`urllib.parse.quote` is not consistent about `encoding` and `errors` defaults

[andrei-korshikov]

Documentation

https://docs.python.org/3/library/urllib.parse.html#urllib.parse.quote reads:

encoding=None, errors=None

At the same time it reads:

encoding defaults to 'utf-8'. errors defaults to 'strict'

At least, it is confusing indeed. If it is not a bug, it should be explained, I think.

---

Off-topic: and thank you for Python:)

Linked PRs

• gh-130578: clarify urllib.parse.quote parameter defaults behavior #130598
• [3.15] gh-130578: clarify urllib.parse.quote parameter defaults behavior (GH-130598) #153383
• [3.14] gh-130578: clarify urllib.parse.quote parameter defaults behavior (GH-130598) #153384
• [3.13] gh-130578: clarify urllib.parse.quote parameter defaults behavior (GH-130598) #153385

[StanFromIreland]

It is partially correct:

cpython/Lib/urllib/parse.py

Line 875 in 43a2a37

def quote(string, safe='/', encoding=None, errors=None):

However:

cpython/Lib/urllib/parse.py

Lines 917 to 920 in 43a2a37

	if encoding is None:
	encoding = 'utf-8'
	if errors is None:
	errors = 'strict'

This should be updated like it is for unquote:

cpython/Lib/urllib/parse.py

Line 692 in 43a2a37

def unquote(string, encoding='utf-8', errors='replace'):

I'm not sure why it was done this way in 2008 in 52dbbb9 by Guido

I'll fix it

[Mr-Sunglasses]

It is partially correct:

cpython/Lib/urllib/parse.py

Line 875 in 43a2a37

def quote(string, safe='/', encoding=None, errors=None):
However:

cpython/Lib/urllib/parse.py

Lines 917 to 920 in 43a2a37

if encoding is None:
encoding = 'utf-8'
if errors is None:
errors = 'strict'
This should be updated like it is for unquote:

cpython/Lib/urllib/parse.py

Line 692 in 43a2a37

def unquote(string, encoding='utf-8', errors='replace'):
I'm not sure why it was done this way in 2008 in 52dbbb9 by Guido

I'll fix it

In the code, encoding and errors parameters default to None, but inside the function body, if the input is a string and these parameters are None, they're set to utf-8 and strict but in the documentation directly states that encoding defaults to utf-8 and errors defaults to strict, which doesn't match the function signature but does match the actual behavior.

I think the function signature uses None as a sentinel value to determine whether the caller explicitly provided these parameters, particularly when handling bytes input. The actual default behavior for string inputs should be indeed utf-8 for encoding and strict for errors.

So I think we can add something like this in the urllib.parse.quote documentation

"The parameters default to None in the signature, but for string inputs, encoding effectively defaults to utf-8 and errors to strict and it should not be specified for string."

[StanFromIreland]

We have quote_from_bytes() for bytes now, unquote already assumes str.

[Mr-Sunglasses]

@StanFromIreland The reason why it should not be update like:

cpython/Lib/urllib/parse.py

Line 692 in 43a2a37

def unquote(string, encoding='utf-8', errors='replace'):

because, if we make it default for the whole function signature but, it should be default to str only because this function accept both str and bytes

[andrei-korshikov]

Related inconsistency is present in the https://docs.python.org/3/library/urllib.parse.html#urllib.parse.urlencode section (bolding is mine):

An alternate function that can be passed as quote_via is quote(), which will encode spaces as %20 and not encode ‘/’ characters.

This is correct description for quote() behaviour, but it not correct in the context of urlencode(): / characters are effectively encoded as %2F with quote_via=urllib.parse.quote because of default safe=''.

I think, / encoding should not be mentioned at all in the context of urlencode(). Or, it may be better to clearly state that both quote() and quote_plus() encode / as %2F for the reason of safe='' default of urlencode().

[vadmium]

I think Issue #90555 is already about the urlencode aspect.