I've been trying to use an XMLHttpRequest to get a binary file created from a C struct.
If I open the file in a hex editor, all characters are encoded as expected.
If I add the file to a static HTML page using <input type='file' /> and then use FileReader.readAsArrayBuffer(), I can see that the ArrayBuffer holds the same expected values.
But, when I try to send it over an XMLHttpRequest, and put its responseText into an ArrayBuffer, there are several characters that are replaced with the Unicode "replacement character".
If I use FileReader.readAsText() with the <input type='file' /> test, I get the same character dropout that happens with XMLHttpRequest.responseText.
I have control over the file server, and I've tried adding a Content-Type header to the response (application/octet-stream?), but it doesn't seem to make a difference.
What am I missing here to ensure that the XMLHttpRequest.responseText has the correct data so the ArrayBuffer is created with the same values as the FileReader way?
Of course, after hours of working on this, right after I rubber duck on SO I find the solution.
XMLHttpRequest has a response property - I shouldn't be using responseText. Also, I can set the responseType property before I send the request to ensure it's returned in an ArrayBuffer.