# Problem with “endfloat” in manuscript option of APA6.cls

Several years ago I commonly wrote papers in MiKTeX using the apa.cls, which allowed you to switch between a two-column formatted "jou" mode and a single-column, double spaced "man" mode.

I have recently tried running these old .tex files in MiKTeX 2.9, and I have found that the endfloat package no longer works properly, either in apa.cls or the new apa6.cls.

I have tried renaming the apaendfloat.cfg file endfloat.cfg and placing it in the working directory.

Here are the errors I get with the shortsample.tex file that comes with apa6.cls. Basically, it completely chokes on the tables, and nothing comes out but gibberish in the table and figure portions of the paper.

Is there a solution for this?

AED endfloat: Processing end Figures and Tables [5]
! Missing number, treated as zero.
\@ef@tttopen
l.83 \end{document}

?
! Missing = inserted for \ifnum.
\@ef@tttopen
l.83 \end{document}

?
! Missing number, treated as zero.
\@ef@tttopen
l.83 \end{document}

?
! Undefined control sequence.
\processtables ...seout \efloat@postttt \ef@setct
{ttt}{0} \clearpage \if@ta...
l.83 \end{document}

?
[6]
! Undefined control sequence.
\section *{\tablesection }...
l.83 \end{document}

?
! Undefined control sequence.
\processtables ...@@spacing } \processtables@hook
\@ifundefined {hrm}{}{\sf ...
l.83 \end{document}

?
(R:\jointtex\SPRINT\apaexample.ttt
\begin{table*}[hbt]
\ifnextchar[{\eatarg}{}
! Undefined control sequence.
\table ...linechar \
\catcode \
=12 \ef@xtable
\fi
l.1 \begin{table}
[tbp]
?
)
! Extra \fi.
\processtables ...ttt}\let \BBAB \normal@BBAB \fi

l.83 \end{document}

?
! Missing number, treated as zero.
\@ef@fffopen
l.83 \end{document}

?
! Missing = inserted for \ifnum.
\@ef@fffopen
l.83 \end{document}

?
! Missing number, treated as zero.
\@ef@fffopen
l.83 \end{document}

?
! Undefined control sequence.
\processfigures ...eout \efloat@postfff \ef@setct
{fff}{0} \clearpage \if@fi...
l.83 \end{document}

?

/b/c4/cOverfull/b/c0/c \hbox (441.54912pt too wide) in paragraph at lines 1--83
[][]\T1//lmr//m//n//12 [tbp]
\caption{Some numbers that could be experimental data.
}
\label{tab:tab1}
\begin{tabular}{lcc}\hline
& \multicolumn{2}{c}{Fa
ctor 2} \\ \cline{2-
/b/c4/cOverfull/b/c0/c \hbox (1151.26506pt too wide) in paragraph at lines 1--83
\T1//lmr//m//n//12 3}
Factor 1  & Condition A  & Condition B   \\ \hline
First
& 586 (231)    & 649 (255)     \\
&    2.2       &    7.5        \\
S
econd    & 590 (195)    & 623 (231)     \\
&    2.8       &    2.5
\\ \hline
\end{tabular}
\end{table}
\efloatseparator
>0
[7]
! Undefined control sequence.
\processfigures ...}{}{\sf } \processfigures@hook
\@input {\jobname .fff} \fi
l.83 \end{document}

?
(R:\jointtex\SPRINT\apaexample.fff)
! Extra }, or forgotten \endgroup.
\processdelayedfloats ...ocessotherdelayedfloats }

l.83 \end{document}

?

! LaTeX /b/c9/cError:/b/c0/c \begin{efloat@float} on input line 1 ended by \end{document}.
See the LaTeX manual or LaTeX Companion for explanation.
Type  H <return>  for immediate help.
...

l.83 \end{document}

?

-
I also received a bug report about the new version of the endfloat package today, it seems that this is my fault. I will examine this tomorrow. So far the best solution is IMHO reverting to endfloat.sty v2.4. –  Axel Sommerfeldt Dec 21 '11 at 21:30

As I just realized the apa6 document class is redefining half of the endfloat package. (Some code was even simply copied from endfloat.sty and modified.) Since some internals needed to be changed from v2.4 to v2.5 of the endfloat package, these redefinitions are not working anymore.
I'm in contact with the maintainer of the apa6 document class anyway, and we need to find a way to make it working again, and this time in a way where the apa6 class is not using internals of the endfloat package so it won't break compatibility again if v2.6 of the endfloat package will be released.
Addendum: Updating to version 2.5c of the endfloat package should help.
I was able to build a version 2.5c of the endfloat package which emulates most of the obsolete internal stuff needed by apa(6), so it should at least work again with the apa and apa6 document classes, but only without using one of the new features added to 2.5. (The package looks now really ugly inside and is even more hard to change in future, but hey, this is not a coding beauty contest.) This version is currently under revision by the maintainer of apa6. If you are in urgent need for this version just drop me an e-mail. –  Axel Sommerfeldt Dec 23 '11 at 14:52