net/url: problem with queryescape?

[vsdutka]

If recieve error when use url.QueryUnescape with query string like "Z_FILTER=%25\u0414\u0430\u0432\u044B\u0434\u043E\u0432%25\u0418\u0432\u0430\u043D%25"
But it is normal query string

[vsdutka]

Sorry/ Queryr string is "Z_FILTER=%25%u0414%u0430%u0432%u044B%u0434%u043E%u0432%25%u0418%u0432%u0430%u043D%25"

[bradfitz]

What's the problem?

[vsdutka]

Test case

func main(){

s, err := url.QueryUnescape("%25%u0414%u0430%u0432%u044B%u0434%u043E%u0432%25%u0418%u0432%u0430%u043D%25")

if err != nil {

    fmt.Println("Error - ", err)

}

fmt.Println(s)

}

Error - invalid URL escape "%u0"

2015-06-17 15:56 GMT+03:00 Brad Fitzpatrick notifications@github.com:

What's the problem?

—
Reply to this email directly or view it on GitHub
#11249 (comment).

[adg]

%u shouldn't appear in any valid urlencoded string: https://en.wikipedia.org/wiki/Percent-encoding#Non-standard_implementations

[vsdutka]

But this implementation massively used by IE.

2015-06-18 9:19 GMT+03:00 Andrew Gerrand notifications@github.com:

%u shouldn't appear in any valid urlencoded string:
https://en.wikipedia.org/wiki/Percent-encoding#Non-standard_implementations

—
Reply to this email directly or view it on GitHub.

[adg]

Can you show me where/how? I can't find information about this.

[vsdutka]

Our application have a lot of pages with code like:

function L_POONAME_W_EKB_FLAG_LOV() {
var filter = prompt("Введите критерий поиска для Лицо - Владелец А/М",
"%");
if (filter == null) { return; }
if (filter == "") { return; }
var the_pathname = location.pathname;
var i = the_pathname.indexOf ('/:');
var j = the_pathname.indexOf ('/', ++i);
if (i != -1)
{
// Syntactically incorrect url so it needs to be corrected
the_pathname = the_pathname.substring (j, the_pathname.length);
}; // (i != -1)
frmLOV = open("axiidvehtr$iidvehtr.l_pooname_w_ekb_flag_listofval" +
"?Z_FILTER=" + *escape(filter) *+ "&Z_MODE=Q" +
"&Z_CALLER_URL=" + *escape(location.protocol + '//' +
location.host + the_pathname + location.search) *+
"&Z_ISSUE_WAIT=Y",
"winLOV",
"scrollbars=yes,resizable=yes,width=400,height=400");
if (frmLOV.opener == null) {
frmLOV.opener = self;
}
}

For unicode chars *escape *produce string like "%uXXXX"

2015-06-18 10:27 GMT+03:00 Andrew Gerrand notifications@github.com:

Can you show me where/how? I can't find information about this.

—
Reply to this email directly or view it on GitHub
#11249 (comment).

[vsdutka]

I use code like

...

r.URL.RawQuery = decodeUnicodeEscapedChars(r.URL.RawQuery)

...

func decodeUnicodeEscapedChars(s string) string {

re := regexp.MustCompile(`%u[0-9ABCDEF]{4}`)

return re.ReplaceAllStringFunc(s, func(s string) string {

    t, err := strconv.Unquote("\"" + `\u` + s[2:] + "\"")

    if err != nil {

        panic(err)

    }

    return t

})

}

But i think its not so good practice.

2015-06-18 13:28 GMT+03:00 Vyacheslav Dutka vsdutka@gmail.com:

Our application have a lot of pages with code like:

function L_POONAME_W_EKB_FLAG_LOV() {
var filter = prompt("Введите критерий поиска для Лицо - Владелец А/М",
"%");
if (filter == null) { return; }
if (filter == "") { return; }
var the_pathname = location.pathname;
var i = the_pathname.indexOf ('/:');
var j = the_pathname.indexOf ('/', ++i);
if (i != -1)
{
// Syntactically incorrect url so it needs to be corrected
the_pathname = the_pathname.substring (j, the_pathname.length);
}; // (i != -1)
frmLOV = open("axiidvehtr$iidvehtr.l_pooname_w_ekb_flag_listofval" +
"?Z_FILTER=" + *escape(filter) *+ "&Z_MODE=Q" +
"&Z_CALLER_URL=" + *escape(location.protocol + '//' +
location.host + the_pathname + location.search) *+
"&Z_ISSUE_WAIT=Y",
"winLOV",
"scrollbars=yes,resizable=yes,width=400,height=400");
if (frmLOV.opener == null) {
frmLOV.opener = self;
}
}

For unicode chars *escape *produce string like "%uXXXX"

2015-06-18 10:27 GMT+03:00 Andrew Gerrand notifications@github.com:

Can you show me where/how? I can't find information about this.

—
Reply to this email directly or view it on GitHub
#11249 (comment).

[bradfitz]

I don't think we should encourage it by enshrining it in the standard library. Anybody can write a package to do this and place it elsewhere on Github.

But such a function probably should be more careful with allocations than decodeUnicodeEscapedChars above is.