I'm using a LaTeX macro to sanitize math expressions for label generation. The macro applies \detokenize to the input and then runs a series of \StrSubstitute commands to remove or replace certain symbols.
I expect that a multiline math expression using \begin{aligned}...\end{aligned} should produce the same sanitized result as the inline version. However, the word aligned still appears at the beginning and end of the result from the multiline version — even though I explicitly remove it using:
\StrSubstitute{\temp}{aligned}{}[\temp]
Example output:
Result of sanitize: [alignedPLpBIndexNullRpCommaForallBLpPLpBRpImpCEqxInBMidForallALpPLpARpImpxInARpRpCommaForallBLpPLpBRpImpDEqxInBMidForallALpPLpARpImpxInARpRpVdashCEqDaligned] Result of sanitize: [PLpBIndexNullRpCommaForallBLpPLpBRpImpCEqxInBMidForallALpPLpARpImpxInARpRpCommaForallBLpPLpBRpImpDEqxInBMidForallALpPLpARpImpxInARpRpVdashCEqD]
Why does the first result still contain aligned? How can I reliably remove it so both variants produce identical output?
My code is below.
\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage{amsmath}
\usepackage{xstring}
\usepackage{xparse}
% -------------------------------
% Label-Sanitizer
% -------------------------------
\newcommand{\sanitize}[1]{%
\edef\temp{\detokenize{#1}}%
\StrSubstitute{\temp}{\string\leftrightarrow}{Equiv}[\temp]%
\StrSubstitute{\temp}{\string\rightarrow}{Imp}[\temp]%
\StrSubstitute{\temp}{\string\eqvdash}{Eqvboth}[\temp]%
\StrSubstitute{\temp}{\string\dashv}{DashV}[\temp]%
\StrSubstitute{\temp}{\string\vdash}{Vdash}[\temp]%
\StrSubstitute{\temp}{\string\iff}{Iff}[\temp]%
\StrSubstitute{\temp}{\string\land}{And}[\temp]%
\StrSubstitute{\temp}{\string\lor}{Or}[\temp]%
\StrSubstitute{\temp}{\string\neg}{Neg}[\temp]%
\StrSubstitute{\temp}{\string\forall}{Forall}[\temp]%
\StrSubstitute{\temp}{\string\exists}{Exists}[\temp]%
\StrSubstitute{\temp}{\string\iota}{Iota}[\temp]%
\StrSubstitute{\temp}{:=}{Defeq}[\temp]%
\StrSubstitute{\temp}{=}{Eq}[\temp]%
\StrSubstitute{\temp}{\string\neq}{Neq}[\temp]%
\StrSubstitute{\temp}{\string\not}{Notin}[\temp]%
\StrSubstitute{\temp}{\string\in}{In}[\temp]%
\StrSubstitute{\temp}{\string\subseteq}{Subseteq}[\temp]%
\StrSubstitute{\temp}{\string\mid}{Mid}[\temp]%
\StrSubstitute{\temp}{\string_}{Index}[\temp]%
\StrSubstitute{\temp}{0}{Null}[\temp]%
\StrSubstitute{\temp}{\string\bigl}{}[\temp]%
\StrSubstitute{\temp}{\string\bigr}{}[\temp]%
\StrSubstitute{\temp}{\string\Bigl}{}[\temp]%
\StrSubstitute{\temp}{\string\Bigr}{}[\temp]%
\StrSubstitute{\temp}{(}{Lp}[\temp]%
\StrSubstitute{\temp}{)}{Rp}[\temp]%
\StrSubstitute{\temp}{\string\begin}{}[\temp]%
\StrSubstitute{\temp}{\string\end}{}[\temp]%
\StrSubstitute{\temp}{aligned}{}[\temp]%
\StrSubstitute{\temp}{\string\{}{}[\temp]%
\StrSubstitute{\temp}{\string\}}{}[\temp]%
\StrSubstitute{\temp}{\char'\{}{}[\temp]%
\StrSubstitute{\temp}{\char'\}}{}[\temp]%
\ExpandArgs{e}\StrSubstitute{\temp}{\string{\iffalse}\fi}{}[\temp]%
\ExpandArgs{e}\StrSubstitute{\temp}{\iffalse{\fi\string}}{}[\temp]%
\StrSubstitute{\temp}{\string\,}{}[\temp]%
\StrSubstitute{\temp}{\string\;}{}[\temp]%
\StrSubstitute{\temp}{\string\\}{}[\temp]
\StrSubstitute{\temp}{\string\ }{}[\temp]%
\StrSubstitute{\temp}{\string&}{}[\temp]%
\StrSubstitute{\temp}{ }{}[\temp]%
\StrSubstitute{\temp}{,}{Comma}[\temp]%
}
\makeatletter
\newcommand{\debugsanitize}[1]{%
\def\tempa{#1}%
\expandafter\sanitize\expandafter{\tempa}%
\typeout{Result of sanitize: [\temp]}%
}
\makeatother
\begin{document}
\debugsanitize{
\begin{aligned}
&P(B_0),\;
\forall B\Bigl(P(B)\rightarrow C= \{ x \in B \mid \forall A (P(A) \rightarrow x \in A) \}\Bigr),\\
&\forall B\Bigl(P(B)\rightarrow D= \{ x \in B \mid \forall A (P(A) \rightarrow x \in A) \}\Bigr)
\vdash C=D
\end{aligned}
}
\debugsanitize{P(B_0),\forall B\Bigl(P(B)\rightarrow C= \{ x \in B \mid \forall A (P(A) \rightarrow x \in A) \}\Bigr),\forall B\Bigl(P(B)\rightarrow D= \{ x \in B \mid \forall A (P(A) \rightarrow x \in A) \}\Bigr) \vdash C=D}
\end{document}
\usepackage[utf8]{inputenc}is redundant and should be omitted. And, if you're using either XeLaTeX or LuaLaTeX, you shouldn't be running\usepackage[utf8]{inputenc}at all.