fewerfloatpages.dtx

% \iffalse meta-comment
%
%% File: fewerfloatpages.dtx (C) Copyright 2019-2021 Frank Mittelbach
%
% It may be distributed and/or modified under the conditions of the
% LaTeX Project Public License (LPPL), either version 1.3c of this
% license or (at your option) any later version.  The latest version
% of this license is in the file
%
%    https://www.latex-project.org/lppl.txt
%
% The development version of the bundle can be found below
%
%    https://github.com/FrankMittelbach/fmitex-fewerfloatpages/
%
% for those people who are interested or want to report an issue.
%

\def\fewerfloatpagesdate   {2021/03/02}
\def\fewerfloatpagesversion{v1.0b}


%<*driver>
\documentclass
    [final]
    {l3doc-TUB}
\hfuzz=1.6pt % not even one tt character, don't worry.

\makeatletter
  \@mparswitchfalse
\makeatother
    
\setcounter{page}{1}       

\newcommand\ctr[1]{\texttt{\upshape#1}}
\newcommand\option[1]{\texttt{\upshape#1}}
\newcommand\mnote[1]{\marginpar{\raggedleft\em #1}}

\usepackage{csquotes}
\EnableCrossrefs
\CodelineIndex


\begin{document}
 \DocInput{fewerfloatpages.dtx}
\addtolength\signaturewidth{42pt}
\makesignature
\end{document}
%</driver>
%
% \fi
%
%
% \title{The \texttt{fewerfloatpages} package\thanks{The current
%   package version is \texttt{\fewerfloatpagesversion} dated
%   \fewerfloatpagesdate.}}
% \author{Frank Mittelbach}
% \address{Mainz, Germany}
% \netaddress{https://www.latex-project.org}
% \personalURL{https://ctan.org/pkg/fewerfloatpages}
%
% \maketitle
%
%
% \begin{abstract}
%
%   \LaTeX{}'s float algorithm has the tendency to produce fairly
%   empty float pages, i.e., pages containing only floats but with a
%   lot of free space remaining that could easily be filled with nearby
%   text. There are good reasons for this behavior; nevertheless, the
%   results look unappealing and in many cases documents are
%   unnecessarily enlarged.
%
%   The \pkg{fewerfloatpages} package provides an extended algorithm
%   that improves on this behavior without the need for manual intervention by
%   the user.
%
% \end{abstract}
%
% \microtypesetup{protrusion=false}
% \tableofcontents
% \microtypesetup{protrusion=true}
%
% \DoNotIndex{\@fpmin}
% \DoNotIndex{\@testwrongwidth}
% \DoNotIndex{\@cons}
% \DoNotIndex{\@currbox}
% \DoNotIndex{\@currtype}
% \DoNotIndex{\@elt}
% \DoNotIndex{\@empty}
% \DoNotIndex{\@gobble}
% \DoNotIndex{\@ne}
% \DoNotIndex{\@next}
% \DoNotIndex{\@spaces}
% \DoNotIndex{\@tempcnta}
% \DoNotIndex{\@tempcntb}
% \DoNotIndex{\@tempdima}
% \DoNotIndex{\@tempdimb}
% \DoNotIndex{\@testtrue}
% \DoNotIndex{\@undefined}
% \DoNotIndex{\@vtryfc}
% \DoNotIndex{\@xxxii}
% \DoNotIndex{\@ztryfc}
% \DoNotIndex{\fl@trace}
% \DoNotIndex{\@bitor}
% \DoNotIndex{\NeedsTeXFormat}
% \DoNotIndex{\TeX}
% \DoNotIndex{\advance}
% \DoNotIndex{\begingroup}
% \DoNotIndex{\bx@B}
% \DoNotIndex{\bx@D}
% \DoNotIndex{\bx@I}
% \DoNotIndex{\count}
% \DoNotIndex{\def}
% \DoNotIndex{\divide}
% \DoNotIndex{\else}
% \DoNotIndex{\endgroup}
% \DoNotIndex{\fi}
% \DoNotIndex{\gdef}
% \DoNotIndex{\global}
% \DoNotIndex{\ht}
% \DoNotIndex{\if@test}
% \DoNotIndex{\if@twocolumn}
% \DoNotIndex{\ifdim}
% \DoNotIndex{\ifnum}
% \DoNotIndex{\ifodd}
% \DoNotIndex{\ifx}
% \DoNotIndex{\let}
% \DoNotIndex{\maxdimen}
% \DoNotIndex{\multiply}
% \DoNotIndex{\newcommand}
% \DoNotIndex{\newcount}
% \DoNotIndex{\newcounter}
% \DoNotIndex{\providecommand}
% \DoNotIndex{\relax}
% \DoNotIndex{\renewcommand}
% \DoNotIndex{\reserved@a}
% \DoNotIndex{\sixt@@n}
% \DoNotIndex{\space}
% \DoNotIndex{\string}
% \DoNotIndex{\the}
% \DoNotIndex{\tw@}
% \DoNotIndex{\typeout}
% \DoNotIndex{\xdef}
% \DoNotIndex{\z@}
%
%
% \section{Introduction}

%    We start by giving a quick overview of \LaTeX{}'s float algorithm
%    and the problems that result from the approach used.
%
%    We then look in some detail into possible alterations and
%    improvements to that algorithm and discuss possible issues that
%    need to be resolved. In this section we also describe all
%    configuration possibilities of the extended algorithm.
%
%    The final section then documents the code changes that are
%    necessary to \LaTeX{} kernel macros to implement the extension.
%
% \subsection{A quick overview of \LaTeX's float algorithm}
%
%    \LaTeX{}'s output routine uses a greedy algorithm to place floats
%    near to their call-outs in the source document. The decision of how
%    to place a float is made when the float is first encountered. If
%    possible it is placed onto the current page, either in mid-text,
%    on top or into the bottom area, depending on what is allowed for
%    the float and how many floats are already placed into those
%    areas.
%
%    If the float can't be placed immediately, it goes into a defer
%    list, and in order to not accumulate too many unplaced floats
%    \LaTeX{} tries to empty that list whenever there is a chance. This
%    chance comes after the next page break: \LaTeX{} then starts a
%    special ``float page'' algorithm in which it examines the defer
%    list and from it forms float pages (i.e., pages that contain only
%    floats). If necessary, it generates several float pages and only
%    stops if there are no floats waiting to be placed, or there are
%    too few floats to form a float page, or there are only floats left
%    that are for one or another reason not allowed to be placed in
%    this way.
% 
%    Finally \LaTeX{} looks at the remaining floats and tries to place
%    as many of them as possible into the top and bottom area of the
%    next page. Then it continues to process further text to fill the
%    text part of that page. Details on the exact behavior of the
%    algorithm are discussed in \cite{fmi:floatplacement}.
%
%
% \subsection{The typical float page and its problems}
%
%   \LaTeX{} considers a float page to be successfully built if its
%   floats take up more than \cs{floatpagefraction} of the whole
%   page. By default this parameter is set to \texttt{.5} which means
%   that such float pages may end up being half empty.
%
%   Many users think that this is not a good value and try to improve on
%   it by enforcing a higher percentage (such as 80\%) only to find
%   that this prevents \LaTeX{} in many cases from successfully
%   generating any float page, with the effect that all floats are
%   suddenly piling up at the end of the document.
%
%   Why is this the case? In a nutshell, because a higher percentage
%   makes it much more likely that a float can't be
%   placed, because it is not big enough to be used on its
%   own and no other nearby floats can be combined with
%   it, because their combination violates some other restriction,
%   e.g., together they are bigger than a page, not all of them are
%   allowed to go on float pages, etc. The moment that happens this
%   float prevents the placement of all later floats of the same
%   class too (i.e., all figures) and disaster is ensured. In most
%   cases these floats will then never get placed, because they need a
%   float of the right size from a different class to appear, which
%   may in theory happen but is, unfortunately, unlikely.
%
%    Thus,\mnote{Tinkering with the parameter settings will usually
%    produce unwanted effects} while tempting, tinkering with this
%    parameter by making it larger is usually not a good idea, unless
%    you are prepared to place most if not all of your floats
%    manually, by overwriting the placement algorithm on the level of
%    individual floats (e.g., using \texttt{!}\@ syntax and/or
%    shifting its position in the source document).
% 
%    Why does the current algorithm have these problems? To some extent,
%    because it offers only global parameters that need to fit
%    different scenarios and thus settings that are suitable when many
%    floats need to be placed result in sub-optimal paginations in
%    document parts that contain only a few floats, and vice versa.
%    To overcome this problem, either one can try to develop
%    algorithms with many more configurable parameters that act
%    differently in different scenarios or one can let the algorithm
%    follow a main strategy, configurable with only a few parameters
%    (like today), but monitor the process and make more local
%    adjustments and corrections depending on the actual outcome of
%    that base strategy and additional knowledge of the actual
%    situation in a given document part. This is the
%    approach taken by the extension implemented in this package.
%
% \section{Improvements to the float page algorithm}
%
%    A simple way to improve on the existing algorithm, without
%    compromising its main goal of placing the floats as fast as
%    possible and as close as possible to their call-outs, is the
%    following: as long as there are many floats waiting to be placed,
%    generate float pages as necessary to get them placed (using the
%    current algorithm and its parameters).

%     Once we are unable to build further float pages, do some level
%    of backtracking by checking if we have actually succeeded in
%    placing all floats.  If there are still floats waiting to be
%    placed then assume that what has been done so far is the best
%    possible way to place as many floats as possible (which it
%    probably is). However, if we have been able to place all floats
%    onto float pages then check if the last float page is
%    sufficiently full; if not, undo that float page and instead
%    redistribute its floats into the top and bottom area of the next
%    upcoming page. This way the floats will be combined with further
%    text and we avoid a possible half-empty float page.
% 
%    This approach will not resolve all the problematic scenarios
%    where we find that \LaTeX{} has decided to favor fairly empty
%    float pages over some tighter type of placement. It will,
%    however, help to improve typical cases that do not involve too
%    many floats.  For\mnote{A typical case where we don't really want
%    \LaTeX{} to make a float page} a example, if a single (larger)
%    float appears near the end of a page, then using the standard
%    algorithm it can't be immediately placed (because there isn't
%    enough free space on the current page). It is therefore moved to
%    the defer list and at the page break it is then placed onto a
%    float page (possibly by itself, if it is large enough to allow
%    for that) even though it could perfectly well go into the top or
%    bottom area of the next page and thus be combined with textual
%    material on that page.
% 
%    With the new algorithm this float page is reexamined and unless
%    it is pretty much filled up already, it is unraveled and its
%    floats are redistributed into the top and bottom areas of the
%    next page.  If, however, we have many floats waiting on the defer
%    list, the normal float page algorithm will first place as many of
%    them as possible into float pages and only the last of these
%    pages will be subject to a closer inspection and a possible
%    unraveling.
% 
%    An extension of this idea (and the one that we actually
%    implement) is to monitor the whole float page generation process
%    and instead of just considering the last float page in the sequence
%    for unraveling, we look at each prospective float page in turn and
%    based on the current situation (e.g., number of floats still
%    being unplaced, free space on the float page, etc.)\@ decide
%    whether this float page should be produced or whether we should
%    stop making float pages and instead place the pending floats into
%    top and bottom areas of the upcoming page.
% 
% 
%  \subsection{Details on the extended algorithm}
% 
%    The main idea of the extended algorithm is to avoid unnecessary cases of
%    float pages especially if those float pages are fairly
%    empty.\mnote{Don't unravel a float page if there are too many
%     floats on the defer list}
%    Natural candidates are single float pages, but even in cases where
%    the current \LaTeX{} algorithm  produces several sequential float
%    pages the extended algorithm may decide to replace them by normal
%    pages under certain conditions.
%    However, the main goal
%    is and should remain to place as many floats as soon as possible
%    and so generating float pages when many floats are waiting is
%    usually essential.
%
% \begin{variable}{floatpagedeferlimit}
%   \begin{syntax}
%   \cs{setcounter}\{floatpagedeferlimit\}\Arg{number}
%   \end{syntax}
%    Whether or not unraveling for a float page is considered at all
%    is guided by the counter \ctr{floatpagedeferlimit}. As long as
%    there are more floats waiting on the defer list than this number,
%    float pages are not considered for unraveling.  The default is
%    \texttt{3} which corresponds to the default value for
%    \ctr{totalnumber}, i.e., with that setting the unraveling of a
%    floating page has a fighting chance to place all floats into the
%    top and bottom areas on the current page.
%    It would also resolve cases for up to three floats, each larger than
%    \cs{floatpagefraction}, where the standard \LaTeX{} algorithm would
%    produce three individual float pages.\strut
% \end{variable}
%
%    If you set the counter to \texttt{1} then only the last float
%    page in a sequence is considered, and only if it contains only a
%    single float and if there are no other floats that are still
%    waiting to be placed.  If you set it to \texttt{0}, then the
%    extension is disabled, because float pages are produced only if
%    there was at least one float on the defer list.
% 
%    Even\mnote{Don't unravel if the float page contains many floats}
%    if we set \ctr{floatpagedeferlimit} to a fairly high value, we
%    may not want to unravel float pages that contain many floats. To
%    support this case there is a second counter that guides the
%    algorithm in this respect.
%
% \begin{variable}{floatpagekeeplimit}
%   \begin{syntax}
%   \cs{setcounter}\{floatpagekeeplimit\}\Arg{number}
%   \end{syntax}
%   Whenever the float page contains at least
%   \ctr{floatpagekeeplimit} floats it will not be unraveled. The
%   default is also \texttt{3} so that float pages with three or more
%   floats are not touched. Obviously the counter can have any effect
%   only if it has a value less than or equal to
%   \ctr{floatpagedeferlimit} because this is tested first.\strut
% \end{variable}
%
%
%    There are, however, a number of other situations in which we
%    shouldn't unravel a float page even if the above checks for the
%    size of the defer list were passed successfully.\mnote{Don't
%    unravel if the float page contains at least one \texttt{\upshape
%    [p]} float} The most important one is the case when the float
%    page contains at least one float that is allowed \emph{only} on
%    float pages (i.e., has a \texttt{[p]} argument).  Such a float
%    would not be placeable in a top/bottom area on any page and thus
%    would be repeatedly sent back to the defer list (possibly forever).
% 
%    The other case where unraveling would normally be
%    counterproductive is when the particular float page is nearly or
%    completely filled up with floats.\mnote{Don't unravel if the
%    float page is nearly filled} If we unravel it, then it is
%    certain that we can place only some of the floats into the top or
%    bottom area of the next page, while some would end up on the defer
%    list. That in turn means that these deferred floats float even
%    further away from their call-out positions than need be.
%
% \begin{variable}{\floatpagekeepfraction}
% \begin{syntax}
%   \cs{renewcommand}\cs{floatpagekeepfraction}\Arg{decimal}
% \end{syntax}
%   So what is a good way to determine if a float page is ``full
%   enough''?  A possible answer is that if the remaining free space
%   on that page is less than \cs{textfraction} we consider it full
%   enough to stay. \cs{textfraction} defines the minimum amount of
%   space that has to be occupied by text on a normal page, thus if
%   all floats together need so much space that this amount of text
%   could not fit, then trying to place all floats onto a
%   normal page can't succeed and some of them would get deferred for
%   sure.  To allow for further flexibility the algorithm uses
%   the variable \cs{floatpagekeepfraction} (defaulting to
%   \cs{textfraction}) so if desired a lower (or even a higher)
%   boundary can be set.\strut
% \end{variable}
%
%   The above parameters give some reasonable configuration
%   possibilities to guide the algorithm as to when and when not to unravel
%   a possible float page and instead produce further normal pages.
%   It should be noted, however, that except for the case of setting
%   \ctr{floatpagedeferlimit} to \texttt{1}, there is always a chance
%   that floats drift further away from their call-outs, because they
%   may not be immediately placeable due to other parameter settings
%   of the float algorithm.  For example, the counter
%   \texttt{topnumber} (default value 2) limits the number of floats
%   that can be placed in the top area on a normal page and if more
%   remain after unraveling only two can immediately go in this area.
%
%
% \subsection{Possible pitfalls and how to avoid them}
%
%    The algorithm detects if a float is allowed only on float pages
%    (i.e., is given in the source as \texttt{[p]}) and it will ensure
%    that float pages containing such floats are not unraveled.
% 
%    However, if you have a float with the default specifier
%    \texttt{[tbp]} whose size is larger than the allowed size of the
%    top or bottom area (e.g., larger than \cs{topfraction}
%    \texttimes{} \cs{textheight}), then this effectively means it can
%    only be placed on a float page.
% 
%    However, according to the specifier the float is allowed to go
%    into the top or bottom area, so the algorithm, as explained so
%    far, would be allowed to unravel and when that float later is
%    considered for top or bottom placement it will get again deferred
%    and thus move from one page to the next, most likely messing up
%    the whole float placement.
% 
%    There\mnote{\option{checktb} (option)} are two possible ways to
%    improve the algorithm to avoid this disaster. One way
%    would be to check the float size when it is initially encountered
%    and remove any specifier that is technically not possible because
%    of the parameter settings and the float size. A possible
%    disadvantage is that this determination will be done once and any
%    later (temporary) change to the float parameters will have no
%    effect.  This is currently the package default. It can be
%    explicitly selected by specifying the option \option{checktb}. In
%    this case you might see warnings like
%\begin{verbatim}
% LaTeX Warning: Float too large for top area: t changed to p on line ...
%\end{verbatim}
%
%    Another\mnote{\option{addbang} (option)} possibility is that we automatically add a
%    \texttt{!}\@  specifier to all floats during unraveling, i.e., when
%    we send them back for reevaluation. This way such floats become
%    placeable into top and bottom areas regardless of their size. This
%    may result in fewer pages at the cost of violating the area size
%    restrictions once in a while. It is specified with the option \option{addbang}.
% 
%    If\mnote{\option{nocheck} (option)} you prefer no automatic
%    adjustment of the specifiers, add the option \option{nocheck}.
%    In this case you might find that floats of certain sizes are
%    unplaceable and thus get delayed to the end of the document.  If
%    that happens, the remedy is either to explicitly specify
%    \texttt{[p]} or \texttt{[hp]} for such a float (to ensure that
%    they aren't subject to unraveling) or to manually add an
%    exclamation specifier, e.g., \texttt{[!tp]} so that \LaTeX{}
%    doesn't use the size restrictions in its algorithm.
%
%
%
% \subsection{Tracing the algorithm}
%
%   The\mnote{\option{trace} (option)} package offers the option
%   \option{trace}, which if used, will result in messages such as
%\begin{verbatim}
%  [1]
%  fewerfloatpages: PAGE: trying to make a float page
%  fewerfloatpages: ----- \@deferlist: \bx@B \bx@D 
%  fewerfloatpages:  starting with \bx@B
%  fewerfloatpages:  --> success:  \bx@B \bx@D 
%  fewerfloatpages: ----- current float page unraveled
%                  (free space 192.50336pt > 109.99832pt)
%  [2]
%\end{verbatim}
%    which means that the algorithm is trying to make a float page
%    from the defer list which at that point contained two floats (the
%    float boxes \cs[no-index]{bx@B} and \cs[no-index]{bx@D}), that it
%    was able to produce a float page containing just
%    \cs[no-index]{bx@B} and \cs[no-index]{bx@D}, and that the extended
%    algorithm then decided to unravel that float page, because it has
%    an unused space of \texttt{192.5pt}, i.e., roughly 16 text lines. With the current
%    \cs{floatpagekeepfraction} that is too much empty space on the
%    page.
%
%   Or it might say
%\begin{verbatim}
%  fewerfloatpages: PAGE: trying to make a float page
%  fewerfloatpages: ----- \@deferlist: \bx@D \bx@F \bx@G \bx@H \bx@I 
%  fewerfloatpages:  starting with \bx@D
%  fewerfloatpages:  --> success: \bx@D \bx@F 
%  fewerfloatpages: ----- too many deferred floats for unraveling (5 > 3)
%  [3]
%\end{verbatim}
%   which means that the algorithm made a float page out of the first
%   two floats from the defer list (i.e., 3 remained).
%   That
%   page was kept regardless of the amount of free space it contained
%   because we have a total of 5 floats on the defer list and the
%   counter \ctr{floatpagedeferlimit} has its default value of \texttt{3}.
%
%   The above tracing messages are both from the same test
%   document. What they also (implicitly)  show is that the unraveling
%   that happened after page~1 resulted in only one float (\cs[no-index]{bx@B})
%   being placed on page~2, because we see the second one (\cs[no-index]{bx@D})
%   reappearing in the defer list after page~2 got finished. In other
%   words it was moved one page further away from its call-out: the
%   price for getting a nicely filled page~2 instead of a fairly empty
%   float page with roughly 200 points left empty.
%   The final part of that test document then exhibits another
%   type of message:
%\begin{verbatim}
%  fewerfloatpages: PAGE: trying to make a float page
%  fewerfloatpages: ----- \@deferlist: \bx@G \bx@H \bx@I 
%  fewerfloatpages:  starting with \bx@G
%  fewerfloatpages:  --> success: \bx@G \bx@H \bx@I 
%  fewerfloatpages: ----- all floats placed on float page(s)
%  fewerfloatpages: ----- current float page kept, full enough
%                        (free space 38.99496pt < 109.99832pt)
%  [4]
%\end{verbatim}
%   This means that the remaining floats (that were left unplaced
%   after float page~3 got constructed) formed a float
%   page and that float page was the last in sequence (i.e., all
%   floats have been placed). However, this time the algorithm decided not
%   to unravel it, because it is nicely full: there are only 39 points
%   of free space left on that page.
%
%   Three other possible messages are shown in this sequence of
%   tracing lines from a second test document (which is using some uncommon
%   settings: \ctr{floatpagedeferlimit} is \texttt{10} and
%   \ctr{floatpagekeeplimit} is \texttt{5}):
%\begin{verbatim}
%  [1]
%  fewerfloatpages: PAGE: trying to make a float page
%  fewerfloatpages: ----- \@deferlist: \bx@B \bx@C \bx@D \bx@E \bx@F \bx@G \bx@H 
%  fewerfloatpages:  starting with \bx@B
%  fewerfloatpages:  --> success: \bx@B \bx@C \bx@D \bx@E \bx@F \bx@G \bx@H 
%  fewerfloatpages: ----- current float page kept (contains at least 5 floats)
%  [2] [3]
%\end{verbatim}
%   In this case 7 floats have been waiting on the defer list and the
%   algorithm was able to construct a float page using all of them.
%   The algorithm then keeps that page because it has 5 or more floats
%   in it (the value of the \ctr{floatpagekeeplimit} counter).
%
%   The next message in that test document shows what happens
%   when there are not enough floats waiting or they are simply too
%   small (to even get past the \cs{floatpagefraction} limit):
%\begin{verbatim}
%  fewerfloatpages: PAGE: trying to make a float page
%  fewerfloatpages: ----- \@deferlist: \bx@I \bx@J 
%  fewerfloatpages:  starting with \bx@I
%  fewerfloatpages:  --> fail
%  fewerfloatpages:  starting with \bx@J
%  fewerfloatpages:  --> fail
%  fewerfloatpages:  --> fail: no float page made
%  [4]
%\end{verbatim}
%   So no float page was made, but for some reason (that becomes clear
%   later) the two floats also didn't got distributed into the top or
%   bottom area of the next page. Instead they remained on the
%   defer list and during processing of page~4 one more float was
%   found so that after that page the defer list had grown to
%   length~3:
%
%\begin{verbatim}
%  fewerfloatpages: PAGE: trying to make a float page
%  fewerfloatpages: ----- \@deferlist: \bx@I \bx@J \bx@K 
%  fewerfloatpages:  starting with \bx@I
%  fewerfloatpages:  --> success: \bx@I \bx@J \bx@K 
%  fewerfloatpages: ----- current float page kept, contains a float
%  fewerfloatpages:       with p but no t or b specifier
%  [5]
%\end{verbatim}
%   This time all floats could be placed, but again the float page
%   wasn't unraveled (even though in the test document it contained a
%   lot of white space) because of the fact that one of its floats (in fact
%   the first though that can only be deduced implicitly) was
%   specified as a \enquote{float page only} float. This explains why
%   on page~4 \cs[no-index]{bx@I} couldn't be placed into the top or bottom
%   area and then all following floats of the same class (the test
%   document contained only \env{figure} floats) couldn't be placed
%   either.
%
%   If\mnote{Detailed tracing of the complete algorithm} you want
%   detailed tracing of the complete algorithm, also load the
%   \pkg{fltrace} package and enable the tracing with
%   \cs{tracefloats} anywhere in your document. Note, however, that
%   the resulting output is very detailed but rather low-level and
%   unpolished.
%
%
%
%
%
% \subsection{Local (manual) adjustments}
%
%   If the extended algorithm is used you will get fewer float pages
%   that contain a noticeable amount of white space. By adjusting
%   \cs{floatpagekeepfraction} and the counters \ctr{floatpagekeeplimit} and
%   \ctr{floatpagedeferlimit} you can direct the algorithm to unravel more or
%   fewer of the otherwise generated float pages. However, in some cases
%   it might happen that redistribution of the floats into the top and
%   bottom areas of the next page(s) may result in some of them
%   drifting too far away from their call-outs. If that happens, you can
%   either try to change the general parameters or you could help the
%   algorithm along by using the optional argument of individual float
%   environments.
%   The two main tools at your disposal are
%   \begin{itemize}
%   \item
%     using the \texttt{[!..]} notation to allow the float to go into
%     the top or bottom area even if it would be normally prevented by
%     other restrictions;
%   \item
%     using \texttt{[p]} to force a float into a float page as that
%     prevents the algorithm from unravelling the float page which contains
%     that float.
%   \end{itemize}
%   As an alternative you can, of course, temporarily alter the
%   definition of the command \cs{floatpagekeepfraction} or the values
%   of two counters in mid-document, but remember that they are not
%   looked at when a float is encountered in the source but when we
%   are at a page break and \LaTeX{} attempts to empty the defer list,
%   which is usually later and unfortunately somewhat asynchronous,
%   i.e., not easy to predict.
%
%
% \StopEventually{
% \begin{thebibliography}{1}
% \raggedright
% \bibitem{fmi:floatplacement}
% Frank Mittelbach.
% \newblock How to influence the position of float environments like figure and table in \LaTeX?
% \newblock \textsl{TUG}boat 35:3, 2014.\\
% \newblock \url{https://www.latex-project.org/publications/indexbytopic/2e-floats/}
% \bibitem{source2e}
% \LaTeX{} Project Team.
% \newblock The \LaTeXe{} Sources (660+ pages), 2020.\\
% \newblock \url{https://www.latex-project.org/help/documentation}
% \end{thebibliography}
% \ifx\thisissuepageref\undefined    ^^A is this TUB production ??? if not gen index
%    \setlength\IndexMin{200pt}  \PrintIndex
% \fi
%}
%
%
%
% \section{The implementation}
%
%    We start off with the package announcement. Requiring a fairly new
%    \LaTeX{} kernel is not absolutely necessary but it will help to
%    ensure that we patch what we think we patch and in the future it
%    means that will can be assured that the rollback functionality of
%    the kernel is available in case will need to support several
%    releases of the package.
%    \begin{macrocode}
%<*package>
\NeedsTeXFormat{LaTeX2e}[2018-04-01]
%    \end{macrocode}
%
%    \begin{macrocode}
\ProvidesPackage{fewerfloatpages}
                [\fewerfloatpagesdate\space \fewerfloatpagesversion\space
                 improve float page generation (FMi)]
%    \end{macrocode}
%
% \subsection{Option handling}
%                
%    This release of the package has four options: \option{trace}
%    for tracing the algorithm, \option{addbang} and
%    \option{checktb} to handle cases where the float size in
%    combination with the float specifiers makes it difficult if not
%    impossible to place the floats, and \option{nocheck} to not make
%    adjustments for that case.
%                
%    The option \option{trace} enables tracing of the algorithm and is
%    implemented by giving the command \cs[no-index]{fl@trace} (which is also
%    used by the \pkg{fltrace} package) a suitable definition.
%
%    To handle the case that the \pkg{fltrace} package is loaded
%    first, we use \cs[no-index]{providecommand}, so that its definition is not
%    overwritten, but used if it is already available. If the package is
%    loaded later everything works fine because it unconditionally
%    defines \cs[no-index]{fl@trace}, i.e., overwrites whatever
%   \pkg{fewerfloatpages} has defined.
%
%    \begin{macrocode}
\DeclareOption{trace}
              {\providecommand\fl@trace[1]%
                {{\let\@elt\@empty\typeout{fewerfloatpages: #1}}}}
%    \end{macrocode}
%
%    The other three options are mutually exclusive so we number them 0
%    to 2 in the command \cs[no-index]{fp@strategy} to ensure that
%    only one is ever active.  Option \option{nocheck} does nothing, with
%    the cost that some floats may float to the end of the
%    document. Option \option{addbang} adds a \texttt{!}\@ to floats
%    that are sent back for reevaluation when a float page gets
%    unraveled.  Option \option{checktb} implements a different
%    approach to handling problematic floats: the vertical
%    size of a float is checked, and if it is too large to be allowed
%    into the top or the bottom area, any \texttt{t} or \texttt{b}
%    specifier is replaced by \texttt{p} (or dropped if \texttt{p} is
%    already specified).
%    \begin{macrocode}
\def\fp@strategy{0}%              
\DeclareOption{nocheck}{\def\fp@strategy{0}}  % better name?
\DeclareOption{addbang}{\def\fp@strategy{1}}
\DeclareOption{checktb}{\def\fp@strategy{2}}
%    \end{macrocode}
%     The actual implementation is done later.  The default is
%    currently \option{checktb} but this may change to
%    \option{addbang} based on user feedback.
%    \begin{macrocode}
\ExecuteOptions{checktb}
\ProcessOptions
%    \end{macrocode}
%
%
% \subsection{Tracing code}
%                
%  \begin{macro}[internal]{\fl@trace}
%    The command \cs[no-index]{fl@trace} is used to output tracing information.
%    By default the tracing of the algorithm is turned off, so
%    \cs[no-index]{fl@trace} will simply swallow its argument. But if
%    \pkg{fltrace} is loaded or the option \option{trace} is given
%    then the command already has a definition so we don't change it
%    here.
%    \begin{macrocode}
\providecommand\fl@trace[1]{}
%    \end{macrocode}
%  \end{macro}
%
%
% \subsection{User-level interfaces}
%
%    For the most part the packages provides internal code that
%    extends the float algorithm of \LaTeX{}. There are, however, also
%    three new parameters that guide this algorithm; they are defined
%    in this section.
%
%  \begin{macro}[var]{\floatpagekeepfraction}
%    The fraction that the algorithm uses to decide whether a given float
%    page is so full that it would be pointless to unravel it for the
%    reasons outlined above. The default is whatever fraction has been
%    chosen as the minimum amount of text that needs to be on a normal
%    page (i.e., \cs{textfraction}).
%    \begin{macrocode}
\newcommand\floatpagekeepfraction{\textfraction}
%    \end{macrocode}
%  \end{macro}
%
%
%  \begin{macro}[var]{floatpagedeferlimit,\c@floatpagedeferlimit}

%    The algorithm will only consider unraveling float pages if there
%    are not too many floats on the defer list. The definition of
%    \enquote{too many} is provided through the counter
%    \ctr{floatpagedeferlimit} if there are more floats waiting to be
%    placed; float pages are generated until their number falls below
%    this level. Thus, a value of \texttt{0} will disable the whole
%    algorithm and a value of \texttt{1} means that only float pages
%    with a single float might get unraveled and only if there aren't
%    others still waiting to be placed.
%    \begin{macrocode}
\newcounter{floatpagedeferlimit}  \setcounter{floatpagedeferlimit}{3}
%    \end{macrocode}
%  \end{macro}



%  \begin{macro}[var]{floatpagekeeplimit,\c@floatpagekeeplimit}
%
%    A float page that contains at least this number of floats will
%    also be kept. The default is \texttt{3} but if you have a lot of
%    small floats it might be better to set this to a higher
%    value.
%    \begin{macrocode}
\newcounter{floatpagekeeplimit}   \setcounter{floatpagekeeplimit}{3}
%    \end{macrocode}
%  \end{macro}





% \subsection{Patching the \LaTeX{} kernel commands}
%                


%
%  \begin{macro}[internal]{\@tryfcolumn}
%    The main macro we have to patch to extend \LaTeX's algorithm is
%    \cs{@tryfcolumn}. That command is changed when \pkg{fltrace}
%    gets loaded, so we make our definition as late as possible to
%    ensure that it will survive.
%    \begin{macrocode}
\AtBeginDocument{%
%    \end{macrocode}
%    
%    \begin{macrocode}
\def \@tryfcolumn #1{%
  \global \@fcolmadefalse
  \ifx #1\@empty
  \else
    \fl@trace{PAGE: trying to make a float
                    \if@twocolumn column/page\else page\fi}%
    \fl@trace{----- \string #1: #1}%
    \xdef\@trylist{#1}%
    \global \let \@failedlist \@empty
    \begingroup
      \let \@elt \@xtryfc \@trylist
    \endgroup
%    \end{macrocode}
%    Up to this point the definition is the same as in the original
%    algorithm. At this point the  switch \cs[no-index]{if@fcolmade}  tells us if making a
%    float page was successful and the original algorithm then called
%    \cs[no-index]{@vtryfc} and removed the floats used for this float page from the defer list.
%
%    In the extended algorithm this is the place where things start to
%    differ as we may not want that float page to actually come into
%    existence.
%    \begin{macrocode}
    \if@fcolmade
%    \end{macrocode}
%    As a first step we count the number of floats in the defer list
%    and save the result in \cs{fp@candidates}.
%    \begin{macrocode}
      \fp@candidates\z@
      \def\@elt##1{\advance\fp@candidates\@ne}%
        #1%
      \let\@elt\relax
%    \end{macrocode}
%    Now we compare this number with the values of the counter \ctr{floatpagedeferlimit}
%    and if it is higher we definitely want to keep the float
%    page. The rationale is that if we unravel now, then all floats from
%    the defer list need to go into the top/bottom areas (or get
%    deferred again but to a later page) and so a high number means the defer list will
%    not get shortened very much and too many floats will get delayed further.
%    \begin{macrocode}
      \ifnum \fp@candidates >\c@floatpagedeferlimit
        \fl@trace{----- too many deferred floats for unraveling
                   (\the\fp@candidates\space> \the\c@floatpagedeferlimit)}%
      \else
%    \end{macrocode}
%
%    Otherwise we do a bit more testing. First we set \cs[no-index]{if@fcolmade}
%    back to false; after all our goal is to not  keep the float page. If
%    during the tests we decide otherwise we set it back to true, which
%    then signals that it should stay.
%
%    We also count the floats on the float page, reusing \cs[no-index]{fp@candidates}
%    for that, which is why we initialize it to zero.
%    \begin{macrocode}
        \global\@fcolmadefalse
        \fp@candidates\z@
%    \end{macrocode}
%    The actual checking is done with
%    \cs{fp@analyse@floats@for@unraveling} and it loops over
%    \cs{@flsucceed}, i.e., the floats for that float page. This
%    checks if any float for that page has only a \texttt{[p]}
%    specifier and if so it sets \cs[no-index]{if@fcolmade} back to
%    \texttt{true} and as a side effect it also does the counting for
%    us. Furthermore, it also changes the switch to true if it finds
%    at least \ctr{floatpagekeeplimit} floats on that page.
%    \begin{macrocode}
        \let\@elt\fp@analyse@floats@for@unraveling
          \@flsucceed
        \let\@elt\relax
%    \end{macrocode}
%    Now we recheck the state of the switch and if it still says
%    \texttt{false}, all tests so far indicate that we don't want the float
%    page.
%    \begin{macrocode}
        \if@fcolmade  \else
%    \end{macrocode}
%    But we aren't done yet: the float page might be nicely filled,
%    in which case it would be a shame to unravel it. During the above
%    loop we also measured the free space on the float page and stored
%    it in \cs{fp@unused@space} (see \cs{@xtryfc} below). We now
%    compare that to the maximum free space that we consider to be
%    still okay and if there is more we finally do the unraveling.
%    \begin{macrocode}
          \@tempdima\floatpagekeepfraction\@colht
          \ifdim \fp@unused@space >\@tempdima
            \fl@trace{----- current float page unraveled^^J%
                      \@spaces\@spaces\@spaces\space\space\space
                      (free space \fp@unused@space\space > \the\@tempdima)}%
%    \end{macrocode}
%    For this we basically return all floats back to the
%    defer list. The switch is still \texttt{false} so it doesn't need
%    changing.
%    \begin{macrocode}
            \xdef #1{\@failedlist\@flsucceed\@flfail}%
%    \end{macrocode}
%    However, we may also want to add a \texttt{!}\@ specifier to each of
%    the floats (if the \option{addbang} option was given)
%    so we loop over all the floats once more to get this
%    done.\footnote{This could have been integrated with
%    \cs{fp@analyse@floats@for@unraveling} but there is not much gain if
%    any and by keeping it separate the processing logic seems clearer to me.}
%    \begin{macrocode}
            \let\@elt\fp@maybe@add@bang
              \@flsucceed
            \let\@elt\relax
          \else
%    \end{macrocode}
%    But if we want to keep the float page after all, we have to set
%    the switch back to \texttt{true} so that the rest of the
%    algorithm proceeds correctly.
%    \begin{macrocode}
            \global \@fcolmadetrue
            \fl@trace{----- current float page kept, full enough^^J%
                      \@spaces\@spaces\@spaces\space\space\space
                      (free space \fp@unused@space\space < \the\@tempdima)}%
          \fi
        \fi
      \fi
%    \end{macrocode}
%    The next \cs[no-index]{else} matches the first \cs[no-index]{if@fcolmade}, i.e., the
%    case that the algorithm wasn't able to make any float page. If
%    we are tracing the algorithm, we want to tell the user about this.
%    \begin{macrocode}
    \else
      \fl@trace{ --> fail: no float page made}%
    \fi
%    \end{macrocode}
%    
%    Finally, at this point we are back in the original algorithm. Now the
%    switch tells the truth about whether or not we want to make a float
%    page, and if so, we go ahead and produce it.
%    \begin{macrocode}
    \if@fcolmade
        \@vtryfc #1%
    \fi
  \fi}%
}% -- END of \AtBeginDocument
%    \end{macrocode}
%  \end{macro}
%
% \pagebreak
%
%
%  \begin{macro}[internal]{\@makefcolumn}
%
%    In contrast to \cs{@tryfcolumn} this macro will always make float
%    pages out of the deferred floats. It is used by \cs{clearpage}
%    when we really need the floats to get out because there is no further
%    text coming up. Thus, in that case we should not unravel the
%    float pages. That would happen with the kernel definition of
%    \cs{@makefcolumn} as that calls \cs{@tryfcolumn} which we just
%    changed above. We therefore modify its definition to include the
%    original code for \cs{@tryfcolumn} instead of calling our updated
%    version.
%
%    Again this change is made at \verb=\begin{document}= so that it
%    is not overwritten in case \pkg{fltrace} is loaded afterwards.
%    \begin{macrocode}
\AtBeginDocument{%
\def\@makefcolumn #1{%
  \begingroup
    \@fpmin -\maxdimen
    \let \@testfp \@gobble
%    \end{macrocode}
%    At this point the original definition called \cs[no-index]{@tryfcolumn} and
%    the lines above ensured that it was always succeeding in making
%    a float page. However, since we have changed that command to do unraveling
%    we had better not use it any more. Instead we replace it by its original
%    definition (with the addition of two tracing lines).
%    \begin{macrocode}
    \global \@fcolmadefalse
    \ifx #1\@empty
    \else
      \fl@trace{PAGE: trying to make a float
                      \if@twocolumn column/page\else page\fi}%
      \fl@trace{----- \string #1: #1}%
      \xdef\@trylist{#1}%
      \global \let \@failedlist \@empty
      \begingroup
        \let \@elt \@xtryfc \@trylist
      \endgroup
      \if@fcolmade
        \@vtryfc #1%
      \fi
    \fi
  \endgroup
}%
}% -- END of \AtBeginDocument
%    \end{macrocode}
%  \end{macro}
%
%
%
%
%  \begin{macro}[internal]{\@xtryfc}
%    The only change to \cs{@xtryfc} is the addition of the
%    \cs[no-index]{fl@trace} calls. But this extra tracing info is generally
%    useful and should also be done in the \pkg{fltrace} package.
%
%    The macro initiates a float page trial starting with the first
%    float in \cs{@trylist}. More detailed explanations can be found
%    in the documented sources of the \LaTeX{} kernel~\cite{source2e}.
%    \begin{macrocode}
\def\@xtryfc #1{%
  \fl@trace{ starting with \string#1}%
  \@next\reserved@a\@trylist{}{}%
  \@currtype \count #1%
  \divide\@currtype\@xxxii
  \multiply\@currtype\@xxxii
  \@bitor \@currtype \@failedlist
  \@testfp #1%
  \@testwrongwidth #1%
  \ifdim \ht #1>\@colht
     \@testtrue
  \fi
  \if@test
    \@cons\@failedlist #1%
    \fl@trace{ --> fail}%
  \else
    \@ytryfc #1%
  \fi
}%
%    \end{macrocode}
%  \end{macro}
%
%
%
%
%
%
%  \begin{macro}[internal]{\@ytryfc}
%
%    The command \cs{@ytryfc}, which is also part of the code in the
%    kernel, loops through the defer list and tries to build a float
%    page starting with the float passed to it in |#1|. If it succeeds,
%    the floats that are part of the float page are listed in
%    \cs{@flsucceed} and the switch \cs[no-index]{if@fcolmade} is set to
%    \texttt{true}. Also of interest to us is that inside the code
%    \cs[no-index]{@tempdima} holds the size taken up by the floats, so we can
%    use this to calculate the unused space on the float page and
%    store it in \cs{fp@unused@space} for use in our extended
%    algorithm.
%    \begin{macrocode}
\def\@ytryfc #1{%
  \begingroup
    \gdef\@flsucceed{\@elt #1}%
    \global\let\@flfail\@empty
    \@tempdima\ht #1%
    \let\@elt\@ztryfc
    \@trylist
    \ifdim \@tempdima >\@fpmin
      \global\@fcolmadetrue
%    \end{macrocode}
%    This branch is executed when the floats together are big enough
%   to form a float page. Thus, this is the right place to calculate
%   the free space by subtracting the used space from the column
%   height (which may not be the full height if there are spanning
%   floats in two column mode).
%    \begin{macrocode}
      \@tempdimb\@colht
      \advance\@tempdimb-\@tempdima
      \xdef\fp@unused@space{\the\@tempdimb}%
%    \end{macrocode}
%    The remaining code is again unchanged except that we added two
%    additional tracing lines (though those should be added to
%    the \pkg{fltrace} package too one of these days).
%    \begin{macrocode}
    \else
      \@cons\@failedlist #1%
      \fl@trace{ --> fail}%
    \fi
  \endgroup
  \if@fcolmade
    \let\@elt\@gobble
    \fl@trace{ --> success: \@flsucceed}%
  \fi}
%    \end{macrocode}
%  \end{macro}
%
%
%
%
%  \begin{macro}[internal]{\@largefloatcheck}
%
%    The final kernel macro we need to patch is
%    \cs{@largefloatcheck}. This is called when a float box is
%    constructed and it checks if that box is larger than the available
%    \cs{textheight}, which would mean it could never be placed
%    anywhere, not even on a float page. The code therefore reduces
%    the box size as necessary and issues a warning.
%
%    This macro is therefore a natural candidate to also check if the
%    float size is too large for the float to go into top or bottom
%    areas (if the option \option{checktb} is used).
%    \begin{macrocode}
\def \@largefloatcheck{%
  \ifdim \ht\@currbox>\textheight
    \@tempdima -\textheight
    \advance \@tempdima \ht\@currbox
    \@latex@warning {Float too large for page by \the\@tempdima}%
    \ht\@currbox \textheight
  \fi    
%    \end{macrocode}
%    The \cs{fp@maybe@check@tb} does the checking (or nothing if the
%    option is not given).
%    \begin{macrocode}
  \fp@maybe@check@tb
}
%    \end{macrocode}
%  \end{macro}
%
%
%
%
%
% \subsection{Internal helper commands and parameters}
%                
%
%  \begin{macro}[internal]{\fp@candidates}
%     We use an internal counter to count the number of floats in the
%    defer list and on a float page under construction.
%    \begin{macrocode}
\newcount\fp@candidates
%    \end{macrocode}
%  \end{macro}
%
%
%
%  \begin{macro}[internal]{\fp@unused@space}
%    In \cs{fp@unused@space} we store the amount of free space
%    on the current float page.
%    \begin{macrocode}
\def\fp@unused@space{}
%    \end{macrocode}
%  \end{macro}
%
%
%
%  \begin{macro}[internal]{\fp@analyse@floats@for@unraveling}
%    With \cs{fp@analyse@floats@for@unraveling} we loop over the
%    floats on the float page, i.e., |#1| will be one such float.
%
%    One of its tasks is to count the floats (in \cs{fp@candidates}) and
%    check if
%    there are at least \ctr{floatpagekeeplimit} of them (which means
%    the float page should definitely be kept).
%
%    Its most important task, however, is to check if one of the floats
%    has only a \texttt{p} specifier but no other. In that case it is
%    essential that we not unravel the float page because such a
%    float would then only go back onto the defer list as it has no
%    place to go except a float page.
%    \begin{macrocode}
\def\fp@analyse@floats@for@unraveling#1{%
  \advance\fp@candidates\@ne
  \ifnum \fp@candidates <\c@floatpagekeeplimit
%    \end{macrocode}
%    So far we haven't got enough floats to know that this float page
%    should be kept so we check the given float specifiers.
%  
%    The test may look a little weird,\footnote{\enquote{Little} might
%    be an understatement. Encoding a lot of information in individual
%    bits of the counter value associated with a float was a great way
%    in the early days of \LaTeX{} to preserve macro space (and
%    absolutely essential back then), but these days \ldots{} Anyway,
%    it is the way it is and that part can't really be changed without
%    breaking a lot of packages.}  but what we want to know is this:
%    is there a \texttt{p} (third bit) but neither a \texttt{b}
%    (second bit) nor a \texttt{t} (first bit). We don't care about
%    \texttt{h} or \texttt{!} which are the next two bits in the float
%    counter nor any of its higher bits (which encode the type of
%    float). So we divide the integer number by 8, which drops
%    the two least significant bits (think of the integer
%    represented in binary format), and then multiply it again by 8. As
%    a result the first two bits are zeroed out.  We then compare the
%    result with the original value and if the two values are the same
%    then the \texttt{b} and \texttt{t} bits must both have been zero
%    from the start. And since the float was on a float page we also
%    know that it had a \texttt{p} specifier.
%    \begin{macrocode}
    \@tempcntb\count#1%
    \divide\@tempcntb 8\relax
    \multiply\@tempcntb 8\relax
    \ifnum \count#1=\@tempcntb
%    \end{macrocode}
%    In that case we set \cs[no-index]{if@fcolmade} to \texttt{true}
%    to signal that this float page should be kept, generate a tracing
%    message and change \cs[no-index]{@elt} to become
%    \cs[no-index]{@gobble} to quickly jump over any remaining floats
%    in the loop without doing further tests or generate further
%    tracing messages.
%    \begin{macrocode}
      \global \@fcolmadetrue
      \fl@trace{----- current float page kept, contains a float}%
      \fl@trace{\@spaces\space\space with p but no t or b specifier}%
      \let\@elt\@gobble
    \fi
%    \end{macrocode}
%    On the other hand, if we have seen enough floats we also know
%    that the float page should be kept, so change the switch, give
%    some tracing info and stop checking:
%    \begin{macrocode}
  \else
    \global \@fcolmadetrue
    \fl@trace{----- current float page kept
              (contains at least \the\fp@candidates\space floats)}%
    \let\@elt\@gobble
  \fi
}
%    \end{macrocode}
%  \end{macro}
%
%
%
%
%  \begin{macro}[internal]{\fp@maybe@add@bang}
%
%    The helper \cs{fp@maybe@add@bang} is used to loop through all of
%    the floats of a float page (receiving each as |#1| in turn) and
%    add a \texttt{!}\@ specifier if there wasn't one before.
%
%    However, we only define it if we implement strategy 1 which is
%    option \option{addbang}.
%    \begin{macrocode}
\ifnum\fp@strategy=1
  \def\fp@maybe@add@bang#1{%
%    \end{macrocode}
%    Find out if the fourth bit is set (which means no \texttt{!}\@)
%    and if so subtract 16 from the float counter which means setting
%    it to zero.
%    \begin{macrocode}
    \@boxfpsbit #1\sixt@@n
    \ifodd \@tempcnta
      \global\advance\count#1-\sixt@@n
    \fi
  }
\else
  \let\fp@maybe@add@bang\@gobble
\fi
%    \end{macrocode}
%  \end{macro}
%
%
%
%
%  \begin{macro}[internal]{\fp@maybe@check@tb}
%    The code in \cs{fp@maybe@check@tb} is used in
%    \cs{@largefloatcheck} to test if the float has a \texttt{t} or
%    \texttt{b} specifier but is too large to fit into the
%    respective area. This test is not made by default but only if the
%    option \option{checktb} is used, i.e., strategy~2.
%    \begin{macrocode}
\ifnum\fp@strategy=2
  \def\fp@maybe@check@tb{%
%    \end{macrocode}
%    Again this is a case of
%    looking at various bits in the float counter value in binary
%    notation.
%    If the specifier contained a \texttt{!}\@ we are ok and it would
%    be wrong to change the specifier, because in
%    that case size restrictions for areas do not apply. For this we
%    have to test the fourth bit which means dividing by 16 and then
%    checking if the result is odd or even (odd means there was no
%    \texttt{!}\@).\footnote{I'm sure we had good reasons to
%    implement it this way in 1992\Dash we probably saved a few bytes
%    which was important back then. But it is certainly odd that for
%    \texttt{!}\@ a value of zero means that it was specified on the
%    float while for all other specifiers a value of \texttt{1}
%    indicates that the specifier was given.} The kernel
%    \cs{@getfpsbit} does this for us and stores the result in
%    \cs[no-index]{@tempcnta} so we can test this with \cs[no-index]{ifodd} to see if the
%    bit was set.
%    \begin{macrocode}
    \@getfpsbit \sixt@@n
    \ifodd \@tempcnta
%    \end{macrocode}
%    If there was no \texttt{!}\@ we check if the height of the
%    float is too large to fit into the top area.
%    \begin{macrocode}
      \ifdim \ht\@currbox>\topfraction\textheight
%    \end{macrocode}
%
%    If that is the case we also check the first bit of the float
%    counter to see if a \texttt{t} was specified. For this we use
%    \cs{@getfpsbit} again but this time with \texttt{2} as the argument
%    since we test the first bit.
%    \begin{macrocode}
        \@getfpsbit \tw@
        \ifodd \@tempcnta
%    \end{macrocode}
%    If \texttt{t}  was specified we need to remove it (next line) and
%    add (if not already present) a \texttt{p} instead. This is done
%    by \cs[no-index]{fp@add@p@bit}. Finally we add a warning for the user about
%    the change.
%    \begin{macrocode}
          \global\advance\count\@currbox -\tw@
          \fp@add@p@bit
          \@latex@warning {Float too large for top area: t changed to p}%
        \fi
      \fi
%    \end{macrocode}
%    A similar test and action is needed for bottom floats; here we
%    need to look at and zero out the second bit (i.e., using 4 as a value).
%    \begin{macrocode}
      \ifdim \ht\@currbox>\bottomfraction\textheight
        \@getfpsbit 4\relax
        \ifodd \@tempcnta 
          \global\advance\count\@currbox -4\relax
          \fp@add@p@bit
          \@latex@warning {Float too large for bottom area:
                           b changed to p}%
        \fi
      \fi
    \fi
  }
%    \end{macrocode}
% In all other cases \cs{fp@maybe@check@tb} does nothing.
%    \begin{macrocode}
\else \let\fp@maybe@check@tb\relax \fi  
%    \end{macrocode}
%  \end{macro}
%
%
%
%  \begin{macro}[internal]{\fp@add@p@bit}
%    The command \cs{fp@add@p@bit} adds the \texttt{p} specifier which
%    means checking the third bit and if not set, adding 8 to the float  counter.
%    \begin{macrocode}
\def\fp@add@p@bit{%
   \@getfpsbit 8\relax
   \ifodd \@tempcnta \else \global\advance\count\@currbox 8\relax \fi}
%    \end{macrocode}
%  \end{macro}
%
%
% \ifx\thisissuepageref\undefined    ^^A is this TUB production ??? if not gen index
% \subsection{Patches that will eventually go into \pkg{fltrace}}
%
%  The \pkg{fewerfloatpages} package added some additional general
%  tracing info into some of the kernel functions which isn't
%  currently available when using only the \pkg{fltrace} package. As
%  that tracing info is generally useful for understanding what the
%  base part of the float algorithm does, it should also be added to
%  the latter package.
%
% \fi
%
%    \begin{macrocode}
%<*package>
%    \end{macrocode}
% \Finale
%


\endinput