Inconsistent casing for percent-escapes in url-encoding

[jeroen]

The examples below are from my bindings, if needed I can translate them to C later.

I did this

When using curl_easy_escape() characters get escaped with uppercase percentage codes:

curl_easy_escape("foo;bar")
# foo%3Bbar

But when using the URL parser with CURLU_URLENCODE and setting e.g. CURLUPART_PATH then we get a lowercase:

curl:::curl_modify_url('https://test', path = 'foo;bar')
# "https://test/foo%3bbar

Is this intentional? The ambiguity is technically not incorrect but makes it more difficult to write tests.

I expected the following

Consistent upper or lowercase when url-encoding.

curl/libcurl version

8.14.1

operating system

MacOS 15

[bagder]

RFC 3986 section 2.1 says

For consistency, URI producers and normalizers should use uppercase hexadecimal digits for all percent- encodings.