Def inside NewDocumentCommand fails?

Question

I'd like to rename arguments like #1 into something like \ies@firstStar to have less error prone programming and less variable renaming when changing the parameters of a function. However doing something like \def\ies@firstStar{#1} fails, any idea what's wrong?

\documentclass{article}
\usepackage{xparse}
\makeatletter
\NewDocumentCommand{\myTest}{s}{
  \bgroup% make definitions local
  \def\ies@firstStar{#1}%
  \IfBooleanTF{\ies@firstStar}{%
    One star
  }{%
    No star
  }
  \egroup
}
\makeatother
\begin{document}
\myTest
\end{document}

Answer 1

Internally, I’m not sure exactly how \IfBooleanTF works, but the following solves the problem. However, at least for this simple example, I fail to see how this makes the code more readable.

\documentclass{article}
\usepackage{xparse}
\makeatletter
\NewDocumentCommand{\myTest}{s}{
  \begingroup% better to use this than \bgroup...\egroup
  \def\ies@firstStar{#1}%
  \expandafter\IfBooleanTF\expandafter{\ies@firstStar}{%
    One star
  }{%
    No star
  }
  \endgroup
}
\makeatother
\begin{document}
\myTest
\myTest*
\end{document}

Answer 2

Use \let instead of \def to avoid the \expandafters. In the case of \def, the replacement text of \ies@firstStar is \BooleanFalse or \BooleanTrue. In the case of \let, \ies@firstStar actually is equivalent to \BooleanFalse or \BooleanTrue.

As Gaussler notes, \let can only assign a single token. Therefore, this approach works because the s argument will never be more than a single token.

\documentclass{article}
\usepackage{xparse}
\makeatletter
\NewDocumentCommand{\myTest}{s}{%
  \begingroup% make definitions local
  \let\ies@firstStar=#1%
  \IfBooleanTF{\ies@firstStar}{%
    One star%
  }{%
    No star%
  }%
  \endgroup
}
\makeatother
\begin{document}
\myTest
\myTest*
\end{document}

Answer 3

It's difficult to understand the need for such an indirection.

Anyway, the current implementation of the s argument type is that, if s is the first argument, the value assigned to #1 in case no * comes along is \BooleanFalse, \BooleanTrue if there is a *.

Then the current implementation of \IfBooleanTF{#1}{True}{False} chooses the true or false branch according on whether #1 is \BooleanTrue or \BooleanFalse. Next you might want to know that currently \BooleanTrue and \BooleanFalse are \chardef tokens, so they're not expandable.

But the user/programmer should not rely on this. In future releases this might change; not likely to happen, but, hey! Who knows?

Besides, I can't see any advantage whatsoever in \IfBooleanTF{\ies@FirstStar} over \IfBooleanTF{#1}. To the contrary, I see several disadvantages, for instance that you need grouping to avoid timing expansions or macro clobbering.

Why doesn't your \def work? Because \ies@FirstStar is not the same as \BooleanTrue or \BooleanFalse: it is a macro expanding to either of them (and expl3 uses \ifx for the check, not \if which wouldn't work, so no expansion is performed). If you use, as suggested, \let instead of \def, the code seems to work, but it might stop to with a future release of expl3. On the contrary, \IfBooleanTF{#1} is guaranteed to work forever.