pdftex primitives

[Chris Rowley]

This discussion of \pdfimage highlights why I asked for a fuller
written description of the intended semantics of the pdftex
primitives.

I have several questions similar to the one about whether multiple
depths are supported; not all of these are about scanning so cannot so
easily be probed empirically.

Although I can get people to attempt to test these empirically, or to
look at the change file so as to work out what may be happening, this
is very inefficient; moreover, it only helps to establish the actual
semantics this week.  It is no help in trying to evaluate whether the
code actually does what Thanh or anyone else intended it to do; and it
is little help in evaluating whether the high-level design spec was
optimal in the first place.

So please, please, please can we have something to read along these lines,
before the merger with etex.


chris