Spring StringUtils.uriDecode decodes strings with non-ASCII characters incorrectly

Affects: 6.1.3

StringUtils.uriDecode will take two different paths depending if "%" is found. One of these does not handle non-ASCII characters.

Fast path:

StringUtils.uriDecode("ü", StandardCharsets.UTF_8) // returns "ü"

Replace path:

StringUtils.uriDecode("%20ü", StandardCharsets.UTF_8) // returns " �"

Comment From: dumbbelloper

Hello, I've submitted a Pull Request that aims to address this issue: PR #32373

Comment From: poutsma

StringUtils::uriDecode expects the input to be encoded, which means that the ü character should have been provided as %C3%BC instead. I will add a documentation note that makes this behavior clear.

Note that StringUtils::uriDecode is a low-level, internal utility method exposed via higher-level components, such as UriComponentsBuilder, that are better at dealing with [en|de]coded URIs.