[tex-k] Incorrect sizes in GFtoDVI output

[Igor Liferenko]

Hi Karl,

> 1) Oddly, I get only one warning from dvitype with io.mf in smoke mode,
> which is what is being discussed on p.32:

The example which I was talking about is not io.mf. It begins on p.31:

    drawdot (35,70);
    drawdot (65,70);
    draw (20,40)..(50,25)..(80,40); shipit; end.

> 2) However, a "naive user" won't even know about the existence of
> dvitype, let alone notice warnings from the result of running it. When I
> view io.dvi (or io.ps or io.pdf), it seems fine to me. That's all that
> page 32 is concerned with.

I'm trying to say that the reader may have an impression that GFtoDVI may
be used with any `gf'-file whatsoever. This will unavoidably lead to confusion,
as the reader will proceed reading the book and do experiments of his
own design.

Probably on p.31 it should be said that `\mode:=proof;' must be typed
after `\relax'.

> 3) If these warnings had been reported in 1984, maybe Knuth would have
> been interested in changing gftodvi's dvi output to avoid them. However,
> I am completely doubtful that he would care to do so in 2029. -k

Alternatively, GFtoDVI in TeXLive may detect the mode and give a warning that
"unintended" mode is being used.

Best,
Igor