How does cprotect work?

Question

We know that cprotect package can be used to include verbatim content in title/section: \cprotect\section{\verb+123+}.

• How does it work?
• If I want to do something similar that the package cannot do by itself, how can I?

Answer 1

See cprotect source code documentation (on CTAN) for details how it works.

In short:

To understand how \cprotect works, we first have to understand how \protect works.

Recall that \protect ensures that the macro following it remains unexpanded, until it's in a normal context.

For \verb and similar, we need a stronger requirement, that is, the argument following it must not be read (tokenized) until \verb is executed in a normal context.

So to do that, \cprotect essentially (*) transforms

\cprotect\section{\verb+abc+}

to

\section{\protect\input{a-1.cpt}}

after writing \verb+abc+ to the file a-1.cpt.

This works, because \protect ensures that the file fed to \input is only tokenized in a normal context.

Of course, this (as well as \protect) relies on the content being eventually executed in a normal context. If the argument is, for example, passed directly to \detokenize after only being expanded/never expanded at all, it will not result in the final expected value. Example.

(*): actually the result is \protect\input a-1.cpt\relax.