summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/simplebnf
diff options
context:
space:
mode:
authorNorbert Preining <norbert@preining.info>2023-11-25 03:03:12 +0000
committerNorbert Preining <norbert@preining.info>2023-11-25 03:03:12 +0000
commitad976e54f48605a66054bc56e945dabc425c956a (patch)
treef3905ec1ab486d1cf12b064887e311f2ea10bebf /macros/latex/contrib/simplebnf
parentac5a5e57a4c9b3c436c3bac327fe885acea28ba4 (diff)
CTAN sync 202311250303
Diffstat (limited to 'macros/latex/contrib/simplebnf')
-rw-r--r--macros/latex/contrib/simplebnf/simplebnf-doc.pdfbin156004 -> 80558 bytes
-rw-r--r--macros/latex/contrib/simplebnf/simplebnf-doc.tex645
-rw-r--r--macros/latex/contrib/simplebnf/simplebnf.sty577
3 files changed, 968 insertions, 254 deletions
diff --git a/macros/latex/contrib/simplebnf/simplebnf-doc.pdf b/macros/latex/contrib/simplebnf/simplebnf-doc.pdf
index f0b419d363..c8fb4b8444 100644
--- a/macros/latex/contrib/simplebnf/simplebnf-doc.pdf
+++ b/macros/latex/contrib/simplebnf/simplebnf-doc.pdf
Binary files differ
diff --git a/macros/latex/contrib/simplebnf/simplebnf-doc.tex b/macros/latex/contrib/simplebnf/simplebnf-doc.tex
index 72fa982e24..1181d3f203 100644
--- a/macros/latex/contrib/simplebnf/simplebnf-doc.tex
+++ b/macros/latex/contrib/simplebnf/simplebnf-doc.tex
@@ -1,18 +1,78 @@
-\documentclass[a4paper]{article}
+\documentclass[11pt]{article}
+\usepackage{geometry}
+\geometry{a4paper,top=1in,bottom=1in,left=1.2in,right=1.2in}
-\usepackage{simplebnf}
-\usepackage{hyperref}
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Font setup %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+% Works for pdfLaTeX, XeLaTeX, and LuaLaTeX
+\usepackage{libertine}
+
+% Math/Mono font setup
+\ExplSyntaxOn
+\bool_if:nTF { \sys_if_engine_xetex_p: || \sys_if_engine_luatex_p: }
+ {
+ % Load mathtools before unicode-math
+ \usepackage{mathtools}
+ \usepackage[
+ math-style=ISO,
+ warnings-off={mathtools-colon, mathtools-overbracket},
+ ]{unicode-math}
+ \setmathfont{LibertinusMath-Regular.otf}[Scale = MatchUppercase]
+ \setmathfont{Garamond-Math.otf}[
+ Scale = MatchUppercase,
+ range = {\Coloneq}
+ ]
+
+ \setmonofont{Inconsolatazi4}[
+ Extension = .otf,
+ UprightFont = *-Regular,
+ BoldFont = *-Bold,
+ Scale=MatchLowercase,
+ AutoFakeSlant = 0.2,
+ ]
+ }
+ {
+ \usepackage[libertine]{newtxmath}
+ \usepackage{inconsolata}
+ }
+\ExplSyntaxOff
+
+% Some fancy symbols (prevent being overwritten by math fonts)
+\AtBeginDocument{%
+ \usepackage[
+ only, fatsemi, fatslash, talloblong, rightarrowtriangle
+ ]{stmaryrd}%
+}
+
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Presentations %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\usepackage{tcolorbox}
\tcbuselibrary{listings,breakable}
-\tcbset{listing engine=listings,colframe=black,colback=white,size=small}
+\tcbset{listing engine=listings, colframe=black, colback=white, size=small}
+
+\NewTCBListing { example } { !O{} !s }
+ {
+ sharp corners,
+ IfBooleanTF = { #2 }
+ { listing above text }
+ { listing side text },
+ fontlower = \small,
+ #1
+ }
+
+\NewTCBListing { listing } { !O{} }
+ { sharp corners, listing only, #1 }
+
\NewDocumentEnvironment {exampleside} {}
- { \tcblisting{listing side text,righthand width=.55\textwidth} }
+ { \tcblisting{listing side text,righthand width=.5\textwidth} }
{ \endtcblisting }
-\NewDocumentCommand \cmd { m } {\texttt{\textbackslash#1}}
-
\NewDocumentEnvironment { presentcommand } { b }
{%
\vspace*{0.5\baselineskip}\noindent\fbox{%
@@ -22,84 +82,432 @@
}
{ }
+
+\NewDocumentCommand \cmd { m } {\texttt{\char`\\#1}}
\NewDocumentCommand \env { m m }
{
\texttt{%
- \textbackslash begin\{#1\} \textrm{#2}%
- \textbackslash end\{#1\}%
+ \char`\\begin\{#1\} \textrm{#2}%
+ \char`\\end\{#1\}%
}%
}
+
+\usepackage{tabularray}
+\UseTblrLibrary{booktabs}
+% \fakeverb
+\usepackage{codehigh}
+
+
+\newcommand*{\secref}[1]{\hyperref[{#1}]{section~\ref*{#1} -- \emph{\nameref*{#1}}}}
+\newcommand*{\subsecref}[1]{\hyperref[{#1}]{subsection~\ref*{#1} -- \emph{\nameref*{#1}}}}
+\newcommand*{\subsubsecref}[1]{\hyperref[{#1}]{\nameref*{#1}}}
+\newcommand*{\parref}[1]{\hyperref[{#1}]{\nameref*{#1}}}
+
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% simplebnf %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+\usepackage{simplebnf}
+
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Miscellaneous %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+\usepackage{hologo}
+\newcommand*{\XeLaTeX}{\hologo{XeLaTeX}}
+\newcommand*{\LuaLaTeX}{\hologo{LuaLaTeX}}
+
+% ``Make sure it comes *last* of your loaded packages"
+\usepackage[pdfborder={0 0 0}, colorlinks=true]{hyperref}
+
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Metadata %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
\title{%
\textsf{simplebnf} --- A simple package to format Backus-Naur form%
- \footnote{This file describes v0.3.2.}}
+ \footnote{This file describes v1.0.0.}}
\author{Jay Lee\footnote{E-mail: %
\href{mailto:jaeho.lee@snu.ac.kr}{\texttt{jaeho.lee@snu.ac.kr}}}}
-\date{2023/01/07}
+\date{2023/11/25}
+
\begin{document}
\maketitle
-This package provides a simple way to typeset grammars written in Backus-Naur form (BNF).
-
-\begin{presentcommand}
- \cmd{SimpleBNFDefEq}
-\end{presentcommand}
-This command is used to typeset the definition symbol separate a nonterminal from its productions. It defaults to \SimpleBNFDefEq. It can be redefined using \verb|RenewDocumentCommand|.
-
-\begin{presentcommand}
- \cmd{SimpleBNFDefOr}
-\end{presentcommand}
-This command is used to typeset the separator symbol between productions. It defaults to \SimpleBNFDefOr. It can be redefined using \verb|RenewDocumentCommand|.
-
-\begin{presentcommand}
- \cmd{SimpleBNFStretch}
-\end{presentcommand}
-This command is used to control the vertical spacing between consecutive rules.
-It defaults to 0.
-It can be redefined using \verb/Renewdocumentcommand/.
-
-\begin{presentcommand}
- \cmd{bnfexpr}
-\end{presentcommand}
-This command is used when typesetting the BNF nonterminal and productions. It defaults to a wrappers around \cmd{texttt}. It can be redefined to customized output using \verb|RenewDocumentCommand|.
-
-\begin{presentcommand}
- \cmd{bnfannot}
-\end{presentcommand}
-This command is used when typesetting the annotations on nonterminals and productions. It defaults to a wrappers around \cmd{textit}. It can be redefined to customized output using \verb|RenewDocumentCommand|.
-
-\begin{presentcommand}
- \env{bnfgrammar}{text}
-\end{presentcommand}
-can be used to typeset BNF grammars. The \textit{text} inside the environment should be formatted as:
-\begin{verbatim}
- term1 ::= rhs1
+
+\vfill
+This package provides a simple way for typesetting grammars in Backus-Naur form (BNF).
+It features a flexible configuration system, allowing for the customization of the domain-specific language (DSL) used in typesetting the grammar.
+Additionally, the package comes with sensible defaults.
+
+Below is the metagrammar of the DSL as defined in this package, which is typeset using the package itself.\footnote{The code is shown in \nameref{sec:appendix}.}
+\begin{tcolorbox}[sharp corners]
+ \begin{center}
+ \begin{bnf}(
+ prod-delim = ;;;,
+ new-line-delim = !,
+ single-line-delim = ?,
+ comment = //,
+ relation = {:::=|:in:},
+ relation-sym-map =
+ {
+ {:::=} = $\Coloneqq$,
+ {:in:} = $\in$,
+ },
+ )[
+ colspec = lrcll,
+ column{2} = {mode=dmath},
+ column{4} = {mode=text, font=\ttfamily},
+ ]
+ G // Gramar :::=
+ ! $P$ // production
+ ! $P \fatsemi {}$ // production w/ a trailing delimiter
+ ! $P \fatsemi G$ // production sequence
+ ;;;
+ P // Production :::= $L \rightarrowtriangle R$
+ ;;;
+ L // LHS :::=
+ ! $v$ // metavariable
+ ! $v\!\fatslash\,c$ // annotated metavariable
+ ;;;
+ R // RHS :::=
+ ! $\talloblong$ // delimiter
+ ! $A \talloblong R$ // alternative sequence
+ ;;;
+ A // Alternative :::=
+ ! $f$ // syntactic form
+ ! $f\!\fatslash\,c$ // annotated syntactic form
+ ;;;
+ \fatsemi // Prod. delimiter :::=
+ ! ;; // default symbol
+ ! $\cdots$ // user-defined
+ ;;;
+ \rightarrowtriangle // Rule relation :::=
+ ! ::= // $\Coloneqq$
+ ! -> // $\to$
+ ! \texttt{\char`\\in} // $\in$
+ ! $\cdots$ // user-defined
+ ;;;
+ \fatslash // Annot. symbol :::=
+ ! : // default symbol
+ ! $\cdots$ // user-defined
+ ;;;
+ \talloblong // Alt. delimiter :::=
+ ! | // new-line delimiter
+ ! || // single-line delimiter
+ ! $\cdots$ // user-defined
+ ;;;
+ v, f, c :in: \textsf{\TeX{} tl} // valid \TeX{} token lists
+ ;;;
+ \end{bnf}
+ \end{center}
+\end{tcolorbox}
+\vfill
+
+\section{Tutorial for the impatient}\label{sec:tutorial}
+Typesetting a grammar is as simple as typing the BNF grammar in the \verb/bnf/ environment:
+\begin{example}[righthand width=.52\linewidth]
+\begin{center}
+\begin{bnf}
+ $\tau$ : \textsf{Type} ::=
+ | \texttt{num} : numbers
+ | \texttt{str} : strings
+ ;;
+ $e$ : \textsf{Expr} ::=
+ | $x$ : variable
+ | $n$ : numeral
+ | \texttt{$e$ + $e$} : addition
+ | \texttt{$e$ * $e$} : multiplication
+ | \texttt{$e$ \textasciicircum{} $e$} : concatenation
+ | \texttt{len($e$)} : length
+ | \texttt{let $x$ = $e_1$ in $e_2$} : definition
+\end{bnf}
+\end{center}
+\end{example}
+
+Typically, each column in the BNF grammar has the same font style.
+In the example above, the comments in the first column, e.g., \textsf{Type}, is typeset in sans-serif, and the second column in math mode, e.g., $\tau$.
+The column for the right-hand side is (mostly) typeset in typewriter font with a bit of non-terminals like $e$ typeset in math mode.
+
+\textsf{simplebnf} provides a straightforward way to customize the font styles for each column\footnote{Note that you must provide a manual alignment specification with the key \texttt{colspec}.}:
+\begin{listing}
+\begin{bnf}[
+ colspec = {llcll},
+ column{1} = {font = \sffamily},
+ column{2} = {mode = dmath},
+ column{4} = {font = \ttfamily},
+]
+ \tau : Type ::=
+ | num : numbers
+ | str : strings
;;
- term2 ::= rhs2
+ e : Expr ::=
+ | $x$ : variable
+ | $n$ : numeral
+ | $e$ + $e$ : addition
+ | $e$ * $e$ : multiplication
+ | $e$ \textasciicircum{} $e$ : concatenation
+ | len($e$) : length
+ | let $x$ = $e_1$ in $e_2$ : definition
+\end{bnf}
+\end{listing}
+If you find yourself using the same configuration repeatedly, you can fix the desired configuration using \cmd{SetBNFLayout}.
+Once set using \cmd{SetBNFLayout}, the configuration is applied to all subsequent \verb/bnf/ environments:
+\begin{listing}
+% Some where above...
+\SetBNFLayout{
+ colspec = {llcll},
+ column{1} = {font = \sffamily},
+ column{2} = {mode = dmath},
+ column{4} = {font = \ttfamily},
+}
+% ...
+\begin{bnf}
+ \tau : Type ::=
+ | num : numbers
+ | str : strings
;;
- ...
- termk ::= rhsk
-\end{verbatim}
-where each of the \textit{rhs} represents alternative syntactic forms of the \textit{term}. An annotation may accompany each alternative in which case the alternative must be separated from its annotation with a colon (\verb/:/). If you don't need annotations, simply omit the colons and annotations altogether. The alternatives themselves are separated using the pipe symbol (\verb/|/).
+ e : Expr ::=
+ | $x$ : variable
+ | $n$ : numeral
+ | $e$ + $e$ : addition
+ | $e$ * $e$ : multiplication
+ | $e$ \textasciicircum{} $e$ : concatenation
+ | len($e$) : length
+ | let $x$ = $e_1$ in $e_2$ : definition
+\end{bnf}
+\end{listing}
-A sample code and the result is shown below:
-\begin{exampleside}
- \begin{bnfgrammar}
- a \in \textit{Vars}
- ;;
- expr ::=
- expr + term : sum
- | term : term
- ;;
- term ::=
- term * a : product
- | a : variable
- \end{bnfgrammar}
-\end{exampleside}
+Since these customizations are provided by the backend \textsf{tabularray}\footnote{\url{https://github.com/lvjr/tabularray}}, consult its documentation for more details.
+The corresponding command is \cmd{SetTblrInner}.
+
+For more advanced customization, such as changing default delimiters and symbols, continue to \secref{sec:config}.
-Annotations can also be provided on left-hand sides, to label the nonterminal instead of a specific production.
-\begin{exampleside}
+\section{Configuration}\label{sec:config}
+While the default configuration should be sufficient for most use cases, some grammars may require using different delimiters and symbols.
+For example, the language itself may contain \verb/->/ as a syntactic form, which conflicts with one of the default symbols for the rule relation.
+
+To this end, \textsf{simplebnf} provides a number of options to configure the DSL used to typeset the grammar.
+Some examples of customizable symbols include the delimiters for productions and alternatives, the relation symbol between the left-hand side and the right-hand side, the annotation symbol, and more.
+
+The default configuration is shown in \autoref{tab:config}.
+\begin{longtblr}[
+ caption = {Default key-values for the configuration of \textsf{simplebnf}.},
+ label = {tab:config}
+]{
+ colspec={ll},
+ row{2-Z}={font=\ttfamily}
+}
+ \toprule
+ Key & Default value \\
+ \midrule
+ prod-delim & ;;\\
+ new-line-delim & \fakeverb{\|}\\
+ single-line-delim & //\\
+ comment & :\\
+ relation & \fakeverb{{::=|->|:in:}}\\
+ relation-sym-map & {
+ \{\\
+ \quad\fakeverb{{::=} = {\ensuremath{\Coloneqq}},}\\
+ \quad\fakeverb{{->} = {\ensuremath{\to}},}\\
+ \quad\fakeverb{{:in:} = {\ensuremath{\in}},}\\
+ \}
+ }\\
+ or-sym & \fakeverb{$|$}\\
+ prod-sep & 2pt\\
+ row-sep & 0pt\\
+ \bottomrule
+\end{longtblr}
+
+The key-value list can be provided as an optional argument to the \verb/bnf/ environment, wrapped in \verb/()/.
+The optional key-value list for the layout---wrapped in \verb/[]/---should be provided after the configuration key-value list.
+See \subsecref{subsec:layout} for more details.
+
+An exmaple of customizing the configuration is shown below:
+\begin{example}*
+\begin{center}
+\begin{bnf}(
+ prod-delim = {--},
+ new-line-delim = {\&},
+ single-line-delim = {\&\&},
+ comment = {//},
+ relation-sym-map =
+ {
+ {->} = {\ensuremath{\hookrightarrow}},
+ {:in:} = {\ensuremath{\in}},
+ },
+ or-sym = {},
+)[
+ colspec = {llcll},
+ column{1} = {font = \sffamily},
+ column{2} = {mode = dmath},
+ column{4} = {font = \ttfamily},
+]
+ \tau // Type ->
+ & num // numbers
+ & str // strings
+ --
+ e // Expr ->
+ & $x$ // variable
+ & $n$ // numeral
+ & $e$ + $e$ // addition
+ & $e$ * $e$ // multiplication
+ & $e$ \textasciicircum{} $e$ // concatenation
+ & len($e$) // length
+ & let $x$ = $e_1$ in $e_2$ // definition
+\end{bnf}
+\end{center}
+\end{example}
+
+
+Furthermore, the configuration can be set globally using \cmd{SetBNFConfig}, similar to \cmd{SetBNFLayout}:
+\begin{listing}
+% Some where above...
+\SetBNFConfig{
+ prod-delim = {--},
+ new-line-delim = {\&},
+ single-line-delim = {\&\&},
+ comment = {//},
+ relation-sym-map =
+ {
+ {->} = {\ensuremath{\hookrightarrow}},
+ {:in:} = {\ensuremath{\in}},
+ },
+ or-sym = {},
+}
+% ...
+\begin{bnf}[
+ colspec = {llcll},
+ column{1} = {font = \sffamily},
+ column{2} = {mode = dmath},
+ column{4} = {font = \ttfamily},
+]
+ \tau // Type ->
+ & num // numbers
+ & str // strings
+ --
+ e // Expr ->
+ & $x$ // variable
+ & $n$ // numeral
+ & $e$ + $e$ // addition
+ & $e$ * $e$ // multiplication
+ & $e$ \textasciicircum{} $e$ // concatenation
+ & len($e$) // length
+ & let $x$ = $e_1$ in $e_2$ // definition
+\end{bnf}
+\end{listing}
+
+
+\subsection{Production delimiter}
+The default production delimiter \verb/prod-delim/ is \verb/;;/.
+This will separate different productions in the same grammar.
+
+\subsection{New-line alternative delimiter}
+The default new-line alternative delimiter \verb/new-line-delim/ is \verb/\|/, which will actually match verbatim \verb/\|/ in the grammar.
+This will separate different alternatives into different lines in the same production.
+
+\subsection{Single-line alternative delimiter}
+The default single-line alternative delimiter \verb/single-line-delim/ is \verb|//|.
+The difference between the single-line alternative delimiter and the new-line alternative delimiter is that the former will not add a new line after the delimiter.
+
+For consecutive alternatives delimited by the single-line alternative delimiter, only the last of them can be annotated.
+
+Note that the single-line alternative delimiter must not contain the new-line alternative delimiter as a substring.
+
+\subsection{Annotation delimiter}
+The default annotation delimiter \verb/comment/ is \verb/:/.
+This will separate the syntactic form and the annotation within the alternative.
+
+\subsection{Rule relations and the symbol map}
+\textsf{simplebnf} provides a \verb/::=/, \verb/->/, and \verb/:in:/ for the rule relation between the left-hand side and the right-hand side.
+Each of them is mapped to a symbol in the \verb/relation-sym-map/ key-value list; By default, they are mapped to $\Coloneqq$ (\verb/\ensuremath{\Coloneqq}/), $\to$ (\verb/\ensuremath{\to}/), and $\in$ (\verb/\ensuremath{\in}/), respectively.
+
+To provide a custom relation symbol, provide the desired symbol and the mapping to the \verb/relation-sym-map/ key-value list and to the \verb/relation/.
+
+\subsection{Or symbol}
+The default or symbol \verb/or-sym/ is \verb/$|$/.
+Probably the most common use case for this option is to remove the or symbol altogether by setting it to \verb/{}/.
+
+\subsection{Production separation}
+The default production separation \verb/prod-sep/ is \verb/2pt/.
+This will add a vertical space between the productions.
+
+\subsection{Row separation}
+The default row separation \verb/row-sep/ is \verb/0pt/.
+This will add a vertical space between the alternatives in the same production.
+
+\subsection{Layout}\label{subsec:layout}
+As mentioned in \secref{sec:tutorial}, the layout of the \verb/bnf/ environment can be customized per-enviroment using the key-value list in the optional argument wrapped in \verb/[]/.
+When provided, these should be provided after the configuration key-value list, which is wrapped in \verb/()/.
+
+The optional key-value list for the layout is passed to the backend \textsf{tabularray} package.
+
+An example of customizing the layout is shown below:
+\begin{example}[righthand width=.53\linewidth]
+\begin{center}
+\begin{bnf}(
+ prod-delim = {--},
+ or-sym = {},
+ prod-sep = 5pt,
+)[
+ colspec = {|[2pt, gray5]llcll},
+ column{1} = {font = \sffamily},
+ column{2} = {mode = dmath},
+ column{4} = {font = \ttfamily},
+ column{5} = {font = \itshape},
+ cell{9}{4} = {cmd = \fbox},
+ row{-} = {bg = azure9},
+]
+ \tau : Type ::=
+ | num : numbers
+ | str : strings
+ --
+ e : Expr ::=
+ | $x$ : variable
+ | $n$ : numeral
+ | $e$ + $e$ : addition
+ | $e$ * $e$ : multiplication
+ | $e$ \textasciicircum{} $e$ : concatenation
+ | len($e$) : length
+ | let $x$ = $e_1$ in $e_2$ : definition
+\end{bnf}
+\end{center}
+\end{example}
+
+
+\section{Caveats}
+The choices of delimiters and symbols should be carefully made, in order to avoid conflicts with the grammar of the target language and our meta-language, \TeX{}.
+
+Another subtlety to consider while customizing delimiters, is that certain delimiters not contain the others as a substring.
+For instance, as mentioned in \secref{sec:config}, the single-line alternative delimiter must not contain the new-line alternative delimiter as a substring.
+
+When using custom OpenType fonts with the \verb/unicode-math/ package in \XeLaTeX{} or \LuaLaTeX{}, beware of the fact that many fonts lack the \verb/\Coloneqq/ symbol, $\Coloneqq$.
+The easiest way to fix this is to actually use a font that provides the symbol, such as \verb/Garamond-Math/, which comes with the \TeX{}Live installation\footnote{\fakeverb{\Coloneq} is not a typo of \fakeverb{\Coloneqq}.}.
+\begin{listing}
+\setmathfont{Garamond-Math.otf}[
+ Scale = MatchUppercase,
+ range = {\Coloneq}
+]
+\end{listing}
+
+
+\section{The deprecated \texttt{bnfgrammar} environment}
+Prior to version 1.0.0, the package provided a sole environment \verb/bnfgrammar/ for typesetting grammars.
+While it is still available, it is now deprecated and will be removed in a future release.
+Using it will result in a deprecation warning.
+
+To migrate to the new \verb/bnf/ environment, simply replace the \verb/bnfgrammar/ environment with \verb/bnf/, and replace \verb/\in/ with \verb/:in:/ and \verb/||/ with \verb|//|.
+Moreover, the new environment does not center the grammar by default, so wrapping it in a \verb/center/ environment is necessary to reproduce the same output.
+While there will be some differences in the spacings and the font styles, it is an easy fix.
+See \secref{sec:tutorial} for more details.
+
+Below is an example of using the deprecated \verb/bnfgrammar/ environment, left here for historical reasons.
+\begin{example}
\begin{bnfgrammar}
a : Variables \in \textit{Vars}
;;
@@ -107,48 +515,71 @@ Annotations can also be provided on left-hand sides, to label the nonterminal in
expr + term
| term
;;
- term ::=
- term * a
- | a
- \end{bnfgrammar}
-\end{exampleside}
-
-You can also provide an optional specification to the grammar environment, to redefine alignment or spacing.
-\begin{tcblisting}{text above listing}
- \begin{bnfgrammar}[lr@{\hspace{4pt}}c@{\hspace{2pt}}ll]
- a : Variables \in \textit{Vars}
- ;;
- expr ::=
- expr + term : sum
- | term : term
- ;;
- term ::=
- term * a : product
- | a : variable
- \end{bnfgrammar}
-\end{tcblisting}
-
-If you want to typeset multiple productions on a single line, you can use double vertical bars by default.
-\begin{exampleside}
- \begin{bnfgrammar}
- a \in \textit{Vars}
- ;;
- expr ::= expr + term || term
- ;;
term ::= term * a || a
\end{bnfgrammar}
-\end{exampleside}
+\end{example}
-The second and third optional arguments specify regular expressions for the line-breaking and non-breaking RHS seperators:
-\begin{tcblisting}{text above listing}
- \begin{bnfgrammar}[llcll][\|\|][\|]
- a \in \textit{Vars}
- ;;
- expr ::= expr + term | term
- ;;
- term ::= term * a
- || a
- \end{bnfgrammar}
-\end{tcblisting}
+
+\section*{Appendix}\label{sec:appendix}
+The following is the code used to typeset the metagrammar of the \textsf{simplebnf} DSL in the first page.
+\begin{listing}[breakable]
+\begin{bnf}(
+ prod-delim = ;;;,
+ new-line-delim = !,
+ single-line-delim = ?,
+ comment = //,
+ relation = {:::=|:in:},
+ relation-sym-map =
+ {
+ {:::=} = $\Coloneqq$,
+ {:in:} = $\in$,
+ },
+)[
+ colspec = lrcll,
+ column{2} = {mode=dmath},
+ column{4} = {mode=text, font=\ttfamily},
+]
+ G // Gramar :::=
+ ! $P$ // production
+ ! $P \fatsemi {}$ // production w/ a trailing delimiter
+ ! $P \fatsemi G$ // production sequence
+;;;
+ P // Production :::= $L \rightarrowtriangle R$
+;;;
+ L // LHS :::=
+ ! $v$ // metavariable
+ ! $v\!\fatslash\,c$ // annotated metavariable
+;;;
+ R // RHS :::=
+ ! $\talloblong$ // delimiter
+ ! $A \talloblong R$ // alternative sequence
+;;;
+ A // Alternative :::=
+ ! $f$ // syntactic form
+ ! $f\!\fatslash\,c$ // annotated syntactic form
+;;;
+ \fatsemi // Prod. delimiter :::=
+ ! ;; // default symbol
+ ! $\cdots$ // user-defined
+;;;
+ \rightarrowtriangle // Rule relation :::=
+ ! ::= // $\Coloneqq$
+ ! -> // $\to$
+ ! \texttt{\char`\\in} // $\in$
+ ! $\cdots$ // user-defined
+;;;
+ \fatslash // Annot. symbol :::=
+ ! : // default symbol
+ ! $\cdots$ // user-defined
+;;;
+ \talloblong // Alt. delimiter :::=
+ ! | // new-line delimiter
+ ! || // single-line delimiter
+ ! $\cdots$ // user-defined
+;;;
+ v, f, c :in: \textsf{\TeX{} tl} // valid \TeX{} token lists
+;;;
+\end{bnf}
+\end{listing}
\end{document}
diff --git a/macros/latex/contrib/simplebnf/simplebnf.sty b/macros/latex/contrib/simplebnf/simplebnf.sty
index 82d39139aa..5ddb566e28 100644
--- a/macros/latex/contrib/simplebnf/simplebnf.sty
+++ b/macros/latex/contrib/simplebnf/simplebnf.sty
@@ -8,191 +8,474 @@
%% Released under the MIT License.
%% ---------------------------------------------------------------------------
%%
+\NeedsTeXFormat{LaTeX2e}[2018-04-01]
+\providecommand\DeclareRelease[3]{}
+\providecommand\DeclareCurrentRelease[2]{}
+\DeclareRelease{0.3.2}{2023-01-07}{simplebnf-2023-01-07.sty}
+\DeclareCurrentRelease{}{2023-11-25}
+
\RequirePackage{expl3,xparse}
-% mathtools is needed for the \Coloneqq simbol
+% The new backend is tabularray
+\RequirePackage{tabularray}
+% mathtools is needed for the \Coloneqq symbol
\RequirePackage{mathtools}
+
\ProvidesExplPackage
{simplebnf}
- {2023/01/07}
- {0.3.2}
+ {2023-11-25}
+ {1.0.0}
{A simple package to format Backus–Naur form}
-\cs_generate_variant:Nn \regex_split:nnNTF {nVNTF}
-\cs_generate_variant:Nn \regex_split:NnN {NVN}
-\NewDocumentCommand\SimpleBNFDefEq{}{\ensuremath{\Coloneqq}}
+\tl_new:N \l__simplebnf_prod_delim_tl
+\tl_new:N \l__simplebnf_new_line_delim_tl
+\tl_new:N \l__simplebnf_single_line_delim_tl
+\tl_new:N \l__simplebnf_comment_tl
+\tl_new:N \l__simplebnf_or_sym_tl
+\tl_new:N \l__simplebnf_relation_tl
+\dim_new:N \l__simplebnf_prod_sep_dim
+\dim_new:N \l__simplebnf_row_sep_dim
+\prop_new:N \l__simplebnf_relation_sym_map_prop
-\NewDocumentCommand\SimpleBNFDefOr{}{\ensuremath{|}}
+\keys_define:nn { simplebnf }
+ {
+ prod-delim .tl_set:N = \l__simplebnf_prod_delim_tl,
+ new-line-delim .tl_set:N = \l__simplebnf_new_line_delim_tl,
+ single-line-delim .tl_set:N = \l__simplebnf_single_line_delim_tl,
+ comment .tl_set:N = \l__simplebnf_comment_tl,
+ relation .tl_set:N = \l__simplebnf_relation_tl,
+ relation-sym-map .code:n =
+ { \prop_set_from_keyval:Nn \l__simplebnf_relation_sym_map_prop { #1 } },
+ or-sym .tl_set:N = \l__simplebnf_or_sym_tl,
+ prod-sep .dim_set:N = \l__simplebnf_prod_sep_dim,
+ row-sep .dim_set:N = \l__simplebnf_row_sep_dim,
+ }
-\NewDocumentCommand\SimpleBNFStretch{}{0em}
+% default key-values
+\keys_set:nn { simplebnf }
+ {
+ prod-delim = ;;,
+ new-line-delim = \|,
+ single-line-delim = //,
+ comment = :,
+ relation = {::=|->|:in:},
+ relation-sym-map =
+ {
+ {::=} = {\ensuremath{\Coloneqq}},
+ {->} = {\ensuremath{\to}},
+ {:in:} = {\ensuremath{\in}},
+ },
+ or-sym = $|$,
+ prod-sep = 2pt,
+ row-sep = 0pt,
+ }
-\seq_new:N \l__input_seq
-\seq_new:N \l__term_seq
-\tl_new:N \l__term_tl
-\tl_new:N \l__keypairs_tl
-\tl_new:N \l__table_tl
-\seq_new:N \l__keypairs_seq
-\bool_new:N \l__first_rhs
-\regex_new:N \g_simplebnf_rhs_newline_r
-\regex_new:N \g_simplebnf_rhs_nb_r
+% bnf environment is internally a tabularray environment
+\NewTblrEnviron{@simplebnf_tblr_env}
+\SetTblrInner[@simplebnf_tblr_env]{rowsep=\l__simplebnf_row_sep_dim}
+
+
+% The main token list that stores the table.
+\tl_new:N \l__simplebnf_table_tl
+
+\seq_new:N \l__simplebnf_lhs_rhs_pair_seq
+\tl_new:N \l__simplebnf_lhs_tl
+\tl_new:N \l__simplebnf_matched_relation_tl
+\tl_new:N \l__simplebnf_rhs_tl
+
+\seq_new:N \l__simplebnf_alternative_seq
+\bool_new:N \l__simplebnf_first_alternative_bool
+
+\regex_new:N \l__simplebnf_new_line_delim_regex
+\regex_new:N \l__simplebnf_single_line_delim_regex
+\regex_new:N \l__simplebnf_comment_regex
+\regex_new:N \l__simplebnf_prod_delim_regex
+\regex_new:N \l__simplebnf_relation_regex
+
+\cs_generate_variant:Nn \regex_split:NnNTF {NVNTF}
+\cs_generate_variant:Nn \regex_split:NnN {NVN}
+\cs_generate_variant:Nn \regex_set:Nn {NV}
+\cs_generate_variant:Nn \regex_replace_all:NnN {NVN}
+
+
+\msg_new:nnn { simplebnf } { lhs } { Malformed~LHS~\msg_line_context: }
+\msg_new:nnn { simplebnf } { alt } { Malformed~alternative~\msg_line_context: }
+\msg_new:nnn { simplebnf } { prod } { Malformed~production~\msg_line_context: }
+\msg_new:nnn { simplebnf } { unk-rel } { Unknown~relation~\msg_line_context: }
+\msg_new:nnn { simplebnf } { no-rel } { No~relation~found~\msg_line_context: }
+\msg_new:nnn { simplebnf } { dep }
+ {
+ Environment~bnfgrammar~is~deprecated.~
+ Consider~using~the~environment~bnf.~
+ Consult~the~documentation~for~more~information.
+ }
-%% Typeset a single rhs of a production.
-%% \l__first_rhs = true => `::=' already typeset
-%% \l__first_rhs = false => move to a newline and typeset `|'
-%% #1 - rhs : annot or rhs
-\cs_new:Nn \simplebnf_typeset_rhs:n
-{
- \bool_if:NTF \l__first_rhs
- {
- \bool_set_false:N \l__first_rhs
- }
- {
- \tl_put_right:Nn \l__table_tl { \\ && \SimpleBNFDefOr & }
- }
-
- \tl_set:Nn \l_tmpa_tl { #1 }
- \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
- \regex_replace_once:nnN { \s+$ } {} \l_tmpa_tl
- \regex_split:nVNTF { : } \l_tmpa_tl \l_tmpa_seq
- {
- \seq_pop_left:NNT \l_tmpa_seq \l_tmpa_tl
- {
- \regex_replace_all:NnN \g_simplebnf_rhs_nb_r { \c{SimpleBNFDefOr} } \l_tmpa_tl
- % Expand only the local temporary variable.
- \tl_put_right:No \l__table_tl
- {
- \exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl} &
- }
- }
-
- \seq_pop_left:NNT \l_tmpa_seq \l_tmpb_tl
- {
- \regex_replace_once:nnN { ^\s+ } {} \l_tmpb_tl
- \tl_put_right:No \l__table_tl
- {
- \exp_after:wN\bnfannot\exp_after:wN{\l_tmpb_tl}
- }
- }
- }
- {
- \regex_replace_all:NnN \g_simplebnf_rhs_nb_r { \c{SimpleBNFDefOr} } \l_tmpa_tl
-
- \tl_put_right:No \l__table_tl
- {
- \exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl}
- }
- }
-}
%% Typeset a single lhs of a production.
%% #1 - lhs : either term or (term : annotation)
-\cs_new:Nn \simplebnf_typeset_lhs:n
-{
- \tl_set:Nx \l_tmpa_tl { #1 }
- \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
- \regex_replace_once:nnN { \s+$ } {} \l_tmpa_tl
-
- \regex_split:nVNTF { : } \l_tmpa_tl \l_tmpa_seq
- {
- \seq_pop_right:NN \l_tmpa_seq \l_tmpa_tl
- \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
- \tl_put_right:No \l__table_tl
- {
- \exp_after:wN\bnfannot\exp_after:wN{\l_tmpa_tl} &
- }
- \seq_pop_left:NN \l_tmpa_seq \l_tmpa_tl
- \tl_put_right:No \l__table_tl
- {
- \exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl}
- }
- }
- {
- \tl_put_right:No \l__table_tl
- {
- \exp_after:wN&\exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl}
- }
- }
-}
+\cs_new:Nn \__simplebnf_typeset_lhs:n
+ {
+ \tl_set:No \l_tmpa_tl { #1 }
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \regex_replace_once:nnN { \s+$ } {} \l_tmpa_tl
+
+ \regex_split:NVNTF \l__simplebnf_comment_regex \l_tmpa_tl \l_tmpa_seq
+ { % annotation
+ \seq_pop_right:NN \l_tmpa_seq \l_tmpa_tl
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+ \tl_put_right:Nn \l__simplebnf_table_tl { & }
+ % metavariable
+ \seq_pop_left:NN \l_tmpa_seq \l_tmpa_tl
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+ }
+ { % not annotated
+ \tl_put_right:Nn \l__simplebnf_table_tl { & }
+ \seq_pop_left:NN \l_tmpa_seq \l_tmpa_tl
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+ }
+
+ % There should be no more tokens in the sequence.
+ \seq_if_empty:NF \l_tmpa_seq
+ { \msg_error:nnn { simplebnf } { lhs } { #1 } }
+ }
+
+
+% Handle a single alternative (or single-line alternatives) of an RHS.
+\cs_new:Nn \__simplebnf_typeset_rhs:n
+ {
+ \tl_if_blank:nF { #1 }
+ {
+ \bool_if:NTF \l__simplebnf_first_alternative_bool
+ { \bool_set_false:N \l__simplebnf_first_alternative_bool }
+ {
+ \tl_put_right:Nn \l__simplebnf_table_tl
+ { \\ & & \l__simplebnf_or_sym_tl & }
+ }
+
+ \tl_set:Nn \l_tmpa_tl { #1 }
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \regex_replace_once:nnN { \s+$ } {} \l_tmpa_tl
+
+ \regex_replace_all:NVN
+ \l__simplebnf_single_line_delim_regex
+ \l__simplebnf_or_sym_tl
+ \l_tmpa_tl
+
+ \regex_split:NVNTF \l__simplebnf_comment_regex \l_tmpa_tl \l_tmpa_seq
+ { % put an alternate syntax form
+ \seq_pop_left:NN \l_tmpa_seq \l_tmpa_tl
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+ \tl_put_right:Nn \l__simplebnf_table_tl { & }
+
+ % put an annotation
+ \seq_pop_left:NN \l_tmpa_seq \l_tmpa_tl
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+
+ % There should be no more tokens in the sequence.
+ \seq_if_empty:NF \l_tmpa_seq
+ { \msg_error:nnn { simplebnf } { alt } { #1 } }
+ }
+ { % no annotation
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+ }
+ }
+ }
+
+
+% Handle a single production P
+\cs_new:Nn \__simplebnf_typeset_production:n
+ {
+ \regex_split:NnNTF
+ \l__simplebnf_relation_regex
+ { #1 }
+ \l__simplebnf_lhs_rhs_pair_seq
+ {
+ % LHS
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_lhs_tl
+ % Rule relation
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_matched_relation_tl
+ % RHS
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_rhs_tl
+
+ \seq_if_empty:NF \l__simplebnf_lhs_rhs_pair_seq
+ { \msg_error:nnn { simplebnf } { prod } { #1 } }
+ % Build the LHS
+ \__simplebnf_typeset_lhs:n \l__simplebnf_lhs_tl
+
+ % Build the rule relation
+ \prop_get:NVNTF
+ \l__simplebnf_relation_sym_map_prop
+ \l__simplebnf_matched_relation_tl
+ \l_tmpa_tl
+ {
+ \tl_put_right:Nn \l__simplebnf_table_tl { & }
+ \tl_put_right:NV \l__simplebnf_table_tl \l_tmpa_tl
+ \tl_put_right:Nn \l__simplebnf_table_tl { & }
+ }
+ {
+ \msg_warning:nnn { simplebnf } { unk-rel } { #1 }
+ \tl_put_right:Nn \l__simplebnf_table_tl { & & }
+ }
+ % Build the RHS
+ %% \l__simplebnf_alternative_seq - (rhs:annot | rhs)...
+ \regex_split:NVN
+ \l__simplebnf_new_line_delim_regex
+ \l__simplebnf_rhs_tl
+ \l__simplebnf_alternative_seq
+
+ \bool_set_true:N \l__simplebnf_first_alternative_bool
+ \seq_map_function:NN \l__simplebnf_alternative_seq \__simplebnf_typeset_rhs:n
+ }
+ { \msg_error:nnn { simplebnf } { no-rel } { #1 } }
+ }
+
+
+% The main control sequence that builds the grammar G into \l__simplebnf_table_tl
+\cs_new:Nn \__simplebnf_build_grammar:n
+ {
+ \regex_set:NV \l__simplebnf_new_line_delim_regex \l__simplebnf_new_line_delim_tl
+ \regex_set:NV \l__simplebnf_single_line_delim_regex \l__simplebnf_single_line_delim_tl
+ \regex_set:NV \l__simplebnf_comment_regex \l__simplebnf_comment_tl
+ \regex_set:NV \l__simplebnf_prod_delim_regex \l__simplebnf_prod_delim_tl
+ \tl_put_left:Nn \l__simplebnf_relation_tl { ( }
+ \tl_put_right:Nn \l__simplebnf_relation_tl { ) }
+ \regex_set:NV \l__simplebnf_relation_regex \l__simplebnf_relation_tl
+
+ % \l_tmpa_seq contains the sequence of productions (P; ...).
+ \regex_split:NnN \l__simplebnf_prod_delim_regex { #1 } \l_tmpa_seq
+
+ % Is this the first production of this grammar?
+ \bool_set_true:N \l_tmpa_bool
+ \seq_map_inline:Nn \l_tmpa_seq
+ {
+ \tl_if_blank:nF { ##1 } % ignore empty productions
+ { % If not first, change row
+ \bool_if:NTF \l_tmpa_bool
+ { \bool_set_false:N \l_tmpa_bool }
+ {
+ \tl_put_right:Nn \l__simplebnf_table_tl { \\[\l__simplebnf_prod_sep_dim] }
+ }
+ \__simplebnf_typeset_production:n { ##1 }
+ }
+ }
+ }
-\NewDocumentCommand \bnfexpr { m } { \texttt { #1 } }
-\NewDocumentCommand \bnfannot { m } { \textit{ #1 } }
%% Typeset a BNF grammar.
-%% #1 - tabular specification (llcll)
-%% #2 - regexp for newline separator for rhses
-%% #2 - regexp for non-breaking separator for rhses
-%% #3 - grammar
-\NewDocumentEnvironment { bnfgrammar } { O{llcll} O{[^\|]\|[^\|]} O{\|\|} +b }
+%% #1 - key-value options for simplebnf
+%% #2 - tabularray specification (default: llcll)
+%% #3 - grammar body
+\NewDocumentEnvironment { bnf } { d() O{llcll} +b }
{
- \regex_gset:Nn \g_simplebnf_rhs_newline_r { #2 }
- \regex_gset:Nn \g_simplebnf_rhs_nb_r { #3 }
+ \IfNoValueF { #1 }
+ { \keys_set:nn { simplebnf } { #1 } }
- %% \l__input_seq is a list of term definitions.
- \regex_split:nnN { ;; } { #4 } \l__input_seq
- \begin{center}
- \tl_set:Nn \l__table_tl
- {
- \begin{tabular}{#1}
- }
+ \__simplebnf_build_grammar:n { #3 }
- \bool_set_true:N \l_tmp_first_term % Is this the first term in this grammar?
- \seq_map_inline:Nn \l__input_seq
+ \begin{@simplebnf_tblr_env}[expand=\l__simplebnf_table_tl]{#2}
+ \tl_use:N \l__simplebnf_table_tl
+ \end{@simplebnf_tblr_env}
+ }
+ { }
+
+
+\NewDocumentCommand \SetBNFLayout { m }
+ {
+ \SetTblrInner[@simplebnf_tblr_env]{#1}
+ }
+
+
+\NewDocumentCommand \SetBNFConfig { m }
+ {
+ \keys_set:nn { simplebnf } { #1 }
+ }
+
+
+% DEPRECATED APIs
+
+% Redefinable user-level variables
+\newcommand*\SimpleBNFDefEq{\ensuremath{\Coloneqq}}
+\newcommand*\SimpleBNFDefOr{\ensuremath{|}}
+\newcommand*\SimpleBNFStretch{0em}
+
+\cs_generate_variant:Nn \regex_split:nnNTF {nVNTF}
+
+%% Typeset a single lhs of a production.
+%% #1 - lhs : either term or (term : annotation)
+\cs_new:Nn \@dep__simplebnf_typeset_lhs:n
+ {
+ \tl_set:Nx \l_tmpa_tl { #1 }
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \regex_replace_once:nnN { \s+$ } {} \l_tmpa_tl
+
+ \regex_split:nVNTF { : } \l_tmpa_tl \l_tmpa_seq
{
- %% If not-first, add newline
- \bool_if:NTF \l_tmp_first_term
+ \seq_pop_right:NN \l_tmpa_seq \l_tmpa_tl
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \tl_put_right:No \l__simplebnf_table_tl
{
- \bool_set_false:N \l_tmp_first_term
+ \exp_after:wN\bnfannot\exp_after:wN{\l_tmpa_tl} &
}
+ \seq_pop_left:NN \l_tmpa_seq \l_tmpa_tl
+ \tl_put_right:No \l__simplebnf_table_tl
{
- \tl_put_right:Nn \l__table_tl { \\[\SimpleBNFStretch] }
+ \exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl}
}
+ }
+ {
+ \tl_put_right:No \l__simplebnf_table_tl
+ {
+ \exp_after:wN&\exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl}
+ }
+ }
+ }
+
+
+%% Typeset a single rhs of a production.
+%% \l__simplebnf_first_alternative_bool = true => `::=' already typeset
+%% \l__simplebnf_first_alternative_bool = false => move to a newline and typeset `|'
+%% #1 - rhs : annot or rhs
+\cs_new:Nn \@dep__simplebnf_typeset_rhs:n
+ {
+ \bool_if:NTF \l__simplebnf_first_alternative_bool
+ {
+ \bool_set_false:N \l__simplebnf_first_alternative_bool
+ }
+ {
+ \tl_put_right:Nn \l__simplebnf_table_tl { \\ && \SimpleBNFDefOr & }
+ }
- \regex_split:nnNTF { ::= } { ##1 } \l__term_seq
- % Parse a ::= definition
+ \tl_set:Nn \l_tmpa_tl { #1 }
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpa_tl
+ \regex_replace_once:nnN { \s+$ } {} \l_tmpa_tl
+ \regex_split:nVNTF { : } \l_tmpa_tl \l_tmpa_seq
+ {
+ \seq_pop_left:NNT \l_tmpa_seq \l_tmpa_tl
{
- %% \l__term_seq - (lhs, rhses)...
- %% \l__term_tl - lhs
- %% \l__keypairs_tl - rhses
- \seq_pop_left:NN \l__term_seq \l__term_tl
- \seq_pop_left:NN \l__term_seq \l__keypairs_tl
-
- \simplebnf_typeset_lhs:n{\l__term_tl}
- \tl_put_right:Nn \l__table_tl
+ \regex_replace_all:NnN \l__simplebnf_single_line_delim_regex { \c{SimpleBNFDefOr} } \l_tmpa_tl
+ % Expand only the local temporary variable.
+ \tl_put_right:No \l__simplebnf_table_tl
{
- & \SimpleBNFDefEq &
+ \exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl} &
}
- %% \l__keypairs_seq - (rhs:annot | rhs)...
- \regex_split:NVN \g_simplebnf_rhs_newline_r \l__keypairs_tl \l__keypairs_seq
+ }
- \bool_set_true:N \l__first_rhs
- \seq_map_function:NN \l__keypairs_seq \simplebnf_typeset_rhs:n
+ \seq_pop_left:NNT \l_tmpa_seq \l_tmpb_tl
+ {
+ \regex_replace_once:nnN { ^\s+ } {} \l_tmpb_tl
+ \tl_put_right:No \l__simplebnf_table_tl
+ {
+ \exp_after:wN\bnfannot\exp_after:wN{\l_tmpb_tl}
+ }
}
+ }
+ {
+ \regex_replace_all:NnN \l__simplebnf_single_line_delim_regex { \c{SimpleBNFDefOr} } \l_tmpa_tl
+
+ \tl_put_right:No \l__simplebnf_table_tl
{
- % Else, parse a \in declaration
- \regex_split:nnNTF { \c{in} } { ##1 } \l__term_seq
+ \exp_after:wN\bnfexpr\exp_after:wN{\l_tmpa_tl}
+ }
+ }
+ }
+
+\cs_new:Nn \@dep__simplebnf_typeset_production:n
+ {
+ \regex_split:nnNTF { ::= } { #1 } \l__simplebnf_lhs_rhs_pair_seq
+ % Parse a ::= definition
+ {
+ %% \l__simplebnf_lhs_rhs_pair_seq - (lhs, rhses)...
+ %% \l__simplebnf_lhs_tl - lhs
+ %% \l__simplebnf_rhs_tl - rhses
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_lhs_tl
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_rhs_tl
+
+ \@dep__simplebnf_typeset_lhs:n{\l__simplebnf_lhs_tl}
+ \tl_put_right:Nn \l__simplebnf_table_tl
+ {
+ & \SimpleBNFDefEq &
+ }
+ %% \l__simplebnf_alternative_seq - (rhs:annot | rhs)...
+ \regex_split:NVN \l__simplebnf_new_line_delim_regex \l__simplebnf_rhs_tl \l__simplebnf_alternative_seq
+
+ \bool_set_true:N \l__simplebnf_first_alternative_bool
+ \seq_map_function:NN \l__simplebnf_alternative_seq \@dep__simplebnf_typeset_rhs:n
+ }
+ {
+ % Else, parse a \in declaration (TODO: There is a huge room for generalization)
+ \regex_split:nnNTF { \c{in} } { #1 } \l__simplebnf_lhs_rhs_pair_seq
+ {
+ %% \l__simplebnf_lhs_rhs_pair_seq - (lhs, rhses)...
+ %% \l__simplebnf_lhs_tl - lhs
+ %% \l__simplebnf_rhs_tl - rhses
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_lhs_tl
+ \seq_pop_left:NN \l__simplebnf_lhs_rhs_pair_seq \l__simplebnf_rhs_tl
+
+ \@dep__simplebnf_typeset_lhs:n{\l__simplebnf_lhs_tl}
+ \tl_put_right:Nn \l__simplebnf_table_tl
{
- %% \l__term_seq - (lhs, rhs)
- \seq_pop_left:NN \l__term_seq \l_tmpa_tl
-
- \simplebnf_typeset_lhs:n{\l_tmpa_tl}
- \tl_put_right:Nn \l__table_tl
- {
- & $\in$ & $
- }
- \seq_pop_left:NN \l__term_seq \l_tmpa_tl
- \tl_put_right:NV \l__table_tl \l_tmpa_tl
- \tl_put_right:Nn \l__table_tl
- {
- $ &
- }
+ & $\in$ &
}
- { \msg_error:nn {simplebnf} { Could not parser ##1 } }
+ %% \l__simplebnf_alternative_seq - (rhs:annot | rhs)...
+ \regex_split:NVN \l__simplebnf_new_line_delim_regex \l__simplebnf_rhs_tl \l__simplebnf_alternative_seq
+
+ \bool_set_true:N \l__simplebnf_first_alternative_bool
+ \seq_map_function:NN \l__simplebnf_alternative_seq \@dep__simplebnf_typeset_rhs:n
+ }
+ {
+ \msg_error:nnn { simplebnf } { error-message }
+ { Could~not~parse~#1 }
}
}
+ }
- \tl_put_right:Nn \l__table_tl { \end{tabular} }
- \tl_use:N \l__table_tl
+
+\seq_new:N \l__input_seq
+\cs_new:Nn \@dep__simplebnf_typeset_grammar:nnn
+ {
+ \regex_set:Nn \l__simplebnf_new_line_delim_regex { #1 }
+ \regex_set:Nn \l__simplebnf_single_line_delim_regex { #2 }
+
+ %% \l__input_seq is a list of term definitions.
+ \regex_split:nnN { ;; } { #3 } \l__input_seq
+
+ \bool_set_true:N \l_tmp_first_term % Is this the first term in this grammar?
+ \seq_map_inline:Nn \l__input_seq
+ {
+ %% If not-first, add newline
+ \bool_if:NTF \l_tmp_first_term
+ {
+ \bool_set_false:N \l_tmp_first_term
+ }
+ {
+ \tl_put_right:Nn \l__simplebnf_table_tl { \\[\SimpleBNFStretch] }
+ }
+
+ \@dep__simplebnf_typeset_production:n { ##1 }
+ }
+ }
+
+
+\NewDocumentCommand \bnfexpr { m } { \texttt { #1 } }
+\NewDocumentCommand \bnfannot { m } { \textit { #1 } }
+
+
+%% Typeset a BNF grammar.
+%% #1 - tabular specification (llcll)
+%% #2 - regexp for newline separator for rhses
+%% #2 - regexp for non-breaking separator for rhses
+%% #3 - grammar
+\NewDocumentEnvironment { bnfgrammar } { O{llcll} O{[^\|]\|[^\|]} O{\|\|} +b }
+ {
+ \msg_warning:nn { simplebnf } { dep }
+ \begin{center}
+ \begin{tabular}{#1}
+ \@dep__simplebnf_typeset_grammar:nnn { #2 } { #3 } { #4 }
+ \tl_use:N \l__simplebnf_table_tl
+ \end{tabular}
\end{center}
}
{ }