by pico303:

In Go 1.3.3, the XML parser for Go is locked into UTF-8 encodings.  In
encoding/xml/xml.go (around line 576), there's the line:

    enc := procInstEncoding(string(data))
    if enc != "" && enc != "utf-8" && enc != "UTF-8" {

For documents with:

    <?xml version="1.0" encoding="ISO-8859-1"?>

you get this error message:

    Invalid body content: xml: encoding "ISO-8859-1" declared but Decoder.CharsetReader is nil

You can override the reader to support alternative encodings, but this means pre-parse
the XML []byte yourself for the proper encoding, setup the reader, then parse the XML. 

Could the package be adapted somehow so you could provide alternate readers ahead of
time, based on the encoding value?  Something like this (pseudocode):

    func init() {
        xml.AddCharsetReader("iso-8859-1", ISO8859Reader)
    }

    func Parse(doc []byte) (SomeStruct, error) {
        var myobj SomeStruct
        if err := xml.Unmarshal(doc, &myobj); err != nil {
            return nil, err
        }
        return myobj, nil
    }