为什么url.Parse不填充URL.RawPath？

Question

While with similar input, URL.RawQuery seems to be properly populated.

Playground

u, err := url.Parse("https://example.com/foo%2fbar")
if err != nil {
    log.Fatal(err)
}
fmt.Println(u.Path)
fmt.Println(u.RawPath)
fmt.Println(u.String(), "
")

u, err = url.Parse("https://example.com/foo%25fbar?q=morefoo%25bar")
if err != nil {
    log.Fatal(err)
}
fmt.Println(u.Path)
fmt.Println(u.RawPath)
fmt.Println(u.RawQuery)
fmt.Println(u.String())

Answer 1

Note that URL.RawPath is not the raw (escaped) path. It may be, but not always. It's just a hint. Its doc says:

RawPath    string // encoded path hint (Go 1.5 and later only; see EscapedPath method)

So when you need the escaped path, always use URL.EscapedPath() and don't rely on the URL.RawPath field. Doc of URL.EscapedPath() says:

EscapedPath returns u.RawPath when it is a valid escaping of u.Path.

This is your case. A raw path is invalid if it contains bytes that need escaping during URL encoding. Your path is such an example because it contains %25 which is the URL escaped text of the percent symbol itself '%', and percent symbols need to be escaped if part of URL paths.

Your first example contained %2f which is the URL escaped text of the slash'/' which if present in a path does not need to be escaped ('/' is a valid character in paths and is treated as a separator).

See this example:

u, err = url.Parse("https://example.com/foo%25fbar?q=morefoo%25bar")
if err != nil {
    log.Fatal(err)
}
fmt.Println("Path:       ", u.Path)
fmt.Println("RawPath:    ", u.RawPath)
fmt.Println("EscapedPath:", u.EscapedPath())
fmt.Println("RawQuery:   ", u.RawQuery)
fmt.Println("String:     ", u.String())

Output (try it on the Go Playground):

Path:        /foo%fbar
RawPath:     
EscapedPath: /foo%25fbar
RawQuery:    q=morefoo%25bar
String:      https://example.com/foo%25fbar?q=morefoo%25bar

RawPath is the empty string (because "/foo%25fbar" is an invalid escaped path), but EscapedPath() returns you the original escaped path.