summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/l3kernel/l3tl.dtx
diff options
context:
space:
mode:
Diffstat (limited to 'macros/latex/contrib/l3kernel/l3tl.dtx')
-rw-r--r--macros/latex/contrib/l3kernel/l3tl.dtx3843
1 files changed, 0 insertions, 3843 deletions
diff --git a/macros/latex/contrib/l3kernel/l3tl.dtx b/macros/latex/contrib/l3kernel/l3tl.dtx
deleted file mode 100644
index 0938e7cd5e..0000000000
--- a/macros/latex/contrib/l3kernel/l3tl.dtx
+++ /dev/null
@@ -1,3843 +0,0 @@
-% \iffalse meta-comment
-%
-%% File: l3tl.dtx
-%
-% Copyright (C) 1990-2024 The LaTeX Project
-%
-% It may be distributed and/or modified under the conditions of the
-% LaTeX Project Public License (LPPL), either version 1.3c of this
-% license or (at your option) any later version. The latest version
-% of this license is in the file
-%
-% https://www.latex-project.org/lppl.txt
-%
-% This file is part of the "l3kernel bundle" (The Work in LPPL)
-% and all files in that bundle must be distributed together.
-%
-% -----------------------------------------------------------------------
-%
-% The development version of the bundle can be found at
-%
-% https://github.com/latex3/latex3
-%
-% for those people who are interested.
-%
-%<*driver>
-\documentclass[full,kernel]{l3doc}
-\begin{document}
- \DocInput{\jobname.dtx}
-\end{document}
-%</driver>
-% \fi
-%
-% \title{^^A
-% The \pkg{l3tl} module\\ Token lists^^A
-% }
-%
-% \author{^^A
-% The \LaTeX{} Project\thanks
-% {^^A
-% E-mail:
-% \href{mailto:latex-team@latex-project.org}
-% {latex-team@latex-project.org}^^A
-% }^^A
-% }
-%
-% \date{Released 2024-03-14}
-%
-% \maketitle
-%
-% \begin{documentation}
-%
-% \TeX{} works with tokens, and \LaTeX3 therefore provides a number of
-% functions to deal with lists of tokens. Token lists may be present
-% directly in the argument to a function:
-% \begin{verbatim}
-% \foo:n { a collection of \tokens }
-% \end{verbatim}
-% or may be stored in a so-called \enquote{token list variable}, which
-% have the suffix \texttt{tl}: a token list variable can also be used as
-% the argument to a function, for example
-% \begin{verbatim}
-% \foo:N \l_some_tl
-% \end{verbatim}
-% In both cases, functions are available to test and manipulate the lists
-% of tokens, and these have the module prefix \texttt{tl}.
-% In many cases, functions which can be applied to token list variables
-% are paired with similar functions for application to explicit lists
-% of tokens: the two \enquote{views} of a token list are therefore collected
-% together here.
-%
-% A token list (explicit, or stored in a variable) can be seen either
-% as a list of \enquote{items},
-% or a list of \enquote{tokens}. An item is whatever \cs{use:n} would
-% grab as its argument: a single non-space token or a brace group,
-% with optional leading explicit space characters (each item is thus
-% itself a token list). A token is either a normal \texttt{N} argument,
-% or \verb*| |, |{|, or |}| (assuming normal \TeX{} category codes).
-% Thus for example
-% \begin{verbatim}
-% { Hello } ~ world
-% \end{verbatim}
-% contains six items (\texttt{Hello}, \texttt{w}, \texttt{o}, \texttt{r},
-% \texttt{l} and \texttt{d}), but thirteen tokens (|{|, \texttt{H}, \texttt{e},
-% \texttt{l}, \texttt{l}, \texttt{o}, |}|, \verb*| |, \texttt{w}, \texttt{o},
-% \texttt{r}, \texttt{l} and \texttt{d}).
-% Functions which act on items are often faster than their analogue acting
-% directly on tokens.
-%
-% \section{Creating and initialising token list variables}
-%
-% \begin{function}{\tl_new:N, \tl_new:c}
-% \begin{syntax}
-% \cs{tl_new:N} \meta{tl~var}
-% \end{syntax}
-% Creates a new \meta{tl~var} or raises an error if the
-% name is already taken. The declaration is global. The
-% \meta{tl~var} is initially empty.
-% \end{function}
-%
-% \begin{function}
-% {
-% \tl_const:Nn, \tl_const:Ne,
-% \tl_const:cn, \tl_const:ce
-% }
-% \begin{syntax}
-% \cs{tl_const:Nn} \meta{tl~var} \Arg{tokens}
-% \end{syntax}
-% Creates a new constant \meta{tl~var} or raises an error
-% if the name is already taken. The value of the
-% \meta{tl~var} is set globally to the \meta{tokens}.
-% \end{function}
-%
-% \begin{function}{\tl_clear:N, \tl_clear:c, \tl_gclear:N, \tl_gclear:c}
-% \begin{syntax}
-% \cs{tl_clear:N} \meta{tl~var}
-% \end{syntax}
-% Clears all entries from the \meta{tl~var}.
-% \end{function}
-%
-% \begin{function}
-% {\tl_clear_new:N, \tl_clear_new:c, \tl_gclear_new:N, \tl_gclear_new:c}
-% \begin{syntax}
-% \cs{tl_clear_new:N} \meta{tl~var}
-% \end{syntax}
-% Ensures that the \meta{tl~var} exists globally by applying
-% \cs{tl_new:N} if necessary, then applies \cs[index=tl_clear:N]{tl_(g)clear:N} to leave
-% the \meta{tl~var} empty.
-% \end{function}
-%
-% \begin{function}
-% {
-% \tl_set_eq:NN, \tl_set_eq:cN, \tl_set_eq:Nc, \tl_set_eq:cc,
-% \tl_gset_eq:NN, \tl_gset_eq:cN, \tl_gset_eq:Nc, \tl_gset_eq:cc
-% }
-% \begin{syntax}
-% \cs{tl_set_eq:NN} \meta{tl~var_1} \meta{tl~var_2}
-% \end{syntax}
-% Sets the content of \meta{tl~var_1} equal to that of
-% \meta{tl~var_2}.
-% \end{function}
-%
-% \begin{function}[added = 2012-05-18]
-% {
-% \tl_concat:NNN, \tl_concat:ccc,
-% \tl_gconcat:NNN, \tl_gconcat:ccc
-% }
-% \begin{syntax}
-% \cs{tl_concat:NNN} \meta{tl~var_1} \meta{tl~var_2} \meta{tl~var_3}
-% \end{syntax}
-% Concatenates the content of \meta{tl~var_2} and \meta{tl~var_3}
-% together and saves the result in \meta{tl~var_1}. The \meta{tl~var_2}
-% is placed at the left side of the new token list.
-% \end{function}
-%
-% \begin{function}[EXP, pTF, added=2012-03-03]{\tl_if_exist:N, \tl_if_exist:c}
-% \begin{syntax}
-% \cs{tl_if_exist_p:N} \meta{tl~var}
-% \cs{tl_if_exist:NTF} \meta{tl~var} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests whether the \meta{tl~var} is currently defined. This does not
-% check that the \meta{tl~var} really is a token list variable.
-% \end{function}
-%
-% \section{Adding data to token list variables}
-%
-% \begin{function}
-% {
-% \tl_set:Nn, \tl_set:NV, \tl_set:Nv, \tl_set:No, \tl_set:Ne, \tl_set:Nf,
-% \tl_set:cn, \tl_set:cV, \tl_set:cv, \tl_set:co, \tl_set:ce, \tl_set:cf,
-% \tl_gset:Nn, \tl_gset:NV, \tl_gset:Nv,
-% \tl_gset:No, \tl_gset:Ne, \tl_gset:Nf,
-% \tl_gset:cn, \tl_gset:cV, \tl_gset:cv,
-% \tl_gset:co, \tl_gset:ce, \tl_gset:cf
-% }
-% \begin{syntax}
-% \cs{tl_set:Nn} \meta{tl~var} \Arg{tokens}
-% \end{syntax}
-% Sets \meta{tl~var} to contain \meta{tokens},
-% removing any previous content from the variable.
-% \end{function}
-%
-% \begin{function}
-% {
-% \tl_put_left:Nn, \tl_put_left:NV, \tl_put_left:Nv, \tl_put_left:Ne,
-% \tl_put_left:No,
-% \tl_put_left:cn, \tl_put_left:cV, \tl_put_left:cv, \tl_put_left:ce,
-% \tl_put_left:co,
-% \tl_gput_left:Nn, \tl_gput_left:NV, \tl_gput_left:Nv, \tl_gput_left:Ne,
-% \tl_gput_left:No,
-% \tl_gput_left:cn, \tl_gput_left:cV, \tl_gput_left:cv, \tl_gput_left:ce,
-% \tl_gput_left:co
-% }
-% \begin{syntax}
-% \cs{tl_put_left:Nn} \meta{tl~var} \Arg{tokens}
-% \end{syntax}
-% Appends \meta{tokens} to the left side of the current content of
-% \meta{tl~var}.
-% \end{function}
-%
-% \begin{function}
-% {
-% \tl_put_right:Nn, \tl_put_right:NV, \tl_put_right:Nv, \tl_put_right:Ne,
-% \tl_put_right:No,
-% \tl_put_right:cn, \tl_put_right:cV, \tl_put_right:cv, \tl_put_right:ce,
-% \tl_put_right:co,
-% \tl_gput_right:Nn, \tl_gput_right:NV, \tl_gput_right:Nv, \tl_gput_right:Ne,
-% \tl_gput_right:No,
-% \tl_gput_right:cn, \tl_gput_right:cV, \tl_gput_right:cv, \tl_gput_right:ce,
-% \tl_gput_right:co
-% }
-% \begin{syntax}
-% \cs{tl_put_right:Nn} \meta{tl~var} \Arg{tokens}
-% \end{syntax}
-% Appends \meta{tokens} to the right side of the current content of
-% \meta{tl~var}.
-% \end{function}
-%
-% \section{Token list conditionals}
-%
-% \begin{function}[EXP,pTF, updated = 2019-09-04]
-% {\tl_if_blank:n, \tl_if_blank:e, \tl_if_blank:V, \tl_if_blank:o}
-% \begin{syntax}
-% \cs{tl_if_blank_p:n} \Arg{token list}
-% \cs{tl_if_blank:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{token list} consists only of blank spaces
-% (\emph{i.e.}~contains no item). The test is \texttt{true} if
-% \meta{token list} is zero or more explicit space characters
-% (explicit tokens with character code~$32$ and category code~$10$),
-% and is \texttt{false} otherwise.
-% \end{function}
-%
-% \begin{function}[EXP,pTF]{\tl_if_empty:N, \tl_if_empty:c}
-% \begin{syntax}
-% \cs{tl_if_empty_p:N} \meta{tl~var}
-% \cs{tl_if_empty:NTF} \meta{tl~var} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{tl~var} is entirely empty
-% (\emph{i.e.}~contains no tokens at all).
-% \end{function}
-%
-% \begin{function}[added = 2012-05-24, updated = 2012-06-05, EXP,pTF]
-% {\tl_if_empty:n, \tl_if_empty:V, \tl_if_empty:o, \tl_if_empty:e}
-% \begin{syntax}
-% \cs{tl_if_empty_p:n} \Arg{token list}
-% \cs{tl_if_empty:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{token list} is entirely empty
-% (\emph{i.e.}~contains no tokens at all).
-% \end{function}
-%
-% \begin{function}[EXP,pTF]
-% {\tl_if_eq:NN, \tl_if_eq:Nc, \tl_if_eq:cN, \tl_if_eq:cc}
-% \begin{syntax}
-% \cs{tl_if_eq_p:NN} \meta{tl~var_1} \meta{tl~var_2}
-% \cs{tl_if_eq:NNTF} \meta{tl~var_1} \meta{tl~var_2} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Compares the content of \meta{tl~var_1} and \meta{tl~var_2} and
-% is logically \texttt{true} if the two contain the same list of
-% tokens (\emph{i.e.}~identical in both the list of characters they
-% contain and the category codes of those characters). Thus for example
-% \begin{verbatim}
-% \tl_set:Nn \l_tmpa_tl { abc }
-% \tl_set:Ne \l_tmpb_tl { \tl_to_str:n { abc } }
-% \tl_if_eq:NNTF \l_tmpa_tl \l_tmpb_tl { true } { false }
-% \end{verbatim}
-% yields \texttt{false}.
-% See also \cs{str_if_eq:nnTF} for a comparison that ignores category codes.
-% \end{function}
-%
-% \begin{function}[TF, added = 2020-07-14]{\tl_if_eq:Nn, \tl_if_eq:cn}
-% \begin{syntax}
-% \cs{tl_if_eq:NnTF} \meta{tl~var_1} \Arg{token list_2} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{tl~var_1} and the \meta{token
-% list_2} contain the same list of tokens, both in respect of
-% character codes and category codes. This conditional is not
-% expandable: see \cs{tl_if_eq:NNTF} for an expandable version when
-% both token lists are stored in variables, or \cs{str_if_eq:nnTF} if
-% category codes are not important.
-% \end{function}
-%
-% \begin{function}[TF]
-% {
-% \tl_if_eq:nn, \tl_if_eq:nV, \tl_if_eq:ne, \tl_if_eq:Vn, \tl_if_eq:en,
-% \tl_if_eq:ee
-% }
-% \begin{syntax}
-% \cs{tl_if_eq:nnTF} \Arg{token list_1} \Arg{token list_2} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if \meta{token list_1} and \meta{token list_2} contain the
-% same list of tokens, both in respect of character codes and category
-% codes. This conditional is not expandable: see \cs{tl_if_eq:NNTF}
-% for an expandable version when token lists are stored in variables,
-% or \cs{str_if_eq:nnTF} if category codes are not important.
-% \end{function}
-%
-% \begin{function}[TF]
-% {
-% \tl_if_in:Nn, \tl_if_in:NV, \tl_if_in:No,
-% \tl_if_in:cn, \tl_if_in:cV, \tl_if_in:co
-% }
-% \begin{syntax}
-% \cs{tl_if_in:NnTF} \meta{tl~var} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{token list} is found in the content of the
-% \meta{tl~var}. The \meta{token list} cannot contain
-% the tokens |{|, |}| or |#|
-% (more precisely, explicit character tokens with category code $1$
-% (begin-group) or $2$ (end-group), and tokens with category code $6$).
-% \end{function}
-%
-% \begin{function}[TF]
-% {
-% \tl_if_in:nn, \tl_if_in:Vn, \tl_if_in:VV, \tl_if_in:on, \tl_if_in:oo,
-% \tl_if_in:nV, \tl_if_in:no
-% }
-% \begin{syntax}
-% \cs{tl_if_in:nnTF} \Arg{token list_1} \Arg{token list_2} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if \meta{token list_2} is found inside \meta{token list_1}.
-% The \meta{token list_2} cannot contain the tokens |{|, |}| or |#|
-% (more precisely, explicit character tokens with category code $1$
-% (begin-group) or $2$ (end-group), and tokens with category code $6$).
-% The search does \emph{not} enter brace (category code $1$/$2$) groups.
-% \end{function}
-%
-% \begin{function}[added = 2017-11-14, EXP,pTF]{\tl_if_novalue:n}
-% \begin{syntax}
-% \cs{tl_if_novalue_p:n} \Arg{token list}
-% \cs{tl_if_novalue:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{token list} and the special \cs{c_novalue_tl} marker
-% contain the same list of tokens, both in respect of character codes and
-% category codes. This means that
-% \cs{exp_args:No} \cs{tl_if_novalue:nTF} \{ \cs{c_novalue_tl} \} is
-% logically \texttt{true} but \cs{tl_if_novalue:nTF} \{ \cs{c_novalue_tl} \}
-% is logically \texttt{false}.
-% This function is intended to allow construction
-% of flexible document interface structures in which missing optional
-% arguments are detected.
-% \end{function}
-%
-% \begin{function}[updated = 2011-08-13, EXP,pTF]
-% {\tl_if_single:N, \tl_if_single:c}
-% \begin{syntax}
-% \cs{tl_if_single_p:N} \meta{tl~var}
-% \cs{tl_if_single:NTF} \meta{tl~var} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the content of the \meta{tl~var} consists of a single \meta{item},
-% \emph{i.e.}~is a single normal token (neither an explicit space
-% character nor a begin-group character) or a single brace group,
-% surrounded by optional spaces on both sides. In other words, such a
-% token list has token count $1$ according to \cs{tl_count:N}.
-% \end{function}
-%
-% \begin{function}[updated = 2011-08-13, EXP,pTF]{\tl_if_single:n}
-% \begin{syntax}
-% \cs{tl_if_single_p:n} \Arg{token list}
-% \cs{tl_if_single:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the \meta{token list} has exactly one \meta{item}, \emph{i.e.}~is
-% a single normal token (neither an explicit space character nor a
-% begin-group character) or a single brace group, surrounded by
-% optional spaces on both sides. In other words, such a token list has
-% token count $1$ according to \cs{tl_count:n}.
-% \end{function}
-%
-% \begin{function}[EXP,pTF]{\tl_if_single_token:n}
-% \begin{syntax}
-% \cs{tl_if_single_token_p:n} \Arg{token list}
-% \cs{tl_if_single_token:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the token list consists of exactly one token, \emph{i.e.}~is
-% either a single space character or a single normal token.
-% Token groups (|{|\ldots|}|) are not single tokens.
-% \end{function}
-%
-% \subsection{Testing the first token}
-%
-% \begin{function}[updated = 2012-07-09, EXP, pTF]
-% {
-% \tl_if_head_eq_catcode:nN, \tl_if_head_eq_catcode:VN,
-% \tl_if_head_eq_catcode:eN, \tl_if_head_eq_catcode:oN
-% }
-% \begin{syntax}
-% \cs{tl_if_head_eq_catcode_p:nN} \Arg{token list} \meta{test token}
-% \cs{tl_if_head_eq_catcode:nNTF} \Arg{token list} \meta{test token}
-% ~~\Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the first \meta{token} in the \meta{token list} has the
-% same category code as the \meta{test token}. In the case where the
-% \meta{token list} is empty, the test is always \texttt{false}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-07-09, EXP, pTF]
-% {
-% \tl_if_head_eq_charcode:nN, \tl_if_head_eq_charcode:VN,
-% \tl_if_head_eq_charcode:eN, \tl_if_head_eq_charcode:fN
-% }
-% \begin{syntax}
-% \cs{tl_if_head_eq_charcode_p:nN} \Arg{token list} \meta{test token}
-% \cs{tl_if_head_eq_charcode:nNTF} \Arg{token list} \meta{test token}
-% ~~\Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the first \meta{token} in the \meta{token list} has the
-% same character code as the \meta{test token}. In the case where the
-% \meta{token list} is empty, the test is always \texttt{false}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-07-09, EXP, pTF]
-% {
-% \tl_if_head_eq_meaning:nN, \tl_if_head_eq_meaning:VN,
-% \tl_if_head_eq_meaning:eN
-% }
-% \begin{syntax}
-% \cs{tl_if_head_eq_meaning_p:nN} \Arg{token list} \meta{test token}
-% \cs{tl_if_head_eq_meaning:nNTF} \Arg{token list} \meta{test token}
-% ~~\Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the first \meta{token} in the \meta{token list} has the
-% same meaning as the \meta{test token}. In the case where
-% \meta{token list} is empty, the test is always \texttt{false}.
-% \end{function}
-%
-% \begin{function}[added = 2012-07-08, EXP, pTF]{\tl_if_head_is_group:n}
-% \begin{syntax}
-% \cs{tl_if_head_is_group_p:n} \Arg{token list}
-% \cs{tl_if_head_is_group:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the first \meta{token} in the \meta{token list}
-% is an explicit begin-group character (with category code~$1$
-% and any character code), in other words, if the \meta{token list}
-% starts with a brace group. In particular, the test is \texttt{false}
-% if the \meta{token list} starts with an implicit token such as
-% \cs{c_group_begin_token}, or if it is empty.
-% This function is useful to implement actions on token lists on
-% a token by token basis.
-% \end{function}
-%
-% \begin{function}[added = 2012-07-08, EXP, pTF]{\tl_if_head_is_N_type:n}
-% \begin{syntax}
-% \cs{tl_if_head_is_N_type_p:n} \Arg{token list}
-% \cs{tl_if_head_is_N_type:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the first \meta{token} in the \meta{token list}
-% is a normal \texttt{N}-type argument. In other words,
-% it is neither an explicit space character
-% (explicit token with character code~$32$ and category code~$10$)
-% nor an explicit begin-group character
-% (with category code~1 and any character code). An empty
-% argument yields \texttt{false}, as it does not have a normal
-% first token.
-% This function is useful to implement actions on token lists on
-% a token by token basis.
-% \end{function}
-%
-% \begin{function}[updated = 2012-07-08, EXP, pTF]{\tl_if_head_is_space:n}
-% \begin{syntax}
-% \cs{tl_if_head_is_space_p:n} \Arg{token list}
-% \cs{tl_if_head_is_space:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
-% \end{syntax}
-% Tests if the first \meta{token} in the \meta{token list}
-% is an explicit space character
-% (explicit token with character code~$32$ and category code~$10$).
-% In particular, the test is \texttt{false} if the \meta{token list}
-% starts with an implicit token such as \cs{c_space_token}, or if it
-% is empty.
-% This function is useful to implement actions on token lists on
-% a token by token basis.
-% \end{function}
-%
-% \section{Working with token lists as a whole}
-%
-% \subsection{Using token lists}
-%
-% \begin{function}[EXP]
-% {\tl_to_str:n, \tl_to_str:o, \tl_to_str:V, \tl_to_str:v, \tl_to_str:e}
-% \begin{syntax}
-% \cs{tl_to_str:n} \Arg{token list}
-% \end{syntax}
-% Converts the \meta{token list} to a \meta{string}, leaving the
-% resulting character tokens in the input stream. A \meta{string}
-% is a series of tokens with category code $12$ (other) with the exception
-% of spaces, which retain category code $10$ (space).
-% The base function requires only a single expansion.
-% Its argument \emph{must} be braced.
-% \begin{texnote}
-% This is the \eTeX{} primitive \tn{detokenize}.
-% Converting a \meta{token list} to a \meta{string} yields a
-% concatenation of the string representations of every token in the
-% \meta{token list}.
-% The string representation of a control sequence is
-% \begin{itemize}
-% \item an escape character, whose character code is given by the
-% internal parameter \tn{escapechar}, absent if the
-% \tn{escapechar} is negative or greater than the largest
-% character code;
-% \item the control sequence name, as defined by \cs{cs_to_str:N};
-% \item a space, unless the control sequence name is a single
-% character whose category at the time of expansion of
-% \cs{tl_to_str:n} is not \enquote{letter}.
-% \end{itemize}
-% The string representation of an explicit character token is that
-% character, doubled in the case of (explicit) macro parameter
-% characters (normally |#|).
-% In particular, the string representation of a token list may
-% depend on the category codes in effect when it is evaluated, and
-% the value of the \tn{escapechar}: for instance |\tl_to_str:n {\a}|
-% normally produces the three character \enquote{backslash},
-% \enquote{lower-case a}, \enquote{space}, but it may also produce a
-% single \enquote{lower-case a} if the escape character is negative
-% and \texttt{a} is currently not a letter.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[EXP]{\tl_to_str:N, \tl_to_str:c}
-% \begin{syntax}
-% \cs{tl_to_str:N} \meta{tl~var}
-% \end{syntax}
-% Converts the content of the \meta{tl~var} into a series of characters
-% with category code $12$ (other) with the exception of spaces, which
-% retain category code $10$ (space). This \meta{string} is then left
-% in the input stream. For low-level details, see the notes given for
-% \cs{tl_to_str:n}.
-% \end{function}
-%
-% \begin{function}[EXP]{\tl_use:N, \tl_use:c}
-% \begin{syntax}
-% \cs{tl_use:N} \meta{tl~var}
-% \end{syntax}
-% Recovers the content of a \meta{tl~var} and places it
-% directly in the input stream. An error is raised if the variable
-% does not exist or if it is invalid. Note that it is possible to use
-% a \meta{tl~var} directly without an accessor function.
-% \end{function}
-%
-% \subsection{Counting and reversing token lists}
-%
-% \begin{function}[added = 2012-05-13, EXP]
-% {\tl_count:n, \tl_count:V, \tl_count:v, \tl_count:e, \tl_count:o}
-% \begin{syntax}
-% \cs{tl_count:n} \Arg{token list}
-% \end{syntax}
-% Counts the number of \meta{items} in the \meta{token list} and leaves this
-% information in the input stream. Unbraced tokens count as one
-% element as do each token group (|{|\ldots|}|). This process
-% ignores any unprotected spaces within the \meta{token list}. See also
-% \cs{tl_count:N}. This function requires three expansions,
-% giving an \meta{integer denotation}.
-% \end{function}
-%
-% \begin{function}[added = 2012-05-13, EXP]{\tl_count:N, \tl_count:c}
-% \begin{syntax}
-% \cs{tl_count:N} \meta{tl~var}
-% \end{syntax}
-% Counts the number of \meta{items} in the \meta{tl~var}
-% and leaves this information in the input stream. Unbraced tokens
-% count as one element as do each token group (|{|\ldots|}|). This
-% process ignores any unprotected spaces within the \meta{tl~var}.
-% See also \cs{tl_count:n}. This function requires three expansions,
-% giving an \meta{integer denotation}.
-% \end{function}
-%
-% \begin{function}[EXP, added = 2019-02-25]{\tl_count_tokens:n}
-% \begin{syntax}
-% \cs{tl_count_tokens:n} \Arg{token list}
-% \end{syntax}
-% Counts the number of \TeX{} tokens in the \meta{token list} and leaves
-% this information in the input stream. Every token, including spaces and
-% braces, contributes one to the total; thus for instance, the token count of
-% |a~{bc}| is $6$.
-% \end{function}
-%
-% \begin{function}[updated = 2012-01-08, EXP]
-% {\tl_reverse:n, \tl_reverse:V, \tl_reverse:o, \tl_reverse:f, \tl_reverse:e}
-% \begin{syntax}
-% \cs{tl_reverse:n} \Arg{token list}
-% \end{syntax}
-% Reverses the order of the \meta{items} in the \meta{token list},
-% so that \meta{item_1}\meta{item_2}\meta{item_3} \ldots \meta{item_n}
-% becomes \meta{item_n}\ldots \meta{item_3}\meta{item_2}\meta{item_1}.
-% This process preserves unprotected space within the
-% \meta{token list}. Tokens are not reversed within braced token
-% groups, which keep their outer set of braces.
-% In situations where performance is important,
-% consider \cs{tl_reverse_items:n}.
-% See also \cs{tl_reverse:N}.
-% \begin{texnote}
-% The result is returned within \tn{unexpanded}, which means that the token
-% list does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[updated = 2012-01-08]
-% {\tl_reverse:N, \tl_reverse:c, \tl_greverse:N, \tl_greverse:c}
-% \begin{syntax}
-% \cs{tl_reverse:N} \meta{tl~var}
-% \end{syntax}
-% Sets the \meta{tl~var} to contain the result of reversing
-% the order of its \meta{items}, so
-% that \meta{item_1}\meta{item_2}\meta{item_3} \ldots \meta{item_n}
-% becomes \meta{item_n}\ldots \meta{item_3}\meta{item_2}\meta{item_1}.
-% This process preserves unprotected spaces within the
-% \meta{tl~var}. Braced token groups are copied without
-% reversing the order of tokens, but keep the outer set of braces.
-% This is equivalent to a combination of an assignment and
-% \cs{tl_reverse:V}. See also \cs{tl_reverse_items:n} for improved
-% performance.
-% \end{function}
-%
-% \begin{function}[added = 2012-01-08, EXP]{\tl_reverse_items:n}
-% \begin{syntax}
-% \cs{tl_reverse_items:n} \Arg{token list}
-% \end{syntax}
-% Reverses the order of the \meta{items} in the \meta{token list},
-% so that \meta{item_1}\meta{item_2}\meta{item_3} \ldots \meta{item_n}
-% becomes \Arg{item_n} \ldots{} \Arg{item_3}\Arg{item_2}\Arg{item_1}.
-% This process removes any unprotected space within the
-% \meta{token list}. Braced token groups are copied without
-% reversing the order of tokens, and keep the outer set of braces.
-% Items which are initially not braced are copied with braces in
-% the result. In cases where preserving spaces is important,
-% consider the slower function \cs{tl_reverse:n}.
-% \begin{texnote}
-% The result is returned within \tn{unexpanded}, which means that the token
-% list does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[added = 2011-07-09, updated = 2012-06-25, EXP]
-% {
-% \tl_trim_spaces:n, \tl_trim_spaces:V, \tl_trim_spaces:v,
-% \tl_trim_spaces:e,
-% \tl_trim_spaces:o
-% }
-% \begin{syntax}
-% \cs{tl_trim_spaces:n} \Arg{token list}
-% \end{syntax}
-% Removes any leading and trailing explicit space characters
-% (explicit tokens with character code~$32$ and category code~$10$)
-% from the \meta{token list} and leaves the result in the input
-% stream.
-% \begin{texnote}
-% The result is returned within \tn{unexpanded}, which means that the token
-% list does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[added = 2018-04-12, EXP]
-% {\tl_trim_spaces_apply:nN, \tl_trim_spaces_apply:oN}
-% \begin{syntax}
-% \cs{tl_trim_spaces_apply:nN} \Arg{token list} \meta{function}
-% \end{syntax}
-% Removes any leading and trailing explicit space characters (explicit
-% tokens with character code~$32$ and category code~$10$) from the
-% \meta{token list} and passes the result to the \meta{function} as an
-% \texttt{n}-type argument.
-% \end{function}
-%
-% \begin{function}[added = 2011-07-09]
-% {
-% \tl_trim_spaces:N, \tl_trim_spaces:c,
-% \tl_gtrim_spaces:N, \tl_gtrim_spaces:c
-% }
-% \begin{syntax}
-% \cs{tl_trim_spaces:N} \meta{tl~var}
-% \end{syntax}
-% Sets the \meta{tl~var} to contain the result of removing any leading
-% and trailing explicit space characters (explicit tokens with
-% character code~$32$ and category code~$10$) from its contents.
-% \end{function}
-%
-% \subsection{Viewing token lists}
-%
-% \begin{function}[updated = 2021-04-29]{\tl_show:N, \tl_show:c}
-% \begin{syntax}
-% \cs{tl_show:N} \meta{tl~var}
-% \end{syntax}
-% Displays the content of the \meta{tl~var} on the terminal.
-% \begin{texnote}
-% This is similar to the \TeX{} primitive \tn{show}, wrapped to a
-% fixed number of characters per line.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[updated = 2015-08-07]{\tl_show:n, \tl_show:e}
-% \begin{syntax}
-% \cs{tl_show:n} \Arg{token list}
-% \end{syntax}
-% Displays the \meta{token list} on the terminal.
-% \begin{texnote}
-% This is similar to the \eTeX{} primitive \tn{showtokens}, wrapped
-% to a fixed number of characters per line.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[added = 2014-08-22, updated = 2021-04-29]{\tl_log:N, \tl_log:c}
-% \begin{syntax}
-% \cs{tl_log:N} \meta{tl~var}
-% \end{syntax}
-% Writes the content of the \meta{tl~var} in the log file. See also
-% \cs{tl_show:N} which displays the result in the terminal.
-% \end{function}
-%
-% \begin{function}[added = 2014-08-22, updated = 2015-08-07]
-% {\tl_log:n, \tl_log:e, \tl_log:x}
-% \begin{syntax}
-% \cs{tl_log:n} \Arg{token list}
-% \end{syntax}
-% Writes the \meta{token list} in the log file. See also
-% \cs{tl_show:n} which displays the result in the terminal.
-% \end{function}
-%
-% \section{Manipulating items in token lists}
-%
-% \subsection{Mapping over token lists}
-%
-% All mappings are done at the current group level, \emph{i.e.}~any
-% local assignments made by the \meta{function} or \meta{code} discussed
-% below remain in effect after the loop.
-%
-% \begin{function}[updated = 2012-06-29, rEXP]
-% {\tl_map_function:NN, \tl_map_function:cN}
-% \begin{syntax}
-% \cs{tl_map_function:NN} \meta{tl~var} \meta{function}
-% \end{syntax}
-% Applies \meta{function} to every \meta{item} in the \meta{tl~var}.
-% The \meta{function} receives one argument for each iteration.
-% This may be a number of tokens if the \meta{item} was stored within
-% braces. Hence the \meta{function} should anticipate receiving
-% \texttt{n}-type arguments. See also \cs{tl_map_function:nN}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29, rEXP]{\tl_map_function:nN}
-% \begin{syntax}
-% \cs{tl_map_function:nN} \Arg{token list} \meta{function}
-% \end{syntax}
-% Applies \meta{function} to every \meta{item} in the \meta{token list},
-% The \meta{function} receives one argument for each iteration.
-% This may be a number of tokens if the \meta{item} was stored within
-% braces. Hence the \meta{function} should anticipate receiving
-% \texttt{n}-type arguments. See also \cs{tl_map_function:NN}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29]
-% {\tl_map_inline:Nn, \tl_map_inline:cn}
-% \begin{syntax}
-% \cs{tl_map_inline:Nn} \meta{tl~var} \Arg{inline function}
-% \end{syntax}
-% Applies the \meta{inline function} to every \meta{item} stored within the
-% \meta{tl~var}. The \meta{inline function} should consist of code which
-% receives the \meta{item} as |#1|. See also \cs{tl_map_function:NN}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29]{\tl_map_inline:nn}
-% \begin{syntax}
-% \cs{tl_map_inline:nn} \Arg{token list} \Arg{inline function}
-% \end{syntax}
-% Applies the \meta{inline function} to every \meta{item} stored within the
-% \meta{token list}. The \meta{inline function} should consist of code which
-% receives the \meta{item} as |#1|. See also \cs{tl_map_function:nN}.
-% \end{function}
-%
-% \begin{function}[rEXP, added = 2019-09-02]
-% {\tl_map_tokens:Nn, \tl_map_tokens:cn, \tl_map_tokens:nn}
-% \begin{syntax}
-% \cs{tl_map_tokens:Nn} \meta{tl~var} \Arg{code}
-% \cs{tl_map_tokens:nn} \Arg{token list} \Arg{code}
-% \end{syntax}
-% Analogue of \cs{tl_map_function:NN} which maps several tokens
-% instead of a single function. The \meta{code} receives each \meta{item} in
-% the \meta{tl~var} or in the \meta{token list} as a trailing brace group. For
-% instance,
-% \begin{verbatim}
-% \tl_map_tokens:Nn \l_my_tl { \prg_replicate:nn { 2 } }
-% \end{verbatim}
-% expands to twice each \meta{item} in the \meta{tl~var}: for each \meta{item} in
-% \cs[no-index]{l_my_tl} the function \cs{prg_replicate:nn} receives |2| and
-% \meta{item} as its two arguments. The function
-% \cs{tl_map_inline:Nn} is typically faster but is not expandable.
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29]
-% {\tl_map_variable:NNn, \tl_map_variable:cNn}
-% \begin{syntax}
-% \cs{tl_map_variable:NNn} \meta{tl~var} \meta{variable} \Arg{code}
-% \end{syntax}
-% Stores each \meta{item} of the \meta{tl~var} in turn in the (token
-% list) \meta{variable} and applies the \meta{code}. The \meta{code}
-% will usually make use of the \meta{variable}, but this is not
-% enforced. The assignments to the \meta{variable} are local. Its
-% value after the loop is the last \meta{item} in the \meta{tl~var},
-% or its original value if the \meta{tl~var} is blank. See also
-% \cs{tl_map_inline:Nn}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29]{\tl_map_variable:nNn}
-% \begin{syntax}
-% \cs{tl_map_variable:nNn} \Arg{token list} \meta{variable} \Arg{code}
-% \end{syntax}
-% Stores each \meta{item} of the \meta{token list} in turn in the
-% (token list) \meta{variable} and applies the \meta{code}. The
-% \meta{code} will usually make use of the \meta{variable}, but this
-% is not enforced. The assignments to the \meta{variable} are local.
-% Its value after the loop is the last \meta{item} in the
-% \meta{tl~var}, or its original value if the \meta{tl~var} is blank.
-% See also \cs{tl_map_inline:nn}.
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29, rEXP]{\tl_map_break:}
-% \begin{syntax}
-% \cs{tl_map_break:}
-% \end{syntax}
-% Used to terminate a \cs[no-index]{tl_map_\ldots} function before all
-% entries in the \meta{token list} have been processed. This
-% normally takes place within a conditional statement, for example
-% \begin{verbatim}
-% \tl_map_inline:Nn \l_my_tl
-% {
-% \str_if_eq:nnT { #1 } { bingo } { \tl_map_break: }
-% % Do something useful
-% }
-% \end{verbatim}
-% See also \cs{tl_map_break:n}.
-% Use outside of a \cs[no-index]{tl_map_\ldots} scenario leads to low
-% level \TeX{} errors.
-% \begin{texnote}
-% When the mapping is broken, additional tokens may be inserted
-% before further items are taken
-% from the input stream. This depends on the design of the mapping
-% function.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[updated = 2012-06-29, rEXP]{\tl_map_break:n}
-% \begin{syntax}
-% \cs{tl_map_break:n} \Arg{code}
-% \end{syntax}
-% Used to terminate a \cs[no-index]{tl_map_\ldots} function before all
-% entries in the \meta{token list} have been processed, inserting
-% the \meta{code} after the mapping has ended. This
-% normally takes place within a conditional statement, for example
-% \begin{verbatim}
-% \tl_map_inline:Nn \l_my_tl
-% {
-% \str_if_eq:nnT { #1 } { bingo }
-% { \tl_map_break:n { <code> } }
-% % Do something useful
-% }
-% \end{verbatim}
-% Use outside of a \cs[no-index]{tl_map_\ldots} scenario leads to low
-% level \TeX{} errors.
-% \begin{texnote}
-% When the mapping is broken, additional tokens may be inserted
-% before the \meta{code} is
-% inserted into the input stream.
-% This depends on the design of the mapping function.
-% \end{texnote}
-% \end{function}
-%
-% \subsection{Head and tail of token lists}
-%
-% Functions which deal with either only the very first item (balanced
-% text or single normal token) in a token list, or the remaining tokens.
-%
-% \begin{function}[updated = 2012-09-09, EXP]
-% {\tl_head:N, \tl_head:n, \tl_head:V, \tl_head:v, \tl_head:f}
-% \begin{syntax}
-% \cs{tl_head:n} \Arg{token list}
-% \end{syntax}
-% Leaves in the input stream the first \meta{item} in the
-% \meta{token list}, discarding the rest of the \meta{token list}.
-% All leading explicit space characters
-% (explicit tokens with character code~$32$ and category code~$10$)
-% are discarded; for example
-% \begin{verbatim}
-% \tl_head:n { abc }
-% \end{verbatim}
-% and
-% \begin{verbatim}
-% \tl_head:n { ~ abc }
-% \end{verbatim}
-% both leave |a| in the input stream. If the \enquote{head} is a
-% brace group, rather than a single token, the braces are removed, and
-% so
-% \begin{verbatim}
-% \tl_head:n { ~ { ~ ab } c }
-% \end{verbatim}
-% yields \verb*| ab|.
-% A blank \meta{token list} (see \cs{tl_if_blank:nTF}) results in
-% \cs{tl_head:n} leaving nothing in the input stream.
-% \begin{texnote}
-% The result is returned within \cs{exp_not:n}, which means that the token
-% list does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[EXP]{\tl_head:w}
-% \begin{syntax}
-% \cs{tl_head:w} \meta{token list} | { } | \cs{q_stop}
-% \end{syntax}
-% Leaves in the input stream the first \meta{item} in the
-% \meta{token list}, discarding the rest of the \meta{token list}.
-% All leading explicit space characters
-% (explicit tokens with character code~$32$ and category code~$10$)
-% are discarded.
-% A blank \meta{token list} (which consists only of space characters)
-% results in a low-level \TeX{} error, which may be avoided by the
-% inclusion of an empty group in the input (as shown), without the need
-% for an explicit test. Alternatively, \cs{tl_if_blank:nF} may be used to
-% avoid using the function with a \enquote{blank} argument.
-% This function requires only a single expansion, and thus is suitable for
-% use within an \texttt{o}-type expansion. In general, \cs{tl_head:n} should
-% be preferred if the number of expansions is not critical.
-% \end{function}
-%
-% \begin{function}[updated = 2012-09-01, EXP]
-% {\tl_tail:N, \tl_tail:n, \tl_tail:V, \tl_tail:v, \tl_tail:f}
-% \begin{syntax}
-% \cs{tl_tail:n} \Arg{token list}
-% \end{syntax}
-% Discards all leading explicit space characters
-% (explicit tokens with character code~$32$ and category code~$10$)
-% and the first \meta{item} in the \meta{token list}, and leaves the
-% remaining tokens in the input stream. Thus for example
-% \begin{verbatim}
-% \tl_tail:n { a ~ {bc} d }
-% \end{verbatim}
-% and
-% \begin{verbatim}
-% \tl_tail:n { ~ a ~ {bc} d }
-% \end{verbatim}
-% both leave \verb*| {bc}d| in the input stream. A blank
-% \meta{token list} (see \cs{tl_if_blank:nTF}) results
-% in \cs{tl_tail:n} leaving nothing in the input stream.
-% \begin{texnote}
-% The result is returned within \cs{exp_not:n}, which means that the
-% token list does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% If you wish to handle token lists where the first token may be a space, and
-% this needs to be treated as the head/tail, this can be accomplished using
-% \cs{tl_if_head_is_space:nTF}, for example
-% \begin{verbatim}
-% \exp_last_unbraced:NNo
-% \cs_new:Npn \__mypkg_gobble_space:w \c_space_tl { }
-% \cs_new:Npn \mypkg_tl_head_keep_space:n #1
-% {
-% \tl_if_head_is_space:nTF {#1}
-% { ~ }
-% { \tl_head:n {#1} }
-% }
-% \cs_new:Npn \mypkg_tl_tail_keep_space:n #1
-% {
-% \tl_if_head_is_space:nTF {#1}
-% { \exp_not:o { \__mypkg_gobble_space:w #1 } }
-% { \tl_tail:n {#1} }
-% }
-% \end{verbatim}
-%
-% \subsection{Items and ranges in token lists}
-%
-% \begin{function}[added = 2014-07-17, EXP]
-% {\tl_item:nn, \tl_item:Nn, \tl_item:cn}
-% \begin{syntax}
-% \cs{tl_item:nn} \Arg{token list} \Arg{integer expression}
-% \end{syntax}
-% Indexing items in the \meta{token list} from~$1$ on the left, this
-% function evaluates the \meta{integer expression} and leaves the
-% appropriate item from the \meta{token list} in the input stream.
-% If the \meta{integer expression} is negative, indexing occurs from
-% the right of the token list, starting at $-1$ for the right-most item.
-% If the index is out of bounds, then the function expands to nothing.
-% \begin{texnote}
-% The result is returned within the \tn{unexpanded}
-% primitive (\cs{exp_not:n}), which means that the \meta{item}
-% does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[EXP, added = 2016-12-06]
-% {\tl_rand_item:N, \tl_rand_item:c, \tl_rand_item:n}
-% \begin{syntax}
-% \cs{tl_rand_item:N} \meta{tl~var}
-% \cs{tl_rand_item:n} \Arg{token list}
-% \end{syntax}
-% Selects a pseudo-random item of the \meta{token list}. If the
-% \meta{token list} is blank, the result is empty.
-% This is not available in older versions of \XeTeX{}.
-% \begin{texnote}
-% The result is returned within the \tn{unexpanded}
-% primitive (\cs{exp_not:n}), which means that the \meta{item}
-% does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[EXP, added = 2017-02-17, updated = 2017-07-15]
-% {\tl_range:Nnn, \tl_range:nnn}
-% \begin{syntax}
-% \cs{tl_range:Nnn} \meta{tl~var} \Arg{start index} \Arg{end index}
-% \cs{tl_range:nnn} \Arg{token list} \Arg{start index} \Arg{end index}
-% \end{syntax}
-% Leaves in the input stream the items from the \meta{start index} to the
-% \meta{end index} inclusive. Spaces and braces are preserved between
-% the items returned (but never at either end of the list).
-% Here \meta{start index} and \meta{end index} should be \meta{integer expressions}.
-% For describing in detail the functions' behavior, let $m$ and $n$ be the start
-% and end index respectively. If either is $0$, the result is empty. A positive
-% index means `start counting from the left end', and a negative index means
-% `from the right end'. Let $l$ be the count of the token list.
-%
-% The \emph{actual start point} is determined as $M=m$ if~$m>0$ and as $M=l+m+1$
-% if~$m<0$. Similarly the \emph{actual end point} is $N=n$ if~$n>0$ and $N=l+n+1$
-% if~$n<0$. If $M>N$, the result is empty. Otherwise it consists of all items from
-% position $M$ to position $N$ inclusive; for the purpose of this rule, we can
-% imagine that the token list extends at infinity on either side, with void items
-% at positions $s$ for $s\le0$ or $s>l$.
-%
-% Spaces in between items in the actual range are preserved. Spaces at either end
-% of the token list will be removed anyway (think to the token list being passed to
-% |\tl_trim_spaces:n| to begin with.
-%
-% Thus, with $l=7$ as in the examples below, all of the following are equivalent
-% and result in the whole token list
-% \begin{verbatim}
-% \tl_range:nnn { abcd~{e{}}fg } { 1 } { 7 }
-% \tl_range:nnn { abcd~{e{}}fg } { 1 } { 12 }
-% \tl_range:nnn { abcd~{e{}}fg } { -7 } { 7 }
-% \tl_range:nnn { abcd~{e{}}fg } { -12 } { 7 }
-% \end{verbatim}
-% Here are some more interesting examples. The calls
-% \begin{verbatim}
-% \iow_term:e { \tl_range:nnn { abcd{e{}}fg } { 2 } { 5 } }
-% \iow_term:e { \tl_range:nnn { abcd{e{}}fg } { 2 } { -3 } }
-% \iow_term:e { \tl_range:nnn { abcd{e{}}fg } { -6 } { 5 } }
-% \iow_term:e { \tl_range:nnn { abcd{e{}}fg } { -6 } { -3 } }
-% \end{verbatim}
-% are all equivalent and will print |bcd{e{}}| on the terminal; similarly
-% \begin{verbatim}
-% \iow_term:e { \tl_range:nnn { abcd~{e{}}fg } { 2 } { 5 } }
-% \iow_term:e { \tl_range:nnn { abcd~{e{}}fg } { 2 } { -3 } }
-% \iow_term:e { \tl_range:nnn { abcd~{e{}}fg } { -6 } { 5 } }
-% \iow_term:e { \tl_range:nnn { abcd~{e{}}fg } { -6 } { -3 } }
-% \end{verbatim}
-% are all equivalent and will print |bcd {e{}}| on the
-% terminal (note the space in the middle). To the contrary,
-% \begin{verbatim}
-% \tl_range:nnn { abcd~{e{}}f } { 2 } { 4 }
-% \end{verbatim}
-% will discard the space after `d'.
-%
-% If we want to get the items from, say, the third to the last in a token
-% list |<tl>|, the call
-% is |\tl_range:nnn { <tl> } { 3 } { -1 }|. Similarly, for discarding
-% the last item, we can do |\tl_range:nnn { <tl> } { 1 } { -2 }|.
-%
-%^^A The behavior of \cs{tl_range:Nnn} is exactly the same, acting on the
-%^^A contents of the tl variable.
-%
-% \begin{texnote}
-% The result is returned within the \tn{unexpanded}
-% primitive (\cs{exp_not:n}), which means that the \meta{item}
-% does not expand further when appearing in an \texttt{e}-type
-% or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \subsection{Sorting token lists}
-%
-% \begin{function}[added = 2017-02-06]
-% {\tl_sort:Nn, \tl_sort:cn, \tl_gsort:Nn, \tl_gsort:cn}
-% \begin{syntax}
-% \cs{tl_sort:Nn} \meta{tl var} \Arg{comparison code}
-% \end{syntax}
-% Sorts the items in the \meta{tl var} according to the
-% \meta{comparison code}, and assigns the result to
-% \meta{tl var}. The details of sorting comparison are
-% described in Section~\ref{sec:l3sort:mech}.
-% \end{function}
-%
-% \begin{function}[added = 2017-02-06, EXP]{\tl_sort:nN}
-% \begin{syntax}
-% \cs{tl_sort:nN} \Arg{token list} \meta{conditional}
-% \end{syntax}
-% Sorts the items in the \meta{token list}, using the
-% \meta{conditional} to compare items, and leaves the result in the
-% input stream. The \meta{conditional} should have signature |:nnTF|,
-% and return \texttt{true} if the two items being compared should be
-% left in the same order, and \texttt{false} if the items should be
-% swapped. The details of sorting comparison are
-% described in Section~\ref{sec:l3sort:mech}.
-% \begin{texnote}
-% The result is returned within \cs{exp_not:n}, which means that the
-% token list does not expand further when appearing in an
-% \texttt{e}-type or \texttt{x}-type argument expansion.
-% \end{texnote}
-% \end{function}
-%
-% \section{Manipulating tokens in token lists}
-%
-% \subsection{Replacing tokens}
-%
-% Within token lists, replacement takes place at the top level: there is
-% no recursion into brace groups (more precisely, within a group defined by
-% a categroy code $1$/$2$ pair).
-%
-% \begin{function}[updated = 2011-08-11]
-% {
-% \tl_replace_once:Nnn, \tl_replace_once:NVn, \tl_replace_once:NnV,
-% \tl_replace_once:Nen, \tl_replace_once:Nne, \tl_replace_once:Nee,
-% \tl_replace_once:cnn, \tl_replace_once:cVn, \tl_replace_once:cnV,
-% \tl_replace_once:cen, \tl_replace_once:cne, \tl_replace_once:cee,
-% \tl_greplace_once:Nnn, \tl_greplace_once:NVn, \tl_greplace_once:NnV,
-% \tl_greplace_once:Nen, \tl_greplace_once:Nne, \tl_greplace_once:Nee,
-% \tl_greplace_once:cnn, \tl_greplace_once:cVn, \tl_greplace_once:cnV,
-% \tl_greplace_once:cen, \tl_greplace_once:cne, \tl_greplace_once:cee
-% }
-% \begin{syntax}
-% \cs{tl_replace_once:Nnn} \meta{tl~var} \Arg{old tokens} \Arg{new tokens}
-% \end{syntax}
-% Replaces the first (leftmost) occurrence of \meta{old tokens} in the
-% \meta{tl~var} with \meta{new tokens}. \meta{Old tokens}
-% cannot contain |{|, |}| or |#|
-% (more precisely, explicit character tokens with category code $1$
-% (begin-group) or $2$ (end-group), and tokens with category code $6$).
-% \end{function}
-%
-% \begin{function}[updated = 2011-08-11]
-% {
-% \tl_replace_all:Nnn, \tl_replace_all:NVn, \tl_replace_all:NnV,
-% \tl_replace_all:Nen, \tl_replace_all:Nne, \tl_replace_all:Nee,
-% \tl_replace_all:cnn, \tl_replace_all:cVn, \tl_replace_all:cnV,
-% \tl_replace_all:cen, \tl_replace_all:cne, \tl_replace_all:cee,
-% \tl_greplace_all:Nnn, \tl_greplace_all:NVn, \tl_greplace_all:NnV,
-% \tl_greplace_all:Nen, \tl_greplace_all:Nne, \tl_greplace_all:Nee,
-% \tl_greplace_all:cnn, \tl_greplace_all:cVn, \tl_greplace_all:cnV,
-% \tl_greplace_all:cen, \tl_greplace_all:cne, \tl_greplace_all:cee
-% }
-% \begin{syntax}
-% \cs{tl_replace_all:Nnn} \meta{tl~var} \Arg{old tokens} \Arg{new tokens}
-% \end{syntax}
-% Replaces all occurrences of \meta{old tokens} in the
-% \meta{tl~var} with \meta{new tokens}. \meta{Old tokens}
-% cannot contain |{|, |}| or |#|
-% (more precisely, explicit character tokens with category code $1$
-% (begin-group) or $2$ (end-group), and tokens with category code $6$).
-% As this function
-% operates from left to right, the pattern \meta{old tokens}
-% may remain after the replacement (see \cs{tl_remove_all:Nn}
-% for an example).
-% \end{function}
-%
-% \begin{function}[updated = 2011-08-11]
-% {
-% \tl_remove_once:Nn, \tl_remove_once:NV, \tl_remove_once:Ne,
-% \tl_remove_once:cn, \tl_remove_once:cV, \tl_remove_once:ce,
-% \tl_gremove_once:Nn, \tl_gremove_once:NV, \tl_gremove_once:Ne,
-% \tl_gremove_once:cn, \tl_gremove_once:cV, \tl_gremove_once:ce
-% }
-% \begin{syntax}
-% \cs{tl_remove_once:Nn} \meta{tl~var} \Arg{tokens}
-% \end{syntax}
-% Removes the first (leftmost) occurrence of \meta{tokens} from the
-% \meta{tl~var}. The \meta{tokens} cannot contain |{|, |}| or |#|
-% (more precisely, explicit character tokens with category code $1$
-% (begin-group) or $2$ (end-group), and tokens with category code $6$).
-% \end{function}
-%
-% \begin{function}[updated = 2011-08-11]
-% {
-% \tl_remove_all:Nn, \tl_remove_all:NV, \tl_remove_all:Ne,
-% \tl_remove_all:cn, \tl_remove_all:cV, \tl_remove_all:ce,
-% \tl_gremove_all:Nn, \tl_gremove_all:NV, \tl_gremove_all:Ne,
-% \tl_gremove_all:cn, \tl_gremove_all:cV, \tl_gremove_all:ce,
-% }
-% \begin{syntax}
-% \cs{tl_remove_all:Nn} \meta{tl~var} \Arg{tokens}
-% \end{syntax}
-% Removes all occurrences of \meta{tokens} from the
-% \meta{tl~var}. The \meta{tokens} cannot contain |{|, |}| or |#|
-% (more precisely, explicit character tokens with category code $1$
-% (begin-group) or $2$ (end-group), and tokens with category code $6$).
-% As this function
-% operates from left to right, the pattern \meta{tokens}
-% may remain after the removal, for instance,
-% \begin{quote}
-% \cs{tl_set:Nn} \cs{l_tmpa_tl} |{abbccd}|
-% \cs{tl_remove_all:Nn} \cs{l_tmpa_tl} |{bc}|
-% \end{quote}
-% results in \cs{l_tmpa_tl} containing \texttt{abcd}.
-% \end{function}
-%
-% \subsection{Reassigning category codes}
-%
-% These functions allow the rescanning of tokens: re-apply \TeX{}'s
-% tokenization process to apply category codes different from those
-% in force when the tokens were absorbed. Whilst this functionality is
-% supported, it is often preferable to find alternative approaches
-% to achieving outcomes rather than rescanning tokens (for example
-% construction of token lists token-by-token with intervening category
-% code changes or using \cs{char_generate:nn}).
-%
-% \begin{function}[updated = 2015-08-11]
-% {
-% \tl_set_rescan:Nnn, \tl_set_rescan:NnV, \tl_set_rescan:Nne,
-% \tl_set_rescan:Nno,
-% \tl_set_rescan:cnn, \tl_set_rescan:cnV, \tl_set_rescan:cne,
-% \tl_set_rescan:cno,
-% \tl_gset_rescan:Nnn, \tl_gset_rescan:NnV, \tl_gset_rescan:Nne,
-% \tl_gset_rescan:Nno,
-% \tl_gset_rescan:cnn, \tl_gset_rescan:cnV, \tl_gset_rescan:cne,
-% \tl_gset_rescan:cno,
-% }
-% \begin{syntax}
-% \cs{tl_set_rescan:Nnn} \meta{tl~var} \Arg{setup} \Arg{tokens}
-% \end{syntax}
-% Sets \meta{tl~var} to contain \meta{tokens}, applying the category
-% code r\'{e}gime specified in the \meta{setup} before carrying out
-% the assignment. (Category codes applied to tokens not explicitly covered
-% by the \meta{setup} are those in force at the point of use of
-% \cs{tl_set_rescan:Nnn}.)
-% This allows the \meta{tl~var} to contain material
-% with category codes other than those that apply when \meta{tokens}
-% are absorbed. The \meta{setup} is run within a group and may
-% contain any valid input, although only changes in category codes,
-% such as uses of \cs{cctab_select:N},
-% are relevant. See also \cs{tl_rescan:nn}.
-% \begin{texnote}
-% The \meta{tokens} are first turned into a string (using
-% \cs{tl_to_str:n}). If the string contains one or more characters
-% with character code \tn{newlinechar} (set equal to
-% \tn{endlinechar} unless that is equal to $32$, before the user
-% \meta{setup}), then it is split into lines at these characters,
-% then read as if reading multiple lines from a file, ignoring
-% spaces (catcode $10$) at the beginning and spaces and tabs
-% (character code $32$ or $9$) at the end of every line.
-% Otherwise, spaces (and tabs) are retained at both ends of the
-% single-line string, as if it appeared in the middle of a line
-% read from a file.
-% \end{texnote}
-% \end{function}
-%
-% \begin{function}[updated = 2015-08-11]{\tl_rescan:nn, \tl_rescan:nV}
-% \begin{syntax}
-% \cs{tl_rescan:nn} \Arg{setup} \Arg{tokens}
-% \end{syntax}
-% Rescans \meta{tokens} applying the category code r\'{e}gime
-% specified in the \meta{setup}, and leaves the resulting tokens in
-% the input stream. (Category codes applied to tokens not explicitly covered
-% by the \meta{setup} are those in force at the point of use of
-% \cs{tl_rescan:nn}.)
-% The \meta{setup} is run within a group and may
-% contain any valid input, although only changes in category codes,
-% such as uses of \cs{cctab_select:N},
-% are relevant. See also \cs{tl_set_rescan:Nnn}, which is more
-% robust than using \cs{tl_set:Nn} in the \meta{tokens} argument of
-% \cs{tl_rescan:nn}.
-% \begin{texnote}
-% The \meta{tokens} are first turned into a string (using
-% \cs{tl_to_str:n}). If the string contains one or more characters
-% with character code \tn{newlinechar} (set equal to
-% \tn{endlinechar} unless that is equal to $32$, before the user
-% \meta{setup}), then it is split into lines at these characters,
-% then read as if reading multiple lines from a file, ignoring
-% spaces (catcode $10$) at the beginning and spaces and tabs
-% (character code $32$ or $9$) at the end of every line.
-% Otherwise, spaces (and tabs) are retained at both ends of the
-% single-line string, as if it appeared in the middle of a line
-% read from a file.
-%
-% Contrarily to the \tn{scantokens} \eTeX{} primitive, \cs{tl_rescan:nn}
-% tokenizes the whole string in the same category code regime rather
-% than one token at a time, so that directives such as \tn{verb}
-% that rely on changing category codes will not function properly.
-% \end{texnote}
-% \end{function}
-%
-% \section{Constant token lists}
-%
-% \begin{variable}{\c_empty_tl}
-% Constant that is always empty.
-% \end{variable}
-%
-% \begin{variable}[added = 2017-11-14]{\c_novalue_tl}
-% A marker for the absence of an argument. This constant |tl| can safely
-% be typeset (\emph{cf.}~\cs{q_nil}), with the result being |-NoValue-|.
-% It is important to note that \cs{c_novalue_tl} is constructed such that it
-% will \emph{not} match the simple text input |-NoValue-|, \emph{i.e.}
-% that
-% \begin{verbatim}
-% \tl_if_eq:NnTF \c_novalue_tl { -NoValue- }
-% \end{verbatim}
-% is logically \texttt{false}. The \cs{c_novalue_tl} marker is intended for
-% use in creating document-level interfaces, where it serves as an indicator
-% that an (optional) argument was omitted. In particular, it is distinct
-% from a simple empty |tl|.
-% \end{variable}
-%
-% \begin{variable}{\c_space_tl}
-% An explicit space character contained in a token list (compare this with
-% \cs{c_space_token}). For use where an explicit space is required.
-% \end{variable}
-%
-% \section{Scratch token lists}
-%
-% \begin{variable}{\l_tmpa_tl, \l_tmpb_tl}
-% Scratch token lists for local assignment. These are never used by
-% the kernel code, and so are safe for use with any \LaTeX3-defined
-% function. However, they may be overwritten by other non-kernel
-% code and so should only be used for short-term storage.
-% \end{variable}
-%
-% \begin{variable}{\g_tmpa_tl, \g_tmpb_tl}
-% Scratch token lists for global assignment. These are never used by
-% the kernel code, and so are safe for use with any \LaTeX3-defined
-% function. However, they may be overwritten by other non-kernel
-% code and so should only be used for short-term storage.
-% \end{variable}
-%
-% \end{documentation}
-%
-% \begin{implementation}
-%
-% \section{\pkg{l3tl} implementation}
-%
-% \begin{macrocode}
-%<*package>
-% \end{macrocode}
-%
-% \begin{macrocode}
-%<@@=tl>
-% \end{macrocode}
-%
-% A token list variable is a \TeX{} macro that holds tokens. By using the
-% \eTeX{} primitive \tn{unexpanded} inside a \TeX{} \tn{edef} it is
-% possible to store any tokens, including |#|, in this way.
-%
-% \subsection{Functions}
-%
-% \begin{macro}{\__kernel_tl_set:Nx,\__kernel_tl_gset:Nx}
-% These two are supplied to get better performance for macros which would
-% otherwise use \cs{tl_set:Ne} or \cs{tl_gset:Ne} internally.
-% \begin{macrocode}
-\cs_new_eq:NN \__kernel_tl_set:Nx \cs_set_nopar:Npe
-\cs_new_eq:NN \__kernel_tl_gset:Nx \cs_gset_nopar:Npe
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_new:N, \tl_new:c}
-% Creating new token list variables is a case of checking for an
-% existing definition and doing the definition.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_new:N #1
- {
- \__kernel_chk_if_free_cs:N #1
- \cs_gset_eq:NN #1 \c_empty_tl
- }
-\cs_generate_variant:Nn \tl_new:N { c }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_const:Nn, \tl_const:Ne, \tl_const:Nx,
-% \tl_const:cn, \tl_const:ce, \tl_const:cx
-% }
-% Constants are also easy to generate. They use \cs{cs_gset_nopar:Npe} instead
-% of \cs{__kernel_tl_gset:Nx} so that the correct scope checking for |c|,
-% instead of for |g|, is applied when
-% \cs{debug_on:n} |{ check-declarations }| is used.
-% Constant assignment functions are patched specially in \pkg{l3debug} to
-% apply such checks.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_const:Nn #1#2
- {
- \__kernel_chk_if_free_cs:N #1
- \cs_gset_nopar:Npe #1 { \__kernel_exp_not:w {#2} }
- }
-\cs_generate_variant:Nn \tl_const:Nn { Ne , c , ce }
-\cs_generate_variant:Nn \tl_const:Nn { Nx , cx }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_clear:N, \tl_clear:c}
-% \begin{macro}{\tl_gclear:N, \tl_gclear:c}
-% Clearing a token list variable means setting it to an empty value.
-% Error checking is sorted out by the parent function.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_clear:N #1
- { \tex_let:D #1 = ~ \c_empty_tl }
-\cs_new_protected:Npn \tl_gclear:N #1
- { \tex_global:D \tex_let:D #1 ~ \c_empty_tl }
-\cs_generate_variant:Nn \tl_clear:N { c }
-\cs_generate_variant:Nn \tl_gclear:N { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_clear_new:N, \tl_clear_new:c}
-% \begin{macro}{\tl_gclear_new:N, \tl_gclear_new:c}
-% Clearing a token list variable means setting it to an empty value.
-% Error checking is sorted out by the parent function.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_clear_new:N #1
- { \tl_if_exist:NTF #1 { \tl_clear:N #1 } { \tl_new:N #1 } }
-\cs_new_protected:Npn \tl_gclear_new:N #1
- { \tl_if_exist:NTF #1 { \tl_gclear:N #1 } { \tl_new:N #1 } }
-\cs_generate_variant:Nn \tl_clear_new:N { c }
-\cs_generate_variant:Nn \tl_gclear_new:N { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_set_eq:NN, \tl_set_eq:Nc, \tl_set_eq:cN, \tl_set_eq:cc}
-% \begin{macro}{\tl_gset_eq:NN, \tl_gset_eq:Nc, \tl_gset_eq:cN, \tl_gset_eq:cc}
-% For setting token list variables equal to each other. To allow for
-% patching, the arguments have to be explicit. In addition this
-% ensures that a braced second argument will not cause problems.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_set_eq:NN #1#2
- { \tex_let:D #1 = ~ #2 }
-\cs_new_protected:Npn \tl_gset_eq:NN #1#2
- { \tex_global:D \tex_let:D #1 = ~ #2 }
-\cs_generate_variant:Nn \tl_set_eq:NN { cN, Nc, cc }
-\cs_generate_variant:Nn \tl_gset_eq:NN { cN, Nc, cc }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_concat:NNN, \tl_concat:ccc}
-% \begin{macro}{\tl_gconcat:NNN, \tl_gconcat:ccc}
-% Concatenating token lists is easy. When checking is turned on, all
-% three arguments must be checked: a token list |#2| or |#3| equal to
-% \cs{scan_stop:} would lead to problems later on.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_concat:NNN #1#2#3
- {
- \__kernel_tl_set:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#2}
- \__kernel_exp_not:w \exp_after:wN {#3}
- }
- }
-\cs_new_protected:Npn \tl_gconcat:NNN #1#2#3
- {
- \__kernel_tl_gset:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#2}
- \__kernel_exp_not:w \exp_after:wN {#3}
- }
- }
-\cs_generate_variant:Nn \tl_concat:NNN { ccc }
-\cs_generate_variant:Nn \tl_gconcat:NNN { ccc }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[pTF]{\tl_if_exist:N, \tl_if_exist:c}
-% Copies of the \texttt{cs} functions defined in \pkg{l3basics}.
-% \begin{macrocode}
-\prg_new_eq_conditional:NNn \tl_if_exist:N \cs_if_exist:N { TF , T , F , p }
-\prg_new_eq_conditional:NNn \tl_if_exist:c \cs_if_exist:c { TF , T , F , p }
-% \end{macrocode}
-% \end{macro}
-%
-% \subsection{Constant token lists}
-%
-% \begin{variable}{\c_empty_tl}
-% Never full. We need to define that constant before using \cs{tl_new:N}.
-% \begin{macrocode}
-\tl_const:Nn \c_empty_tl { }
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{variable}{\c_novalue_tl}
-% A special marker: as we don't have |\char_generate:nn| yet, has to be
-% created the old-fashioned way.
-% \begin{macrocode}
-\group_begin:
-\tex_catcode:D `- = 11 ~
-\tl_const:Ne \c_novalue_tl { - NoValue \token_to_str:N - }
-\group_end:
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{variable}{\c_space_tl}
-% A space as a token list (as opposed to as a character).
-% \begin{macrocode}
-\tl_const:Nn \c_space_tl { ~ }
-% \end{macrocode}
-% \end{variable}
-%
-% \subsection{Adding to token list variables}
-%
-% \begin{macro}
-% {
-% \tl_set:Nn, \tl_set:NV, \tl_set:Nv, \tl_set:No, \tl_set:Ne, \tl_set:Nf, \tl_set:Nx,
-% \tl_set:cn, \tl_set:cV, \tl_set:cv, \tl_set:co, \tl_set:ce, \tl_set:cf, \tl_set:cx
-% }
-% \begin{macro}
-% {
-% \tl_gset:Nn, \tl_gset:NV, \tl_gset:Nv,
-% \tl_gset:No, \tl_gset:Ne, \tl_gset:Nf, \tl_gset:Nx,
-% \tl_gset:cn, \tl_gset:cV, \tl_gset:cv,
-% \tl_gset:co, \tl_gset:ce, \tl_gset:cf, \tl_gset:cx
-% }
-% By using \cs{exp_not:n} token list variables can contain |#| tokens,
-% which makes the token list registers provided by \TeX{}
-% more or less redundant. The \cs{tl_set:No} version is done
-% by hand as it is used quite a lot.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_set:Nn #1#2
- { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w {#2} } }
-\cs_new_protected:Npn \tl_set:No #1#2
- { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w \exp_after:wN {#2} } }
-\cs_new_protected:Npn \tl_gset:Nn #1#2
- { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w {#2} } }
-\cs_new_protected:Npn \tl_gset:No #1#2
- { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w \exp_after:wN {#2} } }
-\cs_generate_variant:Nn \tl_set:Nn { NV , Nv , Ne , Nf }
-\cs_generate_variant:Nn \tl_set:Nn { c, cV , cv , ce , cf }
-\cs_generate_variant:Nn \tl_set:No { c }
-\cs_generate_variant:Nn \tl_set:Nn { Nx , cx }
-\cs_generate_variant:Nn \tl_gset:Nn { NV , Nv , Ne , Nf }
-\cs_generate_variant:Nn \tl_gset:Nn { c, cV , cv , ce , cf }
-\cs_generate_variant:Nn \tl_gset:No { c }
-\cs_generate_variant:Nn \tl_gset:Nn { Nx , cx }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_put_left:Nn, \tl_put_left:NV, \tl_put_left:Nv, \tl_put_left:Ne,
-% \tl_put_left:No, \tl_put_left:Nx,
-% \tl_put_left:cn, \tl_put_left:cV, \tl_put_left:cv, \tl_put_left:ce,
-% \tl_put_left:co, \tl_put_left:cx
-% }
-% \begin{macro}
-% {
-% \tl_gput_left:Nn, \tl_gput_left:NV, \tl_gput_left:Nv, \tl_gput_left:Ne,
-% \tl_gput_left:No, \tl_gput_left:Nx,
-% \tl_gput_left:cn, \tl_gput_left:cV, \tl_gput_left:cv, \tl_gput_left:ce,
-% \tl_gput_left:co, \tl_gput_left:cx
-% }
-% Adding to the left is done directly to gain a little performance.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_put_left:Nn #1#2
- {
- \__kernel_tl_set:Nx #1
- { \__kernel_exp_not:w {#2} \__kernel_exp_not:w \exp_after:wN {#1} }
- }
-\cs_new_protected:Npn \tl_put_left:NV #1#2
- {
- \__kernel_tl_set:Nx #1
- { \exp_not:V #2 \__kernel_exp_not:w \exp_after:wN {#1} }
- }
-\cs_new_protected:Npn \tl_put_left:Nv #1#2
- {
- \__kernel_tl_set:Nx #1
- { \exp_not:v {#2} \__kernel_exp_not:w \exp_after:wN {#1} }
- }
-\cs_new_protected:Npn \tl_put_left:Ne #1#2
- {
- \__kernel_tl_set:Nx #1
- {
- \__kernel_exp_not:w \tex_expanded:D { {#2} }
- \__kernel_exp_not:w \exp_after:wN {#1}
- }
- }
-\cs_new_protected:Npn \tl_put_left:No #1#2
- {
- \__kernel_tl_set:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#2}
- \__kernel_exp_not:w \exp_after:wN {#1}
- }
- }
-\cs_new_protected:Npn \tl_gput_left:Nn #1#2
- {
- \__kernel_tl_gset:Nx #1
- { \__kernel_exp_not:w {#2} \__kernel_exp_not:w \exp_after:wN {#1} }
- }
-\cs_new_protected:Npn \tl_gput_left:NV #1#2
- {
- \__kernel_tl_gset:Nx #1
- { \exp_not:V #2 \__kernel_exp_not:w \exp_after:wN {#1} }
- }
-\cs_new_protected:Npn \tl_gput_left:Nv #1#2
- {
- \__kernel_tl_gset:Nx #1
- { \exp_not:v {#2} \__kernel_exp_not:w \exp_after:wN {#1} }
- }
-\cs_new_protected:Npn \tl_gput_left:Ne #1#2
- {
- \__kernel_tl_gset:Nx #1
- {
- \__kernel_exp_not:w \tex_expanded:D { {#2} }
- \__kernel_exp_not:w \exp_after:wN {#1}
- }
- }
-\cs_new_protected:Npn \tl_gput_left:No #1#2
- {
- \__kernel_tl_gset:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#2}
- \__kernel_exp_not:w \exp_after:wN {#1}
- }
- }
-\cs_generate_variant:Nn \tl_put_left:Nn { c }
-\cs_generate_variant:Nn \tl_put_left:NV { c }
-\cs_generate_variant:Nn \tl_put_left:Nv { c }
-\cs_generate_variant:Nn \tl_put_left:Ne { c }
-\cs_generate_variant:Nn \tl_put_left:No { c }
-\cs_generate_variant:Nn \tl_put_left:Nn { Nx, cx }
-\cs_generate_variant:Nn \tl_gput_left:Nn { c }
-\cs_generate_variant:Nn \tl_gput_left:NV { c }
-\cs_generate_variant:Nn \tl_gput_left:Nv { c }
-\cs_generate_variant:Nn \tl_gput_left:Ne { c }
-\cs_generate_variant:Nn \tl_gput_left:No { c }
-\cs_generate_variant:Nn \tl_gput_left:Nn { Nx , cx }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_put_right:Nn, \tl_put_right:NV, \tl_put_right:Nv, \tl_put_right:Ne,
-% \tl_put_right:No, \tl_put_right:Nx,
-% \tl_put_right:cn, \tl_put_right:cV, \tl_put_right:cv, \tl_put_right:ce,
-% \tl_put_right:co, \tl_put_right:cx
-% }
-% \begin{macro}
-% {
-% \tl_gput_right:Nn, \tl_gput_right:NV, \tl_gput_right:Nv, \tl_gput_right:Ne,
-% \tl_gput_right:No, \tl_gput_right:Nx,
-% \tl_gput_right:cn, \tl_gput_right:cV, \tl_gput_right:cv, \tl_gput_right:ce,
-% \tl_gput_right:co, \tl_gput_right:cx
-% }
-% The same on the right.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_put_right:Nn #1#2
- { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w \exp_after:wN { #1 #2 } } }
-\cs_new_protected:Npn \tl_put_right:NV #1#2
- {
- \__kernel_tl_set:Nx #1
- { \__kernel_exp_not:w \exp_after:wN {#1} \exp_not:V #2 }
- }
-\cs_new_protected:Npn \tl_put_right:Nv #1#2
- {
- \__kernel_tl_set:Nx #1
- { \__kernel_exp_not:w \exp_after:wN {#1} \exp_not:v {#2} }
- }
-\cs_new_protected:Npn \tl_put_right:Ne #1#2
- {
- \__kernel_tl_set:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#1}
- \__kernel_exp_not:w \tex_expanded:D { {#2} }
- }
- }
-\cs_new_protected:Npn \tl_put_right:No #1#2
- {
- \__kernel_tl_set:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#1}
- \__kernel_exp_not:w \exp_after:wN {#2}
- }
- }
-\cs_new_protected:Npn \tl_gput_right:Nn #1#2
- { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w \exp_after:wN { #1 #2 } } }
-\cs_new_protected:Npn \tl_gput_right:NV #1#2
- {
- \__kernel_tl_gset:Nx #1
- { \__kernel_exp_not:w \exp_after:wN {#1} \exp_not:V #2 }
- }
-\cs_new_protected:Npn \tl_gput_right:Nv #1#2
- {
- \__kernel_tl_gset:Nx #1
- { \__kernel_exp_not:w \exp_after:wN {#1} \exp_not:v {#2} }
- }
-\cs_new_protected:Npn \tl_gput_right:Ne #1#2
- {
- \__kernel_tl_gset:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#1}
- \__kernel_exp_not:w \tex_expanded:D { {#2} }
- }
- }
-\cs_new_protected:Npn \tl_gput_right:No #1#2
- {
- \__kernel_tl_gset:Nx #1
- {
- \__kernel_exp_not:w \exp_after:wN {#1}
- \__kernel_exp_not:w \exp_after:wN {#2}
- }
- }
-\cs_generate_variant:Nn \tl_put_right:Nn { c }
-\cs_generate_variant:Nn \tl_put_right:NV { c }
-\cs_generate_variant:Nn \tl_put_right:Nv { c }
-\cs_generate_variant:Nn \tl_put_right:Ne { c }
-\cs_generate_variant:Nn \tl_put_right:No { c }
-\cs_generate_variant:Nn \tl_put_right:Nn { Nx , cx }
-\cs_generate_variant:Nn \tl_gput_right:Nn { c }
-\cs_generate_variant:Nn \tl_gput_right:NV { c }
-\cs_generate_variant:Nn \tl_gput_right:Nv { c }
-\cs_generate_variant:Nn \tl_gput_right:Ne { c }
-\cs_generate_variant:Nn \tl_gput_right:No { c }
-\cs_generate_variant:Nn \tl_gput_right:Nn { Nx, cx }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \subsection{Internal quarks and quark-query functions}
-%
-% \begin{variable}{\q_@@_nil,\q_@@_mark,\q_@@_stop}
-% Internal quarks.
-% \begin{macrocode}
-\quark_new:N \q_@@_nil
-\quark_new:N \q_@@_mark
-\quark_new:N \q_@@_stop
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{variable}{\q_@@_recursion_tail,\q_@@_recursion_stop}
-% Internal recursion quarks.
-% \begin{macrocode}
-\quark_new:N \q_@@_recursion_tail
-\quark_new:N \q_@@_recursion_stop
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{macro}[EXP]{\@@_if_recursion_tail_break:nN}
-% \begin{macro}[pTF]{\@@_if_recursion_tail_stop:n}
-% Functions to query recursion quarks.
-% \begin{macrocode}
-\__kernel_quark_new_test:N \@@_if_recursion_tail_break:nN
-\__kernel_quark_new_conditional:Nn \@@_quark_if_nil:n { TF }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \subsection{Reassigning token list category codes}
-%
-% \begin{variable}{\c_@@_rescan_marker_tl}
-% The rescanning code needs a special token list containing the same
-% character (chosen here to be a colon) with two different category
-% codes: it cannot appear in the tokens being rescanned since all
-% colons have the same category code.
-% \begin{macrocode}
-\tl_const:Ne \c_@@_rescan_marker_tl { : \token_to_str:N : }
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{macro}
-% {
-% \tl_set_rescan:Nnn, \tl_set_rescan:NnV, \tl_set_rescan:Nne,
-% \tl_set_rescan:Nno, \tl_set_rescan:Nnx,
-% \tl_set_rescan:cnn, \tl_set_rescan:cnV, \tl_set_rescan:cne,
-% \tl_set_rescan:cno, \tl_set_rescan:cnx,
-% }
-% \begin{macro}
-% {
-% \tl_gset_rescan:Nnn, \tl_gset_rescan:NnV, \tl_gset_rescan:Nne,
-% \tl_gset_rescan:Nno, \tl_gset_rescan:Nnx,
-% \tl_gset_rescan:cnn, \tl_gset_rescan:cnV, \tl_gset_rescan:cne,
-% \tl_gset_rescan:cno, \tl_gset_rescan:cnx
-% }
-% \begin{macro}{\tl_rescan:nn, \tl_rescan:nV}
-% \begin{macro}{\@@_rescan_aux:}
-% \begin{macro}{\@@_set_rescan:NNnn, \@@_set_rescan_multi:nNN}
-% \begin{macro}[EXP]{\@@_rescan:NNw}
-% In a group, after some initial setup explained below and the user
-% setup~|#3| (followed by \cs{scan_stop:} to be safe), there is a call
-% to \cs{@@_set_rescan:nNN}. This shared auxiliary defined later
-% distinguishes single-line and multi-line ``files''. In the simplest
-% case of multi-line files, it calls (with the same arguments)
-% \cs{@@_set_rescan_multi:nNN}, whose code is included here to help
-% understand the approach. This function rescans its argument |#1|,
-% closes the group, and performs the assignment.
-%
-% One difficulty when rescanning is that \tn{scantokens} treats the
-% argument as a file, and without the correct settings a \TeX{} error
-% occurs:
-% \begin{verbatim}
-% ! File ended while scanning definition of ...
-% \end{verbatim}
-% A related minor issue is a warning due to opening a group before the
-% \tn{scantokens} and closing it inside that temporary file; we avoid
-% that by setting \tn{tracingnesting}. The standard solution to the
-% ``File ended'' error is to grab the rescanned tokens as a delimited
-% argument of an auxiliary, here \cs{@@_rescan:NNw}, that performs the
-% assignment, then let \TeX{} ``execute'' the end of file marker. As
-% usual in delimited arguments we use \cs{prg_do_nothing:} to avoid
-% stripping an outer set braces: this is removed by using
-% \texttt{o}-expanding assignments. The delimiter cannot appear
-% within the rescanned token list because it contains twice the same
-% character, with different catcodes.
-%
-% For \cs{tl_rescan:nn} we cannot simply call \cs{@@_set_rescan:NNnn}
-% \cs{prg_do_nothing:} \cs{use:n} because that would leave the
-% end-of-file marker \emph{after} the result of rescanning. If that
-% rescanned result is code that looks further in the input stream for
-% arguments, it would break.
-%
-% For multi-line files the only subtlety is that \tn{newlinechar}
-% should be equal to \tn{endlinechar} because \tn{newlinechar}
-% characters become new lines and then become \tn{endlinechar}
-% characters when writing to an abstract file and reading back. This
-% equality is ensured by setting \tn{newlinechar} equal to
-% \tn{endlinechar}. Prior to this, \tn{endlinechar} is set to $-1$ if
-% it was $32$ (in particular true after \cs{ExplSyntaxOn}) to avoid
-% unreasonable line-breaks at every space for instance in error
-% messages triggered by the user setup. Another side effect of
-% reading back from the file is that spaces (catcode $10$) are ignored
-% at the beginning of lines, and spaces and tabs (character code $32$
-% and $9$) are ignored at the end of lines.
-%
-% The two \cs{if_false:} \ldots{} \cs{fi:} are there to prevent
-% alignment tabs to cause a change of tabular cell while rescanning.
-% We put the \enquote{opening} one after \cs{group_begin:} so that if
-% one accidentally \texttt{f}-expands \cs{tl_set_rescan:Nnn} braces
-% remain balanced. This is essential in \texttt{e}-type arguments
-% when \tn{expanded} is not available.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_rescan:nn #1#2
- {
- \tl_set_rescan:Nnn \l_@@_internal_a_tl {#1} {#2}
- \exp_after:wN \@@_rescan_aux:
- \l_@@_internal_a_tl
- }
-\cs_generate_variant:Nn \tl_rescan:nn { nV }
-\exp_args:NNo \cs_new_protected:Npn \@@_rescan_aux:
- { \tl_clear:N \l_@@_internal_a_tl }
-\cs_new_protected:Npn \tl_set_rescan:Nnn
- { \@@_set_rescan:NNnn \tl_set:No }
-\cs_new_protected:Npn \tl_gset_rescan:Nnn
- { \@@_set_rescan:NNnn \tl_gset:No }
-\cs_new_protected:Npn \@@_set_rescan:NNnn #1#2#3#4
- {
- \group_begin:
- \if_false: { \fi:
- \int_set_eq:NN \tex_tracingnesting:D \c_zero_int
- \int_compare:nNnT \tex_endlinechar:D = { 32 }
- { \int_set:Nn \tex_endlinechar:D { -1 } }
- \int_set_eq:NN \tex_newlinechar:D \tex_endlinechar:D
- #3 \scan_stop:
- \exp_args:No \@@_set_rescan:nNN { \tl_to_str:n {#4} } #1 #2
- \if_false: } \fi:
- }
-\cs_new_protected:Npn \@@_set_rescan_multi:nNN #1#2#3
- {
- \tex_everyeof:D \exp_after:wN { \c_@@_rescan_marker_tl }
- \exp_after:wN \@@_rescan:NNw
- \exp_after:wN #2
- \exp_after:wN #3
- \exp_after:wN \prg_do_nothing:
- \tex_scantokens:D {#1}
- }
-\exp_args:Nno \use:nn
- { \cs_new:Npn \@@_rescan:NNw #1#2#3 } \c_@@_rescan_marker_tl
- {
- \group_end:
- #1 #2 {#3}
- }
-\cs_generate_variant:Nn \tl_set_rescan:Nnn { NnV , Nne , c , cnV , cne }
-\cs_generate_variant:Nn \tl_set_rescan:Nnn { Nno , Nnx , cno , cnx }
-\cs_generate_variant:Nn \tl_gset_rescan:Nnn { NnV , Nne , c , cnV , cne }
-\cs_generate_variant:Nn \tl_gset_rescan:Nnn { Nno , Nnx , cno , cnx }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\@@_set_rescan:nNN}
-% \begin{macro}{\@@_set_rescan_single:nnNN, \@@_set_rescan_single_aux:nnnNN}
-% \begin{macro}[rEXP]{\@@_set_rescan_single_aux:w}
-% The function \cs{@@_set_rescan:nNN} calls \cs{@@_set_rescan_multi:nNN} or
-% \cs{@@_set_rescan_single:nnNN} |{ ' }| depending on whether its argument
-% is a single-line fragment of code/data or is made of multiple lines
-% by testing for the presence of a \tn{newlinechar} character. If
-% \tn{newlinechar} is out of range, the argument is assumed to be a
-% single line.
-%
-% For a single line, no \tn{endlinechar} should be added, so it is
-% set to $-1$, and spaces should not be removed.
-% Trailing spaces and tabs are a difficult matter, as \TeX{} removes
-% these at a very low level. The only way to preserve them is to
-% rescan not the argument but the argument followed by a character
-% with a reasonable category code. Here, $11$ (letter) and $12$ (other)
-% are accepted, as these are convenient, suitable for
-% delimiting an argument, and it is very unlikely that none of the
-% ASCII characters are in one of these categories. To avoid
-% selecting one particular character to put at the end, whose
-% category code may have been modified, there is a loop through
-% characters from |'| (ASCII $39$) to |~| (ASCII $127$). The choice
-% of starting point was made because this is the start of a very long
-% range of characters whose standard category is letter or other,
-% thus minimizing the number of steps needed by the loop (most often
-% just a single one). If no valid character is found (very rare),
-% fall-back on \cs{@@_set_rescan_multi:nNN}.
-%
-% Otherwise, once a valid character is found (let us use |'| in this
-% explanation) run some code very similar to \cs{@@_set_rescan_multi:nNN}
-% but with |'| added at both ends of the input. Of course, we need to
-% define the auxiliary \cs{@@_set_rescan_single:NNww} on the fly to remove
-% the additional~|'| that is just before |::| (by which we mean
-% \cs{c_@@_rescan_marker_tl}). Note that the argument must be
-% delimited by |'| with the current catcode; this is done thanks to
-% \cs{char_generate:nn}. Yet another issue is that the rescanned
-% token list may contain a comment character, in which case the |'| we
-% expected is not there. We fix this as follows: rather than just
-% |::| we set \tn{everyeof} to |::|\Arg{code1} |'::|\Arg{code2}
-% \cs{s_@@_stop}. The auxiliary \cs{@@_set_rescan_single:NNww} runs the
-% \texttt{o}-expanding assignment, expanding either \meta{code1} or
-% \meta{code2} before its the main argument~|#3|. In the typical case
-% without comment character, \meta{code1} is expanded, removing the
-% leading~|'|. In the rarer case with comment character, \meta{code2}
-% is expanded, calling \cs{@@_set_rescan_single_aux:w}, which removes the
-% trailing |::|\Arg{code1} and the leading~|'|.
-% \begin{macrocode}
-\cs_new_protected:Npn \@@_set_rescan:nNN #1
- {
- \int_compare:nNnTF \tex_newlinechar:D < 0
- { \use_ii:nn }
- {
- \exp_args:Nnf \tl_if_in:nnTF {#1}
- { \char_generate:nn { \tex_newlinechar:D } { 12 } }
- }
- { \@@_set_rescan_multi:nNN }
- {
- \int_set:Nn \tex_endlinechar:D { -1 }
- \@@_set_rescan_single:nnNN { `' }
- }
- {#1}
- }
-\cs_new_protected:Npn \@@_set_rescan_single:nnNN #1
- {
- \int_compare:nNnTF
- { \char_value_catcode:n {#1} / 2 } = 6
- {
- \exp_args:Nof \@@_set_rescan_single_aux:nnnNN
- \c_@@_rescan_marker_tl
- { \char_generate:nn {#1} { \char_value_catcode:n {#1} } }
- }
- {
- \int_compare:nNnTF {#1} < { `\~ }
- {
- \exp_args:Nf \@@_set_rescan_single:nnNN
- { \int_eval:n { #1 + 1 } }
- }
- { \@@_set_rescan_multi:nNN }
- }
- }
-\cs_new_protected:Npn \@@_set_rescan_single_aux:nnnNN #1#2#3#4#5
- {
- \tex_everyeof:D
- {
- #1 \use_none:n
- #2 #1 { \exp:w \@@_set_rescan_single_aux:w }
- \s_@@_stop
- }
- \cs_set:Npn \@@_rescan:NNw ##1##2##3 #2 #1 ##4 ##5 \s_@@_stop
- {
- \group_end:
- ##1 ##2 { ##4 ##3 }
- }
- \exp_after:wN \@@_rescan:NNw
- \exp_after:wN #4
- \exp_after:wN #5
- \tex_scantokens:D { #2 #3 #2 }
- }
-\exp_args:Nno \use:nn
- { \cs_new:Npn \@@_set_rescan_single_aux:w #1 }
- \c_@@_rescan_marker_tl #2
- { \use_i:nn \exp_end: #1 }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \subsection{Modifying token list variables}
-%
-% \begin{macro}
-% {
-% \tl_replace_once:Nnn, \tl_replace_once:NVn, \tl_replace_once:NnV,
-% \tl_replace_once:Nen, \tl_replace_once:Nne, \tl_replace_once:Nee,
-% \tl_replace_once:Nxn, \tl_replace_once:Nnx, \tl_replace_once:Nxx,
-% \tl_replace_once:cnn, \tl_replace_once:cVn, \tl_replace_once:cnV,
-% \tl_replace_once:cen, \tl_replace_once:cne, \tl_replace_once:cee,
-% \tl_replace_once:cxn, \tl_replace_once:cnx, \tl_replace_once:cxx
-% }
-% \begin{macro}
-% {
-% \tl_greplace_once:Nnn, \tl_greplace_once:NVn, \tl_greplace_once:NnV,
-% \tl_greplace_once:Nen, \tl_greplace_once:Nne, \tl_greplace_once:Nee,
-% \tl_greplace_once:Nxn, \tl_greplace_once:Nnx, \tl_greplace_once:Nxx,
-% \tl_greplace_once:cnn, \tl_greplace_once:cVn, \tl_greplace_once:cnV,
-% \tl_greplace_once:cen, \tl_greplace_once:cne, \tl_greplace_once:cee,
-% \tl_greplace_once:cxn, \tl_greplace_once:cnx, \tl_greplace_once:cxx
-% }
-% \begin{macro}
-% {
-% \tl_replace_all:Nnn, \tl_replace_all:NVn, \tl_replace_all:NnV,
-% \tl_replace_all:Nen, \tl_replace_all:Nne, \tl_replace_all:Nee,
-% \tl_replace_all:Nxn, \tl_replace_all:Nnx, \tl_replace_all:Nxx,
-% \tl_replace_all:cnn, \tl_replace_all:cVn, \tl_replace_all:cnV,
-% \tl_replace_all:cen, \tl_replace_all:cne, \tl_replace_all:cee,
-% \tl_replace_all:cxn, \tl_replace_all:cnx, \tl_replace_all:cxx
-% }
-% \begin{macro}
-% {
-% \tl_greplace_all:Nnn, \tl_greplace_all:NVn, \tl_greplace_all:NnV,
-% \tl_greplace_all:Nen, \tl_greplace_all:Nne, \tl_greplace_all:Nee,
-% \tl_greplace_all:Nxn, \tl_greplace_all:Nnx, \tl_greplace_all:Nxx,
-% \tl_greplace_all:cnn, \tl_greplace_all:cVn, \tl_greplace_all:cnV,
-% \tl_greplace_all:cen, \tl_greplace_all:cne, \tl_greplace_all:cee,
-% \tl_greplace_all:cxn, \tl_greplace_all:cnx, \tl_greplace_all:cxx
-% }
-% All of the \texttt{replace} functions call \cs{@@_replace:NnNNNnn}
-% with appropriate arguments. The first two arguments are explained
-% later. The next controls whether the replacement function calls
-% itself (\cs{@@_replace_next:w}) or stops (\cs{@@_replace_wrap:w})
-% after the first replacement. Next comes an \texttt{e}-type
-% assignment function \cs{tl_set:Ne} or \cs{tl_gset:Ne} for local or
-% global replacements. Finally, the three arguments \meta{tl~var}
-% \Arg{pattern} \Arg{replacement} provided by the user. When
-% describing the auxiliary functions below, we denote the contents of
-% the \meta{tl~var} by \meta{token list}.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_replace_once:Nnn
- { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_wrap:w \__kernel_tl_set:Nx }
-\cs_new_protected:Npn \tl_greplace_once:Nnn
- { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_wrap:w \__kernel_tl_gset:Nx }
-\cs_new_protected:Npn \tl_replace_all:Nnn
- { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_next:w \__kernel_tl_set:Nx }
-\cs_new_protected:Npn \tl_greplace_all:Nnn
- { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_next:w \__kernel_tl_gset:Nx }
-\cs_generate_variant:Nn \tl_replace_once:Nnn
- { NnV , Nne , NV , Ne , Nee , c , cnV , cne , cV , ce , cee }
-\cs_generate_variant:Nn \tl_replace_once:Nnn
- { Nx , Nnx , Nxx , cxn , cnx , cxx }
-\cs_generate_variant:Nn \tl_greplace_once:Nnn
- { NnV , Nne , NV , Ne , Nee , c , cnV , cne , cV , ce , cee }
-\cs_generate_variant:Nn \tl_greplace_once:Nnn
- { Nx , Nnx , Nxx , cxn , cnx , cxx }
-\cs_generate_variant:Nn \tl_replace_all:Nnn
- { NnV , Nne , NV , Ne , Nee , c , cnV , cne , cV , ce , cee }
-\cs_generate_variant:Nn \tl_replace_all:Nnn
- { Nx , Nnx , Nxx , cxn , cnx , cxx }
-\cs_generate_variant:Nn \tl_greplace_all:Nnn
- { NnV , Nne , NV , Ne , Nee , c , cnV , cne , cV , ce , cee }
-\cs_generate_variant:Nn \tl_greplace_all:Nnn
- { Nx , Nnx , Nxx , cxn , cnx , cxx }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \@@_replace:NnNNNnn,
-% \@@_replace_auxi:NnnNNNnn,
-% \@@_replace_auxii:nNNNnn,
-% \@@_replace_next:w,
-% \@@_replace_next_aux:w,
-% \@@_replace_wrap:w,
-% }
-% To implement the actual replacement auxiliary
-% \cs{@@_replace_auxii:nNNNnn} we need a \meta{delimiter} with
-% the following properties:
-% \begin{itemize}
-% \item all occurrences of the \meta{pattern}~|#6| in
-% \enquote{\meta{token list} \meta{delimiter}} belong to the
-% \meta{token list} and have no overlap with the \meta{delimiter},
-% \item the first occurrence of the \meta{delimiter} in
-% \enquote{\meta{token list} \meta{delimiter}} is the trailing
-% \meta{delimiter}.
-% \end{itemize}
-% We first find the building blocks for the \meta{delimiter}, namely
-% two tokens \meta{A} and~\meta{B} such that \meta{A} does not appear
-% in~|#6| and |#6| is not~\meta{B} (this condition is trivial if |#6|
-% has more than one token). Then we consider the delimiters
-% \enquote{\meta{A}} and \enquote{\meta{A} \meta{A}$^n$ \meta{B}
-% \meta{A}$^n$ \meta{B}}, for $n\geq 1$, where $\meta{A}^n$ denotes
-% $n$~copies of \meta{A}, and we choose as our \meta{delimiter} the
-% first one which is not in the \meta{token list}.
-%
-% Every delimiter in the set obeys the first condition: |#6|~does not
-% contain~\meta{A} hence cannot be overlapping with the \meta{token
-% list} and the \meta{delimiter}, and it cannot be within the
-% \meta{delimiter} since it would have to be in one of the two
-% \meta{B} hence be equal to this single token (or empty, but this is
-% an error case filtered separately). Given the particular form of
-% these delimiters, for which no prefix is also a suffix, the second
-% condition is actually a consequence of the weaker condition that the
-% \meta{delimiter} we choose does not appear in the \meta{token list}.
-% Additionally, the set of delimiters is such that a \meta{token list}
-% of $n$~tokens can contain at most $O(n^{1/2})$ of them, hence we
-% find a \meta{delimiter} with at most $O(n^{1/2})$ tokens in a time
-% at most $O(n^{3/2})$. Bear in mind that these upper bounds are
-% reached only in very contrived scenarios: we include the case
-% \enquote{\meta{A}} in the list of delimiters to try, so that the
-% \meta{delimiter} is simply \cs{q_@@_mark} in the most common
-% situation where neither the \meta{token list} nor the \meta{pattern}
-% contains \cs{q_@@_mark}.
-%
-% Let us now ahead, optimizing for this most common case. First, two
-% special cases: an empty \meta{pattern}~|#6| is an error, and if
-% |#1|~is absent from both the \meta{token list}~|#5| and the
-% \meta{pattern}~|#6| then we can use it as the \meta{delimiter}
-% through \cs{@@_replace_auxii:nNNNnn} |{#1}|. Otherwise, we end up
-% calling \cs{@@_replace:NnNNNnn} repeatedly with the first two
-% arguments \cs{q_@@_mark} |{?}|, |\?| |{??}|, |\??| |{???}|, and so on,
-% until |#6|~does not contain the control sequence~|#1|, which we take
-% as our~\meta{A}. The argument~|#2| only serves to collect~|?|
-% characters for~|#1|. Note that the order of the tests means that
-% the first two are done every time, which is wasteful (for instance,
-% we repeatedly test for the emptyness of~|#6|). However, this is
-% rare enough not to matter. Finally, choose~\meta{B} to be
-% \cs{q_@@_nil} or~\cs{q_@@_stop} such that it is not equal to~|#6|.
-%
-% The \cs{@@_replace_auxi:NnnNNNnn} auxiliary receives \Arg{A} and
-% |{|\meta{A}$^n$\meta{B}|}| as its arguments, initially with $n=1$.
-% If \enquote{\meta{A} \meta{A}$^n$\meta{B} \meta{A}$^n$\meta{B}} is
-% in the \meta{token list} then increase~$n$ and try again. Once it
-% is not anymore in the \meta{token list} we take it as our
-% \meta{delimiter} and pass this to the \texttt{auxii} auxiliary.
-% \begin{macrocode}
-\cs_new_protected:Npn \@@_replace:NnNNNnn #1#2#3#4#5#6#7
- {
- \tl_if_empty:nTF {#6}
- {
- \msg_error:nne { kernel } { empty-search-pattern }
- { \tl_to_str:n {#7} }
- }
- {
- \tl_if_in:onTF { #5 #6 } {#1}
- {
- \tl_if_in:nnTF {#6} {#1}
- { \exp_args:Nc \@@_replace:NnNNNnn {#2} {#2?} }
- {
- \@@_quark_if_nil:nTF {#6}
- { \@@_replace_auxi:NnnNNNnn #5 {#1} { #1 \q_@@_stop } }
- { \@@_replace_auxi:NnnNNNnn #5 {#1} { #1 \q_@@_nil } }
- }
- }
- { \@@_replace_auxii:nNNNnn {#1} }
- #3#4#5 {#6} {#7}
- }
- }
-\cs_new_protected:Npn \@@_replace_auxi:NnnNNNnn #1#2#3
- {
- \tl_if_in:NnTF #1 { #2 #3 #3 }
- { \@@_replace_auxi:NnnNNNnn #1 { #2 #3 } {#2} }
- { \@@_replace_auxii:nNNNnn { #2 #3 #3 } }
- }
-% \end{macrocode}
-% The auxiliary \cs{@@_replace_auxii:nNNNnn} receives the following
-% arguments:
-% \begin{quote}
-% \Arg{delimiter} \meta{function} \meta{assignment} \\
-% \meta{tl~var} \Arg{pattern} \Arg{replacement}
-% \end{quote}
-% All of its work is done between
-% \cs{group_align_safe_begin:} and \cs{group_align_safe_end:} to avoid
-% issues in alignments. It does the actual replacement within
-% |#3|~|#4|~|{...}|, an \texttt{e}-expanding \meta{assignment}~|#3| to
-% the \meta{tl~var}~|#4|. The auxiliary \cs{@@_replace_next:w} is
-% called, followed by the \meta{token list}, some tokens including the
-% \meta{delimiter}~|#1|, followed by the \meta{pattern}~|#5|.
-% This auxiliary finds an argument delimited by~|#5| (the presence of
-% a trailing~|#5| avoids runaway arguments) and calls
-% \cs{@@_replace_wrap:w} to test whether this |#5| is found within the
-% \meta{token list} or is the trailing one.
-%
-% If on the one hand it is found within the \meta{token list}, then
-% |##1| cannot contain the \meta{delimiter}~|#1| that we worked so
-% hard to obtain, thus \cs{@@_replace_wrap:w} gets~|##1| as its own
-% argument~|##1|, and protects it against
-% the \texttt{e}-expanding assignment. It also finds \cs{exp_not:n}
-% as~|##2| and does nothing to it, thus letting through \cs{exp_not:n}
-% \Arg{replacement} into the assignment. Note that
-% \cs{@@_replace_next:w} and \cs{@@_replace_wrap:w} are always called
-% followed by two empty brace groups. These are safe because no
-% delimiter can match them. They prevent losing braces when grabbing
-% delimited arguments, but require the use of \cs{exp_not:o} and
-% \cs{use_none:nn}, rather than simply \cs{exp_not:n}.
-% Afterwards, \cs{@@_replace_next:w} is called
-% to repeat the replacement, or \cs{@@_replace_wrap:w} if we only want
-% a single replacement. In this second case, |##1| is the
-% \meta{remaining tokens} in the \meta{token list} and |##2| is some
-% \meta{ending code} which ends the assignment and removes the
-% trailing tokens |#5| using some \cs{if_false:} |{| \cs{fi:} |}|
-% trickery because~|#5| may contain any delimiter.
-%
-% If on the other hand the argument~|##1| of \cs{@@_replace_next:w} is
-% delimited by the trailing \meta{pattern}~|#5|, then |##1| is
-% \enquote{\{ \} \{ \} \meta{token list} \meta{delimiter}
-% \Arg{ending code}}, hence \cs{@@_replace_wrap:w} finds
-% \enquote{\{ \} \{ \} \meta{token list}} as |##1| and the
-% \meta{ending code} as~|##2|. It leaves the \meta{token list} into
-% the assignment and unbraces the \meta{ending code} which removes
-% what remains (essentially the \meta{delimiter} and
-% \meta{replacement}).
-% \begin{macrocode}
-\cs_new_protected:Npn \@@_replace_auxii:nNNNnn #1#2#3#4#5#6
- {
- \group_align_safe_begin:
- \cs_set:Npn \@@_replace_wrap:w ##1 #1 ##2
- { \__kernel_exp_not:w \exp_after:wN { \use_none:nn ##1 } ##2 }
- \cs_set:Npe \@@_replace_next:w ##1 #5
- {
- \exp_not:N \@@_replace_wrap:w ##1
- \exp_not:n { #1 }
- \exp_not:n { \exp_not:n {#6} }
- \exp_not:n { #2 { } { } }
- }
- #3 #4
- {
- \exp_after:wN \@@_replace_next_aux:w
- #4
- #1
- {
- \if_false: { \fi: }
- \exp_after:wN \use_none:n \exp_after:wN { \if_false: } \fi:
- }
- #5
- }
- \group_align_safe_end:
- }
-\cs_new:Npn \@@_replace_next_aux:w { \@@_replace_next:w { } { } }
-\cs_new_eq:NN \@@_replace_wrap:w ?
-\cs_new_eq:NN \@@_replace_next:w ?
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_remove_once:Nn, \tl_remove_once:NV, \tl_remove_once:Ne,
-% \tl_remove_once:cn, \tl_remove_once:cV, \tl_remove_once:ce
-% }
-% \begin{macro}
-% {
-% \tl_gremove_once:Nn, \tl_gremove_once:NV,
-% \tl_gremove_once:cn, \tl_gremove_once:cV,
-% }
-% Removal is just a special case of replacement.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_remove_once:Nn #1#2
- { \tl_replace_once:Nnn #1 {#2} { } }
-\cs_new_protected:Npn \tl_gremove_once:Nn #1#2
- { \tl_greplace_once:Nnn #1 {#2} { } }
-\cs_generate_variant:Nn \tl_remove_once:Nn { NV , Ne , c , cV , ce }
-\cs_generate_variant:Nn \tl_gremove_once:Nn { NV , Ne , c , cV , ce }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_remove_all:Nn, \tl_remove_all:NV, \tl_remove_all:Ne,
-% \tl_remove_all:Nx,
-% \tl_remove_all:cn, \tl_remove_all:cV, \tl_remove_all:ce,
-% \tl_remove_all:cx,
-% }
-% \begin{macro}
-% {
-% \tl_gremove_all:Nn, \tl_gremove_all:NV, \tl_gremove_all:Ne,
-% \tl_gremove_all:Nx,
-% \tl_gremove_all:cn, \tl_gremove_all:cV, \tl_gremove_all:ce,
-% \tl_gremove_all:cx,
-% }
-% Removal is just a special case of replacement.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_remove_all:Nn #1#2
- { \tl_replace_all:Nnn #1 {#2} { } }
-\cs_new_protected:Npn \tl_gremove_all:Nn #1#2
- { \tl_greplace_all:Nnn #1 {#2} { } }
-\cs_generate_variant:Nn \tl_remove_all:Nn { NV , Ne , c , cV , ce }
-\cs_generate_variant:Nn \tl_remove_all:Nn { Nx , cx }
-\cs_generate_variant:Nn \tl_gremove_all:Nn { NV , Ne , c , cV , ce }
-\cs_generate_variant:Nn \tl_gremove_all:Nn { Nx , cx }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \subsection{Token list conditionals}
-%
-% \begin{macro}[pTF]{\tl_if_empty:N, \tl_if_empty:c}
-% These functions check whether the token list in the argument is
-% empty and execute the proper code from their argument(s).
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_empty:N #1 { p , T , F , TF }
- {
- \if_meaning:w #1 \c_empty_tl
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_empty:N
- { c } { p , T , F , TF }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[pTF]{\tl_if_empty:n, \tl_if_empty:V, \tl_if_empty:e}
-% The \cs{if:w} triggers the expansion of \cs{tl_to_str:n} which converts the
-% argument to a string: this is empty if and only if the argument is. Then
-% |\if:w \scan_stop: ... \scan_stop:| is \texttt{true} if and only if the
-% string |...| is empty.
-% It could be tempting to use |\if:w \scan_stop: #1 \scan_stop:| directly.
-% But this fails on a token list expanding to anything starting with
-% \cs{scan_stop:} leaving everything that follows in the input stream.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_empty:n #1 { p , TF , T , F }
- {
- \if:w \scan_stop: \tl_to_str:n {#1} \scan_stop:
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_empty:n
- { V , e } { p , TF , T , F }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[pTF,documented-as=\tl_if_empty:nTF]{\tl_if_empty:o}
-% \begin{macro}[EXP]{\@@_if_empty_if:o}
-% The auxiliary function \cs{@@_if_empty_if:o} is for use
-% in various token list conditionals which reduce to testing
-% if a given token list is empty after applying a simple function
-% to it.
-% The test for emptiness is based on \cs{tl_if_empty:nTF}, but
-% the expansion is hard-coded for efficiency, as this auxiliary
-% function is used in several places.
-% We don't put \cs{prg_return_true:} and so on in the definition of
-% the auxiliary, because that would prevent an optimization applied to
-% conditionals that end with this code.
-% Also the |\@@_if_empty_if:o| is expanded once in |\tl_if_empty:oTF| for
-% efficiency as well (and to reduce code doubling).
-% \begin{macrocode}
-\cs_new:Npn \@@_if_empty_if:o #1
- {
- \if:w \scan_stop: \__kernel_tl_to_str:w \exp_after:wN {#1} \scan_stop:
- }
-\exp_args:Nno \use:n
- { \prg_new_conditional:Npnn \tl_if_empty:o #1 { p , TF , T , F } }
- {
- \@@_if_empty_if:o {#1}
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[pTF]{\tl_if_blank:n, \tl_if_blank:V, \tl_if_blank:o}
-% \begin{macro}{\@@_if_blank_p:NNw}
-% \TeX{} skips spaces when reading a non-delimited arguments. Thus,
-% a \meta{token list} is blank if and only if \cs{use_none:n}
-% \meta{token list} |?| is empty after one expansion. The auxiliary
-% \cs{@@_if_empty_if:o} is a fast emptyness test, converting its
-% argument to a string (after one expansion) and using the test
-% \cs{if:w} \cs{scan_stop:} |...| \cs{scan_stop:}.
-% \begin{macrocode}
-\exp_args:Nno \use:n
- { \prg_new_conditional:Npnn \tl_if_blank:n #1 { p , T , F , TF } }
- {
- \@@_if_empty_if:o { \use_none:n #1 ? }
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_blank:n
- { e , V , o } { p , T , F , TF }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[pTF]{\tl_if_eq:NN, \tl_if_eq:Nc, \tl_if_eq:cN, \tl_if_eq:cc}
-% Returns \cs{c_true_bool} if and only if the two token list variables are
-% equal.
-% \begin{macrocode}
-\prg_new_eq_conditional:NNn \tl_if_eq:NN \cs_if_eq:NN { p , T , F , TF }
-\prg_generate_conditional_variant:Nnn \tl_if_eq:NN
- { Nc , c , cc } { p , TF , T , F }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{variable}{\l_@@_internal_a_tl, \l_@@_internal_b_tl}
-% Temporary storage.
-% \begin{macrocode}
-\tl_new:N \l_@@_internal_a_tl
-\tl_new:N \l_@@_internal_b_tl
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{macro}[TF]{\tl_if_eq:Nn}
-% A simple store and compare routine.
-% \begin{macrocode}
-\prg_new_protected_conditional:Npnn \tl_if_eq:Nn #1#2 { T , F , TF }
- {
- \group_begin:
- \tl_set:Nn \l_@@_internal_b_tl {#2}
- \exp_after:wN
- \group_end:
- \if_meaning:w #1 \l_@@_internal_b_tl
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_eq:Nn { c } { TF , T , F }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[TF]
-% {
-% \tl_if_eq:nn, \tl_if_eq:nV, \tl_if_eq:ne, \tl_if_eq:Vn, \tl_if_eq:en,
-% \tl_if_eq:ee,
-% \tl_if_eq:xn, \tl_if_eq:nx, \tl_if_eq:xx,
-% }
-% A simple store and compare routine.
-% \begin{macrocode}
-\prg_new_protected_conditional:Npnn \tl_if_eq:nn #1#2 { T , F , TF }
- {
- \group_begin:
- \tl_set:Nn \l_@@_internal_a_tl {#1}
- \tl_set:Nn \l_@@_internal_b_tl {#2}
- \exp_after:wN
- \group_end:
- \if_meaning:w \l_@@_internal_a_tl \l_@@_internal_b_tl
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_eq:nn
- { nV , ne , nx , V, e , ee , x , xx }
- { TF , T , F }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[TF]
-% {
-% \tl_if_in:Nn, \tl_if_in:NV, \tl_if_in:No,
-% \tl_if_in:cn, \tl_if_in:cV, \tl_if_in:co
-% }
-% See \cs{tl_if_in:nnTF} for further comments. Here we simply
-% expand the token list variable and pass it to \cs{tl_if_in:nnTF}.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_if_in:NnT { \exp_args:No \tl_if_in:nnT }
-\cs_new_protected:Npn \tl_if_in:NnF { \exp_args:No \tl_if_in:nnF }
-\cs_new_protected:Npn \tl_if_in:NnTF { \exp_args:No \tl_if_in:nnTF }
-\prg_generate_conditional_variant:Nnn \tl_if_in:Nn
- { NV , No , c , cV , co } { T , F , TF }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[TF]
-% {
-% \tl_if_in:nn, \tl_if_in:Vn, \tl_if_in:VV, \tl_if_in:on, \tl_if_in:oo,
-% \tl_if_in:nV, \tl_if_in:no
-% }
-% Once more, the test relies on the emptiness test for robustness.
-% The function \cs{@@_tmp:w} removes tokens until the first occurrence
-% of |#2|. If this does not appear in |#1|, then the final |#2| is removed,
-% leaving an empty token list. Otherwise some tokens remain, and the
-% test is \texttt{false}. See \cs{tl_if_empty:nTF} for details on
-% the emptiness test.
-%
-% Treating correctly cases like
-% |\tl_if_in:nnTF {a state}{states}|, where |#1#2| contains |#2| before
-% the end, requires special care.
-% To cater for this case, we insert |{}{}| between the two token
-% lists. This marker may not appear in |#2| because of \TeX{} limitations
-% on what can delimit a parameter, hence we are safe. Using two brace
-% groups makes the test work also for empty arguments.
-% The \cs{if_false:} constructions are a faster way to do
-% \cs{group_align_safe_begin:} and \cs{group_align_safe_end:}.
-% The \cs{scan_stop:} ensures that \texttt{f}-expanding
-% \cs{tl_if_in:nnTF} does not lead to unbalanced braces.
-% \begin{macrocode}
-\prg_new_protected_conditional:Npnn \tl_if_in:nn #1#2 { T , F , TF }
- {
- \scan_stop:
- \if_false: { \fi:
- \cs_set:Npn \@@_tmp:w ##1 #2 { }
- \tl_if_empty:oTF { \@@_tmp:w #1 {} {} #2 }
- { \prg_return_false: } { \prg_return_true: }
- \if_false: } \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_in:nn
- { V , VV , o , oo , nV , no } { T , F , TF }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[pTF, EXP]{\tl_if_novalue:n}
-% \begin{macro}[EXP]{\@@_if_novalue:w}
-% Tests whether |##1| matches |-NoValue-| exactly (with suitable
-% catcodes): this is similar to \cs{quark_if_nil:nTF}. The first
-% argument of \cs{@@_if_novalue:w} is empty if and only if |##1|
-% starts with |-NoValue-|, while the second argument is empty if |##1|
-% is exactly |-NoValue-| or if it has a question mark just following
-% |-NoValue-|. In this second case, however, the material after the
-% first |?!| remains and makes the emptyness test return
-% \texttt{false}.
-% \begin{macrocode}
-\cs_set_protected:Npn \@@_tmp:w #1
- {
- \prg_new_conditional:Npnn \tl_if_novalue:n ##1
- { p , T , F , TF }
- {
- \@@_if_empty_if:o { \@@_if_novalue:w {} ##1 {} ? ! #1 ? ? ! }
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
- \cs_new:Npn \@@_if_novalue:w ##1 #1 ##2 ? ##3 ? ! { ##1 ##2 }
- }
-\exp_args:No \@@_tmp:w { \c_novalue_tl }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[EXP,pTF]{\tl_if_single:N, \tl_if_single:c}
-% Expand the token list and feed it to \cs{tl_if_single:nTF}.
-% \begin{macrocode}
-\cs_new:Npn \tl_if_single_p:N { \exp_args:No \tl_if_single_p:n }
-\cs_new:Npn \tl_if_single:NT { \exp_args:No \tl_if_single:nT }
-\cs_new:Npn \tl_if_single:NF { \exp_args:No \tl_if_single:nF }
-\cs_new:Npn \tl_if_single:NTF { \exp_args:No \tl_if_single:nTF }
-\prg_generate_conditional_variant:Nnn \tl_if_single:N {c} { p , T , F , TF }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}[EXP,pTF]{\tl_if_single:n}
-% \begin{macro}[EXP]{\@@_if_single:nnw}
-% This test is similar to \cs{tl_if_empty:nTF}. Expanding
-% \cs{use_none:nn} |#1| |??| once yields an empty result if |#1| is
-% blank, a single~|?| if |#1| has a single item, and otherwise yields
-% some tokens ending with |??|. Then, \cs{__kernel_tl_to_str:w} makes sure
-% there are no odd category codes. An earlier version would compare
-% the result to a single~|?| using string comparison, but the Lua call
-% is slow in \LuaTeX{}. Instead, \cs{@@_if_single:nnw} picks the
-% second token in front of it. If |#1| is empty, this token is
-% the trailing~|?| and the |\if:w| test yields \texttt{false}. If
-% |#1| has a single item, the token is~|\scan_stop:| and the |\if:w| test
-% yields \texttt{true}. Otherwise, it is one of the characters
-% resulting from \cs{tl_to_str:n}, and the |\if:w| test yields
-% \texttt{false}. Note that \cs{if:w} and
-% \cs{__kernel_tl_to_str:w} are primitives that take care of
-% expansion.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_single:n #1 { p , T , F , TF }
- {
- \if:w \scan_stop: \exp_after:wN \@@_if_single:nnw
- \__kernel_tl_to_str:w
- \exp_after:wN { \use_none:nn #1 ?? } \scan_stop: ? \s_@@_stop
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\cs_new:Npn \@@_if_single:nnw #1#2#3 \s_@@_stop {#2}
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[EXP,pTF]{\tl_if_single_token:n}
-% There are four cases: empty token list, token list starting with a
-% normal token, with a brace group, or with a space token. If the
-% token list starts with a normal token, remove it and check for
-% emptiness. For the next case, an empty token list is not a single
-% token. Finally, we have a non-empty token list starting with a
-% space or a brace group. Applying \texttt{f}-expansion yields an
-% empty result if and only if the token list is a single space.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_single_token:n #1 { p , T , F , TF }
- {
- \tl_if_head_is_N_type:nTF {#1}
- { \@@_if_empty_if:o { \use_none:n #1 } }
- {
- \tl_if_empty:nTF {#1}
- { \if_false: }
- { \@@_if_empty_if:o { \exp:w \exp_end_continue_f:w #1 } }
- }
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-% \end{macrocode}
-% \end{macro}
-%
-% \subsection{Mapping over token lists}
-%
-% \begin{macro}
-% {
-% \tl_map_function:nN, \tl_map_function:NN, \tl_map_function:cN,
-% \@@_map_function:Nnnnnnnnn, \@@_map_function_end:w,
-% \@@_use_none_delimit_by_s_stop:w
-% }
-% Expandable loop macro for token lists. We use the internal scan
-% mark \cs{s_@@_stop} (defined later), which is not allowed to show up
-% in the token list |#1| since it is internal to \pkg{l3tl}. This
-% allows us a very fast test of whether some \meta{item} is the
-% end-marker \cs{s_@@_stop}, namely call
-% \cs{@@_use_none_delimit_by_s_stop:w} \meta{item} \meta{function}
-% \cs{s_@@_stop}, which calls \meta{function} if the \meta{item} is
-% the end-marker. To speed up the loop even more, only test one out
-% of eight items, and once we hit one of the eight end-markers,
-% go more slowly through the last few items of the list using
-% \cs{@@_map_function_end:w}.
-% \begin{macrocode}
-\cs_new:Npn \tl_map_function:nN #1#2
- {
- \@@_map_function:Nnnnnnnnn #2 #1
- \s_@@_stop \s_@@_stop \s_@@_stop \s_@@_stop
- \s_@@_stop \s_@@_stop \s_@@_stop \s_@@_stop
- \prg_break_point:Nn \tl_map_break: { }
- }
-\cs_new:Npn \tl_map_function:NN
- { \exp_args:No \tl_map_function:nN }
-\cs_generate_variant:Nn \tl_map_function:NN { c }
-\cs_new:Npn \@@_map_function:Nnnnnnnnn #1#2#3#4#5#6#7#8#9
- {
- \@@_use_none_delimit_by_s_stop:w
- #9 \@@_map_function_end:w \s_@@_stop
- #1 {#2} #1 {#3} #1 {#4} #1 {#5} #1 {#6} #1 {#7} #1 {#8} #1 {#9}
- \@@_map_function:Nnnnnnnnn #1
- }
-\cs_new:Npn \@@_map_function_end:w \s_@@_stop #1#2
- {
- \@@_use_none_delimit_by_s_stop:w #2 \tl_map_break: \s_@@_stop
- #1 {#2}
- \@@_map_function_end:w \s_@@_stop
- }
-\cs_new:Npn \@@_use_none_delimit_by_s_stop:w #1 \s_@@_stop { }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_map_inline:nn}
-% \begin{macro}{\tl_map_inline:Nn, \tl_map_inline:cn}
-% The inline functions are straight forward by now. We use a little
-% trick with the counter \cs{g__kernel_prg_map_int} to make
-% them nestable. We can also make use of \cs{@@_map_function:Nnnnnnnnn}
-% from before.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_map_inline:nn #1#2
- {
- \int_gincr:N \g__kernel_prg_map_int
- \cs_gset_protected:cpn
- { @@_map_ \int_use:N \g__kernel_prg_map_int :w } ##1 {#2}
- \exp_args:Nc \@@_map_function:Nnnnnnnnn
- { @@_map_ \int_use:N \g__kernel_prg_map_int :w }
- #1
- \s_@@_stop \s_@@_stop \s_@@_stop \s_@@_stop
- \s_@@_stop \s_@@_stop \s_@@_stop \s_@@_stop
- \prg_break_point:Nn \tl_map_break:
- { \int_gdecr:N \g__kernel_prg_map_int }
- }
-\cs_new_protected:Npn \tl_map_inline:Nn
- { \exp_args:No \tl_map_inline:nn }
-\cs_generate_variant:Nn \tl_map_inline:Nn { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_map_tokens:nn, \tl_map_tokens:Nn, \tl_map_tokens:cn,
-% \@@_map_tokens:nnnnnnnnn, \@@_map_tokens_end:w
-% }
-% Much like the function mapping.
-% \begin{macrocode}
-\cs_new:Npn \tl_map_tokens:nn #1#2
- {
- \@@_map_tokens:nnnnnnnnn {#2} #1
- \s_@@_stop \s_@@_stop \s_@@_stop \s_@@_stop
- \s_@@_stop \s_@@_stop \s_@@_stop \s_@@_stop
- \prg_break_point:Nn \tl_map_break: { }
- }
-\cs_new:Npn \tl_map_tokens:Nn
- { \exp_args:No \tl_map_tokens:nn }
-\cs_generate_variant:Nn \tl_map_tokens:Nn { c }
-\cs_new:Npn \@@_map_tokens:nnnnnnnnn #1#2#3#4#5#6#7#8#9
- {
- \@@_use_none_delimit_by_s_stop:w
- #9 \@@_map_tokens_end:w \s_@@_stop
- \use:n {#1} {#2} \use:n {#1} {#3} \use:n {#1} {#4} \use:n {#1} {#5}
- \use:n {#1} {#6} \use:n {#1} {#7} \use:n {#1} {#8} \use:n {#1} {#9}
- \@@_map_tokens:nnnnnnnnn {#1}
- }
-\cs_new:Npn \@@_map_tokens_end:w \s_@@_stop \use:n #1#2
- {
- \@@_use_none_delimit_by_s_stop:w #2 \tl_map_break: \s_@@_stop
- #1 {#2}
- \@@_map_tokens_end:w \s_@@_stop
- }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_map_variable:nNn}
-% \begin{macro}{\tl_map_variable:NNn, \tl_map_variable:cNn}
-% \begin{macro}{\@@_map_variable:Nnn}
-% \cs{tl_map_variable:nNn} \Arg{token list} \meta{tl~var}
-% \Arg{action} assigns \meta{tl~var} to each element and executes
-% \meta{action}. The assignment to \meta{tl~var} is done after the
-% quark test so that this variable does not get set to a quark.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_map_variable:nNn #1#2#3
- { \tl_map_tokens:nn {#1} { \@@_map_variable:Nnn #2 {#3} } }
-\cs_new_protected:Npn \@@_map_variable:Nnn #1#2#3
- { \tl_set:Nn #1 {#3} #2 }
-\cs_new_protected:Npn \tl_map_variable:NNn
- { \exp_args:No \tl_map_variable:nNn }
-\cs_generate_variant:Nn \tl_map_variable:NNn { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_map_break:}
-% \begin{macro}{\tl_map_break:n}
-% The break statements use the general \cs{prg_map_break:Nn}.
-% \begin{macrocode}
-\cs_new:Npn \tl_map_break:
- { \prg_map_break:Nn \tl_map_break: { } }
-\cs_new:Npn \tl_map_break:n
- { \prg_map_break:Nn \tl_map_break: }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \subsection{Using token lists}
-%
-% \begin{macro}{\tl_to_str:n, \tl_to_str:o, \tl_to_str:V, \tl_to_str:v, \tl_to_str:e}
-% Another name for a primitive: defined in \pkg{l3basics}.
-% \begin{macrocode}
-\cs_generate_variant:Nn \tl_to_str:n { o , V , v , e }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_to_str:N, \tl_to_str:c}
-% These functions return the replacement text of a token list as a
-% string.
-% \begin{macrocode}
-\cs_new:Npn \tl_to_str:N #1 { \__kernel_tl_to_str:w \exp_after:wN {#1} }
-\cs_generate_variant:Nn \tl_to_str:N { c }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_use:N, \tl_use:c}
-% Token lists which are simply not defined give a clear \TeX{}
-% error here. No such luck for ones equal to \cs{scan_stop:} so
-% instead a test is made and if there is an issue an error is forced.
-% \begin{macrocode}
-\cs_new:Npn \tl_use:N #1
- {
- \tl_if_exist:NTF #1 {#1}
- {
- \msg_expandable_error:nnn
- { kernel } { bad-variable } {#1}
- }
- }
-\cs_generate_variant:Nn \tl_use:N { c }
-% \end{macrocode}
-% \end{macro}
-%
-% \subsection{Working with the contents of token lists}
-%
-% \begin{macro}{\tl_count:n, \tl_count:V, \tl_count:v, \tl_count:e, \tl_count:o}
-% \begin{macro}{\tl_count:N, \tl_count:c}
-% \begin{macro}{\@@_count:n}
-% Count number of elements within a token list or token list
-% variable. Brace groups within the list are read as a single
-% element. Spaces are ignored.
-% \cs{@@_count:n} grabs the element and replaces it by |+1|.
-% The |0| ensures that it works on an empty list.
-% \begin{macrocode}
-\cs_new:Npn \tl_count:n #1
- {
- \int_eval:n
- { 0 \tl_map_function:nN {#1} \@@_count:n }
- }
-\cs_new:Npn \tl_count:N #1
- {
- \int_eval:n
- { 0 \tl_map_function:NN #1 \@@_count:n }
- }
-\cs_new:Npn \@@_count:n #1 { + 1 }
-\cs_generate_variant:Nn \tl_count:n { V , v , e , o }
-\cs_generate_variant:Nn \tl_count:N { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-%
-% \begin{macro}[EXP]{\tl_count_tokens:n}
-% \begin{macro}[EXP]{\@@_act_count_normal:nN,
-% \@@_act_count_group:nn, \@@_act_count_space:n}
-% The token count is computed through an \cs{int_eval:n} construction.
-% Each \texttt{1+} is output to the \emph{left}, into the integer
-% expression, and the sum is ended by the \cs{exp_end:} inserted by
-% \cs{@@_act_end:wn} (which is technically implemented as \cs{c_zero_int}).
-% Somewhat a hack!
-% \begin{macrocode}
-\cs_new:Npn \tl_count_tokens:n #1
- {
- \int_eval:n
- {
- \@@_act:NNNn
- \@@_act_count_normal:N
- \@@_act_count_group:n
- \@@_act_count_space:
- {#1}
- }
- }
-\cs_new:Npn \@@_act_count_normal:N #1 { 1 + }
-\cs_new:Npn \@@_act_count_space: { 1 + }
-\cs_new:Npn \@@_act_count_group:n #1 { 2 + \tl_count_tokens:n {#1} + }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_reverse_items:n}
-% \begin{macro}{\@@_reverse_items:nwNwn}
-% \begin{macro}{\@@_reverse_items:wn}
-% Reversal of a token list is done by taking one item at a time
-% and putting it after \cs{s_@@_stop}.
-% \begin{macrocode}
-\cs_new:Npn \tl_reverse_items:n #1
- {
- \@@_reverse_items:nwNwn #1 ?
- \s_@@_mark \@@_reverse_items:nwNwn
- \s_@@_mark \@@_reverse_items:wn
- \s_@@_stop { }
- }
-\cs_new:Npn \@@_reverse_items:nwNwn #1 #2 \s_@@_mark #3 #4 \s_@@_stop #5
- {
- #3 #2
- \s_@@_mark \@@_reverse_items:nwNwn
- \s_@@_mark \@@_reverse_items:wn
- \s_@@_stop { {#1} #5 }
- }
-\cs_new:Npn \@@_reverse_items:wn #1 \s_@@_stop #2
- { \__kernel_exp_not:w \exp_after:wN { \use_none:nn #2 } }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {
-% \tl_trim_spaces:n, \tl_trim_spaces:V, \tl_trim_spaces:v,
-% \tl_trim_spaces:e,
-% \tl_trim_spaces:o,
-% }
-% \begin{macro}{\tl_trim_spaces_apply:nN, \tl_trim_spaces_apply:oN}
-% \begin{macro}
-% {
-% \tl_trim_spaces:N, \tl_trim_spaces:c,
-% \tl_gtrim_spaces:N, \tl_gtrim_spaces:c
-% }
-% Trimming spaces from around the input is deferred to an internal
-% function whose first argument is the token list to trim, augmented
-% by an initial \cs{@@_trim_mark:}, and whose second argument is a
-% \meta{continuation}, which receives as a braced argument
-% \cs{@@_trim_mark:} \meta{trimmed token list}. The control sequence
-% \cs{@@_trim_mark:} expands to nothing in a single expansion. In the case
-% at hand, we take \cs{__kernel_exp_not:w} \cs{exp_after:wN} as our
-% continuation, so that space trimming behaves correctly within an
-% \texttt{e}-type or \texttt{x}-type expansion.
-% \begin{macrocode}
-\cs_new:Npn \tl_trim_spaces:n #1
- {
- \@@_trim_spaces:nn
- { \@@_trim_mark: #1 }
- { \__kernel_exp_not:w \exp_after:wN }
- }
-\cs_generate_variant:Nn \tl_trim_spaces:n { V , v , e , o }
-\cs_new:Npn \tl_trim_spaces_apply:nN #1#2
- { \@@_trim_spaces:nn { \@@_trim_mark: #1 } { \exp_args:No #2 } }
-\cs_generate_variant:Nn \tl_trim_spaces_apply:nN { o }
-\cs_new_protected:Npn \tl_trim_spaces:N #1
- { \__kernel_tl_set:Nx #1 { \exp_args:No \tl_trim_spaces:n {#1} } }
-\cs_new_protected:Npn \tl_gtrim_spaces:N #1
- { \__kernel_tl_gset:Nx #1 { \exp_args:No \tl_trim_spaces:n {#1} } }
-\cs_generate_variant:Nn \tl_trim_spaces:N { c }
-\cs_generate_variant:Nn \tl_gtrim_spaces:N { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\@@_trim_spaces:nn}
-% \begin{macro}
-% {
-% \@@_trim_spaces_auxi:w, \@@_trim_spaces_auxii:w,
-% \@@_trim_spaces_auxiii:w, \@@_trim_spaces_auxiv:w
-% }
-% \begin{macro}{\@@_trim_mark:}
-% Trimming spaces from around the input is done using delimited
-% arguments and quarks, and to get spaces at odd places in the
-% definitions, we nest those in \cs{@@_tmp:w}, which then receives
-% a single space as its argument: |#1| is \verb*+ +.
-% Removing leading spaces is done with \cs{@@_trim_spaces_auxi:w},
-% which loops until \cs{@@_trim_mark:}\verb*+ + matches the end of the token
-% list: then |##1| is the token list and |##3| is
-% \cs{@@_trim_spaces_auxii:w}. This hands the relevant tokens to the
-% loop \cs{@@_trim_spaces_auxiii:w}, responsible for trimming
-% trailing spaces. The end is reached when \verb*+ + \cs{s_@@_nil}
-% matches the one present in the definition of \cs{tl_trim_spaces:n}.
-% Then \cs{@@_trim_spaces_auxiv:w} puts the token list into a group,
-% with a lingering \cs{@@_trim_mark:} at the start (which will expand to
-% nothing in one step of expansion), and feeds this to the
-% \meta{continuation}.
-% \begin{macrocode}
-\cs_set_protected:Npn \@@_tmp:w #1
- {
- \cs_new:Npn \@@_trim_spaces:nn ##1
- {
- \@@_trim_spaces_auxi:w
- ##1
- \s_@@_nil
- \@@_trim_mark: #1 { }
- \@@_trim_mark: \@@_trim_spaces_auxii:w
- \@@_trim_spaces_auxiii:w
- #1 \s_@@_nil
- \@@_trim_spaces_auxiv:w
- \s_@@_stop
- }
- \cs_new:Npn
- \@@_trim_spaces_auxi:w ##1 \@@_trim_mark: #1 ##2 \@@_trim_mark: ##3
- {
- ##3
- \@@_trim_spaces_auxi:w
- \@@_trim_mark:
- ##2
- \@@_trim_mark: #1 {##1}
- }
- \cs_new:Npn \@@_trim_spaces_auxii:w
- \@@_trim_spaces_auxi:w \@@_trim_mark: \@@_trim_mark: ##1
- {
- \@@_trim_spaces_auxiii:w
- ##1
- }
- \cs_new:Npn \@@_trim_spaces_auxiii:w ##1 #1 \s_@@_nil ##2
- {
- ##2
- ##1 \s_@@_nil
- \@@_trim_spaces_auxiii:w
- }
- \cs_new:Npn \@@_trim_spaces_auxiv:w ##1 \s_@@_nil ##2 \s_@@_stop ##3
- { ##3 { ##1 } }
- \cs_new:Npn \@@_trim_mark: {}
- }
-\@@_tmp:w { ~ }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}
-% {\tl_sort:Nn, \tl_sort:cn, \tl_gsort:Nn, \tl_gsort:cn, \tl_sort:nN}
-% Implemented in \pkg{l3sort}.
-% \end{macro}
-%
-% \subsection{The first token from a token list}
-%
-% \begin{macro}{\tl_head:N, \tl_head:n, \tl_head:V, \tl_head:v, \tl_head:f}
-% \begin{macro}{\@@_head_auxi:nw, \@@_head_auxii:n}
-% \begin{macro}{\tl_head:w,\@@_tl_head:w}
-% \begin{macro}{\tl_tail:N, \tl_tail:n, \tl_tail:V, \tl_tail:v, \tl_tail:f}
-% Finding the head of a token list expandably always strips braces, which
-% is fine as this is consistent with for example mapping over a list. The
-% empty brace groups in \cs{tl_head:n} ensure that a blank argument gives an
-% empty result. The result is returned within the \tn{unexpanded} primitive.
-% The approach here is to use \cs{if_false:} to allow us to use |}| as
-% the closing delimiter: this is the only safe choice, as any other token
-% would not be able to parse it's own code. More detail in
-% \url{http://tex.stackexchange.com/a/70168}.
-% \begin{macrocode}
-\cs_new:Npn \tl_head:n #1
- {
- \__kernel_exp_not:w \tex_expanded:D
- { { \if_false: { \fi: \@@_head_aux:n #1 { } } } }
- }
-\cs_new:Npn \@@_head_aux:n #1
- {
- \__kernel_exp_not:w {#1}
- \exp_after:wN \use_none:n \exp_after:wN { \if_false: } \fi:
- }
-\cs_generate_variant:Nn \tl_head:n { V , v , f }
-\cs_new:Npn \tl_head:w #1#2 \q_stop {#1}
-\cs_new:Npn \@@_tl_head:w #1#2 \s_@@_stop {#1}
-\cs_new:Npn \tl_head:N { \exp_args:No \tl_head:n }
-% \end{macrocode}
-% To correctly leave the tail of a token list, it's important \emph{not} to
-% absorb any of the tail part as an argument. For example, the simple
-% definition
-% \begin{verbatim}
-% \cs_new:Npn \tl_tail:n #1 { \tl_tail:w #1 \q_stop }
-% \cs_new:Npn \tl_tail:w #1#2 \q_stop
-% \end{verbatim}
-% would give the wrong result for |\tl_tail:n { a { bc } }| (the braces would
-% be stripped). Thus the only safe way to proceed is to first check that
-% there is an item to grab (\emph{i.e.}~that the argument is not blank) and
-% assuming there is to dispose of the first item. As with \cs{tl_head:n},
-% the result is protected from further expansion by \tn{unexpanded}.
-% While we could optimise the test here, this would leave some tokens
-% \enquote{banned} in the input, which we do not have with this definition.
-% \begin{macrocode}
-\exp_args:Nno \use:n { \cs_new:Npn \tl_tail:n #1 }
- {
- \exp_after:wN \__kernel_exp_not:w
- \tl_if_blank:nTF {#1}
- { { } }
- { \exp_after:wN { \use_none:n #1 } }
- }
-\cs_generate_variant:Nn \tl_tail:n { V , v , f }
-\cs_new:Npn \tl_tail:N { \exp_args:No \tl_tail:n }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[pTF]
-% {
-% \tl_if_head_eq_meaning:nN, \tl_if_head_eq_meaning:VN,
-% \tl_if_head_eq_meaning:eN
-% }
-% \begin{macro}[pTF]
-% {
-% \tl_if_head_eq_charcode:nN, \tl_if_head_eq_charcode:VN,
-% \tl_if_head_eq_charcode:eN, \tl_if_head_eq_charcode:fN
-% }
-% \begin{macro}[pTF]
-% {
-% \tl_if_head_eq_catcode:nN, \tl_if_head_eq_catcode:VN,
-% \tl_if_head_eq_catcode:eN, \tl_if_head_eq_catcode:oN
-% }
-% \begin{macro}[EXP]{\@@_head_exp_not:w}
-% \begin{macro}[EXP]{\@@_if_head_eq_empty_arg:w}
-% Accessing the first token of a token list is tricky in three cases:
-% when it has category code $1$ (begin-group token), when it is an
-% explicit space, with category code $10$ and character code $32$, or
-% when the token list is empty (obviously).
-%
-% Forgetting temporarily about this issue we would use the following
-% test in \cs{tl_if_head_eq_charcode:nN}. Here, \cs{tl_head:w} yields
-% the first token of the token list, then passed to \cs{exp_not:N}.
-% \begin{verbatim}
-% \if_charcode:w
-% \exp_after:wN \exp_not:N \tl_head:w #1 \q_nil \q_stop
-% \exp_not:N #2
-% \end{verbatim}
-% The two first special cases are detected by testing if the token
-% list starts with an \texttt{N}-type token (the extra |?| sends empty
-% token lists to the \texttt{true} branch of this test). In those
-% cases, the first token is a character, and since we only care about
-% its character code, we can use \cs{str_head:n} to access it (this
-% works even if it is a space character). An empty argument
-% results in \cs{tl_head:w} leaving two token: |^| and
-% \cs{@@_if_head_eq_empty_arg:w} which will result in the \cs{if_charcode:w}
-% test being false and remove \cs{exp_not:N} and |#2|.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_head_eq_charcode:nN #1#2 { p , T , F , TF }
- {
- \if_charcode:w
- \tl_if_head_is_N_type:nTF { #1 ? }
- { \@@_head_exp_not:w #1 { ^ \@@_if_head_eq_empty_arg:w } \s_@@_stop }
- { \str_head:n {#1} }
- \exp_not:N #2
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_head_eq_charcode:nN
- { V , e , f } { p , TF , T , F }
-% \end{macrocode}
-% For \cs{tl_if_head_eq_catcode:nN}, again we detect special cases
-% with a \cs{tl_if_head_is_N_type:n}. Then we need to test if the
-% first token is a begin-group token or an explicit space token, and
-% produce the relevant token, either \cs{c_group_begin_token} or
-% \cs{c_space_token}. Again, for an empty argument, a hack is used,
-% removing the token given by the user and leaving two tokens in the input
-% stream which will make the \cs{if_catcode:w} test return \texttt{false}.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_head_eq_catcode:nN #1 #2 { p , T , F , TF }
- {
- \if_catcode:w
- \tl_if_head_is_N_type:nTF { #1 ? }
- { \@@_head_exp_not:w #1 { ^ \@@_if_head_eq_empty_arg:w } \s_@@_stop }
- {
- \tl_if_head_is_group:nTF {#1}
- \c_group_begin_token
- \c_space_token
- }
- \exp_not:N #2
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\prg_generate_conditional_variant:Nnn \tl_if_head_eq_catcode:nN
- { V , e , o } { p , TF , T , F }
-% \end{macrocode}
-% For \cs{tl_if_head_eq_meaning:nN}, again, detect special cases. In
-% the normal case, use \cs{tl_head:w}, with no \cs{exp_not:N} this
-% time, since \cs{if_meaning:w} causes no expansion. With an empty
-% argument, the test is \texttt{true}, and \cs{use_none:nnn} removes
-% |#2| and \cs{prg_return_true:} and \cs{else:} (it is safe this way here as
-% in this case \cs{prg_new_conditional:Npnn} didn't optimize these two away).
-% In the special cases, we know that the first token is a character,
-% hence \cs{if_charcode:w} and \cs{if_catcode:w} together are enough.
-% We combine them in some order, hopefully faster than the reverse.
-% Tests are not nested because the arguments may contain unmatched
-% primitive conditionals.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_head_eq_meaning:nN #1#2 { p , T , F , TF }
- {
- \tl_if_head_is_N_type:nTF { #1 ? }
- \@@_if_head_eq_meaning_normal:nN
- \@@_if_head_eq_meaning_special:nN
- {#1} #2
- }
-\prg_generate_conditional_variant:Nnn \tl_if_head_eq_meaning:nN
- { V , e } { p , TF , T , F }
-\cs_new:Npn \@@_if_head_eq_meaning_normal:nN #1 #2
- {
- \exp_after:wN \if_meaning:w
- \@@_tl_head:w #1 { ?? \use_none:nnn } \s_@@_stop #2
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\cs_new:Npn \@@_if_head_eq_meaning_special:nN #1 #2
- {
- \if_charcode:w \str_head:n {#1} \exp_not:N #2
- \exp_after:wN \use_ii:nn
- \else:
- \prg_return_false:
- \fi:
- \use_none:n
- {
- \if_catcode:w \exp_not:N #2
- \tl_if_head_is_group:nTF {#1}
- { \c_group_begin_token }
- { \c_space_token }
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
- }
-% \end{macrocode}
-%
-% Both \cs{tl_if_head_eq_charcode:nN} and \cs{tl_if_head_eq_catcode:nN} will
-% need to get the first token of their argument and apply \cs{exp_not:N} to
-% it. \cs{@@_head_exp_not:w} does exactly that.
-% \begin{macrocode}
-\cs_new:Npn \@@_head_exp_not:w #1 #2 \s_@@_stop
- { \exp_not:N #1 }
-% \end{macrocode}
-%
-% If the argument of \cs{tl_if_head_eq_charcode:nN} and
-% \cs{tl_if_head_eq_catcode:nN} was empty \cs{@@_if_head_eq_empty_arg:w} will
-% be left in the input stream. This macro has to remove \cs{exp_not:N} and
-% the following token from the input stream to make sure no unbalanced
-% if-construct is created and leave tokens there which make the two tests
-% return \texttt{false}.
-% \begin{macrocode}
-\cs_new:Npn \@@_if_head_eq_empty_arg:w \exp_not:N #1
- { ? }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[pTF]{\tl_if_head_is_N_type:n}
-% \begin{macro}[EXP]
-% {
-% \@@_if_head_is_N_type_auxi:w ,
-% \@@_if_head_is_N_type_auxii:n ,
-% }
-% A token list can be empty, can start with an explicit space
-% character (catcode 10 and charcode 32), can start with a begin-group
-% token (catcode 1), or start with an \texttt{N}-type argument. In
-% the first two cases, and when |#1~| starts with |{}~|,
-% \cs{@@_if_head_is_N_type_auxi:w} receives an empty argument hence
-% produces |f| and removes everything before the first
-% \cs{scan_stop:}. In the third case (except when |#1~| starts with
-% |{}~|), the second auxiliary removes the first copy of~|#1| that was
-% used for the space test, then expands \cs{token_to_str:N} which hits
-% the leading begin-group token, leaving a single closing brace to be
-% compared with \cs{scan_stop:}. In the last case,
-% \cs{token_to_str:N} does not change the brace balance so that only
-% \cs{scan_stop:} \cs{scan_stop:} remain, making the character code
-% test true. One cannot optimize by
-% moving one of the \cs{scan_stop:} to the beginning: if |#1| contains
-% primitive conditionals, all of its occurrences must be dealt with before
-% the \cs{if:w} tries to skip the \texttt{true} branch of the
-% conditional.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_head_is_N_type:n #1 { p , T , F , TF }
- {
- \if:w
- \if_false: { \fi: \@@_if_head_is_N_type_auxi:w #1 ~ }
- { \exp_after:wN { \token_to_str:N #1 } }
- \scan_stop: \scan_stop:
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\exp_args:Nno \use:n { \cs_new:Npn \@@_if_head_is_N_type_auxi:w #1 ~ }
- {
- \tl_if_empty:nTF {#1}
- { f \exp_after:wN \use_none:nn }
- { \exp_after:wN \@@_if_head_is_N_type_auxii:n }
- \exp_after:wN { \if_false: } \fi:
- }
-\cs_new:Npn \@@_if_head_is_N_type_auxii:n #1
- { \exp_after:wN \use_none:n \exp_after:wN }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[EXP,pTF]{\tl_if_head_is_group:n}
-% \begin{macro}[EXP]{\@@_if_head_is_group_fi_false:w}
-% Pass the first token of |#1| through \cs{token_to_str:N}, then check
-% for the brace balance. The extra \texttt{?} caters for an empty
-% argument. This could be made faster, but we need all brace tricks
-% to happen in one step of expansion, keeping the token list brace
-% balanced at all times.
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_head_is_group:n #1 { p , T , F , TF }
- {
- \if:w
- \exp_after:wN \use_none:n
- \exp_after:wN { \exp_after:wN { \token_to_str:N #1 ? } }
- \scan_stop: \scan_stop:
- \@@_if_head_is_group_fi_false:w
- \fi:
- \if_true:
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\cs_new:Npn \@@_if_head_is_group_fi_false:w \fi: \if_true: { \fi: \if_false: }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[EXP,pTF]{\tl_if_head_is_space:n}
-% \begin{macro}[EXP]{\@@_if_head_is_space:w}
-% The auxiliary's argument is all that is before the first explicit
-% space in |\prg_do_nothing:#1?~|. If that is a single~|\prg_do_nothing:| the
-% test yields \texttt{true}. Otherwise, that is more than one token, and the
-% test yields \texttt{false}. The work is done within braces (with an
-% |\if_false: { \fi: ... }| construction) both to hide potential
-% alignment tab characters from \TeX{} in a table, and to allow for
-% removing what remains of the token list after its first space. The use of
-% \cs{if:w} ensures that the result of a single step of expansion directly
-% yields a balanced token list (no trailing closing brace).
-% \begin{macrocode}
-\prg_new_conditional:Npnn \tl_if_head_is_space:n #1 { p , T , F , TF }
- {
- \if:w
- \if_false: { \fi: \@@_if_head_is_space:w \prg_do_nothing: #1 ? ~ }
- \scan_stop: \scan_stop:
- \prg_return_true:
- \else:
- \prg_return_false:
- \fi:
- }
-\exp_args:Nno \use:n { \cs_new:Npn \@@_if_head_is_space:w #1 ~ }
- {
- \@@_if_empty_if:o {#1} \else: f \fi:
- \exp_after:wN \use_none:n \exp_after:wN { \if_false: } \fi:
- }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \subsection{Token by token changes}
-%
-% \begin{variable}{\s_@@_act_stop}
-% The \cs[no-index]{@@_act_\ldots{}} functions may be applied to any token list.
-% Hence, we use a private quark, to allow any token, even quarks,
-% in the token list.
-% Only \cs{s_@@_act_stop} may not appear in the token lists manipulated by
-% \cs{@@_act:NNNn} functions.
-% \begin{macrocode}
-\scan_new:N \s_@@_act_stop
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{macro}[EXP]{\@@_act:NNNn}
-% \begin{macro}[EXP]{\@@_act_output:n, \@@_act_reverse_output:n}
-% \begin{macro}[EXP]{\@@_act_loop:w}
-% \begin{macro}[EXP]{\@@_act_normal:NwNNN}
-% \begin{macro}[EXP]{\@@_act_group:nwNNN}
-% \begin{macro}[EXP]{\@@_act_space:wwNNN}
-% \begin{macro}[EXP]{\@@_act_end:wn}
-% \begin{macro}[EXP]
-% {
-% \@@_act_if_head_is_space:nTF,
-% \@@_act_if_head_is_space:w,
-% \@@_act_if_head_is_space_true:w
-% }
-% \begin{macro}[EXP]{\@@_use_none_delimit_by_q_act_stop:w}
-% To help control the expansion, \cs{@@_act:NNNn} should always
-% be preceeded by \cs{exp:w} and ends by producing \cs{exp_end:}
-% once the result has been obtained. This way no internal token of it can be
-% accidentally end up in the input stream.
-% Because \cs{s_@@_act_stop} can't appear without braces around it in the
-% argument~|#1| of \cs{@@_act_loop:w}, we can use this marker to set up a fast
-% test for leading spaces.
-% \begin{macrocode}
-\cs_set_protected:Npn \@@_tmp:w #1
- {
- \cs_new:Npn \@@_act_if_head_is_space:nTF ##1
- {
- \@@_act_if_head_is_space:w
- \s_@@_act_stop ##1 \s_@@_act_stop \@@_act_if_head_is_space_true:w
- \s_@@_act_stop #1 \s_@@_act_stop \use_ii:nn
- }
- \cs_new:Npn \@@_act_if_head_is_space:w
- ##1 \s_@@_act_stop #1 ##2 \s_@@_act_stop
- {}
- \cs_new:Npn \@@_act_if_head_is_space_true:w
- \s_@@_act_stop #1 \s_@@_act_stop \use_ii:nn ##1 ##2
- {##1}
- }
-\@@_tmp:w { ~ }
-% \end{macrocode}
-% (We expand the definition \cs{@@_act_if_head_is_space:nTF} when
-% setting up \cs{@@_act_loop:w}, so we can then undefine the auxiliary.)
-% In the loop, we check how the token list begins and act
-% accordingly. In the \enquote{group} case, we may have
-% reached \cs{s_@@_act_stop}, the end of the list. Then
-% leave \cs{exp_end:} and the result in the input stream,
-% to terminate the expansion of \cs{exp:w}.
-% Otherwise, apply the relevant function to the
-% \enquote{arguments}, |#3|
-% and to the head of the token list. Then repeat the loop.
-% The scheme is the same if the token list starts with an |N|-type
-% or with a space, making sure that
-% \cs{@@_act_space:wwNNN} gobbles the space.
-% \begin{macrocode}
-\exp_args:Nne \use:n { \cs_new:Npn \@@_act_loop:w #1 \s_@@_act_stop }
- {
- \exp_not:o { \@@_act_if_head_is_space:nTF {#1} }
- \exp_not:N \@@_act_space:wwNNN
- {
- \exp_not:o { \tl_if_head_is_group:nTF {#1} }
- \exp_not:N \@@_act_group:nwNNN
- \exp_not:N \@@_act_normal:NwNNN
- }
- \exp_not:n {#1} \s_@@_act_stop
- }
-\cs_undefine:N \@@_act_if_head_is_space:nTF
-\cs_new:Npn \@@_act_normal:NwNNN #1 #2 \s_@@_act_stop #3
- {
- #3 #1
- \@@_act_loop:w #2 \s_@@_act_stop
- #3
- }
-\cs_new:Npn \@@_use_none_delimit_by_s_act_stop:w #1 \s_@@_act_stop { }
-\cs_new:Npn \@@_act_end:wn #1 \@@_act_result:n #2
- { \group_align_safe_end: \exp_end: #2 }
-\cs_new:Npn \@@_act_group:nwNNN #1 #2 \s_@@_act_stop #3#4#5
- {
- \@@_use_none_delimit_by_s_act_stop:w #1 \@@_act_end:wn \s_@@_act_stop
- #5 {#1}
- \@@_act_loop:w #2 \s_@@_act_stop
- #3 #4 #5
- }
-\exp_last_unbraced:NNo
- \cs_new:Npn \@@_act_space:wwNNN \c_space_tl #1 \s_@@_act_stop #2#3
- {
- #3
- \@@_act_loop:w #1 \s_@@_act_stop
- #2 #3
- }
-% \end{macrocode}
-% \cs{@@_act:NNNn} loops over tokens, groups, and spaces in |#4|.
-% |{\s_@@_act_stop}| serves as the end of token list marker, the |?| after it
-% avoids losing outer braces. The result is stored as an argument for the
-% dummy function \cs{@@_act_result:n}.
-% \begin{macrocode}
-\cs_new:Npn \@@_act:NNNn #1#2#3#4
- {
- \group_align_safe_begin:
- \@@_act_loop:w #4 { \s_@@_act_stop } ? \s_@@_act_stop
- #1 #3 #2
- \@@_act_result:n { }
- }
-% \end{macrocode}
-% Typically, the output is done to the right of what was already output,
-% using \cs{@@_act_output:n}, but for the \cs{@@_act_reverse} functions,
-% it should be done to the left.
-% \begin{macrocode}
-\cs_new:Npn \@@_act_output:n #1 #2 \@@_act_result:n #3
- { #2 \@@_act_result:n { #3 #1 } }
-\cs_new:Npn \@@_act_reverse_output:n #1 #2 \@@_act_result:n #3
- { #2 \@@_act_result:n { #1 #3 } }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[EXP]
-% {\tl_reverse:n, \tl_reverse:o, \tl_reverse:V, \tl_reverse:f, \tl_reverse:e}
-% \begin{macro}[EXP]{\@@_reverse_normal:nN}
-% \begin{macro}[EXP]{\@@_reverse_group_preserve:nn}
-% \begin{macro}[EXP]{\@@_reverse_space:n}
-% The goal here is to reverse without losing spaces nor braces.
-% This is done using the general internal function \cs{@@_act:NNNn}.
-% Spaces and \enquote{normal} tokens are output on the left of the current
-% output. Grouped tokens are output to the left but without any reversal
-% within the group.
-% \begin{macrocode}
-\cs_new:Npn \tl_reverse:n #1
- {
- \__kernel_exp_not:w \exp_after:wN
- {
- \exp:w
- \@@_act:NNNn
- \@@_reverse_normal:N
- \@@_reverse_group_preserve:n
- \@@_reverse_space:
- {#1}
- }
- }
-\cs_generate_variant:Nn \tl_reverse:n { o , V , f , e }
-\cs_new:Npn \@@_reverse_normal:N
- { \@@_act_reverse_output:n }
-\cs_new:Npn \@@_reverse_group_preserve:n #1
- { \@@_act_reverse_output:n { {#1} } }
-\cs_new:Npn \@@_reverse_space:
- { \@@_act_reverse_output:n { ~ } }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_reverse:N, \tl_reverse:c, \tl_greverse:N, \tl_greverse:c}
-% This reverses the list, leaving \cs{exp_stop_f:} in front,
-% which stops the \texttt{f}-expansion.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_reverse:N #1
- { \__kernel_tl_set:Nx #1 { \exp_args:No \tl_reverse:n { #1 } } }
-\cs_new_protected:Npn \tl_greverse:N #1
- { \__kernel_tl_gset:Nx #1 { \exp_args:No \tl_reverse:n { #1 } } }
-\cs_generate_variant:Nn \tl_reverse:N { c }
-\cs_generate_variant:Nn \tl_greverse:N { c }
-% \end{macrocode}
-% \end{macro}
-%
-% \subsection{Using a single item}
-%
-% \begin{macro}{\tl_item:nn, \tl_item:Nn, \tl_item:cn}
-% \begin{macro}{\@@_item_aux:nn, \@@_item:nn}
-% The idea here is to find the offset of the item from the left, then use
-% a loop to grab the correct item. If the resulting offset is too large,
-% then \cs{@@_if_recursion_tail_break:nN} terminates the loop, and returns
-% nothing at all.
-% \begin{macrocode}
-\cs_new:Npn \tl_item:nn #1#2
- {
- \exp_args:Nf \@@_item:nn
- { \exp_args:Nf \@@_item_aux:nn { \int_eval:n {#2} } {#1} }
- #1
- \q_@@_recursion_tail
- \prg_break_point:
- }
-\cs_new:Npn \@@_item_aux:nn #1#2
- {
- \int_compare:nNnTF {#1} < 0
- { \int_eval:n { \tl_count:n {#2} + 1 + #1 } }
- {#1}
- }
-\cs_new:Npn \@@_item:nn #1#2
- {
- \@@_if_recursion_tail_break:nN {#2} \prg_break:
- \int_compare:nNnTF {#1} = 1
- { \prg_break:n { \exp_not:n {#2} } }
- { \exp_args:Nf \@@_item:nn { \int_eval:n { #1 - 1 } } }
- }
-\cs_new:Npn \tl_item:Nn { \exp_args:No \tl_item:nn }
-\cs_generate_variant:Nn \tl_item:Nn { c }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_rand_item:n, \tl_rand_item:N, \tl_rand_item:c}
-% Importantly \cs{tl_item:nn} only evaluates its argument once.
-% \begin{macrocode}
-\cs_new:Npn \tl_rand_item:n #1
- {
- \tl_if_blank:nF {#1}
- { \tl_item:nn {#1} { \int_rand:nn { 1 } { \tl_count:n {#1} } } }
- }
-\cs_new:Npn \tl_rand_item:N { \exp_args:No \tl_rand_item:n }
-\cs_generate_variant:Nn \tl_rand_item:N { c }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_range:Nnn, \tl_range:cnn, \tl_range:nnn}
-% \begin{macro}
-% {
-% \@@_range:Nnnn, \@@_range:nnnNn, \@@_range:nnNn, \@@_range_skip:w,
-% \@@_range:w, \@@_range_skip_spaces:n, \@@_range_collect:nn,
-% \@@_range_collect:ff, \@@_range_collect_space:nw,
-% \@@_range_collect_N:nN, \@@_range_collect_group:nN,
-% }
-% To avoid checking for the end of the token list at every step, start
-% by counting the number $l$ of items and \enquote{normalizing} the
-% bounds, namely clamping them to the interval $[0,l]$ and dealing
-% with negative indices. More precisely, \cs{@@_range_items:nnNn}
-% receives the number of items to skip at the beginning of the token
-% list, the index of the last item to keep, a function which is either
-% \cs{@@_range:w} or the token list itself. If nothing should be kept,
-% leave |{}|: this stops the \texttt{f}-expansion of \cs{tl_head:f} and that
-% function produces an empty result. Otherwise, repeatedly call
-% \cs{@@_range_skip:w} to delete |#1| items from the input stream (the
-% extra brace group avoids an off-by-one shift). For the braced
-% version \cs{@@_range_braced:w} sets up
-% \cs{@@_range_collect_braced:w} which stores items one by one in an
-% argument after the semicolon. Depending on the first token of the tail,
-% either just move it (if it is a space) or also decrement the number of
-% items left to find. Eventually, the result is a brace group followed by
-% the rest of the token list, and \cs{tl_head:f} cleans up and gives the
-% result in \cs{exp_not:n}.
-% \begin{macrocode}
-\cs_new:Npn \tl_range:Nnn { \exp_args:No \tl_range:nnn }
-\cs_generate_variant:Nn \tl_range:Nnn { c }
-\cs_new:Npn \tl_range:nnn { \@@_range:Nnnn \@@_range:w }
-\cs_new:Npn \@@_range:Nnnn #1#2#3#4
- {
- \tl_head:f
- {
- \exp_args:Nf \@@_range:nnnNn
- { \tl_count:n {#2} } {#3} {#4} #1 {#2}
- }
- }
-\cs_new:Npn \@@_range:nnnNn #1#2#3
- {
- \exp_args:Nff \@@_range:nnNn
- {
- \exp_args:Nf \@@_range_normalize:nn
- { \int_eval:n { #2 - 1 } } {#1}
- }
- {
- \exp_args:Nf \@@_range_normalize:nn
- { \int_eval:n {#3} } {#1}
- }
- }
-\cs_new:Npn \@@_range:nnNn #1#2#3#4
- {
- \if_int_compare:w #2 > #1 \exp_stop_f: \else:
- \exp_after:wN { \exp_after:wN }
- \fi:
- \exp_after:wN #3
- \int_value:w \int_eval:n { #2 - #1 } \exp_after:wN ;
- \exp_after:wN { \exp:w \@@_range_skip:w #1 ; { } #4 }
- }
-\cs_new:Npn \@@_range_skip:w #1 ; #2
- {
- \if_int_compare:w #1 > \c_zero_int
- \exp_after:wN \@@_range_skip:w
- \int_value:w \int_eval:n { #1 - 1 } \exp_after:wN ;
- \else:
- \exp_after:wN \exp_end:
- \fi:
- }
-\cs_new:Npn \@@_range:w #1 ; #2
- {
- \exp_args:Nf \@@_range_collect:nn
- { \@@_range_skip_spaces:n {#2} } {#1}
- }
-\cs_new:Npn \@@_range_skip_spaces:n #1
- {
- \tl_if_head_is_space:nTF {#1}
- { \exp_args:Nf \@@_range_skip_spaces:n {#1} }
- { { } #1 }
- }
-\cs_new:Npn \@@_range_collect:nn #1#2
- {
- \int_compare:nNnTF {#2} = 0
- {#1}
- {
- \exp_args:No \tl_if_head_is_space:nTF { \use_none:n #1 }
- {
- \exp_args:Nf \@@_range_collect:nn
- { \@@_range_collect_space:nw #1 }
- {#2}
- }
- {
- \@@_range_collect:ff
- {
- \exp_args:No \tl_if_head_is_N_type:nTF { \use_none:n #1 }
- { \@@_range_collect_N:nN }
- { \@@_range_collect_group:nn }
- #1
- }
- { \int_eval:n { #2 - 1 } }
- }
- }
- }
-\cs_new:Npn \@@_range_collect_space:nw #1 ~ { { #1 ~ } }
-\cs_new:Npn \@@_range_collect_N:nN #1#2 { { #1 #2 } }
-\cs_new:Npn \@@_range_collect_group:nn #1#2 { { #1 {#2} } }
-\cs_generate_variant:Nn \@@_range_collect:nn { ff }
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}[EXP]{\@@_range_normalize:nn}
-% This function converts an \meta{index} argument into an explicit
-% position in the token list (a result of $0$ denoting \enquote{out of
-% bounds}). Expects two explicit integer arguments: the \meta{index}
-% |#1| and the string count~|#2|. If |#1| is negative, replace it by
-% $|#1| + |#2| + 1$, then limit to the range $[0, |#2|]$.
-% \begin{macrocode}
-\cs_new:Npn \@@_range_normalize:nn #1#2
- {
- \int_eval:n
- {
- \if_int_compare:w #1 < \c_zero_int
- \if_int_compare:w #1 < -#2 \exp_stop_f:
- 0
- \else:
- #1 + #2 + 1
- \fi:
- \else:
- \if_int_compare:w #1 < #2 \exp_stop_f:
- #1
- \else:
- #2
- \fi:
- \fi:
- }
- }
-% \end{macrocode}
-% \end{macro}
-%
-% \subsection{Viewing token lists}
-%
-% \begin{macro}{\tl_show:N, \tl_show:c, \tl_log:N, \tl_log:c, \@@_show:NN}
-% Showing token list variables is done after checking that the
-% variable is defined (see \cs{__kernel_register_show:N}).
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_show:N { \@@_show:NN \tl_show:n }
-\cs_generate_variant:Nn \tl_show:N { c }
-\cs_new_protected:Npn \tl_log:N { \@@_show:NN \tl_log:n }
-\cs_generate_variant:Nn \tl_log:N { c }
-\cs_new_protected:Npn \@@_show:NN #1#2
- {
- \__kernel_chk_defined:NT #2
- {
- \exp_args:Nf \tl_if_empty:nTF
- { \cs_prefix_spec:N #2 \cs_parameter_spec:N #2 }
- {
- \exp_args:Ne #1
- { \token_to_str:N #2 = \__kernel_exp_not:w \exp_after:wN {#2} }
- }
- {
- \msg_error:nneee { kernel } { bad-type }
- { \token_to_str:N #2 } { \token_to_meaning:N #2 } { tl }
- }
- }
- }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\tl_show:n, \tl_show:e, \tl_show:x, \@@_show:n}
-% \begin{macro}[EXP]{\@@_show:w}
-% Many |show| functions are based on \cs{tl_show:n}.
-% The argument of \cs{tl_show:n} is line-wrapped using
-% \cs{iow_wrap:nnnN} but with a leading |>~| and trailing period, both
-% removed before passing the wrapped text to the \tn{showtokens}
-% primitive. This primitive shows the result with a leading |>~| and
-% trailing period.
-%
-% The token list \cs{l_@@_internal_a_tl} containing the result
-% of all these manipulations is displayed to the terminal using
-% \cs{tex_showtokens:D} and an odd \cs{exp_after:wN} which expand the
-% closing brace to improve the output slightly. The calls to
-% \cs{__kernel_iow_with:Nnn} ensure that the \tn{newlinechar} is set to~$10$
-% so that the \cs{iow_newline:} inserted by the line-wrapping code
-% are correctly recognized by \TeX{}, and that \tn{errorcontextlines}
-% is $-1$ to avoid printing irrelevant context.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_show:n #1
- { \iow_wrap:nnnN { >~ \tl_to_str:n {#1} . } { } { } \@@_show:n }
-\cs_generate_variant:Nn \tl_show:n { e , x }
-\cs_new_protected:Npn \@@_show:n #1
- {
- \tl_set:Nf \l_@@_internal_a_tl { \@@_show:w #1 \s_@@_stop }
- \__kernel_iow_with:Nnn \tex_newlinechar:D { 10 }
- {
- \__kernel_iow_with:Nnn \tex_errorcontextlines:D { -1 }
- {
- \tex_showtokens:D \exp_after:wN \exp_after:wN \exp_after:wN
- { \exp_after:wN \l_@@_internal_a_tl }
- }
- }
- }
-\cs_new:Npn \@@_show:w #1 > #2 . \s_@@_stop {#2}
-% \end{macrocode}
-% \end{macro}
-% \end{macro}
-%
-% \begin{macro}{\tl_log:n, \tl_log:e, \tl_log:x}
-% Logging is much easier, simply line-wrap. The |>~| and trailing
-% period is there to match the output of \cs{tl_show:n}.
-% \begin{macrocode}
-\cs_new_protected:Npn \tl_log:n #1
- { \iow_wrap:nnnN { > ~ \tl_to_str:n {#1} . } { } { } \iow_log:n }
-\cs_generate_variant:Nn \tl_log:n { e , x }
-% \end{macrocode}
-% \end{macro}
-%
-% \begin{macro}{\__kernel_chk_tl_type:NnnT}
-% Helper for checking that |#1| has the correct internal structure to
-% be of a certain type. Make sure that it is defined and that it is a
-% token list, namely a macro with no \tn{long} nor \tn{protected}
-% prefix. Then compare |#1| to an attempt at reconstructing a valid
-% structure of the given type using |#2| (see implementation of
-% \cs{seq_show:N} for instance). If that is successful run the
-% requested code~|#4|.
-% \begin{macrocode}
-\cs_new_protected:Npn \__kernel_chk_tl_type:NnnT #1#2#3#4
- {
- \__kernel_chk_defined:NT #1
- {
- \exp_args:Nf \tl_if_empty:nTF
- { \cs_prefix_spec:N #1 \cs_parameter_spec:N #1 }
- {
- \tl_set:Ne \l_@@_internal_a_tl {#3}
- \tl_if_eq:NNTF #1 \l_@@_internal_a_tl
- {#4}
- {
- \msg_error:nneeee { kernel } { bad-type }
- { \token_to_str:N #1 } { \tl_to_str:N #1 }
- {#2} { \tl_to_str:N \l_@@_internal_a_tl }
- }
- }
- {
- \msg_error:nneee { kernel } { bad-type }
- { \token_to_str:N #1 } { \token_to_meaning:N #1 } {#2}
- }
- }
- }
-% \end{macrocode}
-% \end{macro}
-%
-% \subsection{Internal scan marks}
-%
-% \begin{variable}{\s_@@_nil,\s_@@_mark,\s_@@_stop}
-% Internal scan marks. These are defined here at the end because the
-% code for \cs{scan_new:N} depends on some \pkg{l3tl} functions.
-% \begin{macrocode}
-\scan_new:N \s_@@_nil
-\scan_new:N \s_@@_mark
-\scan_new:N \s_@@_stop
-% \end{macrocode}
-% \end{variable}
-%
-% \subsection{Scratch token lists}
-%
-% \begin{variable}{\g_tmpa_tl, \g_tmpb_tl}
-% Global temporary token list variables.
-% They are supposed to be set and used immediately,
-% with no delay between the definition and the use because you
-% can't count on other macros not to redefine them from under you.
-% \begin{macrocode}
-\tl_new:N \g_tmpa_tl
-\tl_new:N \g_tmpb_tl
-% \end{macrocode}
-% \end{variable}
-%
-% \begin{variable}{\l_tmpa_tl, \l_tmpb_tl}
-% These are local temporary token list variables. Be sure not to assume
-% that the value you put into them will survive for
-% long---see discussion above.
-% \begin{macrocode}
-\tl_new:N \l_tmpa_tl
-\tl_new:N \l_tmpb_tl
-% \end{macrocode}
-% \end{variable}
-%
-% We finally clean up a temporary control sequence that we have used at
-% various points to set up some definitions.
-% \begin{macrocode}
-\cs_undefine:N \@@_tmp:w
-% \end{macrocode}
-%
-% \begin{macrocode}
-%</package>
-% \end{macrocode}
-%
-% \end{implementation}
-%
-% \PrintIndex