LibWeb: Tokenize CSS numbers as doubles

Every later stage uses doubles, so dropping that precision right at the
start of parsing is a little silly. :^)

Userland/Libraries/LibWeb/CSS/Parser/Tokenizer.cpp

@@ -574,11 +574,11 @@ Number Tokenizer::consume_a_number()
 }
 
 // https://www.w3.org/TR/css-syntax-3/#convert-string-to-number
-float Tokenizer::convert_a_string_to_a_number(StringView string)
+double Tokenizer::convert_a_string_to_a_number(StringView string)
 {
     // FIXME: We already found the whole part, fraction part and exponent during
     //        validation, we could probably skip
-    return string.to_float(AK::TrimWhitespace::No).release_value();
+    return string.to_double(AK::TrimWhitespace::No).release_value();
 }
 
 // https://www.w3.org/TR/css-syntax-3/#consume-name

Userland/Libraries/LibWeb/CSS/Parser/Tokenizer.h

@@ -87,7 +87,7 @@ private:
     [[nodiscard]] ErrorOr<Token> consume_a_numeric_token();
     [[nodiscard]] ErrorOr<Token> consume_an_ident_like_token();
     [[nodiscard]] Number consume_a_number();
-    [[nodiscard]] float convert_a_string_to_a_number(StringView);
+    [[nodiscard]] double convert_a_string_to_a_number(StringView);
     [[nodiscard]] ErrorOr<FlyString> consume_an_ident_sequence();
     [[nodiscard]] u32 consume_escaped_code_point();
     [[nodiscard]] ErrorOr<Token> consume_a_url_token();