Uploading PDF using POST and unwanted utf-8 encoding

Dear Community

My name is Sebastian and this forum has been a valuable read-only resource for me in the past months or so. I am developing folder level javascript tools for Acrobat and - I hope that's OK - XChange Editor which is almost entirely conformant with the API.

Now I am having a problem that is driving me nuts and it is this:

I am trying to oupload an entire PDF along with a couple of parameters to a web service under my control.

The server side is functional and is flawlessly handling file upload forms POSTed from browsers.

The server expects Content-Type: 'multipart/form-data' and requires 'basic' authentication.

The Acrobat API offers several options, none of which works for me out of the box:

doc.submitForm({cUrl:..., cSubmitAs: 'PDF'})

- doesn't allow authentication

Net.HTTP.request ({cUrl:..., cVerb: 'POST', oAuthenticate: {...}})

- doesn't *cooperate* with my server when using oAuthenticate

- requires additional custom headers

So I ended up implementing the request from scratch and successfully tested it against my own server and two online test services. (https://ptsv2.com/, https://httpbin.org/post).

The only problem is:

The message body is utf-8 encoded, because I am constructing it in a string object.

That's no problem for most of the message body but it mangles the pdf's binary data, which becomes a utf-8 encoded version of the ASCII-representation of the pdf's binary data.

I will attach the code below but here I am only quoting the message body:

--224061383921529899882446413168
Content-Disposition:form-data; name="Project"

Project-1
--224061383921529899882446413168
Content-Disposition:form-data; name="FILE1"; filename="x.pdf"
Content-Type: application/pdf

%PDF-1.7
%����
[...]
--224061383921529899882446413168--

The characters below %PDF-1.7 are utf-8 encoded as I was able verify by inspecting the server dump of "FILE1". In fact I can *read* the correct ASCII characters in Notepad++ when I view the file in utf-8 mode. However when I switch to Notepad++'s ANSI-mode I see the actual ASCII characters (or bytes if you will) and they differ from the original PDF.

My question probably is:

How can I provide an unencoded binary stream to Net.HTTP.request?

The way I am doing it now requires a string object which interprets the binary data as ASCII charcters and encodes them as utf-8.

Here comes the relevant code

// PDF stream fetched from a local file
var stmPDF = util.readFileIntoStream("/C/temp/eBoxTest/Att.pdf");

// Request headers
var boundary = "224061383921529899882446413168";
var aHeaders = [
    {name:"Content-Type", value: "multipart/form-data;boundary=" +boundary+ "\r\n"},
    {name:"Accept", value: "text/javascript,text/html,application/xml,text/xml\r\n"},
    {name:"Accept-Encoding",value: "gzip,deflate\r\n"},
    {name:"Authorization", value: "Basic " + sAuth64 + "=="}
];

// Request body, including string-converted PDF stream
var strContent = "--" +boundary+ "\r\n"
    + "Content-Disposition:form-data; name=\"Project\""
    + "\r\n\r\n" + "Project-1" + "\r\n"
    + "--" +boundary+ "\r\n"
    + "Content-Disposition:form-data; name=\"FILE1\"; filename=\"x.pdf\"" + "\r\n"
    + "Content-Type: application/pdf" + "\r\n\r\n"
    + util.stringFromStream(stmPDF) + "\r\n"+"--" +boundary+ "--";

// The POST request
Net.HTTP.request ({
    cVerb: 'POST',
    cURL: cUrl,
    aHeaders: aHeaders,
    oRequest: util.streamFromString(strContent),
    oHandler: {
        response: function (response, uri, err) {
            if (err != undefined)
                console.println("== FAIL ==");
            else {
                console.println("== SUCC ==");
                console.println(SOAP.stringFromStream(response));
            }
        }
    }
});

Thanks a lot!