Refactor hash/dgst from one-shot to streaming

[rlm2002]

Addresses issue where hash and digest function encounter hash collisions with large files. Changes from one shot file read to streaming method.

Adds unit tests for hash and digest.

python3 tests/hash/hash-test.py   /* regression test will run with sha256 */
python3 tests/dgst/dgst-test.py.   /* regression test will run with RSA */

/* test other algorithms
 * Valid hash values: md5, sha, sha256, sha384, sha512. Valid key kinds: rsa, ecc
 */
WOLFCLU_LARGE_HASH_ALG=sha512 python3 tests/hash/hash-test.py
WOLFCLU_LARGE_DGST_KEY=ecc   python3 tests/dgst/dgst-test.py

*Note: each algorithm hashes a 4.5 GiB sparse file, so `all` is slow —
budget several minutes per algorithm. The env var only accepts a single algorithm name or `all`.

Fixes zd21765

[wolfSSL-Fenrir-bot]

Fenrir Automated Review — PR #225

Scan targets checked: wolfclu-bugs, wolfclu-src

No new issues found in the changed files. ✅

[lealem47]

@rlm2002 Could you extract the hash streaming loop in clu_dgst_setup.c and clu_hash.c into one shared function? You could put it in src/tools/clu_funcs.c

[lealem47]

Thanks @rlm2002 !