diff options
| -rw-r--r-- | src/pkg/http/request.go | 2 | ||||
| -rw-r--r-- | src/pkg/http/url.go | 38 | ||||
| -rw-r--r-- | src/pkg/http/url_test.go | 19 |
3 files changed, 41 insertions, 18 deletions
diff --git a/src/pkg/http/request.go b/src/pkg/http/request.go index 10dc08c2b..884fe48fa 100644 --- a/src/pkg/http/request.go +++ b/src/pkg/http/request.go @@ -134,7 +134,7 @@ const defaultUserAgent = "Go http package" // // If Body is present, "Transfer-Encoding: chunked" is forced as a header. func (req *Request) Write(w io.Writer) os.Error { - uri := URLEscape(req.URL.Path) + uri := urlEscape(req.URL.Path, false) if req.URL.RawQuery != "" { uri += "?" + req.URL.RawQuery } diff --git a/src/pkg/http/url.go b/src/pkg/http/url.go index f879f8d8f..40ea86549 100644 --- a/src/pkg/http/url.go +++ b/src/pkg/http/url.go @@ -128,7 +128,13 @@ func CanonicalPath(path string) string { // converting %AB into the byte 0xAB and '+' into ' ' (space). // It returns an error if any % is not followed // by two hexadecimal digits. -func URLUnescape(s string) (string, os.Error) { +func URLUnescape(s string) (string, os.Error) { return urlUnescape(s, true) } + +// urlUnescape is like URLUnescape but can be told not to +// convert + into space. URLUnescape implements what is +// called "URL encoding" but that only applies to query strings. +// Elsewhere in the URL, + does not mean space. +func urlUnescape(s string, doPlus bool) (string, os.Error) { // Count %, check that they're well-formed. n := 0 hasPlus := false @@ -145,7 +151,7 @@ func URLUnescape(s string) (string, os.Error) { } i += 3 case '+': - hasPlus = true + hasPlus = doPlus i++ default: i++ @@ -165,7 +171,11 @@ func URLUnescape(s string) (string, os.Error) { j++ i += 3 case '+': - t[j] = ' ' + if doPlus { + t[j] = ' ' + } else { + t[j] = '+' + } j++ i++ default: @@ -178,12 +188,14 @@ func URLUnescape(s string) (string, os.Error) { } // URLEscape converts a string into URL-encoded form. -func URLEscape(s string) string { +func URLEscape(s string) string { return urlEscape(s, true) } + +func urlEscape(s string, doPlus bool) string { spaceCount, hexCount := 0, 0 for i := 0; i < len(s); i++ { c := s[i] if shouldEscape(c) { - if c == ' ' { + if c == ' ' && doPlus { spaceCount++ } else { hexCount++ @@ -199,7 +211,7 @@ func URLEscape(s string) string { j := 0 for i := 0; i < len(s); i++ { switch c := s[i]; { - case c == ' ': + case c == ' ' && doPlus: t[j] = '+' j++ case shouldEscape(c): @@ -314,16 +326,16 @@ func ParseURL(rawurl string) (url *URL, err os.Error) { url.Userinfo, url.Host = split(url.Authority, '@', true) } - if url.Path, err = URLUnescape(path); err != nil { + if url.Path, err = urlUnescape(path, false); err != nil { goto Error } // Remove escapes from the Authority and Userinfo fields, and verify // that Scheme and Host contain no escapes (that would be illegal). - if url.Authority, err = URLUnescape(url.Authority); err != nil { + if url.Authority, err = urlUnescape(url.Authority, false); err != nil { goto Error } - if url.Userinfo, err = URLUnescape(url.Userinfo); err != nil { + if url.Userinfo, err = urlUnescape(url.Userinfo, false); err != nil { goto Error } if strings.Index(url.Scheme, "%") >= 0 { @@ -349,7 +361,7 @@ func ParseURLReference(rawurlref string) (url *URL, err os.Error) { if url, err = ParseURL(rawurl); err != nil { return nil, err } - if url.Fragment, err = URLUnescape(frag); err != nil { + if url.Fragment, err = urlUnescape(frag, false); err != nil { return nil, &URLError{"parse", rawurl, err} } return url, nil @@ -368,16 +380,16 @@ func (url *URL) String() string { if url.Host != "" || url.Userinfo != "" { result += "//" if url.Userinfo != "" { - result += URLEscape(url.Userinfo) + "@" + result += urlEscape(url.Userinfo, false) + "@" } result += url.Host } - result += URLEscape(url.Path) + result += urlEscape(url.Path, false) if url.RawQuery != "" { result += "?" + url.RawQuery } if url.Fragment != "" { - result += "#" + URLEscape(url.Fragment) + result += "#" + urlEscape(url.Fragment, false) } return result } diff --git a/src/pkg/http/url_test.go b/src/pkg/http/url_test.go index 542ad0a38..a16bbbddf 100644 --- a/src/pkg/http/url_test.go +++ b/src/pkg/http/url_test.go @@ -54,7 +54,7 @@ var urltests = []URLTest{ "www.google.com", "", "www.google.com", "/file one&two", "", "", }, - "http://www.google.com/file+one%26two", + "http://www.google.com/file%20one%26two", }, // user URLTest{ @@ -76,7 +76,7 @@ var urltests = []URLTest{ "john doe@www.google.com", "john doe", "www.google.com", "/", "", "", }, - "ftp://john+doe@www.google.com/", + "ftp://john%20doe@www.google.com/", }, // query URLTest{ @@ -100,6 +100,17 @@ var urltests = []URLTest{ }, "", }, + // %20 outside query + URLTest{ + "http://www.google.com/a%20b?q=c+d", + &URL{ + "http://www.google.com/a%20b?q=c+d", + "http", "//www.google.com/a%20b?q=c+d", + "www.google.com", "", "www.google.com", + "/a b", "q=c+d", "", + }, + "", + }, // path without /, so no query parsing URLTest{ "http:www.google.com/?q=go+language", @@ -107,9 +118,9 @@ var urltests = []URLTest{ "http:www.google.com/?q=go+language", "http", "www.google.com/?q=go+language", "", "", "", - "www.google.com/?q=go language", "", "", + "www.google.com/?q=go+language", "", "", }, - "http:www.google.com/%3fq%3dgo+language", + "http:www.google.com/%3fq%3dgo%2blanguage", }, // non-authority URLTest{ |