[tex-eplain] Another bug in index proofing macros

[geolsoft at mail.ru]

On Tue, Aug 31, 2004 at 04:49:38PM -0400, Dorai Sitaram wrote:
> Hmm, I ran the experiment, and the addition of the \nointerlineskip
> doesn't seem to remove the vertical shift...


Please find below input file on which I ran my experiments
(excerpt from Elements of Calculus and Analytic Geometry by
George~B.\ Thomas,~Jr.)

Run it with the command:

  tex -interact=nonstopmode testidx && makeindex -L testidx.idx && tex -interact=nonstopmode testidx && xdvi testidx

Note that I include local copy of eplain.tex (`\include
./eplain.tex') in which I make my modifications.

Try to run it with the original `\indexproofunbox', and
compare, e.g., positions of page numbers on the first page
(which contains an indexing command) and on the second page
(which has none).  You'll notice the vertical shift.  Now if
you change `\indexproofunbox' as I described earlier, no
shift is present.

Many thanks.


Best regards,
Oleg Katsitadze


--------------------start of testidx.tex--------------------
\input ./eplain.tex

\hsize=2.7in
\vsize=4in
\indexproofingtrue
%
% \allowhyphens
%
% This is from The TeXbook, p.~395.  It allows following (preceding) word to be
% hyphenated
%
\newdimen\zzz
\def\allowhyphens{\nobreak\hskip\zzz}%
\hookaction{afterindexterm}{\allowhyphens}%

Let us note that the division in Eq.~(1) can only be {\it indicated\/} when we
are talking about a general function~$f(x)$, but for any specific equation such
as $f(x)=x^3-3x+3$ in Eq.~(2) of Article~2.6, this division of~$\Delta y$
by~$\Delta x$ is actually to be carried out [as we did in going from Eq.~(4) to
Eq.~(5)] {\it before\/} we do the next operation.

Having performed the division indicated in Eq.~(1), we now investigate what
happens if we hold~$x_1$ fixed and take~$\Delta x$ to be smaller and smaller,
approaching zero.  If $m_{\sec}$ approaches a constant value, we call this value
its {\it \sidx{Limit}limit\/} and define this to be the slope~$m_{\tan}$ of the
tangent to the curve at~$P$.  The mathematical symbols which summarize this
discussion are
$$\eqalignno{m_{\tan}&=\lim_{Q\to P}m_{\sec}=\lim_{\Delta x\to0}{\Delta y\over\Delta x}\cr
             &=\lim_{\Delta x\to0}{f(x_1+\Delta x)-f(x_1)\over\Delta x}.&(2)}$$
The symbol ``$\lim$'' with ``$\Delta x\to0$'' written beneath it is read ``the
limit, as~$\Delta x$ approaches zero, of \dots''

Chapter~3 will treat the theory of limits in greater depth.  For the remainder
of this chapter, the reader is encouraged to use the following intuitive idea:
``$L$ is the limit of~$m(h)$ as~$h$ approaches zero'' means that when~$h$ is
very close to zero, but not equal to zero, then~$m(h)$ is very close to~$L$, and
possibly equal to~$L$.

\bye
--------------------end of testidx.tex--------------------