On some pages requests detect encoding incorrectly

[Euphorbium]

For example it detects http://www.15min.lt encoding as ISO-8859-1, but the encoding is set as utf8

[sigmavirus24]

This is a duplicate of #1087. You can see there that we don't read the content to determine the encoding. If you want you can use apparent_encoding() to get an encoding from charade. You can also use from requests.utils import get_encodings_from_content, but we don't parse the HTML anymore. We simply report encoding from the Headers and if there is none, we default to ISO-8859-1.

[Lukasa]

To clarify @sigmavirus24's point, we only default to ISO-8859-1 in some cases.