This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
In discussion of test case mdocs09 at today's telcon (see http://www.w3.org/Member/bugzilla/show_bug.cgi?id=612) the WG decided that an implementation of unparsed-text() should be allowed to apply implementation-dependent heuristics to determine the encoding of the external file between steps 3 and 4 (this might include reference to a BOM, or for example auto-detection of a file as HTML with a META tag giving the encoding). The spec will be changed accordingly.
In this connection, it was also pointed out that the set of encodings accepted by the processor is an implementation-defined feature, but is not listed as such in the (non-normative) Appendix F.
Spec now updated.