Set encoding to UTF-8 when no encoding is specified

[phihag]

PayPal's actual responses just say Content-Type: application/json, causing requests to guess the encoding. This is wrong; the data from PayPal is always UTF-8.
Therefore, default to UTF-8 when no encoding is set.

Demonstration of the problem:

import paypalhttp

class PseudoRequest:
    verb = 'GET'
    path = '/2020/paypal.json'
    headers = {'Content-Type': 'application/json'}
    body = None

class PseudoEnvironment:
    base_url = 'https://phihag.de'

environment = PseudoEnvironment()
client = paypalhttp.HttpClient(environment)
res = client.execute(PseudoRequest())
s = res.result.payer.address.admin_area_1

assert 'ü' in s, s

[cschlueter]

I'm running into the same problem. While executing an OrdersGetRequest with an Umlaut inside the shipping address. This should warrant an urgent fix, because I can imagine a lot of people run into this problem.

Furthermore, I can validate the proposed fix on my side, after changing the line manually, locally in the braintree http library it works as expected. Obviously not a solution for production.

@phihag any way to work around this issue without a fix in the library? I'm a little lost.

[phihag]

@cschlueter Instead of making an instance of PayPalHttpClient, make an instance of the following class:

class FixedEncodingPayPalHttpClient(PayPalHttpClient):
    # Workaround for https://www.paypal-support.com/s/case/5002E00001bw3sxQAA/ /
    # https://github.com/paypal/paypalhttp_python/pull/5
    def parse_response(self, response):
        response.encoding = response.encoding or 'utf-8'
        return super().parse_response(response)

[cschlueter]

@cschlueter Instead of making an instance of PayPalHttpClient, make an instance of the following class:

class FixedEncodingPayPalHttpClient(PayPalHttpClient):
    # Workaround for https://www.paypal-support.com/s/case/5002E00001bw3sxQAA/ /
    # https://github.com/paypal/paypalhttp_python/pull/5
    def parse_response(self, response):
        response.encoding = response.encoding or 'utf-8'
        return super().parse_response(response)

Ah, perfect, thanks so much for the workaround.