Normalize hashing algorithm names

[tiran]

BPO	38153
Nosy	@gpshead, @tiran, @miss-islington, @iritkatriel
PRs	bpo-38153: Normalize hashing algo names #16083 bpo-38153: detect shake independently from sha3 (GH-16143) #16143 [3.8] bpo-38153: Normalize hashlib algorithm names (GH-16083) (GH-16144) #16144 [3.8] bpo-38153: detect shake independently from sha3 (GH-16143) #16179

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/tiran'
closed_at = <Date 2020-10-10.01:36:14.198>
created_at = <Date 2019-09-13.09:50:32.516>
labels = ['extension-modules', 'type-bug', '3.8', '3.9']
title = 'Normalize hashing algorithm names'
updated_at = <Date 2020-10-10.01:36:14.197>
user = 'https://github.com/tiran'

bugs.python.org fields:

activity = <Date 2020-10-10.01:36:14.197>
actor = 'gregory.p.smith'
assignee = 'christian.heimes'
closed = True
closed_date = <Date 2020-10-10.01:36:14.198>
closer = 'gregory.p.smith'
components = ['Extension Modules']
creation = <Date 2019-09-13.09:50:32.516>
creator = 'christian.heimes'
dependencies = []
files = []
hgrepos = []
issue_num = 38153
keywords = ['patch']
message_count = 8.0
messages = ['352270', '352314', '352324', '352436', '352539', '352541', '378355', '378360']
nosy_count = 4.0
nosy_names = ['gregory.p.smith', 'christian.heimes', 'miss-islington', 'iritkatriel']
pr_nums = ['16083', '16143', '16144', '16179']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue38153'
versions = ['Python 3.8', 'Python 3.9']

[tiran]

The hashlib module / PEP-452 and OpenSSL use slightly different conventions for hashing algorithm names. The old and common algorithms like md5 to sha512 use the same strings (all lower case, no dashes or underscores). But new algorithms like sha3_512, shake, sha512_256, and blake2 use different conventions.

The inconsistency bloats the list of available algorithms. Also the builtin OpenSSL constructor does not support Python's preferred names.

>>> import hashlib, _hashlib
>>> sorted(hashlib.algorithms_available)
['blake2b', 'blake2b512', 'blake2s', 'blake2s256', 'md4', 'md5', 'md5-sha1', 'ripemd160', 'sha1', 'sha224', 'sha256', 'sha3-224', 'sha3-256', 'sha3-384', 'sha3-512', 'sha384', 'sha3_224', 'sha3_256', 'sha3_384', 'sha3_512', 'sha512', 'sha512-224', 'sha512-256', 'shake128', 'shake256', 'shake_128', 'shake_256', 'sm3', 'whirlpool']
>>> _hashlib.new("sha3_512")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: unsupported hash type
>>> _hashlib.new("sha3-512")
<sha3-512 HASH object @ 0x7f1387890840>

I propose to normalize names to Python standard names for HASH.name, repr, list of available algorithms, and for the new() constructor.

[gpshead]

as discussed, not pushing this back into 3.7 as we aren't hearing users call this a problem.

[gpshead]

New changeset 995b5d3 by Gregory P. Smith (Christian Heimes) in branch 'master':
bpo-38153: Normalize hashlib algorithm names (GH-16083)
995b5d3

[tiran]

New changeset eb2b0c6 by Christian Heimes in branch 'master':
bpo-38153: detect shake independently from sha3 (GH-16143)
eb2b0c6

[tiran]

New changeset e8d7fa2 by Christian Heimes in branch '3.8':
[3.8] bpo-38153: Normalize hashlib algorithm names (GH-16083) (GH-16144)
e8d7fa2

[miss-islington]

New changeset 0067fc2 by Miss Islington (bot) in branch '3.8':
bpo-38153: detect shake independently from sha3 (GH-16143)
0067fc2

[iritkatriel]

Can this be closed?

[gpshead]

looks like it, thanks!