Since dir_index is unconditionaly incremented outside of the lambda function where the vulnerable pattern occurs, this results in an access out of bounds issue under certain scenarios. For example, if /tmp/x is a directory that only contains a single file y, then the following scenario will cause a crash due to the out of bounds read:

>>> tf.io.gfile.glob('/tmp/x/')
Segmentation fault

There are multiple invariants and preconditions that are assumed by the parallel implementation of GetMatchingPaths but are not verified by the PRs introducing it (#40861 and #44310). Thus, we are completely rewriting the implementation to fully specify and validate these.