Re: What about this grammar?

> An explicit statement on character zero (NUL) may be helpful, if there
> isn't one already (i didn't find one). It should be permissable for an
> implementation nor to allow NUL anywhere in its input.

It’s straightforward to write an iXML grammar that will parse a file
that contains NULL characters.

Given out.ixml:

S: -#0, 'a', #a?.

And a file, out.bin, that contains a null followed by an ‘a’ followed
perhaps by a newline:

$ coffeepot -g:/tmp/out.ixml -i:/tmp/out.bin
<S>a</S>

It parses just fine.

Obviously, an implementation can observe that there are files it cannot
read, but I don’t think we should enshrine those limits in the spec.

> PS: i had no idea there was an audio spec called iXML :)

Nor I, until yesterday when I searched for “ixml” instead of “invisible
xml” looking for something.

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica