Environment "alignat" incorrectly parsed

[tdegeus]

I have ran into a bug concerning the following piece of LaTeX code:

\begin{alignat}{2}
  ...
\end{alignat}

In my case there is no different between the two TeX-files that I compare. However, this piece of code is parsed to the following output:

\begin{align}at{2}
  ...
\end{align}

which is obviously incorrect (and problematic).

Thanks for the effort. And thanks for the great program, BTW!
Tom

[ftilmann]

Hi with the extremely simple MWE for both old and new file:

\begin{alignat}{2}
ipsem loret
\end{alignat}

I failed to reproduce your bug. Can you construct a full MWE and attach and the exact command you used for calling.

[tdegeus]

@ftilmann that's quick. I have been able to reproduce the following minimal example.

old_min.tex

\documentclass{goose-article}

\begin{document}

\begin{alignat}{2}
  ...
\end{alignat}

\begin{equation}\label{eq:elas:C}
\end{equation}

\begin{equation}\label{eq:elas:lam-mu}
  \lambda = \frac{E \nu}{ (1+\nu)(1-2\nu)} \qquad
\end{equation}

\end{document}

new_min.tex

\documentclass{goose-article}

\begin{document}

\begin{alignat}{2}
  ...
\end{alignat}

\begin{equation}\label{eq:elas:C}
\end{equation}

\begin{equation}
  \mu     = \frac{E    }{2(1+\nu)        }
\end{equation}

\end{document}

I run the following command:

$ latexdiff --exclude-textcmd="section,subsection" old_min.tex new_min.tex > diff.tex

To find diff.tex

\documentclass{goose-article}
%DIF LATEXDIFF DIFFERENCE FILE
%DIF DEL old_min.tex   Fri Nov 25 12:39:38 2016
%DIF ADD new_min.tex   Fri Nov 25 12:39:41 2016
%DIF PREAMBLE EXTENSION ADDED BY LATEXDIFF
%DIF UNDERLINE PREAMBLE %DIF PREAMBLE
\RequirePackage[normalem]{ulem} %DIF PREAMBLE
\RequirePackage{color}\definecolor{RED}{rgb}{1,0,0}\definecolor{BLUE}{rgb}{0,0,1} %DIF PREAMBLE
\providecommand{\DIFadd}[1]{{\protect\color{blue}\uwave{#1}}} %DIF PREAMBLE
\providecommand{\DIFdel}[1]{{\protect\color{red}\sout{#1}}}                      %DIF PREAMBLE
%DIF SAFE PREAMBLE %DIF PREAMBLE
\providecommand{\DIFaddbegin}{} %DIF PREAMBLE
\providecommand{\DIFaddend}{} %DIF PREAMBLE
\providecommand{\DIFdelbegin}{} %DIF PREAMBLE
\providecommand{\DIFdelend}{} %DIF PREAMBLE
%DIF FLOATSAFE PREAMBLE %DIF PREAMBLE
\providecommand{\DIFaddFL}[1]{\DIFadd{#1}} %DIF PREAMBLE
\providecommand{\DIFdelFL}[1]{\DIFdel{#1}} %DIF PREAMBLE
\providecommand{\DIFaddbeginFL}{} %DIF PREAMBLE
\providecommand{\DIFaddendFL}{} %DIF PREAMBLE
\providecommand{\DIFdelbeginFL}{} %DIF PREAMBLE
\providecommand{\DIFdelendFL}{} %DIF PREAMBLE
%DIF END PREAMBLE EXTENSION ADDED BY LATEXDIFF

\begin{document}

\begin{align}at{2}
  ...
\end{align}

\begin{equation}\label{eq:elas:C}
\end{equation}

\begin{equation}\DIFdelbegin %DIFDELCMD < %DIFDELCMD < \label{eq:elas:lam-mu}%%%
%DIFDELCMD <   %%%
\DIFdel{\lambda }\DIFdelend 
  \DIFaddbegin \DIFadd{\mu     }\DIFaddend = \DIFdelbegin \DIFdel{\frac{E \nu}{ (1+\nu)(1-2\nu)} }%DIFDELCMD < \qquad
%DIFDELCMD < %%%
\DIFdelend \DIFaddbegin \DIFadd{\frac{E    }{2(1+\nu)        }
}\DIFaddend \end{equation}

\end{document}

Note:

$ latexdiff --version
This is LATEXDIFF 1.1.1  (Algorithm::Diff 1.15 so, Perl v5.18.2)
  (c) 2004-2015 F J Tilmann

Hope this helps!

[ftilmann]

Thanks. I can confirm that I can reproduce this also with the latest github version. It is truly bizarre, as removing either one of the following equation environments lets it process correctly. Offhand, I don't know the reason but I will look into it.

[tdegeus]

I agree that it is truly bizar. So much so that I had to write a Python script to iteratively trim down my file to this minimal example.

[ftilmann]

Found the problem. The bug turned out to be in a contributed subroutine, so I had to get my head around how that one worked first. The faulty line is only activated if there is an equation with a \frac command in any equation, and the bug (currently) only affects alignat (because align starts with the same letters).
Your carefully constructed minimal example helped a lot in nailing this quickly (I am not normally that responsive, just had another task that I wanted to avoid :-) ). If you don't want to update, there are only a few lines that need changing, which could implemented by hand (the commit also contains some cosmetic changes, though).
PS: I know the equation in your example from elasticity theory relating Lame Parameters to Young's modulus (?) and Poisson ratio.
Fixed with commit e39bf4e

[tdegeus]

Thanks! Really appreciate it!

Yes, it is indeed from elasticity theory. We have developed an FFT-based solver for mechanics, and are about the submit the review (see https://arxiv.org/abs/1603.08893).