[Tex/LaTex] Removing spurious pagebreak in index

indexingpage-breaking

EDIT: To clarify my question: I am not looking for an index that will start without a page break; I, rather, want an index where I manually would take care of the page break.

Say you've inserted a ton of \clearpages, to make sure that the "next" page starts on pg 37. Now you want to insert the index "here", at page 37 – so you type \printindex[names] there – and, surprise – the index now starts at page 38, because the index decided to enter one more \clearpage :/ And that means now you have to go back, and re-enter ammount of \clearpages, so they match to page N-1; so that along with the extra pagebreak, the index will start where at page N. This I'd see as an unnecessary complication.

Second, let's say you don't want to use the default section title of \printindex; you'd rather use \section*. The easiest thing to think of (without knowing any better), is to suppress the default index heading (in this case, since the example is for \printindex from index package, that would be in \newindex command): 

\newindex{names}{ndx}{nnd}{}  % instead of: {names}{ndx}{nnd}{Name Index}

… (which does indeed suppress the default index heading) – and then, instead of that heading, one would type \section*{Name Index} followed by a \printindex[names]. Obviously, if \printindex inserts its own pagebreak, then this approach won't work, since the section title and the first page of the index will be on different pages – which is precisely what the MWE in this question demonstrates….

Damn it, I'm really starting to hate Latex :/ naah… :)

Take a look at the following MWE:

\documentclass[a4paper,twoside,11pt]{book}


% must load index before hyperref, else no hyperlinks!
\usepackage{index}
\usepackage{hyperref}
\usepackage{trace}

% declare new index files
\newindex{names}{ndx}{nnd}{Name Index}
\newindex{titles}{tdx}{tnd}{Title Index}

\usepackage{forloop}
\usepackage{trace}

% expandable modulo operation
\def\truncdiv#1#2{((#1-(#2-1)/2)/#2)}
\def\moduloop#1#2{(#1-\truncdiv{#1}{#2}*#2)}
\def\modulo#1#2{\number\numexpr\moduloop{#1}{#2}\relax}


% needed to generate index files
\newcounter{ct}
\def\modme{10}

% generate names (.nnd) index file
\newwrite\tempnnd
\immediate\openout\tempnnd=\jobname.nnd
\immediate\write\tempnnd{\protect\noexpand\begin{theindex}}

\forLoop[1]{0}{255}{ct}{
    \immediate\write\tempnnd{%
      \protect\noexpand\item Surname\thect, Name\thect, %
      %\protect\noexpand\hyperpage{\thect}
      % hyperlink all to page 1, 
      % to avoid Latex complaining:
      \protect\noexpand\hyperpage{1} %
    }
    \ifnum\modulo\thect\modme=0
      \immediate\write\tempnnd{}
      \immediate\write\tempnnd{\protect\noexpand\indexspace}
      \immediate\write\tempnnd{}
    \else
    \fi
}
\immediate\write\tempnnd{\protect\noexpand\end{theindex}}
\immediate\closeout\tempnnd

% generate titles (.tnd) index file
\newwrite\temptnd
\immediate\openout\temptnd=\jobname.tnd
\immediate\write\temptnd{\protect\noexpand\begin{theindex}}

\forLoop{0}{255}{ct}{
    \immediate\write\temptnd{%
      \protect\noexpand\item \noexpand\emph {Title of Work\thect}, %
      %\protect\noexpand\hyperpage{\thect}
      % hyperlink all to page 1, 
      % to avoid Latex complaining:
      \protect\noexpand\hyperpage{1} %
    }
    \ifnum\modulo\thect\modme=0
      \immediate\write\temptnd{}
      \immediate\write\temptnd{\protect\noexpand\indexspace}
      \immediate\write\temptnd{}
    \else
    \fi
}
\immediate\write\temptnd{\protect\noexpand\end{theindex}}
\immediate\closeout\temptnd


\begin{document}

\clearpage

\section*{Test 1}
\addcontentsline{toc}{section}{Test 1}

\phantomsection % NOTE NO PAGEBREAK PRESENT AROUND HERE!
\addcontentsline{toc}{chapter}{Name Index} \traceon
\printindex[names] \traceoff

\phantomsection
\addcontentsline{toc}{chapter}{Title Index} % master TOC and PDF bookmark
\printindex[titles] % the title index

\end{document}

The first two pages of the output are:

test.png

Generated with: 

montage -bordercolor LimeGreen -border 1 -density 100 test.pdf[0] test.pdf[1] -geometry +2+2 -tile 2x1 test.png

My question is:

Can someone explain to me why there a pagebreak before the Name Index section – and how do I remove it?

 

As seen in the MWE, I even tried tracing for it: 

$ grep -C10 'Surname0' b.log
{into \Hy@temp=macro:->}
{\ifx}
{true}
{\else}
{blank space  }
{\relax}

\item ->\par \hangindent 40\p@                 <= \item
{\par}                                         <= \par (?)
@firstpass
[]\OT1/cmr/m/n/10.95 Surname0, Name0, []0[]    <= typeset Surname0
@\par via @@0 b=0 p=-10000 d=100
@@1: line 1.2- t=100 -> @@0

{changing \hangindent=40.0pt}
{into \hangindent=0.0pt}
%% goal height=476.5225, max depth=5.5         <= pagebreak?!
% t=0.0 g=476.5225 b=10000 p=0 c=100000#       <=    WHERE FROM??
{vertical mode: \hangindent}
{changing \hangindent=0.0pt}
{into \hangindent=40.0pt}

… but I simply don't understand which command would cause a new page after the first name is typeset ?! And if the pagebreak happened after Surname0 got typeset, why isn't Surname0 on first page ??!

Best Answer

The problem seems quite abstract, because it's quite rare to have an index start without a page break. However, here is a possibility. Since you don't want page breaks, it's necessary to avoid any automatism in index generation.

\documentclass[a4paper,twoside,11pt]{book}

\usepackage[noautomatic,nonewpage]{imakeidx}
\usepackage{hyperref}

% declare new index files
\makeindex[name=names,title=Name Index,intoc]
\makeindex[name=titles,title=Title Index,intoc]
\indexsetup{level=\section*} % no page break with sections

\begin{document}

\tableofcontents

% generate fake index entries
\count255=0
\loop\ifnum\count255<200
  \index[names]{Surname\number\count255, Name\number\count255}
  \index[titles]{Title of Work\number\count255}
  \advance\count255 1
\repeat

\clearpage

\phantomsection
\section*{Test 1}
\addcontentsline{toc}{section}{Test 1}

\printindex[names]
\printindex[titles]

\end{document}
Related Question