i may have found the problem in LFCdhtml.js.
it has code to check if the request is a POST and doesn't have "content-type"
header set:
===
if (method_$0 == "POST" && headers_$2["content-type"] == null) {
headers_$2["content-type"] = "application/x-www-form-urlencoded"
};
===
however, my code is setting "Content-Type" header:
===
setHeader('Content-Type',
'application/x-www-form-urlencoded; charset=UTF-8');
===
my server sees the two HTTP headers and is combining them most likely.
the bug is that LFCdhtml.js is not treating HTTP header field name as
case-insensitive:
http://www.w3.org/Protocols/rfc2616/rfc2616-sec4.html
my patch consists of lower-casing the setHeader implementation in LFCdhtml.js:
===
--- LFCdhtml.js 2010-10-22 08:21:57.000000000 -0700
+++ LFCdhtml-new.js 2012-02-28 18:01:23.000000000 -0800
@@ -6494,7 +6494,7 @@
}},"setHeader",function($0,$1){
if(!this.headers){
this.headers=new (lz.Param)()
-};this.headers.setValue($0,$1)
+};this.headers.setValue($0.toLowerCase(),$1)
},"getRequestHeaderParams",function(){
return this.headers
},"clearRequestHeaderParams",function(){
===
i will be filling a bug on this issue.
thanks,
augusto.
On Feb 28, 2012, at 5:21 PM, augusto callejas wrote:
> hi-
>
> i have code below that performs a POST request, where i'm setting the
> Content-Type HTTP header.
>
> ===
> <dataset name="stopWS" querytype="POST" type="http"
> datafromchild="true" proxied="false" queuerequests="true">
> <method name="doRequest">
> setHeader('Content-Type',
> 'application/x-www-form-urlencoded; charset=UTF-8');
> super.doRequest();
> </method>
> </dataset>
> ===
>
> ===
> stopWS.setAttribute('src',
> '/api/stop');
> stopWS.doRequest();
> ===
>
> i'm seeing two different values in the HTTP request header Content-Type on
> the server, depending on the version of the client:
>
> swf8: application/x-www-form-urlencoded; charset=UTF-8
> html5: application/x-www-form-urlencoded; charset=UTF-8
> application/x-www-form-urlencoded
>
> the html5 version of my app is appending "application/x-www-form-urlencoded"
> to the Content-Type header.
> this appears to be a bug, correct?
>
> thanks,
> augusto.
>
> ps: i'm running openlaszlo 4.9 in proxied mode.
