summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/nicematrix
diff options
context:
space:
mode:
authorNorbert Preining <norbert@preining.info>2020-02-29 03:02:30 +0000
committerNorbert Preining <norbert@preining.info>2020-02-29 03:02:30 +0000
commit4736ac744a8c914e0d5ced1d157e017a385bbf43 (patch)
treecc6b754ea0d704573dfffffdbf327e14057ce864 /macros/latex/contrib/nicematrix
parent02193cc5ca03abf234d3bc9d77b4821d85f58c62 (diff)
CTAN sync 202002290302
Diffstat (limited to 'macros/latex/contrib/nicematrix')
-rw-r--r--macros/latex/contrib/nicematrix/nicematrix-french.pdfbin312439 -> 332466 bytes
-rw-r--r--macros/latex/contrib/nicematrix/nicematrix-french.tex352
-rw-r--r--macros/latex/contrib/nicematrix/nicematrix.dtx2516
-rw-r--r--macros/latex/contrib/nicematrix/nicematrix.pdfbin693305 -> 723168 bytes
4 files changed, 1856 insertions, 1012 deletions
diff --git a/macros/latex/contrib/nicematrix/nicematrix-french.pdf b/macros/latex/contrib/nicematrix/nicematrix-french.pdf
index 31d18ea42e..3b7adbbf92 100644
--- a/macros/latex/contrib/nicematrix/nicematrix-french.pdf
+++ b/macros/latex/contrib/nicematrix/nicematrix-french.pdf
Binary files differ
diff --git a/macros/latex/contrib/nicematrix/nicematrix-french.tex b/macros/latex/contrib/nicematrix/nicematrix-french.tex
index 05af9a1b73..5a8d655762 100644
--- a/macros/latex/contrib/nicematrix/nicematrix-french.tex
+++ b/macros/latex/contrib/nicematrix/nicematrix-french.tex
@@ -18,17 +18,6 @@
\usepackage{unicode-math}
-\newbox\ExampleOne
-
-\setbox\ExampleOne\hbox{%
-$\begin{NiceArray}{|*{4}{C|}}[hlines,first-row,first-col]
- & e & a & b & c \\
-e & e & a & b & c \\
-a & a & e & c & b \\
-b & b & c & e & a \\
-c & c & b & a & e
-\end{NiceArray}$}
-
\usepackage{arydshln}
\usepackage{verbatim}
@@ -53,6 +42,8 @@ c & c & b & a & e
\usepackage{footnote}
+\usepackage{booktabs}
+
\usepackage[hyperfootnotes = false]{hyperref}
\hypersetup
@@ -99,8 +90,9 @@ tableau.
Cette extension peut être utilisée avec |xelatex|, |lualatex| et |pdflatex| mais aussi avec le cheminement
classique |latex|-|dvips|-|ps2pdf| (ou Adobe Distiller). Deux ou trois compilations successives peuvent être
nécessaires. Cette extension nécessite et charge les extensions \pkg{expl3}, \pkg{l3keys2e}, \pkg{xparse},
-\pkg{array}, \pkg{amsmath} et \pkg{tikz}. Elle charge aussi la bibliothèque Tikz \pkg{fit}. L'utilisateur final
-n'a qu'à charger l'extension \pkg{nicematrix} avec l'instruction habituelle : |\usepackage{nicematrix}|.
+\pkg{array}, \pkg{amsmath} et \pkg{tikz}. Elle charge aussi la bibliothèque Tikz \pkg{fit}. L'utilisateur final n'a
+qu'à charger l'extension \pkg{nicematrix} avec l'instruction habituelle :
+|\usepackage{nicematrix}|.
@@ -224,12 +216,10 @@ doivent pas être employées. Voir p.~\pageref{NiceArray} la partie concernant |
\section{Les lignes en pointillés continues}
-
-
À l'intérieur des environnements de l'extension \pkg{nicematrix}, de nouvelles commandes sont définies : |\Ldots|,
-|\Cdots|, |\Vdots|, |\Ddots|, and |\Iddots|. Ces commandes sont conçues pour être utilisées à la place de |\dots|,
+|\Cdots|, |\Vdots|, |\Ddots|, et |\Iddots|. Ces commandes sont conçues pour être utilisées à la place de |\dots|,
|\cdots|, |\vdots|, |\ddots| et |\iddots|.\footnote{La commande |\iddots|, définie dans \pkg{nicematrix}, est une
- variante de |\ddots| avec les points allant vers le haut : \smash{$\iddots$}. Si \pkg{mathdots} est chargée, la
+ variante de |\ddots| avec les points allant vers le haut. Si \pkg{mathdots} est chargée, la
version de \pkg{mathdots} est utilisée. Elle correspond à la commande |\adots| de \pkg{unicode-math}.}
\newcounter{fniddots}
\setcounter{fniddots}{\thefootnote}
@@ -574,7 +564,7 @@ L'extension \pkg{nicematrix} crée un nœud Tikz pour chaque case du tableau con
tracer les lignes en pointillés entre les cases du tableau. Toutefois, l'utilisateur peut aussi utiliser
directement ces nœuds. On commence par donner un nom au tableau (avec l'option |name|). Cela étant fait, les nœuds
sont accessibles à travers les noms «\textsl{nom}-$i$-$j$» où \textsl{nom} est le nom donné au tableau et $i$ et
-$j$ les numéros de ligne et de colonne de la case considérée.
+$j$ les numéros de rangée et de colonne de la case considérée.
\medskip
\begin{BVerbatim}[baseline=c,boxwidth=11cm]
@@ -646,8 +636,8 @@ a & a & a
\medskip
Les noms des «nœuds larges» s'obtiennent en ajoutant le suffixe «|-large|» au nom des nœuds normaux. Dans l'exemple
suivant, on a surligné tous les «nœuds larges». Nous considérons que cet exemple se suffit à lui-même comme
-définition de ces nœuds.\footnote{Il n'y a pas de «nœuds larges» créés dans les lignes et colonnes extérieures
- (pour ces lignes et colonnes, voir p.~\pageref{exterior}).}
+définition de ces nœuds.\footnote{Il n'y a pas de «nœuds larges» créés dans les rangées et colonnes extérieures
+ (pour ces rangées et colonnes, voir p.~\pageref{exterior}).}
\[\begin{pNiceMatrix}[
create-large-nodes,
@@ -785,19 +775,118 @@ L'environnement |{NiceArray}| est similaire à l'environnement |{array}|. Comme
obligatoire est le préambule du tableau. Néanmoins, pour des raisons techniques, l'utilisateur doit utiliser les
lettres |L|, |C| et |R|\footnote{Les types de colonnes |L|, |C| et |R| sont définis localement à l'intérieur de
|{NiceArray}| avec la commande |\newcolumntype| de \pkg{array}. Cette définition masque une éventuelle définition
- précédente. En fait, les types de colonnes |w| and |W| sont également redéfinis.} au lieu de |l|, |c| et |r|.
+ précédente. En fait, les types de colonnes |w| et |W| sont également redéfinis.} au lieu de |l|, |c| et |r|.
Il est possible d'utiliser les constructions |w{...}{...}|, |W{...}{...}|, \verb+|+, |>{...}|, |<{...}|, |@{...}|,
|!{...}| et |*{n}{...}| mais les lettres |p|, |m| et |b| ne doivent pas être employées.\footnote{Dans une commande
|\multicolumn|, on doit également utiliser les lettres |L|, |C| et |R|.}
\medskip
-L'environnement |{NiceArray}| accepte les options classiques |t|, |c| et |b| de |{array}| mais aussi d'autres
-options définies par \pkg{nicematrix} (|renew-dots|, |columns-width|, etc.).
+En plus des options proposées pour |{pNiceMatrix}| et ses variantes, l'environnement |{NiceArray}| propose une
+option |baseline| qui prend en argument un entier qui indique le numéro de rangée dont la ligne de base servira de
+ligne de base pour l'environnement |{NiceArray}|.
+
+\medskip
+\begin{BVerbatim}[baseline=c,boxwidth=11cm]
+$A =
+\begin{NiceArray}{CCCC}[hvlines,~emphase#baseline=2@]
+1 & 2 & 3 & 4 \\
+1 & 2 & 3 & 4 \\
+1 & 2 & 3 & 4 \\
+\end{NiceArray}$
+\end{BVerbatim}
+%
+$A =
+\begin{NiceArray}{CCCC}[hvlines,baseline=2]
+1 & 2 & 3 & 4 \\
+1 & 2 & 3 & 4 \\
+1 & 2 & 3 & 4 \\
+\end{NiceArray}$
+
+\smallskip
+L'option |hvlines| est présentée plus loin (cf. p. \pageref{hvlines}).
+
+\medskip
+L'option |baseline| peut aussi prendre les trois valeurs spéciales |t|, |c| et |b|. Ces trois lettres peuvent aussi
+être utilisées de manière absolue comme pour l'option de l'environnement |{array}| de \pkg{array}. La valeur
+initiale de |baseline| est~|c|.
+
+
+\medskip
+Dans l'exemple suivant, on utilise l'option |t| (synonyme de |baseline=t|) immédiatement après un |\item| de liste.
+On remarquera que la présence d'un |\hline| initial n'empêche pas l'alignement sur la ligne de base de la première
+rangée (avec |{array}| de {array}, il faut utiliser |\firsthline|\footnote{On peut aussi utiliser |\firsthline|
+ avec |{NiceArray}|.}).
+
+
+\begin{BVerbatim}[baseline=c,boxwidth=9cm]
+\begin{enumerate}
+\item un item
+\smallskip
+\item \renewcommand{\arraystretch}{1.2}
+$\begin{NiceArray}[t]{LCCCCCC}
+\hline
+n & 0 & 1 & 2 & 3 & 4 & 5 \\
+u_n & 1 & 2 & 4 & 8 & 16 & 32
+\hline
+\end{NiceArray}$
+\end{enumerate}
+\end{BVerbatim}
+%
+\begin{minipage}{5cm}
+\begin{enumerate}
+\item un item
+\smallskip
+\item \renewcommand{\arraystretch}{1.2}
+$\begin{NiceArray}[t]{LCCCCCC}
+\hline
+n & 0 & 1 & 2 & 3 & 4 & 5 \\
+u_n & 1 & 2 & 4 & 8 & 16 & 32 \\
+\hline
+\end{NiceArray}$
+\end{enumerate}
+\end{minipage}
+
+
+\medskip
+Il est également possible d'utiliser les outils de \pkg{booktabs}: |\toprule|,
+|\bottomrule| et |\midrule|.\par\nobreak
+
+\smallskip
+\begin{BVerbatim}[baseline=c,boxwidth=9cm]
+\begin{enumerate}
+\item an item
+\smallskip
+\item
+$\begin{NiceArray}[t]{LCCCCCC}
+~emphase#\toprule@
+n & 0 & 1 & 2 & 3 & 4 & 5 \\
+~emphase#\midrule@
+u_n & 1 & 2 & 4 & 8 & 16 & 32
+~emphase#\bottomrule@
+\end{NiceArray}$
+\end{enumerate}
+\end{BVerbatim}
+\begin{minipage}{5cm}
+\begin{enumerate}
+\item an item
+\smallskip
+\item
+$\begin{NiceArray}[t]{LCCCCCC}
+\toprule
+n & 0 & 1 & 2 & 3 & 4 & 5 \\
+\midrule
+u_n & 1 & 2 & 4 & 8 & 16 & 32 \\
+\bottomrule
+\end{NiceArray}$
+\end{enumerate}
+\end{minipage}
+
+
\vspace{1cm}
-Un exemple avec un système linéaire (on a besoin de |{NiceArray}| pour le trait vertical)
+L'utilisation de |{NiceArray}| permet de tracer des filets verticaux :\par\nobreak
\bigskip
\begin{BVerbatim}[baseline=c,boxwidth=10cm]
@@ -817,7 +906,8 @@ a_1 & ? & \Cdots & ? & ? \\
\vspace{1cm}
Il existe également des variantes pour l'environnement |{NiceArray}| : |{pNiceArray}|, |{bNiceArray}|,
-|{BNiceArray}|, |{vNiceArray}| et |{VNiceArray}|.
+|{BNiceArray}|, |{vNiceArray}| et |{VNiceArray}|. La clé |baseline| n'est \emph{pas} disponible pour ces
+environnements.
Dans l'exemple suivant, on utilise un environnement |{pNiceArray}| (on n'utilise pas |{pNiceMatrix}| car on
@@ -846,19 +936,19 @@ a_{n-1,1} & \Cdots & a_{n-1,n}
En fait, l'environnement |{pNiceArray}| et ses variantes sont fondés sur un environnement plus général, appelé
|{NiceArrayWithDelims}|. Les deux premiers arguments obligatoires de cet environnement sont les délimiteurs gauche
et droit qui seront utilisés dans la construction de la matrice. Il est possible d'utiliser |{NiceArrayWithDelims}|
-si on a besoin de délimiteurs atypiques.
+si on a besoin de délimiteurs atypiques ou asymétriques.
\medskip
\begin{BVerbatim}[baseline=c,boxwidth=11cm]
$\begin{~emphase#NiceArrayWithDelims@}
- {\downarrow}{\downarrow}{CCC}
+ {\downarrow}{\uparrow}{CCC}[margin]
1 & 2 & 3 \\
4 & 5 & 6 \\
7 & 8 & 9 \\
\end{~emphase#NiceArrayWithDelims@}$
\end{BVerbatim}
$\begin{NiceArrayWithDelims}
- {\downarrow}{\downarrow}{CCC}
+ {\downarrow}{\uparrow}{CCC}[margin]
1 & 2 & 3 \\
4 & 5 & 6 \\
7 & 8 & 9 \\
@@ -874,7 +964,7 @@ Les environnements de \pkg{nicematrix} permettent de composer des rangées et de
Si elle est présente, la première rangée est numérotée par $0$ (et non $1$). Il en est de même pour la première
colonne. Dans le cas général, on doit spécifier le numéro de la dernière rangée et de la dernière colonne comme
-valeurs des options |last-row| and |last-col| (si elles sont présentes).
+valeurs des options |last-row| et |last-col| (si elles sont présentes).
\begin{Verbatim}
$\begin{pNiceMatrix}[~emphase#first-row,last-row=5,first-col,last-col=5@]
@@ -902,8 +992,8 @@ Il y a plusieurs remarques à formuler.
\begin{itemize}[beginpenalty=10000]
\item Si on utilise un environnement avec préambule explicite (c'est-à-dire |{NiceArray}| ou l'une de ses
variantes), on ne doit pas mettre dans ce préambule de spécification de colonne pour les éventuelles première et
-dernière colonne : la première colonne sera automatiquement (et nécessairement) de type |R| et la dernière de type
-|L|.
+dernière colonne : ce sera automatiquement (et nécessairement) une colonne |R| pour la première colonne et une
+colonne |L| pour la dernière.
\item Si on utilise un environnement avec préambule explicite et une dernière colonne, on doit utiliser l'option
|last-col| \emph{sans} valeur. En effet le numéro de la dernière colonne est, dans ce cas, déduit du préambule (qui
@@ -1079,6 +1169,12 @@ $\begin{pNiceArray}{C|C:CVC}
\end{pNiceArray}$
\end{scope}
+\bigskip
+\emph{Remarque} : Quand l'extension \pkg{array} (sur laquelle s'appuie \pkg{nicematrix}) est chargée, les traits
+verticaux et horizontaux que l'on insère rendent le tableau plus large ou plus long d'une quantité égale à la
+largeur du trait\footnote{En fait, cela est vrai pour |\hline| et ``\verb+|+'' mais pas pour |\cline|.}. Avec
+\pkg{nicematrix}, les lignes en pointillés tracées par |\hdottedline| et ``|:|'' ont le même effet.
+
\section{La largeur des colonnes}
\label{width}
@@ -1176,8 +1272,10 @@ c & d
\bigskip
Mais il est aussi possible de fixer une zone dans laquelle toutes les matrices auront leurs colonnes de la même
largeur, égale à la largeur de la case la plus large de toutes les matrices de la zone. Cette construction utilise
-l'environnement |{NiceMatrixBlock}| avec l'option |auto-columns-width|.\footnote{Pour le moment, c'est le seul
- usage de l'environnement |{NiceMatrixBlock}| mais il pourrait y en avoir davantage dans le futur.}
+l'environnement |{NiceMatrixBlock}| avec l'option |auto-columns-width|\footnote{Pour le moment, c'est le seul
+ usage de l'environnement |{NiceMatrixBlock}| mais il pourrait y en avoir davantage dans le futur.}.
+L'environnement |{NiceMatrixBlock}| n'a pas de rapport direct avec la commande |\Block| présentée juste ci-dessous
+(cf.~p.~\pageref{Block}).
\medskip
\begin{BVerbatim}[baseline=c,boxwidth=8.5cm]
@@ -1206,6 +1304,8 @@ $\begin{pNiceMatrix}
\section{Les matrices par blocs}
+\label{Block}
+
Cette partie, qui introduit une commande |\Block|, n'a pas de rapport direct avec l'environnement
|{NiceMatrixBlock}| présenté dans la section précédente.
@@ -1216,8 +1316,12 @@ d'un rectangle de cases fusionnées.
La commande |\Block| doit être utilisée dans la case supérieure gauche du bloc avec deux arguments. Le premier
argument est la taille de ce bloc avec la syntaxe $i$-$j$ où $i$ est le nombre de rangées et $j$ le nombre de
colonnes du bloc. Le deuxième argument, est, sans surprise, le contenu du bloc (en mode mathématique). Un nœud Tikz
-correspondant à l'ensemble des cellules fusionnées est créé sous le nom «\textsl{nom}-$i$-$j$» où \textsl{nom} est
-le nom donné au tableau.
+correspondant à l'ensemble des cases fusionnées est créé sous le nom «$i$-$j$-block» où \textsl{nom} est le
+nom donné au tableau. Si on a demandé la création des nœuds |medium|, alors un nœud de ce type est aussi créé pour
+ce bloc avec un nom suffixé par |-medium|.
+
+\medskip
+Dans les exemples qui suivent, on utilise la commande |\arrayrulecolor| de \pkg{colortbl}.
\medskip
\begin{BVerbatim}[baseline=c,boxwidth=10.6cm]
@@ -1281,6 +1385,43 @@ Pour des raisons techniques, il n'est pas possible d'écrire |\Block{|$i$|-|$j$|
\section{Fonctionnalités avancées}
+\subsection{La commande \textbackslash rotate}
+
+Utilisée au début d'une case, la commande |\rotate| (fournie par \pkg{nicematrix}) compose le contenu après une
+rotation de 90° dans le sens direct.
+
+Dans l'exemple suivant, on l'utilise dans le |code-for-first-row|.
+
+\bigskip
+
+\begin{BVerbatim}[baseline=c,boxwidth=12cm]
+\NiceMatrixOptions%
+ {code-for-first-row = \scriptstyle ~emphase#\rotate@ \text{image de },
+ code-for-last-col = \scriptstyle }
+$\begin{pNiceMatrix}[first-row,last-col=4]
+e_1 & e_2 & e_3 \\
+1 & 2 & 3 & e_1 \\
+4 & 5 & 6 & e_2 \\
+7 & 8 & 9 & e_3 \\
+\end{pNiceMatrix}$
+\end{BVerbatim}
+%
+\begin{scope}
+\NiceMatrixOptions%
+ {code-for-first-row = \scriptstyle\rotate \text{image de },
+ code-for-last-col = \scriptstyle }
+\begin{minipage}{3cm}
+$\begin{pNiceMatrix}[first-row,last-col=4]
+e_1 & e_2 & e_3 \\
+1 & 2 & 3 & e_1 \\
+4 & 5 & 6 & e_2 \\
+7 & 8 & 9 & e_3 \\
+\end{pNiceMatrix}$
+\end{minipage}
+\end{scope}
+
+
+
\subsection{L'option small}
Avec l'option |small|, les environnements de l'extension \pkg{nicematrix} sont composés d'une manière proche de ce
@@ -1365,8 +1506,8 @@ motif général. Ces commandes sont nommées |\pAutoNiceMatrix|, |\bAutoNiceMatr
|\VAutoNiceMatrix| et |\BAutoNiceMatrix|.
Chacune de ces commandes prend deux arguments obligatoires : le premier est la taille de la matrice, sous la forme
-$n$-$p$, où $n$ est le nombre de lignes et $p$ est le nombre de colonnes et le deuxième est le motif (c'est-à-dire
-simplement des tokens qui seront insérés dans chaque case de la matrice, exceptées celles des éventuelles lignes et
+$n$-$p$, où $n$ est le nombre de rangées et $p$ est le nombre de colonnes et le deuxième est le motif (c'est-à-dire
+simplement des tokens qui seront insérés dans chaque case de la matrice, exceptées celles des éventuelles rangées et
colonnes extérieures).
\medskip
@@ -1377,24 +1518,93 @@ $C = ~emphase#\pAutoNiceMatrix@{3-3}{C_{\arabic{iRow},\arabic{jCol}}}$
$C = \pAutoNiceMatrix{3-3}{C_{\arabic{iRow},\arabic{jCol}}}$
-\subsection{L'option hlines}
+\subsection{Les options hlines, vlines et hvlines}
+
+\label{hvlines}
-Dans les environnements de \pkg{nicematrix}, on peut bien entendu ajouter des filets horizontaux entre les lignes
-avec la commande~|\hline|. Par souci de commodité, l'extension \pkg{nicematrix} fournit l'option |hlines| qui impose
-directement que tous les filets horizontaux soient tracés (à l'exception, très naturelle, du filet avant
-l'éventuelle «première rangée» et après l'éventuelle «dernière rangée»).
+Dans les environnements de \pkg{nicematrix}, on peut bien entendu ajouter des filets horizontaux entre les rangées
+avec la commande~|\hline| et des filets verticaux avec le spécificateur ``\verb+|+'' dans le préambule de
+l'environnement. Par souci de commodité, l'extension \pkg{nicematrix} fournit aussi l'option |hlines| (resp.
+|vlines|) qui impose directement que tous les filets horizontaux (resp. verticaux) soient tracés (à l'exception,
+très naturelle, des filets extérieurs aux rangées et colonnes extérieures). L'option |lines| est la conjonction des
+options |hlines| et |vlines|.
\medskip
-\begin{BVerbatim}[baseline=c,boxwidth=10.6cm]
-$\begin{NiceArray}{|*{4}{C|}}[~emphase#hlines@,first-row,first-col]
+Dans l'exemple suivant, on utilise la commande |\arrayrulecolor| de \pkg{colortbl}.
+
+\medskip
+\begin{BVerbatim}[baseline=c,boxwidth=11cm]
+\arrayrulecolor{cyan}
+$\begin{NiceArray}{CCCC}%
+ [~emphase#hvlines@,first-row,first-col]
& e & a & b & c \\
e & e & a & b & c \\
a & a & e & c & b \\
b & b & c & e & a \\
c & c & b & a & e
\end{NiceArray}$
+\arrayrulecolor{black}
\end{BVerbatim}
-\raisebox{-5mm}{\box\ExampleOne}
+%
+\arrayrulecolor{cyan}
+$\begin{NiceArray}{CCCC}[hvlines,first-row,first-col]
+ & e & a & b & c \\
+e & e & a & b & c \\
+a & a & e & c & b \\
+b & b & c & e & a \\
+c & c & b & a & e
+\end{NiceArray}$
+\arrayrulecolor{black}
+
+
+\bigskip
+Il y a néanmoins une différence entre l'utilisation de l'option |vlines| et du spécificateur ``\verb+|+'' dans le
+préambule de l'environnement : les filets tracés par |vlines| traversent les double-filets horizontaux tracés par
+|\hline\hline|.
+
+\medskip
+\begin{BVerbatim}[baseline=c,boxwidth=11.5cm]
+$\begin{NiceArray}{CCCC}[vlines] \hline
+a & b & c & d \\ \hline \hline
+1 & 2 & 3 & 4 \\
+1 & 2 & 3 & 4 \\ \hline
+\end{NiceArray}$
+\end{BVerbatim}
+%
+$\begin{NiceArray}{CCCC}[vlines]
+\hline
+a & b & c & d \\
+\hline \hline
+1 & 2 & 3 & 4 \\
+1 & 2 & 3 & 4 \\
+\hline
+\end{NiceArray}$
+
+\bigskip
+Dans le cas d'un environnement avec délimiteurs (par exemple |{pNiceArray}| ou |pNiceMatrix|), l'option |vlines| ne
+trace pas de filets verticaux au niveau des deux délimiteurs (bien entendu).
+
+\medskip
+\begin{BVerbatim}[baseline=c,boxwidth=10.6cm]
+\setlength{\arrayrulewidth}{0.2pt}
+$\begin{pNiceMatrix}[vlines]
+1 & 2 & 3 & 4 & 5 & 6 \\
+1 & 2 & 3 & 4 & 5 & 6 \\
+1 & 2 & 3 & 4 & 5 & 6 \\
+\end{pNiceMatrix}$
+\end{BVerbatim}
+%
+\begin{scope}
+\setlength{\arrayrulewidth}{0.2pt}
+$\begin{pNiceMatrix}[vlines]
+1 & 2 & 3 & 4 & 5 & 6 \\
+1 & 2 & 3 & 4 & 5 & 6 \\
+1 & 2 & 3 & 4 & 5 & 6 \\
+\end{pNiceMatrix}$
+\end{scope}
+
+
+
\subsection{L'option light-syntax}
@@ -1470,8 +1680,8 @@ L'extension \pkg{nicematrix} fournit la commande |\OnlyMainNiceMatrix| qui est d
définitions de nouveaux types de colonnes. Son argument n'est exécuté que si on se place dans la partie principale
du tableau, c'est-à-dire que l'on n'est pas dans l'une des éventuelles rangées extérieures.
-Par exemple, si on souhaite définir un type de colonne |?| pour tracer un trait fort d'épaisseur 1~pt, on pourra
-écrire :
+Par exemple, si on souhaite définir un type de colonne |?| pour tracer un trait fort (noir) d'épaisseur 1~pt, on
+pourra écrire\footnote{La commande |\vrule| est une commande de TeX (et non de LaTeX).} :
\begin{Verbatim}
\newcolumntype{?}{!{\OnlyMainNiceMatrix{\vrule width 1 pt}}}
\end{Verbatim}
@@ -1513,7 +1723,7 @@ ne peuvent pas se croiser entre elles.\footnote{En revanche, les lignes créées
Cela signifie qu'une ligne en pointillés créée par l'une des ces commandes s'arrête automatiquement quand elle
arrive à une autre ligne pointillée déjà tracée par l'une de ces commandes. Par conséquent, l'ordre dans lequel les
lignes sont tracées a son importance pour le résultat final. Voici cet ordre (c'est à dessein qu'il a été choisi
-ainsi) : |\Hdotsfor|, |\Vdots|, |\Ddots|, |\Iddots|, |\Cdots| and |\Ldots|.
+ainsi) : |\Hdotsfor|, |\Vdots|, |\Ddots|, |\Iddots|, |\Cdots| et |\Ldots|.
De ce fait, on peut tracer la matrice suivante :
@@ -1550,8 +1760,8 @@ Les environnements créés par \pkg{nicematrix} sont numérotés par un compteur
complètement développable — et non d'un compteur).
Si l'environnement concerné a le numéro $n$, alors le nœud de la rangée~$i$ et de la colonne~$j$ a pour nom
-|nm-|$n$|-|$i$|-|$j$. Les noms des nœuds |medium| et |large| correspondants s'obtiennent suffixant par |-medium| et
-|-large|.
+|nm-|$n$|-|$i$|-|$j$. Les noms des nœuds |medium| et |large| correspondants s'obtiennent en suffixant par |-medium|
+et |-large|.
\subsection{Lignes diagonales}
@@ -1774,10 +1984,9 @@ obsolètes :
\end{itemize}
\medskip
-Depuis la version 3.8 de \pkg{nicematrix}, une erreur est levée dès qu'un de ces environnements est utilisé.
-Néanmoins, pour le moment, il reste possible de les utiliser si on charge l'extension \pkg{nicematrix} avec
-l'option |obsolete-environments|. Il faut toutefois avoir conscience que ces environnements seront sans doute
-supprimés dans une version future de \pkg{nicematrix}.
+Depuis la version 3.12 de \pkg{nicematrix}, on ne peut utiliser ces environnements que si on a chargé l'extension
+\pkg{nicematrix} avec l'option |obsolete-environments|. Il faut toutefois avoir conscience que ces environnements
+seront certainement supprimés dans une prochaine version de \pkg{nicematrix}.
@@ -2093,7 +2302,11 @@ a_{21} & a_{22} & a_{23} & a_{24} \\
a_{31} & a_{32} & a_{33} & a_{34} \\
a_{41} & a_{42} & a_{43} & a_{44}
\end{pNiceArray}\]
-
+%
+On remarquera que les traits que l'on vient de tracer sont dessinés \emph{après} la matrice sans modifier la
+position des composantes de celle-ci. En revanche, les traits tracés par |\hline| ou le spécificateur ``\verb+|+"''
+«écartent» les composantes de la matrice (quand l'extension \pkg{array} est chargée, ce qui est toujours le cas
+avec \pkg{nicematrix}).\footnote{En revanche les traits tracés par |\ncline| n'écartent pas les lignes de la matrice.}
\vspace{1cm}
@@ -2159,15 +2372,20 @@ les \emph{blend modes}. Néanmoins, le code suivant, dans le préambule du docum
\end{scope}
\vspace{1cm}
-On rappelle que dans le cas d'un ensemble de cellules fusionnées (avec la commande |\Block|), un nœud Tikz est créé
-pour l'ensemble des cellules avec pour nom $i$|-|$j$ où $i$ et $j$ sont les numéros de ligne et de colonne de
-la cellule en haut à gauche (où a été utilisée la commande |\Block|).
+On rappelle que dans le cas d'un ensemble de cases fusionnées (avec la commande |\Block|), un nœud Tikz est créé
+pour l'ensemble des cases avec pour nom $i$|-|$j$|-block| où $i$ et $j$ sont les numéros de ligne et de colonne de
+la case en haut à gauche (où a été utilisée la commande |\Block|). Si on a demandé la création des nœuds |medium|,
+alors un nœud de ce type est aussi créé pour ce bloc avec un nom suffixé par |-medium|.
\medskip
\begin{BVerbatim}[baseline=c,boxwidth=11.6cm]
$\begin{pNiceMatrix}%
- [margin,
- code-after = { \tikz \node [~emphase#highlight = (1-1)@] {} ; } ]
+ [
+ margin,
+ create-medium-nodes,
+ code-after =
+ { \tikz \node [~emphase#highlight = (1-1-block-medium)@] {} ; }
+ ]
\Block{3-3}<\Large>{A} & & & 0 \\
& \hspace*{1cm} & & \Vdots \\
& & & 0 \\
@@ -2175,8 +2393,12 @@ $\begin{pNiceMatrix}%
\end{pNiceMatrix}$
\end{BVerbatim}
$\begin{pNiceMatrix}%
- [margin,
- code-after = { \tikz \node [highlight = (1-1)] {} ; } ]
+ [
+ margin,
+ create-medium-nodes,
+ code-after =
+ { \tikz \node [highlight = (1-1-block-medium)] {} ; }
+ ]
\Block{3-3}<\Large>{A} & & & 0 \\
& \hspace*{1cm} & & \Vdots \\
& & & 0 \\
@@ -2431,5 +2653,11 @@ L_i & a_{i1} & \Cdots & a_{ik} & \Cdots & a_{in} \\
\draw [color = gray] (A-3-3) to [bend left] (B-3-3) ;
\end{tikzpicture}
+\section*{Autre documentation}
+
+Le document |nicematrix.pdf| (fourni avec l'extension \pkg{nicematrix}) contient une traduction anglaise de la
+documentation ici présente, ainsi que le code source commenté et un historique des versions.
+
+
\end{document}
diff --git a/macros/latex/contrib/nicematrix/nicematrix.dtx b/macros/latex/contrib/nicematrix/nicematrix.dtx
index 2752f0dd6f..9e446e9d48 100644
--- a/macros/latex/contrib/nicematrix/nicematrix.dtx
+++ b/macros/latex/contrib/nicematrix/nicematrix.dtx
@@ -15,8 +15,8 @@
%
% \fi
% \iffalse
-\def\myfileversion{3.11}
-\def\myfiledate{2020/02/07}
+\def\myfileversion{3.12}
+\def\myfiledate{2020/02/28}
%
%
%<*batchfile>
@@ -55,16 +55,6 @@ version 2005/12/01 or later.
\usepackage{nicematrix}
\usepackage{siunitx}
-\newbox\ExampleOne
-
-\setbox\ExampleOne\hbox{$\begin{NiceArray}{|*{4}{C|}}[hlines,first-row,first-col]
- & e & a & b & c \\
-e & e & a & b & c \\
-a & a & e & c & b \\
-b & b & c & e & a \\
-c & c & b & a & e
-\end{NiceArray}$}
-
\NewDocumentEnvironment {scope} {} {} {}
\def\interitem{\vspace{7mm plus 2 mm minus 3mm}}
@@ -118,7 +108,7 @@ c & c & b & a & e
% three compilations may be necessary. This package requires and \textbf{loads}
% the packages \pkg{expl3}, \pkg{l3keys2e}, \pkg{xparse}, \pkg{array},
% \pkg{amsmath} and \pkg{tikz}. It also loads the Tikz library \pkg{fit}. The
-% final user only has to load the extension with |\usepackage{nicematrix}|.
+% final user only has to load the extension with |\usepackage{nicematrix}|.
%
%
% \medskip
@@ -240,10 +230,10 @@ c & c & b & a & e
% defined: |\Ldots|, |\Cdots|, |\Vdots|, |\Ddots|, and |\Iddots|. These commands
% are intended to be used in place of |\dots|, |\cdots|, |\vdots|, |\ddots| and
% |\iddots|.\footnote{The command |\iddots|, defined in \pkg{nicematrix}, is a
-% variant of |\ddots| with dots going forward: \smash{$\iddots$}. If |mathdots|
-% is loaded, the version of |mathdots| is used. It corresponds to the command
-% |\adots| of \pkg{unicode-math}.}
-% \newcounter{fniddots}
+% variant of |\ddots| with dots going forward. If |mathdots| is loaded, the
+% version of |mathdots| is used. It corresponds to the command |\adots| of
+% \pkg{unicode-math}.}
+% \newcounter{fniddots}
% \setcounter{fniddots}{\thefootnote}
%
% \smallskip
@@ -590,7 +580,7 @@ c & c & b & a & e
%
% The distance between a node and the end of a dotted line is set by
% |dotted-lines-margin|. The initial value of this key is $0.3$~em (it's
-% recommended to use a unit dependant of the current font). For an example,
+% recommended to use a unit dependent of the current font). For an example,
% cf.~p.~\pageref{permutation}.
%
%
@@ -834,13 +824,109 @@ c & c & b & a & e
% the letters |L|, |C|, |R|.}
%
% \medskip
-% The environment |{NiceArray}| accepts the classical options |t|, |c| and |b|
-% of |{array}| but also other options defined by \pkg{nicematrix} (|renew-dots|,
-% |columns-width|, etc.).
+% The environment |{NiceArray}| accepts the options available for
+% |{pNiceMatrix}| and its variants but also a option |baseline| whose value is
+% an integer which indicates the number of the row whose baseline is used
+% as baseline for the environment |{NiceArray}|.
+%
+% \medskip
+% \begin{BVerbatim}[baseline=c,boxwidth=11cm]
+% $A =
+% \begin{NiceArray}{CCCC}[hvlines,~emphase#baseline=2@]
+% 1 & 2 & 3 & 4 \\
+% 1 & 2 & 3 & 4 \\
+% 1 & 2 & 3 & 4 \\
+% \end{NiceArray}$
+% \end{BVerbatim}
+% $A =
+% \begin{NiceArray}{CCCC}[hvlines,baseline=2]
+% 1 & 2 & 3 & 4 \\
+% 1 & 2 & 3 & 4 \\
+% 1 & 2 & 3 & 4 \\
+% \end{NiceArray}$
+%
+% \small
+% (The option |hvlines| is presented further: cf. p. \pageref{hvlines}.)
+%
+% \medskip
+% It's also possible to use the option |baseline| with one of the special values
+% |t|, |c| or |b|. These letters may also be used absolutely like the option of
+% the environment |{array}| of \pkg{array}. The initial value of |baseline| is~|c|.
+%
+%
+% \medskip
+% In the following example, we use the option |t| (equivalent to |baseline=t|)
+% immediately after an |\item| of list. One should remark that the presence of
+% a |\hline| at the beginning of the array doesn't prevent the alignment of the
+% baseline with the baseline of the first row (with |{array}| of \pkg{array},
+% one must use |\firsthline|\footnote{It's also possible to use |\firsthline|
+% with |{NiceArray}|.}).
+%
+% \smallskip
+% \begin{BVerbatim}[baseline=c,boxwidth=9cm]
+% \begin{enumerate}
+% \item an item
+% \smallskip
+% \item \renewcommand{\arraystretch}{1.2}
+% $\begin{NiceArray}~emphase#[t]@{LCCCCCC}
+% \hline
+% n & 0 & 1 & 2 & 3 & 4 & 5 \\
+% u_n & 1 & 2 & 4 & 8 & 16 & 32
+% \hline
+% \end{NiceArray}$
+% \end{enumerate}
+% \end{BVerbatim}
+% \begin{minipage}{5cm}
+% \begin{enumerate}
+% \item an item
+% \smallskip
+% \item \renewcommand{\arraystretch}{1.2}
+% $\begin{NiceArray}[t]{LCCCCCC}
+% \hline
+% n & 0 & 1 & 2 & 3 & 4 & 5 \\
+% u_n & 1 & 2 & 4 & 8 & 16 & 32 \\
+% \hline
+% \end{NiceArray}$
+% \end{enumerate}
+% \end{minipage}
+%
+% \medskip
+% However, it's also possible to use the tools of \pkg{booktabs}: |\toprule|,
+% |\bottomrule| and |\midrule|.\par\nobreak
%
+% \smallskip
+% \begin{BVerbatim}[baseline=c,boxwidth=9cm]
+% \begin{enumerate}
+% \item an item
+% \smallskip
+% \item
+% $\begin{NiceArray}[t]{LCCCCCC}
+% ~emphase#\toprule@
+% n & 0 & 1 & 2 & 3 & 4 & 5 \\
+% ~emphase#\midrule@
+% u_n & 1 & 2 & 4 & 8 & 16 & 32
+% ~emphase#\bottomrule@
+% \end{NiceArray}$
+% \end{enumerate}
+% \end{BVerbatim}
+% \begin{minipage}{5cm}
+% \begin{enumerate}
+% \item an item
+% \smallskip
+% \item
+% $\begin{NiceArray}[t]{LCCCCCC}
+% \toprule
+% n & 0 & 1 & 2 & 3 & 4 & 5 \\
+% \midrule
+% u_n & 1 & 2 & 4 & 8 & 16 & 32 \\
+% \bottomrule
+% \end{NiceArray}$
+% \end{enumerate}
+% \end{minipage}
+%
%
% \vspace{1cm}
-% An example with a linear system (we need |{NiceArray}| for the vertical rule):
+% With |{NiceArray}|, it's possible to draw vertical rules:\par\nobreak
%
% \bigskip
% \begin{BVerbatim}[baseline=c,boxwidth=10cm]
@@ -861,7 +947,7 @@ c & c & b & a & e
% \bigskip
% In fact, there is also variants for the environment |{NiceArray}|:
% |{pNiceArray}|, |{bNiceArray}|, |{BNiceArray}|, |{vNiceArray}| and
-% |{VNiceArray}|.
+% |{VNiceArray}|. The key |baseline| is not available for these environments.
%
% In the following example, we use an environment |{pNiceArray}| (we don't use
% |{pNiceMatrix}| because we want to use the types |L| and |R|
@@ -889,19 +975,19 @@ c & c & b & a & e
% more general environment, called |{NiceArrayWithDelims}|. The first two
% mandatory arguments of this environment are the left and right delimiters used
% in the construction of the matrix. It's possible to use
-% |{NiceArrayWithDelims}| if we want to use atypical delimiters.
+% |{NiceArrayWithDelims}| if we want to use atypical or asymetrical delimiters.
%
% \medskip
% \begin{BVerbatim}[baseline=c,boxwidth=11cm]
% $\begin{~emphase#NiceArrayWithDelims@}
-% {\downarrow}{\downarrow}{CCC}
+% {\downarrow}{\uparrow}{CCC}[margin]
% 1 & 2 & 3 \\
% 4 & 5 & 6 \\
% 7 & 8 & 9 \\
% \end{~emphase#NiceArrayWithDelims@}$
% \end{BVerbatim}
% $\begin{NiceArrayWithDelims}
-% {\downarrow}{\downarrow}{CCC}
+% {\downarrow}{\uparrow}{CCC}[margin]
% 1 & 2 & 3 \\
% 4 & 5 & 6 \\
% 7 & 8 & 9 \\
@@ -944,9 +1030,8 @@ c & c & b & a & e
% \begin{itemize}
% \item For the environments with an explicit preamble (i.e. |{NiceArray}| and
% its variants), no letter must be given in that preamble for the potential
-% first column and the potential last column: the first column will be
-% automatically (and necessarily) of type |R| and the last column will be
-% automatically of % type |L|.
+% first column and the potential last column: they will automatically (and
+% necessarily) be of type |R| for the first column and |L| for the last one.
% \item In an environment with an explicit preamble, the option |last-col| must
% be used \emph{without} value: the number of columns will be automatically
% computed from the preamble of the array.
@@ -1104,7 +1189,7 @@ c & c & b & a & e
% It's possible to change in \pkg{nicematrix} the letter used to specify a
% vertical dotted line with the option |letter-for-dotted-lines| available in
% |\NiceMatrixOptions|. For example, in this document, we have loaded the
-% extension \pkg{arydshln} which uses the letter ``:'' to specify a vertical
+% extension \pkg{arydshln} which uses the letter ``|:|'' to specify a vertical
% dashed line. Thus, by using |letter-for-dotted-lines|, we can use the
% vertical lines of both \pkg{arydshln} and \pkg{nicematrix}.
%
@@ -1126,6 +1211,16 @@ c & c & b & a & e
% \end{NiceArray}\right)$
% \end{scope}
%
+% \bigskip
+% \emph{Remark} : In the extension \pkg{array} (on which the extension
+% \pkg{nicematrix} relies), horizontal and vertical rules make the array larger
+% or wider by a quantity equal to the width of the rule\footnote{In fact, this
+% is true only for |\hline| and ``"|"'' but not for |\cline|.}. In
+% \pkg{nicematrix}, the dotted lines drawn |\hdottedline| and ``|:|'' do
+% likewise.
+%
+%
+%
% \section{The width of the columns}
% \label{width}
%
@@ -1225,9 +1320,11 @@ c & c & b & a & e
% But it's also possible to fix a zone where all the matrices will have their
% columns of the same width, equal to the widest cell of all the matrices. This
% construction uses the environment |{NiceMatrixBlock}| with the option
-% |auto-columns-width|.\footnote{At this time, this is the only usage of the
-% environment |{NiceMatrixBlock}| but it may have other usages in the future.}
-%
+% |auto-columns-width|\footnote{At this time, this is the only usage of the
+% environment |{NiceMatrixBlock}| but it may have other usages in the future.}.
+% The environment |{NiceMatrixBlock}| has no direct link with the command
+% |\Block| presented just below (cf.~p.~\pageref{Block}).
+%
% \medskip
% \begin{BVerbatim}[baseline=c,boxwidth=8.5cm]
% ~emphase#\begin{NiceMatrixBlock}[auto-columns-width]@
@@ -1255,6 +1352,7 @@ c & c & b & a & e
%
%
% \section{Block matrices}
+% \label{Block}
%
% This section has no direct link with the previous one where an environment
% |{NiceMatrixBlock}| was introduced.
@@ -1268,8 +1366,13 @@ c & c & b & a & e
% $i$\verb|-|$j$ where $i$ is the number of rows of the block and $j$ its number
% of columns. The second argument is the content of the block (composed in math
% mode). A Tikz node corresponding to the merged cells is created with the name
-% ``\textsl{name}-$i$-$j$'' where \textsl{name} is the name given to the array.
+% ``$i$-$j$-block''. If the user has required the creation of the ``medium
+% nodes'', a node of this type is also created with a name suffixed by
+% |-medium|.
%
+% \medskip
+% In the following examples, we use the command |\arrayrulecolor| of
+% \pkg{colortbl}.
%
% \medskip
% \begin{BVerbatim}[baseline=c,boxwidth=10.6cm]
@@ -1300,7 +1403,7 @@ c & c & b & a & e
% previous example. Since this element is composed in math mode, it's not
% possible to use directly a command like |\large|, |\Large| and |\LARGE|.
% That's why the command |\Block| provides an option between angle brackets to
-% specificy some TeX code which will be inserted before the beginning of the
+% specify some TeX code which will be inserted before the beginning of the
% math mode.
%
% \medskip
@@ -1334,6 +1437,41 @@ c & c & b & a & e
%
% \section{Advanced features}
%
+% \subsection{The command \textbackslash rotate}
+%
+% The package \pkg{nicematrix} provides a command |\rotate|. When used in the
+% beginning of a cell, this command composes the contents of the cell after a
+% rotation of 90° in the direct sens.
+%
+% In the following command, we use that command in the |code-for-first-row|.
+%
+%\bigskip
+%
+% \begin{BVerbatim}[baseline=c,boxwidth=12cm]
+% \NiceMatrixOptions%
+% {code-for-first-row = \scriptstyle ~emphase#\rotate@ \text{image of },
+% code-for-last-col = \scriptstyle }
+% $\begin{pNiceMatrix}[first-row,last-col=4]
+% e_1 & e_2 & e_3 \\
+% 1 & 2 & 3 & e_1 \\
+% 4 & 5 & 6 & e_2 \\
+% 7 & 8 & 9 & e_3 \\
+% \end{pNiceMatrix}$
+% \end{BVerbatim}
+% \begin{scope}
+% \NiceMatrixOptions%
+% {code-for-first-row = \scriptstyle\rotate \text{image of },
+% code-for-last-col = \scriptstyle }
+% \begin{minipage}{3cm}
+% $\begin{pNiceMatrix}[first-row,last-col=4]
+% e_1 & e_2 & e_3 \\
+% 1 & 2 & 3 & e_1 \\
+% 4 & 5 & 6 & e_2 \\
+% 7 & 8 & 9 & e_3 \\
+% \end{pNiceMatrix}$
+% \end{minipage}
+% \end{scope}
+%
% \subsection{The option small}
%
% With the option |small|, the environments of the extension \pkg{nicematrix}
@@ -1383,8 +1521,8 @@ c & c & b & a & e
%
% In the cells of the array, it's possible to use the LaTeX counters |iRow| and
% |jCol| which represent the number of the current row and the number of the
-% current col\footnote{We recall that the first row (if it exists) has the
-% number~$0$ and that the first col (if it exists) has also the number~$0$.}.
+% current column\footnote{We recall that the first row (if it exists) has the
+% number~$0$ and that the first column (if it exists) has also the number~$0$.}.
% Of course, the user must not change the value of these counters which are used
% internally by \pkg{nicematrix}.
%
@@ -1436,27 +1574,90 @@ c & c & b & a & e
%
% $C = \pAutoNiceMatrix{3-3}{C_{\arabic{iRow},\arabic{jCol}}}$
%
-% \subsection{The option hlines}
+% \subsection{The options hlines, vlines and hvlines}
%
+% \label{hvlines}
% You can add horizontal rules between rows in the environments of
-% \pkg{nicematrix} with the usual command |\hline|. But, by convenience, the
-% extension \pkg{nicematrix} also provides the option |hlines|. With this
-% option, all the horizontal rules will be drawn (excepted, of course, the rule
-% before the potential ``first row'' and the rule after the potential ``last
-% row'').
+% \pkg{nicematrix} with the usual command |\hline| and you can use the specifier
+% ``"|"'' to add vertical rules. However, by convenience, the extension
+% \pkg{nicematrix} also provides the option |hlines| (resp. |vlines|) which will
+% draw all the horizontal (resp. vertical) rules (excepted, of course, the
+% exterior rules corresponding to the exterior rows and columns). The key
+% |hvlines| is an alias for the conjonction for the keys |hlines| et |vlines|.
+%
+% % \medskip
+% In the following example, we use the command |\arrayrulecolor| of
+% \pkg{colortbl}.
%
% \medskip
-% \begin{BVerbatim}[baseline=c,boxwidth=10.6cm]
-% $\begin{NiceArray}{|*{4}{C|}}[~emphase#hlines@,first-row,first-col]
+% \begin{BVerbatim}[baseline=c,boxwidth=11cm]
+% \arrayrulecolor{cyan}
+% $\begin{NiceArray}{CCCC}%
+% [~emphase#hvlines@,first-row,first-col]
+% % & e & a & b & c \\
+% e & e & a & b & c \\
+% a & a & e & c & b \\
+% b & b & c & e & a \\
+% c & c & b & a & e
+% \end{NiceArray}$
+% \arrayrulecolor{black}
+% \end{BVerbatim}
+% \arrayrulecolor{cyan}
+% $\begin{NiceArray}{CCCC}[hvlines,first-row,first-col]
% & e & a & b & c \\
% e & e & a & b & c \\
% a & a & e & c & b \\
% b & b & c & e & a \\
% c & c & b & a & e
% \end{NiceArray}$
+% \arrayrulecolor{black}
+%
+% \bigskip
+% However, there is a difference between the key |vlines| and the use of
+% the specifier ``"|"'' in the preamble of the environment: the rules drawn by
+% |vlines| completely cross the double-rules drawn by |\hline\hline|.
+%
+% \medskip
+% \begin{BVerbatim}[baseline=c,boxwidth=11.5cm]
+% $\begin{NiceArray}{CCCC}[vlines] \hline
+% a & b & c & d \\ \hline \hline
+% 1 & 2 & 3 & 4 \\
+% 1 & 2 & 3 & 4 \\ \hline
+% \end{NiceArray}$
% \end{BVerbatim}
-% \raise-5mm\box\ExampleOne
-%
+% $\begin{NiceArray}{CCCC}[vlines]
+% \hline
+% a & b & c & d \\
+% \hline \hline
+% 1 & 2 & 3 & 4 \\
+% 1 & 2 & 3 & 4 \\
+% \hline
+% \end{NiceArray}$
+%
+% \bigskip
+% For the environments with delimiters (for example |{pNiceArray}| or
+% |{pNiceMatrix}|), the option |vlines| don't draw vertical rules on both sides,
+% where are the delimiters (fortunately).
+%
+% \medskip
+% \begin{BVerbatim}[baseline=c,boxwidth=10.6cm]
+% \setlength{\arrayrulewidth}{0.2pt}
+% $\begin{pNiceMatrix}[vlines]
+% 1 & 2 & 3 & 4 & 5 & 6 \\
+% 1 & 2 & 3 & 4 & 5 & 6 \\
+% 1 & 2 & 3 & 4 & 5 & 6 \\
+% \end{pNiceMatrix}$
+% \end{BVerbatim}
+% \begin{scope}
+% \setlength{\arrayrulewidth}{0.2pt}
+% $\begin{pNiceMatrix}[vlines]
+% 1 & 2 & 3 & 4 & 5 & 6 \\
+% 1 & 2 & 3 & 4 & 5 & 6 \\
+% 1 & 2 & 3 & 4 & 5 & 6 \\
+% \end{pNiceMatrix}$
+% \end{scope}
+%
+%
% \subsection{The option light-syntax}
%
% The option |light-syntax|\footnote{This option is inspired by the extension
@@ -1494,7 +1695,7 @@ c & c & b & a & e
% It's possible to change the character used to mark the end of rows with the
% option |end-of-row|. As said before, the initial value is a semicolon.
%
-% \subsection{Utilisation of the column type S of siunitx}
+% \subsection{Use of the column type S of siunitx}
%
% If the package \pkg{siunitx} is loaded (before or after \pkg{nicematrix}),
% it's possible to use the |S| column type of \pkg{siunitx} in the environments
@@ -1538,7 +1739,8 @@ c & c & b & a & e
% not in an eventual exterior row.
%
% For example, one may wish to define a new column type |?| in order to draw a
-% thick rule of width 1~pt. The following definition will do the job:
+% (black) thick rule of width 1~pt. The following definition will do the
+% job\footnote{The command |\vrule| is a TeX (and not LaTeX) command.}:
%
% \begin{Verbatim}
% \newcolumntype{?}{!{\OnlyMainNiceMatrix{\vrule width 1 pt}}}
@@ -1859,11 +2061,10 @@ c & c & b & a & e
% |{VNiceArrayRC}|.
% \end{itemize}
%
-% Since the version 3.8, an error is raised when one of these environments is
-% used. It's still possible to use these environments by loading
+% Since the version 3.12, the only way to use these environments is loading
% \pkg{nicematrix} with the option |obsolete-environments|.
%
-% However, these environments will probably be completely deleted in a future
+% However, these environments will certainly be completely deleted in a future
% version of \pkg{nicematrix}.
%
% \section{Examples}
@@ -2185,6 +2386,14 @@ c & c & b & a & e
% a_{41} & a_{42} & a_{43} & a_{44}
% \end{pNiceArray}\]
%
+% We should remark that the rules we have drawn are drawn \emph{after} the
+% construction of the array and, hence, they don't spread the cells of the
+% array. We recall that, on the other side, the command |\hline| and the
+% specifier ``"|"'' spread the cells (when the package \pkg{array} is loaded
+% but, when the package \pkg{nicematrix} is loaded, \pkg{array} is always
+% loaded).\footnote{On the other side, the command |\cline| doesn't spread the
+% rows of the array.}
+%
%
% \vspace{1cm}
% The package \pkg{nicematrix} is constructed upon the environment |{array}|
@@ -2249,25 +2458,35 @@ c & c & b & a & e
% |\ExplSyntaxOff|
% \end{scope}
%
-% \vspace{1cm}
+% \vspace{1cm}
% We recall that, for a rectangle of merged cells (with the command |\Block|), a
-% Tikz node is created for the set of merged cells with the name $i$|-|$j$ where
-% $i$ and $j$ are the number of the row and the number of the column of the
-% upper left cell (where the command |\Block| has been issued).
+% Tikz node is created for the set of merged cells with the name
+% $i$|-|$j$-|block| where $i$ and $j$ are the number of the row and the number
+% of the column of the upper left cell (where the command |\Block| has been
+% issued). If the user has required the creation of the |medium| nodes, a node
+% of this type is also created with a name suffixed by |-medium|.
%
% \medskip
% \begin{BVerbatim}[baseline=c,boxwidth=11.6cm]
% $\begin{pNiceMatrix}%
-% [margin,
-% code-after = { \tikz \node [~emphase#highlight = (1-1)@] {} ; } ]
+% [
+% margin,
+% create-medium-nodes,
+% code-after =
+% { \tikz \node [~emphase#highlight = (1-1-block-medium)@] {} ; }
+% ]
% \Block{3-3}<\Large>{A} & & & 0 \\
% & \hspace*{1cm} & & \Vdots \\
% & & & 0 \\
% 0 & \Cdots& 0 & 0
% \end{pNiceMatrix}$
% \end{BVerbatim}
-% $\begin{pNiceMatrix}[margin,
-% code-after = { \tikz \node [highlight = (1-1)] {} ; } ]
+% $\begin{pNiceMatrix}[
+% margin,
+% create-medium-nodes,
+% code-after =
+% { \tikz \node [highlight = (1-1-block-medium)] {} ; }
+% ]
% \Block{3-3}<\Large>{A} & & & 0 \\
% & \hspace*{1cm} & & \Vdots \\
% & & & 0 \\
@@ -2416,13 +2635,13 @@ c & c & b & a & e
% \end{pNiceArray}\]
%
%
-% \subsection{Direct utilisation of the Tikz nodes}
+% \subsection{Direct use of the Tikz nodes}
%
% In the following example, we illustrate the mathematical product of two
% matrices.
%
% \medskip
-% The utilisation of |{NiceMatrixBlock}| with the option |auto-columns-width|
+% The use of |{NiceMatrixBlock}| with the option |auto-columns-width|
% gives the same width for all the columns and, therefore, a perfect alignment
% of the two superposed matrices.
% \begin{Verbatim}
@@ -2550,8 +2769,14 @@ c & c & b & a & e
% |\halign|.
%
%
-% \subsection{Declaration of the package and extensions loaded}
+% \bigskip
+% \subsection*{Declaration of the package and extensions loaded}
+%
%
+% The prefix |nicematrix| has been registred for this extension.
+%
+% See: |http://mirrors.ctan.org/macros/latex/contrib/l3kernel/l3prefixes.pdf|
+%
%<@@=nicematrix>
%
% \bigskip
@@ -2563,7 +2788,7 @@ c & c & b & a & e
% \begin{macrocode}
\RequirePackage{tikz}
\usetikzlibrary{fit}
-\RequirePackage{expl3}[2019/07/01]
+\RequirePackage{expl3}[2020/02/08]
% \end{macrocode}
%
% We give the traditionnal declaration of a package written with |expl3|:
@@ -2576,10 +2801,28 @@ c & c & b & a & e
{Mathematical matrices with TikZ}
% \end{macrocode}
%
+%
+% \medskip
+% The version of 2020/02/08 of \pkg{expl3} has replaced |\l_keys_key_tl| by
+% |\l_keys_key_str|. We have immediately changed in this file. Now, you test the
+% existence of |\l_keys_key_str| in order to detect whether the version of LaTeX
+% used by the final user is up to date.
+% \begin{macrocode}
+\msg_new:nnn { nicematrix } { expl3~too~old }
+ {
+ Your~version~of~LaTeX~(especially~expl3)~is~too~old.~
+ You~can~go~on~but~you~will~probably~have~other~errors~
+ if~you~use~the~functionalities~of~nicematrix.
+ }
+\cs_if_exist:NF \l_keys_key_str
+ { \msg_error:nn { nicematrix } { expl3~too~old } }
+% \end{macrocode}
+%
+%
% \bigskip
-% We test if the class option |draft| has been used. In this case, we raise the
-% flag |\c_@@_draft_bool| because we won't draw the dotted lines if the option
-% |draft| is used.
+% We test the class option |draft|. In this case, we raise the flag
+% |\c_@@_draft_bool| because we won't draw the dotted lines if the option
+% |draft| is used.
% \begin{macrocode}
\bool_new:N \c_@@_draft_bool
\DeclareOption { draft } { \bool_set_true:N \c_@@_draft_bool }
@@ -2606,7 +2849,7 @@ c & c & b & a & e
\cs_new_protected:Npn \@@_error:nn { \msg_error:nnn { nicematrix } }
\cs_new_protected:Npn \@@_error:nnn { \msg_error:nnnn { nicematrix } }
\cs_new_protected:Npn \@@_fatal:n { \msg_fatal:nn { nicematrix } }
-\cs_new_protected:Npn \@@_fatal:nn { \msg_fatal:nn { nicematrix } }
+\cs_new_protected:Npn \@@_fatal:nn { \msg_fatal:nnn { nicematrix } }
\cs_new_protected:Npn \@@_msg_new:nn { \msg_new:nnn { nicematrix } }
\cs_new_protected:Npn \@@_msg_new:nnn { \msg_new:nnnn { nicematrix } }
% \end{macrocode}
@@ -2617,7 +2860,8 @@ c & c & b & a & e
% \end{macrocode}
%
%
-% \subsection{Technical definitions}
+% \bigskip
+% \subsection*{Technical definitions}
%
% We test whether the current class is \cls{revtex4-1} or \cls{revtex4-2}
% because these classes redefines |\array| (of \pkg{array}) in a way
@@ -2657,12 +2901,14 @@ c & c & b & a & e
{
\mathinner
{
- \mkern 1 mu
- \raise \p@ \hbox:n { . }
- \mkern 2 mu
- \raise 4 \p@ \hbox:n { . }
- \mkern 2 mu
- \raise 7 \p@ \vbox { \kern 7 pt \hbox:n { . } } \mkern 1 mu
+ \tex_mkern:D 1 mu
+ \box_move_up:nn { 1 pt } { \hbox:n { . } }
+ \tex_mkern:D 2 mu
+ \box_move_up:nn { 4 pt } { \hbox:n { . } }
+ \tex_mkern:D 2 mu
+ \box_move_up:nn { 7 pt }
+ { \vbox:n { \kern 7 pt \hbox:n { . } } }
+ \tex_mkern:D 1 mu
}
}
% \end{macrocode}
@@ -2756,10 +3002,10 @@ c & c & b & a & e
\bool_new:N \l_@@_impossible_line_bool
% \end{macrocode}
%
-%
+%
% \bigskip
% We have to know whether \pkg{colortbl} is loaded for the redefinition of
-% |\everycr| and for |\vline|.
+% |\everycr| and |\vline| and for the options |hlines| and |vlines|.
% \begin{macrocode}
\bool_new:N \c_@@_colortbl_loaded_bool
\AtBeginDocument
@@ -2773,6 +3019,14 @@ c & c & b & a & e
}
% \end{macrocode}
%
+% \bigskip
+% We have put a argument |w| (\emph{weird}) for the following function because
+% its argument should be a specifier of \pkg{pgf} point between rounded
+% brackets.
+% \begin{macrocode}
+\cs_set_protected:Npn \@@_extract_coords:w
+ { \tikz@parse@node \pgfutil@firstofone }
+% \end{macrocode}
%
% \bigskip
% The length |\l_@@_inter_dots_dim| is the distance between two dots for the
@@ -2795,20 +3049,40 @@ c & c & b & a & e
%
% \bigskip
% The length |\l_@@_radius_dim| is the radius of the dots for the dotted lines.
-% The default value is 0.34~pt but it will be changed if the option |small| is
-% used.
+% The default value is 0.53~pt but it will be changed if the option |small| is
+% used (to 0.37~pt).
% \begin{macrocode}
\dim_new:N \l_@@_radius_dim
\dim_set:Nn \l_@@_radius_dim { 0.53 pt }
% \end{macrocode}
%
% \bigskip
-% The name of the current environment or the current command (will be used only
-% in the error messages).
+% The name of the current environment or the current command (despite the name).
% \begin{macrocode}
-\str_new:N \g_@@_type_env_str
+\str_new:N \g_@@_name_env_str
% \end{macrocode}
-%
+%
+% The string |\g_@@_com_or_env_str| will contain the word \emph{command} or
+% \emph{environment} whether we are in a command of \pkg{nicematrix} or a an
+% environment of \pkg{nicematrix}. The default value is \emph{environment}.
+% \begin{macrocode}
+\str_new:N \g_@@_com_or_env_str
+\str_set:Nn \g_@@_com_or_env_str { environment }
+% \end{macrocode}
+%
+% The following control sequence will be able to reconstruct the full name of
+% the current command or environment (despite the name). This command must
+% \emph{not} be protected since it's used in error messages.
+% \begin{macrocode}
+\cs_new:Npn \@@_full_name_env:
+ {
+ \str_if_eq:VnTF \g_@@_com_or_env_str { command }
+ { command \space \c_backslash_str \g_@@_name_env_str }
+ { environment \space \{ \g_@@_name_env_str \} }
+ }
+% \end{macrocode}
+%
+% \bigskip
% \begin{macrocode}
\tl_new:N \g_@@_code_after_tl
% \end{macrocode}
@@ -2824,11 +3098,17 @@ c & c & b & a & e
% \end{macrocode}
%
% The TeX counters |\c@iRow| and |\c@jCol| will be created in the beginning of
-% the environment |{NiceArrayWithDelims}| (if they don't exist previously).
+% |{NiceArrayWithDelims}| (if they don't exist previously).
%
+% \begin{macrocode}
+\bool_new:N \g_@@_row_of_col_done_bool
+% \end{macrocode}
+%
+%
% \bigskip
-% \subsubsection{Variables for the exterior rows and columns}
+% \textbf{Variables for the exterior rows and columns}\par\nobreak
%
+% \medskip
% The keys for the exterior rows and columns are |first-row|, |first-col|,
% |last-row| and |last-col|. However, internally, these keys are not coded in a
% similar way.
@@ -2907,9 +3187,9 @@ c & c & b & a & e
% \end{itemize}
%
% \bigskip
-% \subsubsection{The column S of siunitx}
+% \textbf{The column S of siunitx}\par\nobreak
%
-%
+% \medskip
% We want to know whether the package \pkg{siunitx} is loaded and, if it is
% loaded, we redefine the |S| columns of \pkg{siunitx}.
% \begin{macrocode}
@@ -2965,7 +3245,7 @@ c & c & b & a & e
% when \pkg{siunitx} is loaded (and it may be loaded after \pkg{nicematrix})
% and, in fact, after the beginning of the document --- because some
% instructions of \pkg{siunitx} are executed in a |\AtBeginDocument|). That's
-% why this extraction will be done only at the first utilisation of an
+% why this extraction will be done only at the first use of an
% environment of \pkg{nicematrix} with the command |\@@_adapt_S_column:|.
% \begin{macrocode}
\cs_set_protected:Npn \@@_adapt_S_column:
@@ -2975,9 +3255,8 @@ c & c & b & a & e
% won't be known. That's why we test the existence of
% |\c_@@_siunitx_loaded_bool| and not its value.\footnote{Indeed,
% \pkg{nicematrix} may be used in the preamble of the LaTeX document. For
-% example, in this document, we compose a matrix in the box |\ExampleOne| before
-% loading \pkg{arydshln} (because \pkg{arydshln} is not totally compatible with
-% \pkg{nicematrix}).}
+% example, you can compose a matrix in a box before the loading of
+% \pkg{arydshln} (not totally compatible with \pkg{nicematrix}).}
% \begin{macrocode}
\bool_if:NT \c_@@_siunitx_loaded_bool
{
@@ -3036,7 +3315,7 @@ c & c & b & a & e
%
%
% \bigskip
-% \subsection{The options}
+% \subsection*{The options}
%
% The boolean |\l_@@_light_syntax_bool| corresponds to the option |light-syntax|.
% \begin{macrocode}
@@ -3044,15 +3323,15 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% The token list |\l_@@_pos_env_str| will contain one of the three values |t|,
+% The token list |\l_@@_baseline_str| will contain one of the three values |t|,
% |c| or |b| and will indicate the position of the environment as in the option
% of the environment |{array}|. For the environment |{pNiceMatrix}|,
% |{pNiceArray}| and their variants, the value will programmatically be fixed to
% |c|. For the environment |{NiceArray}|, however, the three values |t|, |c| and
% |b| are possible.
% \begin{macrocode}
-\str_new:N \l_@@_pos_env_str
-\str_set:Nn \l_@@_pos_env_str c
+\str_new:N \l_@@_baseline_str
+\str_set:Nn \l_@@_baseline_str c
% \end{macrocode}
%
% \bigskip
@@ -3073,9 +3352,11 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% The flag |\l_@@_hlines_bool| corresponds to the option |\hlines|.
+% The flag |\l_@@_hlines_bool| corresponds to the option |\hlines| and the flag
+% |\l_@@_vlines_bool| to the option |\vlines|.
% \begin{macrocode}
\bool_new:N \l_@@_hlines_bool
+\bool_new:N \l_@@_vlines_bool
% \end{macrocode}
%
% \bigskip
@@ -3117,11 +3398,17 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% The dimensions |\l_@@_left_margin_dim| and |\l_@@_right_margin_dim| correspond
-% to the options |left-margin| and |right-margin|.
+% The dimension |\l_@@_left_margin_dim| correspondto the option |left-margin|
+% (idem for the right margin).
% \begin{macrocode}
\dim_new:N \l_@@_left_margin_dim
\dim_new:N \l_@@_right_margin_dim
+% \end{macrocode}
+%
+% \bigskip
+% The following dimensions will be used internally to compute the width of the
+% potential ``first column'' and ``last column''.
+% \begin{macrocode}
\dim_new:N \g_@@_width_last_col_dim
\dim_new:N \g_@@_width_first_col_dim
% \end{macrocode}
@@ -3144,6 +3431,20 @@ c & c & b & a & e
\tl_set:Nn \l_@@_end_of_row_tl { ; }
% \end{macrocode}
%
+% \medskip
+% Sometimes, we want to have several arrays vertically juxtaposed in order to
+% have an alignment of the columns of these arrays. To acheive this goal, one
+% may wish to use the same width for all the columns (for example with the
+% option |columns-width| or the option |auto-columns-width| of the environment
+% |{NiceMatrixBlock}|). However, even if we use the same type of delimiters, the
+% width of the delimiters may be different from an array to another because the
+% width of the delimiter is fonction of its size. That's why we create an option
+% called |max-delimiter-width| which will give to the delimiters the width of
+% a delimiter (of the same type) of big size. The following boolean corresponds
+% to this option.
+% \begin{macrocode}
+\bool_new:N \l_@@_max_delimiter_width_bool
+% \end{macrocode}
%
% \bigskip
% First, we define a set of keys ``|NiceMatrix / Global|'' which will be used
@@ -3151,6 +3452,7 @@ c & c & b & a & e
% \begin{macrocode}
\keys_define:nn { NiceMatrix / Global }
{
+ max-delimiter-width .bool_set:N = \l_@@_max_delimiter_width_bool ,
dotted-lines-margin .dim_set:N = \l_@@_dotted_lines_margin_dim ,
dotted-lines-margin .value_required:n = true ,
light-syntax .bool_set:N = \l_@@_light_syntax_bool ,
@@ -3167,6 +3469,8 @@ c & c & b & a & e
code-for-last-row .value_required:n = true ,
small .bool_set:N = \l_@@_small_bool ,
hlines .bool_set:N = \l_@@_hlines_bool ,
+ vlines .bool_set:N = \l_@@_vlines_bool ,
+ hvlines .meta:n = { hlines , vlines } ,
parallelize-diags .bool_set:N = \l_@@_parallelize_diags_bool ,
% \end{macrocode}
%
@@ -3213,13 +3517,14 @@ c & c & b & a & e
{ \dim_set:Nn \l_@@_columns_width_dim { #1 } } ,
columns-width .value_required:n = true ,
name .code:n =
- \unless \ifmeasuring@
- \str_set:Nn \l_tmpa_str { #1 }
- \seq_if_in:NVTF \g_@@_names_seq \l_tmpa_str
- { \@@_error:nn { Duplicate~name } { #1 } }
- { \seq_gput_left:NV \g_@@_names_seq \l_tmpa_str }
- \str_set_eq:NN \l_@@_name_str \l_tmpa_str
- \fi ,
+ \legacy_if:nF { measuring@ }
+ {
+ \str_set:Nn \l_tmpa_str { #1 }
+ \seq_if_in:NVTF \g_@@_names_seq \l_tmpa_str
+ { \@@_error:nn { Duplicate~name } { #1 } }
+ { \seq_gput_left:NV \g_@@_names_seq \l_tmpa_str }
+ \str_set_eq:NN \l_@@_name_str \l_tmpa_str
+ } ,
name .value_required:n = true ,
code-after .tl_gset:N = \g_@@_code_after_tl ,
code-after .value_required:n = true ,
@@ -3275,10 +3580,8 @@ c & c & b & a & e
% \begin{macrocode}
renew-matrix .code:n = \@@_renew_matrix: ,
renew-matrix .value_forbidden:n = true ,
- RenewMatrix .code:n = \@@_error:n { Option~RenewMatrix~suppressed } ,
transparent .meta:n = { renew-dots , renew-matrix } ,
transparent .value_forbidden:n = true,
- Transparent .code:n = \@@_error:n { Option~Transparent~suppressed } ,
% \end{macrocode}
%
%
@@ -3321,7 +3624,7 @@ c & c & b & a & e
% \begin{macrocode}
letter-for-dotted-lines .code:n =
{
- \int_compare:nTF { \tl_count:n { #1 } = \c_one_int }
+ \int_compare:nTF { \tl_count:n { #1 } = 1 }
{ \str_set:Nx \l_@@_letter_for_dotted_lines_str { #1 } }
{ \@@_error:n { Bad~value~for~letter~for~dotted~lines } }
} ,
@@ -3378,9 +3681,11 @@ c & c & b & a & e
% The options |c|, |t| and |b| of the environment |{NiceArray}| have the same
% meaning as the option of the classical environment |{array}|.
% \begin{macrocode}
- c .code:n = \str_set:Nn \l_@@_pos_env_str c ,
- t .code:n = \str_set:Nn \l_@@_pos_env_str t ,
- b .code:n = \str_set:Nn \l_@@_pos_env_str b ,
+ c .code:n = \str_set:Nn \l_@@_baseline_str c ,
+ t .code:n = \str_set:Nn \l_@@_baseline_str t ,
+ b .code:n = \str_set:Nn \l_@@_baseline_str b ,
+ baseline .tl_set:N = \l_@@_baseline_str ,
+ baseline .value_required:n = true ,
% \end{macrocode}
%
% In the environments |{NiceArray}| and its variants, the option |last-col| must
@@ -3411,7 +3716,8 @@ c & c & b & a & e
% \end{macrocode}
%
%
-% \subsection{Important code used by \{NiceArrayWithDelims\} }
+% \bigskip
+% \subsection*{Important code used by \{NiceArrayWithDelims\} }
%
% The pseudo-environment |\@@_Cell:|--|\@@_end_Cell:| will be used to format the
% cells of the array. In the code, the affectations are global because this
@@ -3419,7 +3725,7 @@ c & c & b & a & e
% environment |{array}|).
%
% \begin{macrocode}
-\cs_new_protected:Nn \@@_Cell:
+\cs_new_protected:Npn \@@_Cell:
{
% \end{macrocode}
% We increment |\c@jCol|, which is the counter of the columns.
@@ -3430,29 +3736,30 @@ c & c & b & a & e
% the |\everycr| because some packages, like \pkg{arydshln}, create special rows
% in the |\halign| that we don't want to take into account.
% \begin{macrocode}
- \int_compare:nNnT \c@jCol = \c_one_int
+ \int_compare:nNnT \c@jCol = 1
{
- \int_compare:nNnT \l_@@_first_col_int = \c_one_int
+ \int_compare:nNnT \l_@@_first_col_int = 1
\@@_begin_of_row:
}
\int_gset:Nn \g_@@_col_total_int
{ \int_max:nn \g_@@_col_total_int \c@jCol }
% \end{macrocode}
-% The content of the cell is composed in the box |\l_tmpa_box| because we want
+% The content of the cell is composed in the box |\l_@@_cell_box| because we want
% to compute some dimensions of the box. The |\hbox_set_end:| corresponding to
% this |\hbox_set:Nw| will be in the |\@@_end_Cell:| (and the
% |\c_math_toggle_token| also).
% \begin{macrocode}
- \hbox_set:Nw \l_tmpa_box
+ \hbox_set:Nw \l_@@_cell_box
\c_math_toggle_token
\bool_if:NT \l_@@_small_bool \scriptstyle
% \end{macrocode}
%
% We will call \emph{corners} of the matrix the cases which are at the
% intersection of the exterior rows and exterior columns (of course, the four
-% corners doesn't always exist simultaneously). The codes
-% |\l_@@_code_for_first_row_tl| and \emph{al} don't apply in the corners of the
-% matrix.
+% corners doesn't always exist simultaneously).
+%
+% The codes |\l_@@_code_for_first_row_tl| and \emph{al} don't apply in the
+% corners of the matrix.
% \begin{macrocode}
\int_compare:nNnTF \c@iRow = 0
{ \int_compare:nNnT \c@jCol > 0 \l_@@_code_for_first_row_tl }
@@ -3465,15 +3772,17 @@ c & c & b & a & e
%
% \interitem
% The following macro |\@@_begin_of_row| is usually used in the cell
-% number~$1$ of the array. However, when the key |first-col| is used,
-% |\@@_begin_of_row| is executed in the cell number~$0$ of the array.
+% number~$1$ of the row. However, when the key |first-col| is used,
+% |\@@_begin_of_row| is executed in the cell number~$0$ of the row.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_begin_of_row:
+\cs_new_protected:Npn \@@_begin_of_row:
{
\int_gincr:N \c@iRow
\dim_gset_eq:NN \g_@@_dp_ante_last_row_dim \g_@@_dp_last_row_dim
\dim_gset:Nn \g_@@_dp_last_row_dim { \box_dp:N \@arstrutbox }
\dim_gset:Nn \g_@@_ht_last_row_dim { \box_ht:N \@arstrutbox }
+ \tikz [ remember~picture , baseline ] \coordinate
+ ( nm - \int_use:N \g_@@_env_int - row - \int_use:N \c@iRow - base ) ;
}
% \end{macrocode}
%
@@ -3481,31 +3790,31 @@ c & c & b & a & e
% \interitem
% The following code is used in each cell of the array. It actualises quantities
% that, at the end of the array, will give informations about the vertical
-% dimension of the two first rows and the two last rows.
+% dimension of the two first rows and the two last rows. If the user uses the
+% |last-row|, some lines will be dynamically added to this command.
% \begin{macrocode}
-\cs_new_protected:Npn \@@_actualization_for_first_and_last_row:
+\cs_new_protected:Npn \@@_update_for_first_and_last_row:
{
- \int_compare:nNnT \c@iRow = 0
+ \int_compare:nNnTF \c@iRow = 0
{
\dim_gset:Nn \g_@@_dp_row_zero_dim
- { \dim_max:nn \g_@@_dp_row_zero_dim { \box_dp:N \l_tmpa_box } }
+ { \dim_max:nn \g_@@_dp_row_zero_dim { \box_dp:N \l_@@_cell_box } }
\dim_gset:Nn \g_@@_ht_row_zero_dim
- { \dim_max:nn \g_@@_ht_row_zero_dim { \box_ht:N \l_tmpa_box } }
+ { \dim_max:nn \g_@@_ht_row_zero_dim { \box_ht:N \l_@@_cell_box } }
}
- \int_compare:nNnT \c@iRow = \c_one_int
- {
- \dim_gset:Nn \g_@@_ht_row_one_dim
- { \dim_max:nn \g_@@_ht_row_one_dim { \box_ht:N \l_tmpa_box } }
+ {
+ \int_compare:nNnT \c@iRow = 1
+ {
+ \dim_gset:Nn \g_@@_ht_row_one_dim
+ { \dim_max:nn \g_@@_ht_row_one_dim { \box_ht:N \l_@@_cell_box } }
+ }
}
- \dim_gset:Nn \g_@@_ht_last_row_dim
- { \dim_max:nn \g_@@_ht_last_row_dim { \box_ht:N \l_tmpa_box } }
- \dim_gset:Nn \g_@@_dp_last_row_dim
- { \dim_max:nn \g_@@_dp_last_row_dim { \box_dp:N \l_tmpa_box } }
}
% \end{macrocode}
-%
+%
+%
% \begin{macrocode}
-\cs_new_protected:Nn \@@_end_Cell:
+\cs_new_protected:Npn \@@_end_Cell:
{
\c_math_toggle_token
\hbox_set_end:
@@ -3517,12 +3826,12 @@ c & c & b & a & e
%
% \begin{macrocode}
\dim_gset:Nn \g_@@_max_cell_width_dim
- { \dim_max:nn \g_@@_max_cell_width_dim { \box_wd:N \l_tmpa_box } }
+ { \dim_max:nn \g_@@_max_cell_width_dim { \box_wd:N \l_@@_cell_box } }
% \end{macrocode}
%
% The following computations are for the ``first row'' and the ``last row''.
% \begin{macrocode}
- \@@_actualization_for_first_and_last_row:
+ \@@_update_for_first_and_last_row:
% \end{macrocode}
%
% Now, we can create the Tikz node of the cell.
@@ -3548,9 +3857,7 @@ c & c & b & a & e
\int_use:N \c@jCol
}
]
- \bgroup
- \box_use:N \l_tmpa_box
- \egroup ;
+ { \box_use_drop:N \l_@@_cell_box } ;
}
% \end{macrocode}
%
@@ -3561,19 +3868,18 @@ c & c & b & a & e
%
% \bigskip
% In the environment |{NiceArrayWithDelims}|, we will have to redefine the
-% column types |w| and |W|. These definitions are rather long because we have to
-% construct the |w|-nodes in these columns. The redefinition of these two column
-% types are very close and that's why we use a macro |\@@_renewcolumntype:nn|.
-% The first argument is the type of the column (|w| or |W|) and the second
-% argument is a code inserted at a special place and which is the only
-% difference between the two definitions.
-% \begin{macrocode}
-\cs_new_protected:Nn \@@_renewcolumntype:nn
+% column types |w| and |W|. The redefinition of these two column types are very
+% close and that's why we use a macro |\@@_renewcolumntype:nn|. The first
+% argument is the type of the column (|w| or |W|) and the second argument is a
+% code inserted at a special place and which is the only difference between the
+% two definitions.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_renewcolumntype:nn #1 #2
{
\newcolumntype #1 [ 2 ]
{
> {
- \hbox_set:Nw \l_tmpa_box
+ \hbox_set:Nw \l_@@_cell_box
\@@_Cell:
}
c
@@ -3581,23 +3887,7 @@ c & c & b & a & e
\@@_end_Cell:
\hbox_set_end:
#2
- \tikz [ remember~picture ]
- \node
- [
- name = nm - \int_use:N \g_@@_env_int -
- \int_use:N \c@iRow -
- \int_use:N \c@jCol - w,
- alias =
- \str_if_empty:NF \l_@@_name_str
- {
- \l_@@_name_str -
- \int_use:N \c@iRow -
- \int_use:N \c@jCol - w
- } ,
- inner~sep = \c_zero_dim ,
- ]
- { \makebox [ ##2 ] [ ##1 ] { \box_use:N \l_tmpa_box } }
- ;
+ \makebox [ ##2 ] [ ##1 ] { \box_use_drop:N \l_@@_cell_box }
}
}
}
@@ -3608,7 +3898,7 @@ c & c & b & a & e
% The argument of the following command |\@@_instruction_of_type:n| defined
% below is the type of the instruction (|Cdots|, |Vdots|, |Ddots|, etc.). This
% command writes in the corresponding |\g_@@_|\textsl{type}|_lines_tl| the
-% instruction which will really draw the line after the construction of the
+% instruction which will actually draw the line after the construction of the
% matrix.
%
% \medskip
@@ -3684,25 +3974,44 @@ c & c & b & a & e
}
\array
% \end{macrocode}
-% |\l_@@_pos_env_str| may have the value |t|, |c| or |b|.
+% |\l_@@_baseline_str| may have the value |t|, |c| or |b|. However, if the value
+% is |b|, we compose the |\array| (of \pkg{array}) with the option |t| and the
+% right translation will be done further.
% \begin{macrocode}
- [ \l_@@_pos_env_str ]
+ [ \str_if_eq:VnTF \l_@@_baseline_str c c t ]
}
% \end{macrocode}
%
-%
+% \medskip
+% We keep in memory the standard version of |\ialign| because we will redefine
+% |\ialign| in the environment |{NiceArrayWithDelims}| but restore the standard
+% version for use in the cells of the array.
+% \begin{macrocode}
+\cs_set_eq:NN \@@_standard_ialign: \ialign
+% \end{macrocode}
%
% \bigskip
% The following must \emph{not} be protected because it begins with |\noalign|.
% \begin{macrocode}
-\cs_new:Npn \@@_everycr:
- { \noalign { \@@_everycr_i: } }
+\cs_new:Npn \@@_everycr: { \noalign { \@@_everycr_i: } }
% \end{macrocode}
%
% \begin{macrocode}
\cs_new_protected:Npn \@@_everycr_i:
{
\int_gzero:N \c@jCol
+% \end{macrocode}
+% The |\hbox:n| is mandatory.
+% \begin{macrocode}
+ \hbox:n
+ {
+ \tikz [ remember~picture ]
+ \coordinate
+ ( nm - \int_use:N \g_@@_env_int - row - \int_eval:n { \c@iRow + 1 } ) ;
+ }
+% \end{macrocode}
+% We add the potential horizontal lines specified by the option |hlines|.
+% \begin{macrocode}
\bool_if:NT \l_@@_hlines_bool
{
% \end{macrocode}
@@ -3712,10 +4021,14 @@ c & c & b & a & e
% \begin{macrocode}
\int_compare:nNnT \c@iRow > { -1 }
{
- \int_compare:nNnF \c@iRow = \l_@@_last_row_int
- {
- \hrule \@height \arrayrulewidth
- \skip_vertical:n { - \arrayrulewidth }
+ \bool_if:NF \g_@@_row_of_col_done_bool
+ {
+ \int_compare:nNnF \c@iRow = \l_@@_last_row_int
+ {
+ \bool_if:NTF \c_@@_colortbl_loaded_bool
+ { { \CT@arc@ \hrule height \arrayrulewidth } }
+ { \hrule height \arrayrulewidth }
+ }
}
}
}
@@ -3730,6 +4043,7 @@ c & c & b & a & e
% \begin{macrocode}
\cs_new_protected:Npn \@@_pre_array:
{
+ \box_clear_new:N \l_@@_cell_box
\cs_if_exist:NT \theiRow
{ \int_set_eq:NN \l_@@_save_iRow_int \c@iRow }
\int_gzero_new:N \c@iRow
@@ -3804,18 +4118,13 @@ c & c & b & a & e
\dim_gzero_new:N \g_@@_dp_last_row_dim
\dim_gset:Nn \g_@@_dp_last_row_dim { \box_dp:N \@arstrutbox }
% \end{macrocode}
-% After its first utilisation, the definition of |\ialign| will revert
+% After its first use, the definition of |\ialign| will revert
% automatically to its default definition. With this programmation, we will
% have, in the cells of the array, a clean version of |\ialign|.\footnote{The
% user will probably not employ directly |\ialign| in the array... but more
% likely environments that utilize |\ialign| internally (e.g.: |{substack}|).}
% \begin{macrocode}
- \cs_set:Npn \ialign
- {
- \everycr { }
- \tabskip = \c_zero_skip
- \halign
- }
+ \cs_set_eq:NN \ialign \@@_standard_ialign:
\halign
}
% \end{macrocode}
@@ -3829,6 +4138,8 @@ c & c & b & a & e
% \end{macrocode}
%
% \begin{macrocode}
+ \cs_set_eq:NN \firsthline \hline
+ \cs_set_eq:NN \lasthline \hline
\cs_set_eq:NN \Ldots \@@_Ldots
\cs_set_eq:NN \Cdots \@@_Cdots
\cs_set_eq:NN \Vdots \@@_Vdots
@@ -3839,6 +4150,7 @@ c & c & b & a & e
\cs_set_eq:NN \Hdotsfor \@@_Hdotsfor:
\cs_set_eq:NN \multicolumn \@@_multicolumn:nnn
\cs_set_eq:NN \Block \@@_Block:
+ \cs_set_eq:NN \rotate \@@_rotate:
\cs_set_eq:NN \OnlyMainNiceMatrix \@@_OnlyMainNiceMatrix:n
\bool_if:NT \l_@@_renew_dots_bool
{
@@ -3919,18 +4231,18 @@ c & c & b & a & e
{
!
{
- \skip_horizontal:n { 0.53 pt }
% \end{macrocode}
-% If the array is an array with all the columns of the same width, we don't ask
-% for the creation of the extra nodes because we will use the ``|col|'' nodes
-% for the vertical dotted line.
+%
+% \medskip
+% The following code because we want the dotted line to have exactly the same
+% position as a vertical rule drawn by ``"|"'' (considering the rule having a
+% width equal to the diameter of the dots).
% \begin{macrocode}
- \bool_if:nF
+ \int_compare:nNnF \c@iRow = 0
{
- \l_@@_auto_columns_width_bool
- || \dim_compare_p:nNn \l_@@_columns_width_dim > \c_zero_dim
+ \int_compare:nNnF \c@iRow = \l_@@_last_row_int
+ { \skip_horizontal:n { 2 \l_@@_radius_dim } }
}
- { \bool_gset_true:N \g_@@_large_nodes_bool }
% \end{macrocode}
%
% \medskip
@@ -3986,24 +4298,24 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% \subsection{The environment \{NiceArrayWithDelims\}}
+% \subsection*{The environment \{NiceArrayWithDelims\}}
%
% \begin{macrocode}
\NewDocumentEnvironment { NiceArrayWithDelims } { m m O { } m ! O { } }
{
- \str_if_empty:NT \g_@@_type_env_str
- {
- \str_gset:Nn \g_@@_type_env_str
- { environment ~ { NiceArrayWithDelims } }
- }
+ \bool_gset_false:N \g_@@_row_of_col_done_bool
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { NiceArrayWithDelims } }
\@@_adapt_S_column:
\@@_test_if_math_mode:
\bool_if:NT \l_@@_in_env_bool { \@@_fatal:n { Yet~in~env } }
\bool_set_true:N \l_@@_in_env_bool
% \end{macrocode}
%
-% We deactivate Tikz externalization (since we use Tikz pictures with the
-% options |overlay| and |remember picture|, there would be errors).
+% We deactivate Tikz externalization.
+%
+% Since we use Tikz pictures with the options |overlay| and |remember picture|,
+% there would be errors).
% \begin{macrocode}
\cs_if_exist:NT \tikz@library@external@loaded
{
@@ -4038,13 +4350,24 @@ c & c & b & a & e
{ \keys_set:nn { NiceMatrix / pNiceArray } }
{ #3 , #5 }
% \end{macrocode}
-%
+%
+%
% A value of $-1$ for the counter |\l_@@_last_row_int| means that the user has
% used the option |last-row| without value, that is to say without specifying
% the number of that last row. In this case, we try to read that value from the
% |aux| file (if it has been written on a previous run).
%
% \begin{macrocode}
+ \int_compare:nNnT \l_@@_last_row_int > { -2 }
+ {
+ \tl_put_right:Nn \@@_update_for_first_and_last_row:
+ {
+ \dim_gset:Nn \g_@@_ht_last_row_dim
+ { \dim_max:nn \g_@@_ht_last_row_dim { \box_ht:N \l_@@_cell_box } }
+ \dim_gset:Nn \g_@@_dp_last_row_dim
+ { \dim_max:nn \g_@@_dp_last_row_dim { \box_dp:N \l_@@_cell_box } }
+ }
+ }
\int_compare:nNnT \l_@@_last_row_int = { -1 }
{
\bool_set_true:N \l_@@_last_row_without_value_bool
@@ -4070,8 +4393,7 @@ c & c & b & a & e
}
% \end{macrocode}
%
-% The code in |\@@_pre_array:| is used only by |{NiceArrayWithDelims}|. It
-% exists only for historical reasons. We should change that.
+% The code in |\@@_pre_array:| is used only by |{NiceArrayWithDelims}|.
% \begin{macrocode}
\@@_pre_array:
% \end{macrocode}
@@ -4079,32 +4401,30 @@ c & c & b & a & e
% \medskip
% We compute the width of the two delimiters.
% \begin{macrocode}
- \dim_gzero_new:N \g_@@_left_delim_dim
- \dim_gzero_new:N \g_@@_right_delim_dim
+ \dim_zero_new:N \l_@@_left_delim_dim
+ \dim_zero_new:N \l_@@_right_delim_dim
\bool_if:NTF \l_@@_NiceArray_bool
{
- \dim_gset:Nn \g_@@_left_delim_dim { 2 \arraycolsep }
- \dim_gset:Nn \g_@@_right_delim_dim { 2 \arraycolsep }
+ \dim_gset:Nn \l_@@_left_delim_dim { 2 \arraycolsep }
+ \dim_gset:Nn \l_@@_right_delim_dim { 2 \arraycolsep }
}
{
- \group_begin:
- \dim_set_eq:NN \nulldelimiterspace \c_zero_dim
\hbox_set:Nn \l_tmpa_box
{
\c_math_toggle_token
\left #1 \vcenter to 3 cm { } \right.
\c_math_toggle_token
}
- \dim_gset:Nn \g_@@_left_delim_dim { \box_wd:N \l_tmpa_box }
+ \dim_set:Nn \l_@@_left_delim_dim
+ { \box_wd:N \l_tmpa_box - \nulldelimiterspace }
\hbox_set:Nn \l_tmpa_box
{
- \dim_set_eq:NN \nulldelimiterspace \c_zero_dim
\c_math_toggle_token
\left. \vcenter to 3 cm { } \right #2
\c_math_toggle_token
}
- \dim_gset:Nn \g_@@_right_delim_dim { \box_wd:N \l_tmpa_box }
- \group_end:
+ \dim_set:Nn \l_@@_right_delim_dim
+ { \box_wd:N \l_tmpa_box - \nulldelimiterspace }
}
% \end{macrocode}
%
@@ -4122,27 +4442,44 @@ c & c & b & a & e
\int_compare:nNnTF \l_@@_first_col_int = 0
{ \tl_put_left:NV \l_tmpa_tl \c_@@_preamble_first_col_tl }
{
- \bool_if:NT \l_@@_NiceArray_bool
- {
- \bool_if:NF \l_@@_exterior_arraycolsep_bool
- { \tl_put_left:Nn \l_tmpa_tl { @ { } } }
+ \bool_if:nT
+ {
+ \l_@@_NiceArray_bool
+ && ! \l_@@_vlines_bool
+ && ! \l_@@_exterior_arraycolsep_bool
}
+ { \tl_put_left:Nn \l_tmpa_tl { @ { } } }
}
\int_compare:nNnTF \l_@@_last_col_int > { -1 }
{ \tl_put_right:NV \l_tmpa_tl \c_@@_preamble_last_col_tl }
{
- \bool_if:NT \l_@@_NiceArray_bool
- {
- \bool_if:NF \l_@@_exterior_arraycolsep_bool
- { \tl_put_right:Nn \l_tmpa_tl { @ { } } }
+ \bool_if:nT
+ {
+ \l_@@_NiceArray_bool
+ && ! \l_@@_vlines_bool
+ && ! \l_@@_exterior_arraycolsep_bool
}
+ { \tl_put_right:Nn \l_tmpa_tl { @ { } } }
}
+ \tl_put_right:Nn \l_tmpa_tl { > { \@@_error_too_much_cols: } l }
% \end{macrocode}
% Here is the beginning of the box which will contain the array. The
% |\hbox_set_end:| corresponding to this |\hbox_set:Nw| will be in the second
% part of the environment (and the closing |\c_math_toggle_token| also).
% \begin{macrocode}
\hbox_set:Nw \l_@@_the_array_box
+% \end{macrocode}
+% If the key |\vlines| is used, we increase |\arraycolsep| by
+% |0.5\arrayrulewidth| in order to reserve space for the width of the vertical
+% rules drawn with Tikz after the end of the array. However, the first
+% |\arraycolsep| is used once (between columns, |\arraycolsep| is used twice).
+% That's why we add a |0.5\arrayrulewidth| more.
+% \begin{macrocode}
+ \bool_if:NT \l_@@_vlines_bool
+ {
+ \dim_add:Nn \arraycolsep { 0.5 \arrayrulewidth }
+ \skip_horizontal:n { 0.5 \arrayrulewidth }
+ }
\skip_horizontal:n \l_@@_left_margin_dim
\skip_horizontal:n \l_@@_extra_left_margin_dim
\c_math_toggle_token
@@ -4161,13 +4498,22 @@ c & c & b & a & e
\c_math_toggle_token
\skip_horizontal:n \l_@@_right_margin_dim
\skip_horizontal:n \l_@@_extra_right_margin_dim
+% \end{macrocode}
+% If the key |\vlines| is used, we have increased |\arraycolsep| by
+% |0.5\arrayrulewidth| in order to reserve space for the width of the vertical
+% rules drawn with Tikz after the end of the array. However, the last
+% |\arraycolsep| is used once (between columns, |\arraycolsep| is used twice).
+% That's we add a |0.5 \arrayrulewidth| more.
+% \begin{macrocode}
+ \bool_if:NT \l_@@_vlines_bool { \skip_horizontal:n { 0.5 \arrayrulewidth } }
\hbox_set_end:
% \end{macrocode}
-% Now, the box containing the array is constructed (in |\l_@@_the_array_box|).
-% However, we have some computations to do before inserting that box in the TeX
-% flow (mainly because the exterior columns are in overlapping positions).
+% End of the construction of the array (in the box |\l_@@_the_array_box|).
%
% \bigskip
+% It the user has used the key |last-row| with a value, we control that the
+% given value is correct (since we have just contructed the array, we know the
+% real number of rows of the array).
% \begin{macrocode}
\int_compare:nNnT \l_@@_last_row_int > { -2 }
{
@@ -4181,6 +4527,25 @@ c & c & b & a & e
}
}
% \end{macrocode}
+%
+% Now, the definition of |\c@jCol| and
+% |\g_@@_col_total_int| change: |\c@jCol| will be the number of
+% columns without the ``last column''; |\g_@@_col_total_int| will be the
+% number of columns with this ``last column''.\footnote{We remind that the
+% potential ``first column'' has the number~$0$.}
+% \begin{macrocode}
+ \int_gset_eq:NN \c@jCol \g_@@_col_total_int
+ \bool_if:nT \g_@@_last_col_found_bool { \int_gdecr:N \c@jCol }
+% \end{macrocode}
+%
+% We fix also the value of |\c@iRow| and |\g_@@_row_total_int| with the
+% same principle.
+% \begin{macrocode}
+ \int_gset_eq:NN \g_@@_row_total_int \c@iRow
+ \int_compare:nNnT \l_@@_last_row_int > { -1 } { \int_gdecr:N \c@iRow }
+% \end{macrocode}
+%
+%
%
% \bigskip
% \textbf{Now, we begin the real construction in the output flow of TeX}. First, we take
@@ -4203,24 +4568,76 @@ c & c & b & a & e
% \begin{macrocode}
\bool_if:NTF \l_@@_NiceArray_bool
{
- \int_compare:nNnTF \l_@@_first_row_int = 0
+% \end{macrocode}
+% Remember that, when the key |b| is used, the |\array| (of \pkg{array}) is
+% constructed with the option |t| (and not |b|). Now, we do the translation to
+% take into account the option |b|.
+% \begin{macrocode}
+ \str_if_eq:VnTF \l_@@_baseline_str { b }
{
- \str_if_eq:VnT \l_@@_pos_env_str { t }
- {
- \box_move_up:nn { \g_@@_dp_row_zero_dim + \g_@@_ht_row_one_dim }
- }
+ \begin { tikzpicture }
+ \@@_extract_coords:w ( nm - \int_use:N \g_@@_env_int - row - 1 )
+ \dim_gset_eq:NN \g_tmpa_dim \pgf@y
+ \@@_extract_coords:w
+ (
+ nm - \int_use:N \g_@@_env_int -
+ row - \int_use:N \c@iRow - base
+ )
+ \dim_gsub:Nn \g_tmpa_dim \pgf@y
+ \end { tikzpicture }
+ \int_compare:nNnT \l_@@_first_row_int = 0
+ {
+ \dim_gadd:Nn \g_tmpa_dim
+ { \g_@@_ht_row_zero_dim + \g_@@_dp_row_zero_dim }
+ }
+ \box_move_up:nn \g_tmpa_dim { \box_use_drop:N \l_@@_the_array_box }
}
- {
- \int_compare:nNnT \l_@@_last_row_int > 0
+ {
+ \str_if_eq:VnTF \l_@@_baseline_str { c }
+ { \box_use_drop:N \l_@@_the_array_box }
{
- \str_if_eq:VnT \l_@@_pos_env_str { b }
+% \end{macrocode}
+% We convert a value of |t| to a value of |1|.
+% \begin{macrocode}
+ \str_if_eq:VnT \l_@@_baseline_str { t }
+ { \str_set:Nn \l_@@_baseline_str { 1 } }
+% \end{macrocode}
+% Now, we convert the value of |\l_@@_baseline_str| (which should represent an
+% integer) to an integer stored in |\l_tmpa_int|.
+% \begin{macrocode}
+ \int_set:Nn \l_tmpa_int \l_@@_baseline_str
+ \bool_if:nT
+ {
+ \int_compare_p:nNn \l_tmpa_int < 0
+ || \int_compare_p:nNn \l_tmpa_int > \g_@@_row_total_int
+ }
+ {
+ \@@_error:n { bad~value~for~baseline }
+ \int_set:Nn \l_tmpa_int 1
+ }
+% \end{macrocode}
+% We use a |{tikzpicture}| to extract coordinates (nothing is drawn).
+% \begin{macrocode}
+ \begin { tikzpicture }
+ \@@_extract_coords:w
+ ( nm - \int_use:N \g_@@_env_int - row - 1 )
+ \dim_gset_eq:NN \g_tmpa_dim \pgf@y
+ \@@_extract_coords:w
+ (
+ nm - \int_use:N \g_@@_env_int -
+ row - \int_use:N \l_tmpa_int- base
+ )
+ \dim_gsub:Nn \g_tmpa_dim \pgf@y
+ \end { tikzpicture }
+ \int_compare:nNnT \l_@@_first_row_int = 0
{
- \box_move_down:nn
- { \g_@@_ht_last_row_dim + \g_@@_dp_ante_last_row_dim }
+ \dim_gadd:Nn \g_tmpa_dim
+ { \g_@@_ht_row_zero_dim + \g_@@_dp_row_zero_dim }
}
+ \box_move_up:nn \g_tmpa_dim
+ { \box_use_drop:N \l_@@_the_array_box }
}
}
- { \box_use_drop:N \l_@@_the_array_box }
}
% \end{macrocode}
%
@@ -4236,6 +4653,7 @@ c & c & b & a & e
}
{ \dim_zero:N \l_tmpa_dim }
% \end{macrocode}
+%
% We compute |\l_tmpb_dim| which is the total height of the ``last row''
% below the array (when the key |last-row| is used). A value of $-2$ for
% |\l_@@_last_row_int| means that there is no ``last row''.\footnote{A value of
@@ -4278,25 +4696,96 @@ c & c & b & a & e
\right #2
\c_math_toggle_token
}
- \box_set_ht:Nn \l_tmpa_box { \box_ht:N \l_tmpa_box + \l_tmpa_dim }
- \box_set_dp:Nn \l_tmpa_box { \box_dp:N \l_tmpa_box + \l_tmpb_dim }
- \box_use_drop:N \l_tmpa_box
+% \end{macrocode}
+% Now, the box |\l_tmpa_box| is created with the correct delimiters.
+%
+% \smallskip
+% We will put the box in the TeX flow. However, we have a small work to do
+% when the option |max-delimiter-width| is used.
+% \begin{macrocode}
+ \bool_if:NTF \l_@@_max_delimiter_width_bool
+ { \@@_put_box_in_flow_bis:nn { #1 } { #2 } }
+ \@@_put_box_in_flow:
}
% \end{macrocode}
+%
% We take into account a potential ``last column'' (this ``last column'' has
% been constructed in an overlapping position and we have computed its width in
% |\g_@@_width_last_col_dim|: see p.~\pageref{overlap-right}).
% \begin{macrocode}
\bool_if:NT \g_@@_last_col_found_bool
- {
- \skip_horizontal:n \g_@@_width_last_col_dim
- \skip_horizontal:n \arraycolsep
- }
+ { \skip_horizontal:n { \g_@@_width_last_col_dim + \arraycolsep } }
\@@_after_array:
}
% \end{macrocode}
% This is the end of the environment |{NiceArrayWithDelims}|.
%
+% \bigskip
+% The command |\@@_put_box_in_flow:| puts the box |\l_tmpa_box| (which contains
+% the array) in the flow. It is used for the environments with delimiters.
+% First, we have to modify the height and the depth to take back into account
+% the potential exterior rows (the total height of the first row has been
+% computed in |\l_tmpa_dim| and the total height of the potential last row in
+% |\l_tmpb_dim|).
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_put_box_in_flow:
+ {
+ \box_set_ht:Nn \l_tmpa_box { \box_ht:N \l_tmpa_box + \l_tmpa_dim }
+ \box_set_dp:Nn \l_tmpa_box { \box_dp:N \l_tmpa_box + \l_tmpb_dim }
+ \box_use_drop:N \l_tmpa_box
+ }
+% \end{macrocode}
+%
+% \bigskip
+% The command |\@@_put_box_in_flow_bis:| is used when the option
+% |max-delimiter-width| is used because, in this case, we have to adjust the
+% widths of the delimiters. The arguments |#1| and |#2| are the delimiters
+% specified by the user.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_put_box_in_flow_bis:nn #1 #2
+ {
+% \end{macrocode}
+% We will compute the real width of both delimiters used.
+% \begin{macrocode}
+ \dim_zero_new:N \l_@@_real_left_delim_dim
+ \dim_zero_new:N \l_@@_real_right_delim_dim
+ \hbox_set:Nn \l_tmpb_box
+ {
+ \c_math_toggle_token
+ \left #1
+ \vcenter
+ {
+ \vbox_to_ht:nn
+ { \box_ht:N \l_tmpa_box + \box_dp:N \l_tmpa_box }
+ { }
+ }
+ \right .
+ \c_math_toggle_token
+ }
+ \dim_set:Nn \l_@@_real_left_delim_dim
+ { \box_wd:N \l_tmpb_box - \nulldelimiterspace }
+ \hbox_set:Nn \l_tmpb_box
+ {
+ \c_math_toggle_token
+ \left .
+ \vbox_to_ht:nn
+ { \box_ht:N \l_tmpa_box + \box_dp:N \l_tmpa_box }
+ { }
+ \right #2
+ \c_math_toggle_token
+ }
+ \dim_set:Nn \l_@@_real_right_delim_dim
+ { \box_wd:N \l_tmpb_box - \nulldelimiterspace }
+% \end{macrocode}
+%
+% Now, we can put the box in the TeX flow with the horizontal adjustments on
+% both sides.
+% \begin{macrocode}
+ \skip_horizontal:n { \l_@@_left_delim_dim - \l_@@_real_left_delim_dim }
+ \@@_put_box_in_flow:
+ \skip_horizontal:n { \l_@@_right_delim_dim - \l_@@_real_right_delim_dim }
+ }
+% \end{macrocode}
%
% \interitem
% The construction of the array in the environment |{NiceArrayWithDelims}| is,
@@ -4307,21 +4796,32 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { @@-normal-syntax } { }
% \end{macrocode}
+% First, we test whether the environment is empty. If it is empty, we raise a
+% fatal error (it's only a security). In order to detect whether it is empty, we
+% test whether the next token is |\end| and, if it's the case, we test if this
+% is the end of the environment (if it is not, an standard error will be raised
+% by LaTeX for incorrect nested environments).
+% \begin{macrocode}
+ {
+ \peek_meaning_ignore_spaces:NTF \end
+ { \@@_analyze_end:Nn }
+% \end{macrocode}
% Here is the call to |\array| (we have a dedicated macro |\@@_array:| because
% of compatibility with the classes \cls{revtex4-1} and \cls{revtex4-2}).
% \begin{macrocode}
- { \exp_args:NV \@@_array: \l_tmpa_tl }
+ { \exp_args:NV \@@_array: \l_tmpa_tl }
+ }
{
% \end{macrocode}
%
% If all the columns must have the same width (if the user has used the option
-% |columns-width| or the option |auto-column-width| of the environment
+% |columns-width| or the option |auto-columns-width| of the environment
% |{NiceMatrixBlock}|), we will add a row in the array to fix the width of the
% columns and construct the ``|col|'' nodes |nm-|\textsl{a}|-col-|\textsl{j}
% (these nodes will be used by the horizontal open dotted lines and by the
% commands |\@@_vdottedline:n|). We have written a dedicated function for that job.
% \begin{macrocode}
- \@@_create_row_of_col_nodes:
+ \@@_create_col_nodes:
\endarray
}
% \end{macrocode}
@@ -4334,6 +4834,12 @@ c & c & b & a & e
\NewDocumentEnvironment { @@-light-syntax } { b }
{
% \end{macrocode}
+% First, we test whether the environment is empty. It's only a security. Of
+% course, this test is more easy than the similar test for the ``normal syntax''
+% because we have the whole body of the environment in |#1|.
+% \begin{macrocode}
+ \tl_if_empty:nT { #1 } { \@@_fatal:n { empty~environment } }
+% \end{macrocode}
% Here is the call to |\array| (we have a dedicated macro |\@@_array:| because
% of compatibility with the classes \cls{revtex4-1} and \cls{revtex4-2}).
% \begin{macrocode}
@@ -4352,7 +4858,7 @@ c & c & b & a & e
\seq_gpop_left:NN \g_@@_rows_seq \l_tmpa_tl
\exp_args:NV \@@_line_with_light_syntax_i:n \l_tmpa_tl
\seq_map_function:NN \g_@@_rows_seq \@@_line_with_light_syntax:n
- \@@_create_row_of_col_nodes:
+ \@@_create_col_nodes:
\endarray
}
% \end{macrocode}
@@ -4387,102 +4893,106 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% The command |\@@_create_row_of_col_nodes:| will construct the potential last
-% row. That last row (when it is created) is a false row used to fix the width
-% of the columns (when the array is constructed with an option whhich specify
-% the width of the columns) and create the |col|-nodes (that nodes will be used
-% by |\vdottedline| for instance).
+% The following command is used by the code which detects whether the
+% environment is empty (we raise a fatal error in this case: it's only a
+% security).
% \begin{macrocode}
-\cs_new:Npn \@@_create_row_of_col_nodes:
+\cs_new_protected:Npn \@@_analyze_end:Nn #1 #2
{
- \bool_if:nT
- {
- \l_@@_auto_columns_width_bool
- || \dim_compare_p:nNn \l_@@_columns_width_dim > \c_zero_dim
- }
- {
- \crcr
- \int_compare:nNnT \l_@@_first_col_int = 0 { \omit & }
- \omit
+ \str_if_eq:VnT \g_@@_name_env_str { #2 }
+ { \@@_fatal:n { empty~environment } }
+% \end{macrocode}
+% We reput in the stream the |\end{...}| we have extracted and the user will
+% have an error for incorrect nested environments.
+% \begin{macrocode}
+ \end { #2 }
+ }
+% \end{macrocode}
+%
+% \bigskip
+% The command |\@@_create_col_nodes:| will construct a special last row.
+% That last row is a false row used to create the |col|-nodes and to fix the
+% width of the columns (when the array is constructed with an option which
+% specify the width of the columns).
+% \begin{macrocode}
+\cs_new:Npn \@@_create_col_nodes:
+ {
+ \crcr
+ \int_compare:nNnT \c@iRow = 0 { \@@_fatal:n { Zero~row } }
+ \int_compare:nNnT \l_@@_first_col_int = 0 { \omit & }
+ \omit
+% \end{macrocode}
+% The following instruction must be put after the instructions |\omit|.
+% \begin{macrocode}
+ \bool_gset_true:N \g_@@_row_of_col_done_bool
% \end{macrocode}
% First, we put a ``|col|'' node on the left of the first column (of course, we
% have to do that \emph{after} the |\omit|).
% \begin{macrocode}
- \skip_horizontal:N \arraycolsep
- \tikz [ remember~picture , overlay ]
- \coordinate [ name = nm - \int_use:N \g_@@_env_int - col - 0 ] ;
- \skip_horizontal:n { - \arraycolsep }
+ \tikz [ remember~picture ]
+ \coordinate ( nm - \int_use:N \g_@@_env_int - col - 1 ) ;
% \end{macrocode}
-% We compute in |\g_tmpa_dim| the common width of the columns. We use a global
-% variable because we are in a cell of an |\halign| and because we have to use this
-% variable in other cells (of the same row). The affectation of |\g_tmpa_dim|,
-% like all the affectations, must be done after the |\omit| of the cell.
+% We compute in |\g_tmpa_skip| the common width of the columns (it's a skip and
+% not a dimension). We use a global variable because we are in a cell of an
+% |\halign| and because we have to use this variable in other cells (of the same
+% row). The affectation of |\g_tmpa_skip|, like all the affectations, must be
+% done after the |\omit| of the cell.
% \begin{macrocode}
+ \bool_if:nTF
+ {
+ \l_@@_auto_columns_width_bool
+ || \dim_compare_p:nNn \l_@@_columns_width_dim > \c_zero_dim
+ }
+ {
\bool_if:nTF
{
\l_@@_auto_columns_width_bool
&& ! \l_@@_block_auto_columns_width_bool
}
{
- \dim_gset:Nn \g_tmpa_dim
+ \skip_gset:Nn \g_tmpa_skip
{ \g_@@_max_cell_width_dim + 2 \arraycolsep }
}
{
- \dim_gset:Nn \g_tmpa_dim
+ \skip_gset:Nn \g_tmpa_skip
{ \l_@@_columns_width_dim + 2 \arraycolsep }
- }
- \skip_horizontal:N \g_tmpa_dim
- \tikz [ remember~picture , overlay ]
- \coordinate [ name = nm - \int_use:N \g_@@_env_int - col - 1 ] ;
+ }
+ }
+ { \skip_gset:Nn \g_tmpa_skip { 0 pt~plus 1 fill } }
+ \skip_horizontal:N \g_tmpa_skip
+ \hbox:n
+ {
+ \tikz [ remember~picture ]
+ \coordinate ( nm - \int_use:N \g_@@_env_int - col - 2 ) ;
+ }
% \end{macrocode}
% We begin a loop over the columns. The integer |\g_tmpa_int| will be the
-% number of the current column. This integer is not used to fix the
-% width of the column (since all the columns have the same width equal to
-% |\g_@@_tmpa_dim|) but for the Tikz nodes.
-% \begin{macrocode}
- \int_gset:Nn \g_tmpa_int 1
- \bool_if:nTF \g_@@_last_col_found_bool
- { \prg_replicate:nn { \g_@@_col_total_int - 3 } }
- { \prg_replicate:nn { \g_@@_col_total_int - 2 } }
- {
- &
- \omit
+% number of the current column. This integer is used for the Tikz nodes.
+% \begin{macrocode}
+ \int_gset:Nn \g_tmpa_int 1
+ \bool_if:NTF \g_@@_last_col_found_bool
+ { \prg_replicate:nn { \g_@@_col_total_int - 2 } }
+ { \prg_replicate:nn { \g_@@_col_total_int - 1 } }
+ {
+ &
+ \omit
% \end{macrocode}
% The incrementation of the counter |\g_tmpa_int| must be done after the |\omit|
% of the cell.
% \begin{macrocode}
- \int_gincr:N \g_tmpa_int
- \skip_horizontal:N \g_tmpa_dim
-% \end{macrocode}
-% We create a ``|col|'' node on the right of the current column.
-% \begin{macrocode}
- \@@_create_col_node:
- }
+ \int_gincr:N \g_tmpa_int
+ \skip_horizontal:N \g_tmpa_skip
% \end{macrocode}
-% For the last column, we want a special treatment because of the final
-% |\arraycolsep|.
+% We create the ``|col|'' node on the right of the current column.
% \begin{macrocode}
- &
- \omit
- \int_gincr:N \g_tmpa_int
- \skip_horizontal:N \g_tmpa_dim
- \skip_horizontal:n { - \arraycolsep }
- \@@_create_col_node:
- \skip_horizontal:N \arraycolsep
+ \tikz [ remember~picture ]
+ \coordinate
+ (
+ nm - \int_use:N \g_@@_env_int -
+ col - \int_eval:n { \g_tmpa_int + 1 }
+ ) ;
}
- }
-% \end{macrocode}
-%
-% \bigskip
-% \begin{macrocode}
-\cs_new_protected:Npn \@@_create_col_node:
- {
- \tikz [ remember~picture , overlay ]
- \coordinate
- [
- name = nm - \int_use:N \g_@@_env_int -
- col - \int_use:N \g_tmpa_int
- ] ;
+ \cr
}
% \end{macrocode}
%
@@ -4497,10 +5007,10 @@ c & c & b & a & e
{
\@@_begin_of_row:
% \end{macrocode}
-% The contents of the cell is constructed in the box |\l_tmpa_box| because we
+% The contents of the cell is constructed in the box |\l_@@_cell_box| because we
% have to compute some dimensions of this box.
% \begin{macrocode}
- \hbox_set:Nw \l_tmpa_box
+ \hbox_set:Nw \l_@@_cell_box
\c_math_toggle_token
\bool_if:NT \l_@@_small_bool \scriptstyle
% \end{macrocode}
@@ -4519,12 +5029,16 @@ c & c & b & a & e
}
{ \l_@@_code_for_first_col_tl }
}
+% \end{macrocode}
+% Be careful: despite this letter |l| the cells of the ``first column'' are
+% composed in a |R| manner since they are composed in a |\hbox_overlap_left:n|.
+% \begin{macrocode}
l
<
{
\c_math_toggle_token
\hbox_set_end:
- \@@_actualization_for_first_and_last_row:
+ \@@_update_for_first_and_last_row:
% \end{macrocode}
% We actualise the width of the ``first column'' because we will use this width
% after the construction of the array.
@@ -4533,7 +5047,7 @@ c & c & b & a & e
{
\dim_max:nn
\g_@@_width_first_col_dim
- { \box_wd:N \l_tmpa_box }
+ { \box_wd:N \l_@@_cell_box }
}
% \end{macrocode}
% The content of the cell is inserted in an overlapping position.
@@ -4564,10 +5078,10 @@ c & c & b & a & e
0
}
]
- { \box_use:N \l_tmpa_box } ;
+ { \box_use_drop:N \l_@@_cell_box } ;
\skip_horizontal:n
{
- \g_@@_left_delim_dim +
+ \l_@@_left_delim_dim +
\l_@@_left_margin_dim +
\l_@@_extra_left_margin_dim
}
@@ -4597,7 +5111,7 @@ c & c & b & a & e
% The contents of the cell is constructed in the box |\l_tmpa_box| because we
% have to compute some dimensions of this box.
% \begin{macrocode}
- \hbox_set:Nw \l_tmpa_box
+ \hbox_set:Nw \l_@@_cell_box
\c_math_toggle_token
\bool_if:NT \l_@@_small_bool \scriptstyle
% \end{macrocode}
@@ -4621,7 +5135,7 @@ c & c & b & a & e
{
\c_math_toggle_token
\hbox_set_end:
- \@@_actualization_for_first_and_last_row:
+ \@@_update_for_first_and_last_row:
% \end{macrocode}
% We actualise the width of the ``last column'' because we will use this width
% after the construction of the array.
@@ -4630,7 +5144,7 @@ c & c & b & a & e
{
\dim_max:nn
\g_@@_width_last_col_dim
- { \box_wd:N \l_tmpa_box }
+ { \box_wd:N \l_@@_cell_box }
}
\skip_horizontal:n { - 2 \arraycolsep }
% \end{macrocode}
@@ -4641,7 +5155,7 @@ c & c & b & a & e
{
\skip_horizontal:n
{
- \g_@@_right_delim_dim +
+ \l_@@_right_delim_dim +
\l_@@_right_margin_dim +
\l_@@_extra_right_margin_dim
}
@@ -4668,7 +5182,7 @@ c & c & b & a & e
\int_use:N \c@jCol
}
]
- { \box_use:N \l_tmpa_box } ;
+ { \box_use_drop:N \l_@@_cell_box } ;
}
}
}
@@ -4684,8 +5198,8 @@ c & c & b & a & e
\NewDocumentEnvironment { NiceArray } { }
{
\bool_set_true:N \l_@@_NiceArray_bool
- \str_if_empty:NT \g_@@_type_env_str
- { \str_gset:Nn \g_@@_type_env_str { environment ~ { NiceArray } } }
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { NiceArray } }
% \end{macrocode}
% We put . and . for the delimiters but, in fact, that doesn't matter because
% these arguments won't be used in |{NiceArrayWithDelims}| (because the flag
@@ -4704,8 +5218,8 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { pNiceArray } { }
{
- \str_if_empty:NT \g_@@_type_env_str
- { \str_gset:Nn \g_@@_type_env_str { environment ~ { pNiceArray } } }
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { pNiceArray } }
\@@_test_if_math_mode:
\NiceArrayWithDelims ( )
}
@@ -4715,8 +5229,8 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { bNiceArray } { }
{
- \str_if_empty:NT \g_@@_type_env_str
- { \str_gset:Nn \g_@@_type_env_str { environment ~ { bNiceArray } } }
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { bNiceArray } }
\@@_test_if_math_mode:
\NiceArrayWithDelims [ ]
}
@@ -4726,8 +5240,8 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { BNiceArray } { }
{
- \str_if_empty:NT \g_@@_type_env_str
- { \str_gset:Nn \g_@@_type_env_str { environment ~ { BNiceArray } } }
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { BNiceArray } }
\@@_test_if_math_mode:
\NiceArrayWithDelims \{ \}
}
@@ -4737,8 +5251,8 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { vNiceArray } { }
{
- \str_if_empty:NT \g_@@_type_env_str
- { \str_gset:Nn \g_@@_type_env_str { environment ~ { vNiceArray } } }
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { vNiceArray } }
\@@_test_if_math_mode:
\NiceArrayWithDelims | |
}
@@ -4748,8 +5262,8 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { VNiceArray } { }
{
- \str_if_empty:NT \g_@@_type_env_str
- { \str_gset:Nn \g_@@_type_env_str { environment ~ { VNiceArray } } }
+ \str_if_empty:NT \g_@@_name_env_str
+ { \str_gset:Nn \g_@@_name_env_str { VNiceArray } }
\@@_test_if_math_mode:
\NiceArrayWithDelims \| \|
}
@@ -4757,14 +5271,14 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% \subsection{The environment \{NiceMatrix\} and its variants}
+% \subsection*{The environment \{NiceMatrix\} and its variants}
%
% \begin{macrocode}
\cs_new_protected:Npn \@@_define_env:n #1
{
\NewDocumentEnvironment { #1 NiceMatrix } { ! O { } }
{
- \str_gset:Nn \g_@@_type_env_str { environment ~ { #1 NiceMatrix } }
+ \str_gset:Nn \g_@@_name_env_str { #1 NiceMatrix }
\keys_set:nn { NiceMatrix / NiceMatrix } { ##1 }
\begin { #1 NiceArray }
{
@@ -4792,7 +5306,7 @@ c & c & b & a & e
%
%
% \bigskip
-% \subsection{How to know whether a cell is ``empty''}
+% \subsection*{How to know whether a cell is ``empty''}
%
% The conditionnal |\@@_if_not_empty_cell:nnT| tests whether a cell is empty.
% The first two arguments must be LaTeX3 counters for the row and the column of
@@ -4893,49 +5407,22 @@ c & c & b & a & e
% \end{macrocode}
%
%
-% \subsection{After the construction of the array}
-%
-%
-% \begin{macrocode}
-\cs_new_protected:Nn \@@_after_array:
- {
- \int_compare:nNnTF \c@iRow > 0
- \@@_after_array_i:
- {
- \@@_error:n { Zero~row }
- \@@_restore_iRow_jCol:
- }
- }
-% \end{macrocode}
+% \bigskip
+% \subsection*{After the construction of the array}
%
%
% \medskip
-% We deactivate Tikz externalization (since we use Tikz pictures with the
-% options |overlay| and |remember picture|, there would be errors).
+% We deactivate Tikz externalization.
+%
+% Since we use Tikz pictures with the options |overlay| and |remember picture|,
+% there would be errors.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_after_array_i:
+\cs_new_protected:Npn \@@_after_array:
{
\group_begin:
\cs_if_exist:NT \tikz@library@external@loaded
{ \tikzset { external / export = false } }
% \end{macrocode}
-% Now, the definition of |\c@jCol| and
-% |\g_@@_col_total_int| change: |\c@jCol| will be the number of
-% columns without the ``last column''; |\g_@@_col_total_int| will be the
-% number of columns with this ``last column''.\footnote{We remind that the
-% potential ``first column'' has the number~$0$.}
-% \begin{macrocode}
- \int_gset_eq:NN \c@jCol \g_@@_col_total_int
- \bool_if:nT \g_@@_last_col_found_bool { \int_gdecr:N \c@jCol }
-% \end{macrocode}
-%
-% We fix also the value of |\c@iRow| and |\g_@@_row_total_int| with the
-% same principle.
-% \begin{macrocode}
- \int_gset_eq:NN \g_@@_row_total_int \c@iRow
- \int_compare:nNnT \l_@@_last_row_int > { -1 }
- { \int_gsub:Nn \c@iRow \c_one_int }
-% \end{macrocode}
%
% In the user has used the option |last-row| without value, we write in the
% |aux| file the number of that last row for the next run.
@@ -4995,10 +5482,9 @@ c & c & b & a & e
% The booleans |\g_@@_medium_nodes_bool| and |\g_@@_large_nodes_bool| may be
% raised directly in cells of the array (for example in commands |\Block|) but
% also because the user has used the options |create-medium-nodes| and
-% |create-large-nodes| (these options raise the booleans
-% |\l_@@_medium_nodes_bool| and |\l_@@_large_nodes_bool| but theses booleans are
-% converted into the global version |\g_@@_medium_nodes_bool| and
-% |\g_@@_large_nodes_bool| before the creation of the array).
+% |create-large-nodes| (these options raise |\l_@@_medium_nodes_bool| and
+% |\l_@@_large_nodes_bool| but theses booleans are converted into the global
+% versions before the creation of the array).
% \begin{macrocode}
\bool_if:nTF \g_@@_medium_nodes_bool
{
@@ -5037,6 +5523,8 @@ c & c & b & a & e
\g_@@_Ldots_lines_tl
% \end{macrocode}
%
+%
+%
% \bigskip
% Now, the |code-after|.
% \begin{macrocode}
@@ -5049,17 +5537,71 @@ c & c & b & a & e
name~prefix = nm - \int_use:N \g_@@_env_int -
}
}
+ \bool_if:NT \l_@@_vlines_bool \@@_draw_vlines:
\cs_set_eq:NN \line \@@_line:nn
\g_@@_code_after_tl
\tl_gclear:N \g_@@_code_after_tl
\group_end:
- \str_gclear:N \g_@@_type_env_str
+ \str_gclear:N \g_@@_name_env_str
\@@_restore_iRow_jCol:
}
% \end{macrocode}
%
+% \bigskip
% \begin{macrocode}
-\cs_new_protected:Nn \@@_restore_iRow_jCol:
+\cs_new_protected:Npn \@@_draw_vlines:
+ {
+ \group_begin:
+% \end{macrocode}
+% The command |\CT@arc@| is a command of color from \pkg{colortbl}.
+% \begin{macrocode}
+ \bool_if:NT \c_@@_colortbl_loaded_bool \CT@arc@
+ \begin { tikzpicture } [ line~width = \arrayrulewidth ]
+% \end{macrocode}
+% First, we compute in |\l_tmpa_dim| the height of the rules we have to draw.
+% \begin{macrocode}
+ \@@_extract_coords:w ( row - 1 )
+ \dim_set_eq:NN \l_tmpa_dim \pgf@y
+ \@@_extract_coords:w ( row - \int_eval:n { \c@iRow + 1 } )
+ \dim_set:Nn \l_tmpa_dim { \l_tmpa_dim - \pgf@y }
+% \end{macrocode}
+% We translate vertically to take into account the potential ``last row''.
+% \begin{macrocode}
+ \dim_zero:N \l_tmpb_dim
+ \int_compare:nNnT \l_@@_last_row_int > { -1 }
+ {
+ \dim_set:Nn \l_tmpb_dim
+ { \g_@@_dp_last_row_dim + \g_@@_ht_last_row_dim }
+% \end{macrocode}
+% We adjust the value of |\l_tmpa_dim| by the width of the horizontal rule just
+% before the ``last row''.
+% \begin{macrocode}
+ \@@_extract_coords:w ( row - \int_eval:n { \c@iRow + 1 } )
+ \dim_add:Nn \l_tmpa_dim \pgf@y
+ \@@_extract_coords:w ( row - \int_eval:n { \g_@@_row_total_int + 1 } )
+ \dim_sub:Nn \l_tmpa_dim \pgf@y
+ \dim_sub:Nn \l_tmpa_dim \l_tmpb_dim
+ }
+% \end{macrocode}
+%
+% Now, we can draw the lines with a loop.
+% \begin{macrocode}
+ \int_step_inline:nnn
+ { \bool_if:NTF \l_@@_NiceArray_bool 1 2 }
+ { \bool_if:NTF \l_@@_NiceArray_bool { \c@jCol + 1 } \c@jCol }
+ {
+ \draw ( col - ##1 ) ++ ( 0 , \dim_use:N \l_tmpb_dim )
+ -- ++ ( 0 , \dim_use:N \l_tmpa_dim ) ;
+ }
+ \end { tikzpicture }
+ \group_end:
+ }
+% \end{macrocode}
+%
+%
+% \bigskip
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_restore_iRow_jCol:
{
\cs_if_exist:NT \theiRow { \int_gset_eq:NN \c@iRow \l_@@_save_iRow_int }
\cs_if_exist:NT \thejCol { \int_gset_eq:NN \c@jCol \l_@@_save_jCol_int }
@@ -5076,7 +5618,7 @@ c & c & b & a & e
% a & a+b & a+b+c
% \end{pNiceMatrix}\]
% For a closed extremity, we use the normal node and for a open one, we use the
-% ``medium node'' or, if it exists, the |w|-node.
+% ``medium node'' or the |col|-node.
% \[ \begin{pNiceMatrix}[create-medium-nodes,
% code-after = {\begin{tikzpicture}
% \node [highlight = (2-1)] {} ;
@@ -5114,7 +5656,7 @@ c & c & b & a & e
% \end{itemize}
%
% \begin{macrocode}
-\cs_new_protected:Nn \@@_find_extremities_of_line:nnnn
+\cs_new_protected:Npn \@@_find_extremities_of_line:nnnn #1 #2 #3 #4
{
% \end{macrocode}
% First, we declare the current cell as ``dotted'' because we forbide
@@ -5203,9 +5745,8 @@ c & c & b & a & e
% \end{macrocode}
% We test wether the initial extremity of the dotted line is an implicit cell
% already dotted (by another dotted line). In this case, we can't draw the line
-% because we have no Tikz node at the extremity of the arrow (and we can't use
-% the ``medium node'' or the ``large node'' because we should use the normal
-% node since the extremity is not open).
+% because we have no Tikz node at the extremity of the arrow.
+%
% \begin{macrocode}
\cs_if_free:cT
{
@@ -5275,11 +5816,9 @@ c & c & b & a & e
}
}
% \end{macrocode}
-% We test wether the initial extremity of the dotted line is an implicit cell
+% We test whether the initial extremity of the dotted line is an implicit cell
% already dotted (by another dotted line). In this case, we can't draw the line
-% because we have no Tikz node at the extremity of the arrow (and we can't use
-% the ``medium node'' or the ``large node'' because we should use the normal
-% node since the extremity is not open).
+% because we have no Tikz node at the extremity of the arrow.
% \begin{macrocode}
\cs_if_free:cT
{
@@ -5298,12 +5837,12 @@ c & c & b & a & e
}
% \end{macrocode}
% If we have at least one open extremity, we create the ``medium nodes'' in the
-% matrix\footnote{We should change this. Indeed, for an open extremity of an
-% \emph{horizontal} dotted line, we use the |w|-node, if, it exists, and not the
-% ``medium node''.}. We remind that, when used once, the command
-% |\@@_create_medium_nodes:| becomes no-op in the current TeX group.
+% matrix but we should change that because, for an open extremity on the left or
+% the right side of the array, we actually use the |col|-nodes.
+% We remind that, when used once, the command |\@@_create_medium_nodes:| becomes
+% no-op in the current TeX group.
% \begin{macrocode}
- \bool_if:nT \l_@@_initial_open_bool \@@_create_medium_nodes:
+ \bool_if:NT \l_@@_initial_open_bool \@@_create_medium_nodes:
\bool_if:NT \l_@@_final_open_bool \@@_create_medium_nodes:
}
% \end{macrocode}
@@ -5321,25 +5860,31 @@ c & c & b & a & e
% The two arguments of the command |\@@_retrieve_coords:nn| are the suffix and
% the anchor that must be used for the two nodes.
%
-% The coordinates are stored in |\g_@@_x_initial_dim|, |\g_@@_y_initial_dim|,
-% |\g_@@_x_final_dim|, |\g_@@_y_final_dim|. These variables are global for
-% technical reasons: we have to do an affectation in an environment
-% |{tikzpicture}|.
+% The coordinates are stored in four variables:
+% \begin{itemize}
+% \item |\g_@@_x_initial_dim|
+% \item |\g_@@_y_initial_dim|
+% \item |\g_@@_x_final_dim|
+% \item |\g_@@_y_final_dim|.
+% \end{itemize}
+%
+% These variables are global for technical reasons: we have to do an affectation
+% in an environment |{tikzpicture}|.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_retrieve_coords:nn
+\cs_new_protected:Npn \@@_retrieve_coords:nn #1 #2
{
\dim_gzero_new:N \g_@@_x_initial_dim
\dim_gzero_new:N \g_@@_y_initial_dim
\dim_gzero_new:N \g_@@_x_final_dim
\dim_gzero_new:N \g_@@_y_final_dim
\begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node \pgfutil@firstofone
+ \@@_extract_coords:w
( nm - \int_use:N \g_@@_env_int -
\int_use:N \l_@@_initial_i_int -
\int_use:N \l_@@_initial_j_int #1 )
\dim_gset:Nn \g_@@_x_initial_dim \pgf@x
\dim_gset:Nn \g_@@_y_initial_dim \pgf@y
- \tikz@parse@node \pgfutil@firstofone
+ \@@_extract_coords:w
( nm - \int_use:N \g_@@_env_int -
\int_use:N \l_@@_final_i_int -
\int_use:N \l_@@_final_j_int #2 )
@@ -5351,96 +5896,69 @@ c & c & b & a & e
% \end{macrocode}
%
% \interitem
-% For the horizontal lines with open extremities, we must take into account the
-% ``|col|'' nodes created in the environments which have a fixed width of the
-% columns. The following command will recompute the $x$-value of the extremities
-% in this case (erasing the value computed in |\@@_retrieve_coords:nn|).
+% For the horizontal lines with open extremities, we take into account the
+% ``|col|'' nodes. The following command will recompute the $x$-value of the
+% extremities in this case (erasing the value computed in
+% |\@@_retrieve_coords:nn|).
% \begin{macrocode}
-\cs_new_protected:Nn \@@_adjust_with_col_nodes:
+\cs_new_protected:Npn \@@_adjust_with_col_nodes:
{
\bool_if:NT \l_@@_initial_open_bool
{
- \cs_if_exist:cT
- { pgf@sh@ns@nm - \int_use:N \g_@@_env_int - col - 0 }
- {
- \begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node \pgfutil@firstofone
- ( nm - \int_use:N \g_@@_env_int - col - 0 )
- \dim_gset:Nn \g_@@_x_initial_dim \pgf@x
- \end { tikzpicture }
- }
- }
+ \begin { tikzpicture } [ remember~picture ]
+ \@@_extract_coords:w ( nm - \int_use:N \g_@@_env_int - col - 1 )
+ \dim_gset:Nn \g_@@_x_initial_dim \pgf@x
+ \end { tikzpicture }
+ }
\bool_if:NT \l_@@_final_open_bool
{
- \cs_if_exist:cT
- {
- pgf@sh@ns@nm - \int_use:N \g_@@_env_int - col -
- \int_use:N \c@jCol
- }
- {
- \begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node \pgfutil@firstofone
- ( nm - \int_use:N \g_@@_env_int - col - \int_use:N \c@jCol )
- \dim_gset:Nn \g_@@_x_final_dim \pgf@x
- \end { tikzpicture }
- }
- }
+ \begin { tikzpicture } [ remember~picture ]
+ \@@_extract_coords:w
+ ( nm - \int_use:N \g_@@_env_int - col - \int_eval:n { \c@jCol + 1 } )
+ \dim_gset:Nn \g_@@_x_final_dim \pgf@x
+ \end { tikzpicture }
+ }
}
% \end{macrocode}
%
% \interitem
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_Ldots:nn
+\cs_new_protected:Npn \@@_draw_Ldots:nn #1 #2
{
\cs_if_free:cT { @@ _ dotted _ #1 - #2 }
{
\bool_set_false:N \l_@@_impossible_line_bool
- \@@_find_extremities_of_line:nnnn { #1 } { #2 } 0 \c_one_int
+ \@@_find_extremities_of_line:nnnn { #1 } { #2 } 0 1
\bool_if:NF \l_@@_impossible_line_bool \@@_actually_draw_Ldots:
}
}
% \end{macrocode}
%
% \medskip
-% The command |\@@_actually_draw_Ldots:| draws the |Ldots| line using
-% |\l_@@_initial_i_int|, |\l_@@_initial_j_int|, |\l_@@_initial_open_bool|,
-% |\l_@@_final_i_int|, |\l_@@_final_j_int| and |\l_@@_final_open_bool|. We have
-% a dedicated command because if is used also by |\Hdotsfor|.
+% The command |\@@_actually_draw_Ldots:| draws the |Ldots| line using the
+% following variables:
+% \begin{itemize}
+% \item |\l_@@_initial_i_int|
+% \item |\l_@@_initial_j_int|
+% \item |\l_@@_initial_open_bool|
+% \item |\l_@@_final_i_int|
+% \item |\l_@@_final_j_int|
+% \item |\l_@@_final_open_bool|.
+% \end{itemize}
+%
+% We have a dedicated command because it is used also by |\Hdotsfor|.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_actually_draw_Ldots:
+\cs_new_protected:Npn \@@_actually_draw_Ldots:
{
\@@_retrieve_coords:xx
{
\bool_if:NTF \l_@@_initial_open_bool
- {
-% \end{macrocode}
-% If a |w|-node exists we use the |w|-node for the extremity.
-% \begin{macrocode}
- \cs_if_exist:cTF
- {
- pgf@sh@ns@nm
- - \int_use:N \g_@@_env_int
- - \int_use:N \l_@@_initial_i_int
- - \int_use:N \l_@@_initial_j_int - w
- }
- { - w.base~west }
- { - medium.base~west }
- }
- { .base~east }
+ { - medium.base~west }
+ { .base~east }
}
{
\bool_if:NTF \l_@@_final_open_bool
- {
- \cs_if_exist:cTF
- {
- pgf@sh@ns@nm
- - \int_use:N \g_@@_env_int
- - \int_use:N \l_@@_final_i_int
- - \int_use:N \l_@@_final_j_int - w
- }
- { - w.base~east }
- { - medium.base~east }
- }
+ { - medium.base~east }
{ .base~west }
}
\@@_adjust_with_col_nodes:
@@ -5460,43 +5978,23 @@ c & c & b & a & e
%
% \bigskip
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_Cdots:nn
+\cs_new_protected:Npn \@@_draw_Cdots:nn #1 #2
{
\cs_if_free:cT { @@ _ dotted _ #1 - #2 }
{
\bool_set_false:N \l_@@_impossible_line_bool
- \@@_find_extremities_of_line:nnnn { #1 } { #2 } 0 \c_one_int
+ \@@_find_extremities_of_line:nnnn { #1 } { #2 } 0 1
\bool_if:NF \l_@@_impossible_line_bool
{
\@@_retrieve_coords:xx
{
\bool_if:NTF \l_@@_initial_open_bool
- {
- \cs_if_exist:cTF
- {
- pgf@sh@ns@nm
- - \int_use:N \g_@@_env_int
- - \int_use:N \l_@@_initial_i_int
- - \int_use:N \l_@@_initial_j_int - w
- }
- { - w.mid~west }
- { - medium.mid~west }
- }
+ { - medium.mid~west }
{ .mid~east }
}
{
\bool_if:NTF \l_@@_final_open_bool
- {
- \cs_if_exist:cTF
- {
- pgf@sh@ns@nm
- - \int_use:N \g_@@_env_int
- - \int_use:N \l_@@_final_i_int
- - \int_use:N \l_@@_final_j_int - w
- }
- { - w.mid~east }
- { - medium.mid~east }
- }
+ { - medium.mid~east }
{ .mid~west }
}
\@@_adjust_with_col_nodes:
@@ -5516,12 +6014,12 @@ c & c & b & a & e
% command |\@@_retrieve_coords:nn| in the arguments |T| and |F| of the
% \pkg{expl3} commands (why?).
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_Vdots:nn
+\cs_new_protected:Npn \@@_draw_Vdots:nn #1 #2
{
\cs_if_free:cT { @@ _ dotted _ #1 - #2 }
{
\bool_set_false:N \l_@@_impossible_line_bool
- \@@_find_extremities_of_line:nnnn { #1 } { #2 } \c_one_int 0
+ \@@_find_extremities_of_line:nnnn { #1 } { #2 } 1 0
\bool_if:NF \l_@@_impossible_line_bool
{
\@@_retrieve_coords:xx
@@ -5578,12 +6076,12 @@ c & c & b & a & e
% default, we parallelize the diagonals lines. The first diagonal line is drawn
% and then, all the other diagonal lines are drawn parallel to the first one.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_Ddots:nn
+\cs_new_protected:Npn \@@_draw_Ddots:nn #1 #2
{
\cs_if_free:cT { @@ _ dotted _ #1 - #2 }
{
\bool_set_false:N \l_@@_impossible_line_bool
- \@@_find_extremities_of_line:nnnn { #1 } { #2 } \c_one_int \c_one_int
+ \@@_find_extremities_of_line:nnnn { #1 } { #2 } 1 1
\bool_if:NF \l_@@_impossible_line_bool
{
\@@_retrieve_coords:xx
@@ -5609,7 +6107,7 @@ c & c & b & a & e
% We test if the diagonal line is the first one (the counter |\l_@@_ddots_int|
% is created for this usage).
% \begin{macrocode}
- \int_compare:nNnTF \l_@@_ddots_int = \c_one_int
+ \int_compare:nNnTF \l_@@_ddots_int = 1
% \end{macrocode}
% If the diagonal line is the first one, we have no adjustment of the line to do
% but we store the $\Delta_x$ and the $\Delta_y$ of the line because these
@@ -5648,7 +6146,7 @@ c & c & b & a & e
% \bigskip
% We draw the |\Iddots| diagonals in the same way.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_Iddots:nn
+\cs_new_protected:Npn \@@_draw_Iddots:nn #1 #2
{
\cs_if_free:cT { @@ _ dotted _ #1 - #2 }
{
@@ -5670,7 +6168,7 @@ c & c & b & a & e
\bool_if:NT \l_@@_parallelize_diags_bool
{
\int_incr:N \l_@@_iddots_int
- \int_compare:nNnTF \l_@@_iddots_int = \c_one_int
+ \int_compare:nNnTF \l_@@_iddots_int = 1
{
\dim_set:Nn \l_@@_delta_x_two_dim
{ \g_@@_x_final_dim - \g_@@_x_initial_dim }
@@ -5705,7 +6203,7 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% \subsection{The actual instructions for drawing the dotted line with Tikz}
+% \subsection*{The actual instructions for drawing the dotted line with Tikz}
%
% The command |\@@_draw_tikz_line:| draws the line using four implicit
% arguments:
@@ -5715,7 +6213,7 @@ c & c & b & a & e
% first affectation was in an instruction |\tikz|.
%
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_tikz_line:
+\cs_new_protected:Npn \@@_draw_tikz_line:
{
% \end{macrocode}
% The dimension |\l_@@_l_dim| is the length $\ell$ of the line to draw. We use
@@ -5777,7 +6275,7 @@ c & c & b & a & e
\int_set:Nn \l_tmpa_int
{
\dim_ratio:nn
- { \l_@@_l_dim - ( \l_@@_dotted_lines_margin_dim * 2 ) }
+ { \l_@@_l_dim - ( 2 \l_@@_dotted_lines_margin_dim ) }
\l_@@_inter_dots_dim
}
}
@@ -5811,10 +6309,9 @@ c & c & b & a & e
{ \bool_if:NTF \l_@@_final_open_bool 2 1 }
}
% \end{macrocode}
-% In the loop over the dots (|\int_step_inline:nnnn|), the dimensions
-% |\g_@@_x_initial_dim| and |\g_@@_y_initial_dim| will be used for the
-% coordinates of the dots. But, before the loop, we must move until the
-% first dot.
+% In the loop over the dots, the dimensions |\g_@@_x_initial_dim| and
+% |\g_@@_y_initial_dim| will be used for the coordinates of the dots. But,
+% before the loop, we must move until the first dot.
%
% \begin{macrocode}
\dim_gadd:Nn \g_@@_x_initial_dim
@@ -5822,20 +6319,16 @@ c & c & b & a & e
( \g_@@_x_final_dim - \g_@@_x_initial_dim ) *
\dim_ratio:nn
{ \l_@@_l_dim - \l_@@_inter_dots_dim * \l_tmpa_int }
- { \l_@@_l_dim * 2 }
+ { 2 \l_@@_l_dim }
* \l_tmpb_int
}
-% \end{macrocode}
-% (In a multiplication of a dimension and an integer, the integer must always be
-% put in second position.)
-% \begin{macrocode}
\dim_gadd:Nn \g_@@_y_initial_dim
{
( \g_@@_y_final_dim - \g_@@_y_initial_dim ) *
\dim_ratio:nn
{ \l_@@_l_dim - \l_@@_inter_dots_dim * \l_tmpa_int }
- { \l_@@_l_dim * 2 } *
- \l_tmpb_int
+ { 2 \l_@@_l_dim }
+ * \l_tmpb_int
}
\begin { tikzpicture } [ overlay ]
\int_step_inline:nnn 0 \l_tmpa_int
@@ -5852,7 +6345,8 @@ c & c & b & a & e
}
% \end{macrocode}
%
-% \subsection{User commands available in the new environments}
+% \bigskip
+% \subsection*{User commands available in the new environments}
%
% We give new names for the commands |\ldots|, |\cdots|, |\vdots| and |\ddots|
% because these commands will be redefined (if the option |renew-dots| is used).
@@ -5865,12 +6359,10 @@ c & c & b & a & e
% \end{macrocode}
%
% \interitem
-% The command |\@@_add_to_empty_cells:| adds the current cell to
-% |\g_@@_empty_cells_seq| which is the list of the empty cells (the cells
-% explicitly declared ``empty'': there may be, of course, other empty cells in
-% the matrix).
+% The command |\@@_add_to_empty_cells:| declares the current cell as empty. For
+% efficiency, this is done by creating a special control sequence.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_add_to_empty_cells:
+\cs_new_protected:Npn \@@_add_to_empty_cells:
{
\cs_gset:cpx
{ @@ _ empty _ \int_use:N \c@iRow - \int_use:N \c@jCol }
@@ -5939,7 +6431,7 @@ c & c & b & a & e
% \bigskip
% The command |\@@_Hspace:| will be linked to |\hspace| in |{NiceArray}|.
% \begin{macrocode}
-\cs_new_protected:Nn \@@_Hspace:
+\cs_new_protected:Npn \@@_Hspace:
{
\@@_add_to_empty_cells:
\hspace
@@ -5967,14 +6459,15 @@ c & c & b & a & e
%
%
% \bigskip
-% The command |\@@_Hdotsfor| will be linked to |\Hdotsfor| in |{NiceArray}|.
-% This command uses an optional argument like |\hdotsfor| but this argument is
-% discarded (in |\hdotsfor|, this argument is used for fine tuning of the space
-% between two consecutive dots). Tikz nodes are created for all the cells of
-% the array, even the implicit cells of the |\Hdotsfor|.
+% The command |\@@_Hdotsfor| will be linked to |\Hdotsfor| in
+% |{NiceArrayWithDelims}|. This command uses an optional argument (as does
+% |\hdotsfor|) but this argument is discarded (in |\hdotsfor|, this argument is
+% used for fine tuning of the space between two consecutive dots). Tikz nodes
+% are created for all the cells of the array, even the implicit cells of the
+% |\Hdotsfor|.
%
% \medskip
-% This command must not be protected since it begins with |\multicolumn|.
+% This command must \emph{not} be protected since it begins with |\multicolumn|.
% \begin{macrocode}
\cs_new:Npn \@@_Hdotsfor:
{
@@ -5984,7 +6477,7 @@ c & c & b & a & e
% \end{macrocode}
%
% The command |\@@_Hdotsfor_i| is defined with the tools of \pkg{xparse} because
-% it has an optionnal argument. Note that such a command defined by
+% it has an optional argument. Note that such a command defined by
% |\NewDocumentCommand| is protected and that's why we have put the
% |\multicolumn| before (in the definition of |\@@_Hdotsfor:|).
% \begin{macrocode}
@@ -5998,10 +6491,10 @@ c & c & b & a & e
{
\tl_gput_right:Nx \g_@@_Hdotsfor_lines_tl
{
- \@@_draw_Hdotsfor:nnn
- { \int_use:N \c@iRow }
+ \@@_Hdotsfor:nnn
+ { \int_use:N \c@iRow }
{ \int_use:N \c@jCol }
- { #2 }
+ { #2 }
}
\prg_replicate:nn { #2 - 1 } { & \multicolumn { 1 } { C } { } }
}
@@ -6010,7 +6503,7 @@ c & c & b & a & e
%
% \medskip
% \begin{macrocode}
-\cs_new_protected:Nn \@@_draw_Hdotsfor:nnn
+\cs_new_protected:Npn \@@_Hdotsfor:nnn #1 #2 #3
{
\bool_set_false:N \l_@@_initial_open_bool
\bool_set_false:N \l_@@_final_open_bool
@@ -6032,7 +6525,7 @@ c & c & b & a & e
\@@_if_not_empty_cell:nnTF \l_@@_initial_i_int \l_tmpa_int
{ \int_set:Nn \l_@@_initial_j_int { #2 - 1 } }
{
- \int_set:Nn \l_@@_initial_j_int {#2}
+ \int_set:Nn \l_@@_initial_j_int { #2 }
\bool_set_true:N \l_@@_initial_open_bool
}
}
@@ -6052,8 +6545,23 @@ c & c & b & a & e
}
\bool_if:nT { \l_@@_initial_open_bool || \l_@@_final_open_bool }
\@@_create_medium_nodes:
- \@@_actually_draw_Ldots:
% \end{macrocode}
+%
+% \medskip
+% The command |\@@_adjust_with_col_nodes:| is used in the command
+% |\@@_actually_draw_Ldots:| in order to recompute the $x$-value of
+% the initial point and the $x$-value of the final point with the exact position
+% of the left side and the right side of the array (these informations are in
+% the |col|-node. However, we don't want this operation with the dotted lines
+% drawn by |\Hdotsfor|. That's why we deactivate locally this command.
+% \begin{macrocode}
+ \group_begin:
+ \cs_set:Npn \@@_adjust_with_col_nodes: { }
+ \@@_actually_draw_Ldots:
+ \group_end:
+% \end{macrocode}
+%
+% \medskip
% We declare all the cells concerned by the |\Hdotsfor| as ``dotted'' (for the
% dotted lines created by |\Cdots|, |\Ldots|, etc., this job is done by
% |\@@_find_extremities_of_line:nnnn|). This declaration is done by defining a
@@ -6065,11 +6573,28 @@ c & c & b & a & e
% \end{macrocode}
%
%
+% \bigskip
+% The control sequence |\@@_rotate:| will be linked to |\rotate| in
+% |{NiceArrayWithDelims}|.
+%
+% The command will exit three levels of groups in order
+% to execute the command
+%
+% \qquad ``|\box_rotate:Nn \l_@@_cell_box { 90 }|''
+%
+% just after the construction of the box |\l_@@_cell_box|.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_rotate: { \group_insert_after:N \@@_rotate_i: }
+\cs_new_protected:Npn \@@_rotate_i: { \group_insert_after:N \@@_rotate_ii: }
+\cs_new_protected:Npn \@@_rotate_ii: { \group_insert_after:N \@@_rotate_iii: }
+\cs_new_protected:Npn \@@_rotate_iii: { \box_rotate:Nn \l_@@_cell_box { 90 } }
+% \end{macrocode}
%
-% \subsection{The command \textbackslash line accessible in code-after}
+% \bigskip
+% \subsection*{The command \textbackslash line accessible in code-after}
%
% In the |code-after|, the command |\@@_line:nn| will be linked to |\line|. This
-% command takes two arguments which are the specification of two cells in the
+% command takes two arguments which are the specifications of two cells in the
% array (in the format $i$-$j$) and draws a dotted line between these cells.
%
% \medskip
@@ -6101,7 +6626,7 @@ c & c & b & a & e
% \end{macrocode}
%
% \begin{macrocode}
-\cs_new_protected:Nn \@@_line_i:nn
+\cs_new_protected:Npn \@@_line_i:nn #1 #2
{
\bool_if:NF \c_@@_draft_bool
{
@@ -6120,10 +6645,10 @@ c & c & b & a & e
{
\begin { tikzpicture }
\path~(#1)~--~(#2)~node[at~start]~(i)~{}~node[at~end]~(f)~{} ;
- \tikz@parse@node \pgfutil@firstofone ( i )
+ \@@_extract_coords:w ( i )
\dim_gset:Nn \g_@@_x_initial_dim \pgf@x
\dim_gset:Nn \g_@@_y_initial_dim \pgf@y
- \tikz@parse@node \pgfutil@firstofone ( f )
+ \@@_extract_coords:w ( f )
\dim_gset:Nn \g_@@_x_final_dim \pgf@x
\dim_gset:Nn \g_@@_y_final_dim \pgf@y
\end { tikzpicture }
@@ -6142,251 +6667,93 @@ c & c & b & a & e
% this command because they have to do other settings (for example, the diagonal
% lines must be parallelized).
%
-% \bigskip
-% \subsection{The commands to draw dotted lines to separate columns and rows}
%
-% The command |\hdottedline| draws an horizontal dotted line to separate two
-% rows. Similarly, the letter ``:'' in the preamble draws a vertical dotted line
-% (the letter can be changed with the option |letter-for-dotted-lines|). Both
-% mechanisms write instructions in the |code-after|. The actual instructions in
-% the |code-after| use the commands |\@@_hdottedline:n| and |\@@_vdottedline:n|.
-%
% \bigskip
-% We want the horizontal lines at the same position\footnote{In fact, almost the
-% same position because of the width of the line: the width of a dotted line is
-% not the same as the width of a line created by |\hline|.} as the line created
-% by |\hline| (or |\hdashline| of \pkg{arydshln}). That's why we use a
-% |\noalign| to insert a box with a |\dotfill|.
+% \subsection*{The vertical rules}
%
-%
-%
-% \medskip
-% Some extensions, like the extension \pkg{doc}, do a redefinition of the
-% command |\dotfill| of LaTeX. That's why we define a command |\@@_dotfill:| as
-% we wish. We test whether we are in draft mode because, in this case, we don't
-% draw the dotted lines.
-% \begin{macrocode}
-\bool_if:NTF \c_@@_draft_bool
- { \cs_set_eq:NN \@@_dotfill: \prg_do_nothing: }
- {
- \cs_set:Npn \@@_dotfill:
- {
-% \end{macrocode}
-% If the option |small| is used, we change the space between dots (we can't
-% use |\l_@@_inter_dots_dim| which will be set after the construction of the
-% array). We can't put the |\bool_if:NT| in the first argument of
-% |\hbox_to_wd:nn| because |\cleaders| is a special TeX primitive.
+% We give to the user the possibility to define new types of columns (with
+% |\newcolumntype| of \pkg{array}) for special vertical rules (\emph{e.g.} rules
+% thicker than the standard ones) which will not extend in the potential
+% exterior rows of the array.
+%
+% We provide the command |\OnlyMainNiceMatrix| in that goal. However, that
+% command must be no-op outside the environments of \pkg{nicematrix} (and so the
+% user will be allowed to use the same new type of column in the environments
+% of \pkg{nicematrix} and in the standard environments of \pkg{array}).
+%
+% That's why we provide first a global definition of |\OnlyMainNiceMatrix|.
% \begin{macrocode}
- \bool_if:NT \l_@@_small_bool
- { \dim_set:Nn \l_@@_inter_dots_dim { 0.25 em } }
- \cleaders
- \hbox_to_wd:nn
- { \l_@@_inter_dots_dim }
- {
- \c_math_toggle_token
- \bool_if:NT \l_@@_small_bool \scriptstyle
- \hss . \hss
- \c_math_toggle_token
- }
- \hfill
- }
- }
+\cs_set_eq:NN \OnlyMainNiceMatrix \use:n
% \end{macrocode}
-%
-% This command must \emph{not} be protected because it starts with |\noalign|.
+%
+% \medskip
+% Another definition of |\OnlyMainNiceMatrix| will be linked to the command in
+% the environments of \pkg{nicematrix}. Here is that definition, called
+% |\@@_OnlyMainNiceMatrix:n|.
+%
% \begin{macrocode}
-\cs_new:Npn \@@_hdottedline:
- {
- \noalign
+\cs_new_protected:Npn \@@_OnlyMainNiceMatrix:n #1
+ {
+ \int_compare:nNnTF \l_@@_first_col_int = 0
+ { \@@_OnlyMainNiceMatrix_i:n { #1 } }
{
- \bool_gset_true:N \g_@@_large_nodes_bool
- \cs_if_exist:cTF { @@_width_ \int_use:N \g_@@_env_int }
- { \dim_set_eq:Nc \l_tmpa_dim { @@_width_ \int_use:N \g_@@_env_int } }
- { \dim_set:Nn \l_tmpa_dim { 5 mm } }
- \hbox_overlap_right:n
+ \int_compare:nNnTF \c@jCol = 0
{
- \bool_if:nT
- {
- \l_@@_NiceArray_bool
- &&
- ! \l_@@_exterior_arraycolsep_bool
- &&
- \int_compare_p:nNn \l_@@_first_col_int > 0
- }
- { \skip_horizontal:n { - \arraycolsep } }
- \hbox_to_wd:nn
- {
- \l_tmpa_dim + 2 \arraycolsep
- - \l_@@_left_margin_dim - \l_@@_right_margin_dim
- }
- \@@_dotfill:
+ \int_compare:nNnF \c@iRow = { -1 }
+ { \int_compare:nNnF \c@iRow = { \l_@@_last_row_int - 1 } { #1 } }
}
- }
- }
+ { \@@_OnlyMainNiceMatrix_i:n { #1 } }
+ }
+ }
% \end{macrocode}
+% This definition may seem complicated by we must remind that the number of row
+% |\c@iRow| is incremented in the first cell of the row, \emph{after} an
+% potential vertical rule on the left side of the first cell.
%
-% \bigskip
-% \begin{macrocode}
-\cs_new_protected:Nn \@@_vdottedline:n
+% \smallskip
+% The command |\@@_OnlyMainNiceMatrix_i:n| is only a short-cut which is used
+% twice in the above command. This command must \emph{not} be protected.
+% % \begin{macrocode}
+\cs_new_protected:Npn \@@_OnlyMainNiceMatrix_i:n #1
{
-% \end{macrocode}
-% We should allow the letter ``|:|'' in the first position of the preamble
-% but that would need a special programmation.
-% \begin{macrocode}
- \int_compare:nNnTF #1 = 0
- { \@@_error:n { Use~of~:~in~first~position } }
- {
- \bool_if:NF \c_@@_draft_bool
- {
- \dim_zero_new:N \g_@@_x_initial_dim
- \dim_zero_new:N \g_@@_y_initial_dim
- \dim_zero_new:N \g_@@_x_final_dim
- \dim_zero_new:N \g_@@_y_final_dim
- \bool_set_true:N \l_@@_initial_open_bool
- \bool_set_true:N \l_@@_final_open_bool
-% \end{macrocode}
-% If a ``|col|'' node exists (if the array has been constructed with a fixed
-% width of column), we use it.
-% \begin{macrocode}
- \cs_if_exist:cTF
- { pgf@sh@ns@nm -\int_use:N \g_@@_env_int - col - #1 }
- {
- \begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node\pgfutil@firstofone
- ( col - #1 )
- \dim_gset:Nn \g_@@_x_initial_dim \pgf@x
- \dim_gset:Nn \g_@@_x_final_dim \pgf@x
- \dim_gset:Nn \g_@@_y_final_dim \pgf@y
- \end { tikzpicture }
- \dim_gset:Nn \g_@@_y_initial_dim { - \c_max_dim }
- \int_step_inline:nn \c@jCol
- {
- \begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node\pgfutil@firstofone
- ( 1 - ##1 . north~east )
- \dim_gset:Nn \g_@@_y_initial_dim
- { \dim_max:nn \g_@@_y_initial_dim \pgf@y }
- \end { tikzpicture }
- }
- }
-% \end{macrocode}
-% If not, we use the ``large node''.
-% \begin{macrocode}
- {
- \begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node\pgfutil@firstofone
- ( 1 - #1 - large .north~east )
- \dim_gset:Nn \g_@@_x_initial_dim \pgf@x
- \dim_gset:Nn \g_@@_y_initial_dim \pgf@y
- \tikz@parse@node\pgfutil@firstofone
- ( \int_use:N \c@iRow - #1 - large .south~east )
- \dim_gset:Nn \g_@@_x_final_dim \pgf@x
- \dim_gset:Nn \g_@@_y_final_dim \pgf@y
- \end { tikzpicture }
-% \end{macrocode}
-% However, if the previous column was constructed with a letter |w|, we use the
-% |w|-nodes (and we erase the previous computation of the $x$-value of the
-% vertical dotted line).
-% \begin{macrocode}
- \cs_if_exist:cT
- { pgf@sh@ns@nm -\int_use:N \g_@@_env_int - 1 - #1 - w }
- {
- \begin { tikzpicture } [ remember~picture ]
- \tikz@parse@node\pgfutil@firstofone
- ( 1 - #1 - w .north~east )
- \dim_gset:Nn \g_@@_x_initial_dim \pgf@x
- \tikz@parse@node\pgfutil@firstofone
- ( \int_use:N \c@iRow - #1 - w .south~east )
- \dim_gset:Nn \g_@@_x_final_dim \pgf@x
- \end { tikzpicture }
- \dim_gadd:Nn \g_@@_x_initial_dim \arraycolsep
- \dim_gadd:Nn \g_@@_x_final_dim \arraycolsep
- }
- }
- \@@_draw_tikz_line:
- }
- }
+ \int_compare:nNnF \c@iRow = 0
+ { \int_compare:nNnF \c@iRow = \l_@@_last_row_int { #1 } }
}
% \end{macrocode}
+% Remember that |\c@iRow| is not always inferior to |\l_@@_last_row_int| because
+% |\l_@@_last_row_int| may be equal to $-2$ or $-1$ (we can't write
+% |\int_compare:nNnT \c@iRow < \l_@@_last_row_int|).
%
+%
+% \medskip
+% In fact, independently of |\OnlyMainNiceMatrix|, which is a convenience given
+% to the user, we have to modify the behaviour of the standard specifier ``"|"''.
%
-% \bigskip
-% \subsection{The vertical rules}
-%
-% We don't want that a vertical rule drawn by the specifier ``"|"'' extends
-% in the eventual ``first row'' and ``last row'' of the array.
%
-% The natural way to do that would be to redefine the specifier ``"|"''
-% with |\newcolumntype|:
+% Remark first that the natural way to do that would be to redefine the
+% specifier ``"|"'' with |\newcolumntype|:
% \begin{Verbatim}[commandchars=\~\#\+]
-% \newcolumntype { | }
-% { ! { \int_compare:nNnF \c@iRow = 0 \vline } }
+% \newcolumntype { | } { ! { \OnlyMainNiceMatrix \vline } }
% \end{Verbatim}
%
% However, this code fails if the user uses "\DefineShortVerb{\|}" of
% \pkg{fancyvrb}. Moreover, it would not be able to deal correctly with two
% consecutive specifiers ``"|"'' (in a preambule like "ccc||ccc").
%
-% That's why we will do a redefinition of the macro |\@arrayrule| of \pkg{array}
-% and this redefinition will add |\@@_vline:| instead of |\vline| to the
-% preamble.
+% That's why we have done a redefinition of the macro |\@arrayrule| of
+% \pkg{array} and this redefinition will add |\@@_vline:| instead of |\vline| to
+% the preamble (that definition is in the beginning of |{NiceArrayWithDelims}|).
%
% Here is the definition of |\@@_vline:|. This definition \emph{must} be
% protected because you don't want that macro expanded during the construction
-% of the preamble (the tests must be effective in each row and not once when the
-% preamble is constructed).
-% \begin{macrocode}
-\cs_new_protected:Npn \@@_vline:
- {
- \int_compare:nNnTF \l_@@_first_col_int = 0
- {
- \int_compare:nNnTF \c@jCol = 0
- {
- \int_compare:nNnTF \l_@@_first_row_int = 0
- {
- \int_compare:nNnF \c@iRow = 0
- {
- \int_compare:nNnF \c@iRow = \l_@@_last_row_int
- \@@_vline_i:
- }
- }
- {
- \int_compare:nNnF \c@iRow = 0
- {
- \int_compare:nNnF \c@iRow = \l_@@_last_row_int
- \@@_vline_i:
- }
- }
- }
- {
- \int_compare:nNnF \c@iRow = 0
- {
- \int_compare:nNnF \c@iRow = \l_@@_last_row_int
- \@@_vline_i:
- }
- }
- }
- {
- \int_compare:nNnTF \c@jCol = 0
- {
- \int_compare:nNnF \c@iRow = { -1 }
- {
- \int_compare:nNnF \c@iRow = { \l_@@_last_row_int - 1 }
- \@@_vline_i:
- }
- }
- {
- \int_compare:nNnF \c@iRow = 0
- {
- \int_compare:nNnF \c@iRow = \l_@@_last_row_int
- \@@_vline_i:
- }
- }
- }
- }
-% \end{macrocode}
+% of the preamble (the tests in |\@@_OnlyMainNiceMatrix:n| must be effective in
+% each row and not once for all when the preamble is constructed).
%
% \medskip
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_vline: { \@@_OnlyMainNiceMatrix:n { \@@_vline_i: } }
+% \end{macrocode}
+%
% If \pkg{colortbl} is loaded, the following macro will be redefined (in a
% |\AtBeginDocument|) to take into account the color fixed by |\arrayrulecolor|
% of \pkg{colortbl}.
@@ -6395,35 +6762,129 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% We give now the definition of |\OnlyMainNiceMatrix|. Internally, it is not
-% used by \pkg{nicematrix}. It's only a facility given to the final user, which
-% may be useful in the definitions of new columns types (with |\newcolumntype|).
+% \subsection*{The commands to draw dotted lines to separate columns and rows}
%
-% \smallskip
-% First, we give the definition of |\OnlyMainNiceMatrix| in the general case:
-% it's no-op (thus, a definition of column type may be used outside the
-% environments of \pkg{nicematrix}, in |{array}|, etc.).
+% These commands don't use the normal nodes, the medium nor the large nodes.
+% They only use the |col|-nodes and the |row|-nodes.
+%
+% \textbf{Horizontal dotted lines}\par\nobreak
+%
+% \medskip
+% The following command must \emph{not} be protected because it's meant to be
+% expanded in a |\noalign|.
% \begin{macrocode}
-\cs_set_eq:NN \OnlyMainNiceMatrix \use:n
+\cs_new:Npn \@@_hdottedline:
+ {
+ \bool_if:NF \c_@@_draft_bool
+ {
+ \noalign{ \skip_vertical:n { 2 \l_@@_radius_dim } }
+ \@@_hdottedline_i:
+ }
+ }
% \end{macrocode}
%
% \medskip
-% Now, we give the definition of |\OnlyMainNiceMatrix| which will be used
-% in the environments of \pkg{nicematrix}. This command
-% |\@@_OnlyMainNiceMatrix:n| will be linked to |\OnlyMainNiceMatrix| in
-% |\@@_pre_array:|. This command is ``fully expandable'' and that's why we have
-% not protected it, even tough this characteristic will probably not be used.
+% On the other side, the following command should be protected.
% \begin{macrocode}
-\cs_new:Npn \@@_OnlyMainNiceMatrix:n #1
+\cs_new_protected:Npn \@@_hdottedline_i:
+ {
+% \end{macrocode}
+% We write in the code-after the instruction that will eventually draw the
+% dotted line. It's not possible to draw this dotted line now because we don't
+% known the length of the line (we don't even know the number of columns).
+% \begin{macrocode}
+ \tl_gput_right:Nx \g_@@_code_after_tl
+ { \@@_hdottedline:n { \int_use:N \c@iRow } }
+ }
+% \end{macrocode}
+%
+%
+% \medskip
+% The command |\@@_hdottedline:n| is the command written in the |code-after|
+% that will actually draw the dotted line. Its argument is the number of the row
+% \emph{before} which we will draw the row.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_hdottedline:n #1
{
- \int_compare:nNnF \c@iRow = 0
- { \int_compare:nNnF \c@iRow = \l_@@_last_row_int { #1 } }
+ \dim_zero_new:N \g_@@_x_initial_dim
+ \dim_zero_new:N \g_@@_y_initial_dim
+ \dim_zero_new:N \g_@@_x_final_dim
+ \dim_zero_new:N \g_@@_y_final_dim
+ \bool_set_true:N \l_@@_initial_open_bool
+ \bool_set_true:N \l_@@_final_open_bool
+ \begin { tikzpicture } [ remember~picture ]
+ \@@_extract_coords:w ( row - #1 )
+% \end{macrocode}
+% We do a translation par |-\l_@@_radius_dim| because we want the dotted line to
+% have exactly the same position as a vertical rule drawn by ``"|"''
+% (considering the rule having a width equal to the diameter of the dots).
+% \begin{macrocode}
+ \dim_gset:Nn \g_@@_y_initial_dim { \pgf@y - \l_@@_radius_dim }
+ \dim_gset:Nn \g_@@_y_final_dim { \pgf@y - \l_@@_radius_dim }
+ \@@_extract_coords:w ( col - 1 )
+ \dim_gset:Nn \g_@@_x_initial_dim \pgf@x
+ \@@_extract_coords:w ( col - \int_eval:n { \c@jCol + 1 } )
+ \dim_gset:Nn \g_@@_x_final_dim \pgf@x
+ \end { tikzpicture }
+ \@@_draw_tikz_line:
+ }
+% \end{macrocode}
+%
+%
+%
+% \bigskip
+% \textbf{Vertical dotted lines}\par\nobreak
+%
+% \medskip
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_vdottedline:n #1
+ {
+% \end{macrocode}
+% We should allow the letter ``|:|'' in the first position of the preamble
+% but that would need a special programmation.
+% \begin{macrocode}
+ \int_compare:nNnTF #1 = 0
+ { \@@_error:n { Use~of~:~in~first~position } }
+ { \bool_if:NF \c_@@_draft_bool { \@@_vdottedline_i:n { #1 } } }
+ }
+% \end{macrocode}
+%
+% \medskip
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_vdottedline_i:n #1
+ {
+ \dim_zero_new:N \g_@@_x_initial_dim
+ \dim_zero_new:N \g_@@_y_initial_dim
+ \dim_zero_new:N \g_@@_x_final_dim
+ \dim_zero_new:N \g_@@_y_final_dim
+ \bool_set_true:N \l_@@_initial_open_bool
+ \bool_set_true:N \l_@@_final_open_bool
+ \begin { tikzpicture } [ remember~picture ]
+ \@@_extract_coords:w ( col - \int_eval:n { #1 + 1 } )
+% \end{macrocode}
+% We do a translation par |-\l_@@_radius_dim| because we want the dotted line to
+% have exactly the same position as a vertical rule drawn by ``"|"''
+% (considering the rule having a width equal to the diameter of the dots).
+% \begin{macrocode}
+ \dim_gset:Nn \g_@@_x_initial_dim { \pgf@x - \l_@@_radius_dim }
+ \dim_gset:Nn \g_@@_x_final_dim { \pgf@x - \l_@@_radius_dim }
+ \@@_extract_coords:w ( row - 1 )
+% \end{macrocode}
+% We arbitrary decrease the height of the dotted line by a quantity equal to
+% |\l_@@_inter_dots_dim| in order to improve the visual impact.
+% \begin{macrocode}
+ \dim_gset:Nn \g_@@_y_initial_dim { \pgf@y - ( \l_@@_inter_dots_dim / 2 ) }
+ \@@_extract_coords:w ( row - \int_eval:n { \c@iRow + 1 } )
+ \dim_gset:Nn \g_@@_y_final_dim { \pgf@y + ( \l_@@_inter_dots_dim / 2 ) }
+ \end { tikzpicture }
+ \@@_draw_tikz_line:
}
% \end{macrocode}
+%
%
%
% \bigskip
-% \subsection{The environment \{NiceMatrixBlock\}}
+% \subsection*{The environment \{NiceMatrixBlock\}}
%
% The following flag will be raised when all the columns of the environments of
% the block must have the same width in ``auto'' mode.
@@ -6485,8 +6946,8 @@ c & c & b & a & e
}
% \end{macrocode}
%
-%
-% \subsection{The extra nodes}
+% \bigskip
+% \subsection*{The extra nodes}
%
% First, two variants of the functions |\dim_min:nn| and |\dim_max:nn|.
% \begin{macrocode}
@@ -6495,10 +6956,10 @@ c & c & b & a & e
% \end{macrocode}
%
% \bigskip
-% We have three macros of creation of nodes: |\@@_create_medium_nodes:|, |\@@_create_large_nodes:| and
-% |\@@_create_medium_and_large_nodes:|. They must \emph{not} be used in the
-% |code-after| because the |code-after| is executed in a scope of |prefix name|
-% of Tikz.
+% We have three macros of creation of nodes: |\@@_create_medium_nodes:|,
+% |\@@_create_large_nodes:| and |\@@_create_medium_and_large_nodes:|. They must
+% \emph{not} be used in the |code-after| because the |code-after| is executed in
+% a scope of |prefix name| of Tikz.
%
% \bigskip
% We have to compute the mathematical coordinates of the ``medium nodes''. These
@@ -6509,6 +6970,7 @@ c & c & b & a & e
% \bigskip
% The command |\@@_computations_for_medium_nodes:| must be used in a
% |{tikzpicture}|.
+%
% \medskip
% For each row $i$, we compute two dimensions
% \texttt{l_@@_row_\textsl{i}_min_dim} and \texttt{l_@@_row_\textsl{i}_max_dim}.
@@ -6555,14 +7017,15 @@ c & c & b & a & e
% implicit ampersands |&|). In this case, of course, we don't update the
% dimensions we want to compute.
% \begin{macrocode}
- { \cs_if_exist:cT
+ {
+ \cs_if_exist:cT
{ pgf@sh@ns@nm - \int_use:N \g_@@_env_int - \@@_i: - \@@_j: }
% \end{macrocode}
% We retrieve the coordinates of the anchor |south west| of the (normal) node of
% the cell ($i$-$j$). They will be stored in |\pgf@x| and |\pgf@y|.
% \begin{macrocode}
{
- \tikz@parse@node \pgfutil@firstofone
+ \@@_extract_coords:w
( nm - \int_use:N \g_@@_env_int
- \@@_i: - \@@_j: .south~west )
\dim_set:cn { l_@@_row_\@@_i: _min_dim}
@@ -6576,7 +7039,7 @@ c & c & b & a & e
% We retrieve the coordinates of the anchor |north east| of the (normal) node of
% the cell ($i$-$j$). They will be stored in |\pgf@x| and |\pgf@y|.
% \begin{macrocode}
- \tikz@parse@node \pgfutil@firstofone
+ \@@_extract_coords:w
( nm - \int_use:N \g_@@_env_int - \@@_i: - \@@_j: .north~east )
\dim_set:cn { l_@@_row _ \@@_i: _ max_dim }
{ \dim_max:vn { l_@@_row _ \@@_i: _ max_dim } \pgf@y }
@@ -6620,10 +7083,10 @@ c & c & b & a & e
% \medskip
% The command |\@@_create_large_nodes:| must be used when we want to create only
% the ``large nodes'' and not the medium ones (if we want to create both, we
-% have to use the command |\@@_create_medium_and_large_nodes:|). However, the
+% have to use |\@@_create_medium_and_large_nodes:|). However, the
% computation of the mathematical coordinates of the ``large nodes'' needs the
-% computation of the mathematical coordinates of the ``medium nodes''. That's
-% why we use first |\@@_computations_for_medium_nodes:| and then the command
+% computation of the mathematical coordinates of the ``medium nodes''. Hence,
+% we use first |\@@_computations_for_medium_nodes:| and then the command
% |\@@_computations_for_large_nodes:|.
% \begin{macrocode}
\cs_new_protected:Npn \@@_create_large_nodes:
@@ -6634,7 +7097,6 @@ c & c & b & a & e
\tikzset { name~suffix = -large }
\@@_create_nodes:
\end { tikzpicture }
- \@@_compute_width_of_array:
\cs_set_protected:Npn \@@_create_large_nodes: { }
\cs_set_protected:Npn \@@_create_medium_and_large_nodes:
{ \@@_create_medium_nodes: }
@@ -6656,9 +7118,7 @@ c & c & b & a & e
\@@_computations_for_large_nodes:
\tikzset { name~suffix = -large }
\@@_create_nodes:
- \@@_compute_width_of_array:
\end { tikzpicture }
- \@@_compute_width_of_array:
\cs_set_protected:Npn \@@_create_medium_and_large_nodes: { }
\cs_set_protected:Npn \@@_create_medium_nodes: { }
\cs_set_protected:Npn \@@_create_large_nodes: { }
@@ -6780,37 +7240,10 @@ c & c & b & a & e
}
% \end{macrocode}
%
-%
-%
-% \medskip
-% We can now compute the width of the array (used by |\hdottedline|). We should
-% modify this point because it's a waste to construct all the ``large nodes'' only
-% for computing the width of the array.
-% \begin{macrocode}
-\cs_new_protected:Npn \@@_compute_width_of_array:
- {
- \begin { tikzpicture } [ remember~picture , overlay ]
- \tikz@parse@node \pgfutil@firstofone
- ( nm - \int_use:N \g_@@_env_int - 1 - 1 - large .north~west )
- \dim_gset:Nn \g_tmpa_dim \pgf@x
- \tikz@parse@node \pgfutil@firstofone
- ( nm - \int_use:N \g_@@_env_int - 1 -
- \int_use:N \c@jCol - large .north~east )
- \dim_gset:Nn \g_tmpb_dim \pgf@x
- \end { tikzpicture }
- \iow_now:Nn \@mainaux \ExplSyntaxOn
- \iow_now:Nx \@mainaux
- {
- \cs_gset:cpn { @@_width_ \int_use:N \g_@@_env_int }
- { \dim_eval:n { \g_tmpb_dim - \g_tmpa_dim } }
- }
- \iow_now:Nn \@mainaux \ExplSyntaxOff
- }
-% \end{macrocode}
%
% \bigskip
% \begin{macrocode}
-\cs_new_protected:Npn \@@_extract_coords: #1 - #2 \q_stop
+\cs_new_protected:Npn \@@_extract_coords_values: #1 - #2 \q_stop
{
\cs_set:Npn \@@_i: { #1 }
\cs_set:Npn \@@_j: { #2 }
@@ -6822,9 +7255,9 @@ c & c & b & a & e
% issued in the format $i$|-|$j$ and the second is the value of~$n$ (the length
% of the ``multi-cell'').
% \begin{macrocode}
-\cs_new_protected:Nn \@@_node_for_multicolumn:nn
+\cs_new_protected:Npn \@@_node_for_multicolumn:nn #1 #2
{
- \@@_extract_coords: #1 \q_stop
+ \@@_extract_coords_values: #1 \q_stop
\coordinate ( @@~south~west ) at
(
\dim_use:c { l_@@_column _ \@@_j: _ min _ dim } ,
@@ -6849,58 +7282,40 @@ c & c & b & a & e
}
% \end{macrocode}
%
-%
-% \subsection{Block matrices}
+% \bigskip
+% \subsection*{Block matrices}
%
% The code in this section if for the construction of \emph{block matrices}. It
% has no direct link with the environment |{NiceMatrixBlock}|.
+%
%
% The following command will be linked to |\Block| in the environments of
% \pkg{nicematrix}. We define it with |\NewDocumentCommand| of \pkg{xparse}
-% because it has an optionnal argument between |<| and |>| (for TeX instructions
+% because it has an optional argument between |<| and |>| (for TeX instructions
% put before the math mode of the label)
%
% \begin{macrocode}
-\NewDocumentCommand \@@_Block: { m D < > { } m }
- { \@@_Block_i #1 \q_stop { #2 } { #3 } }
+\NewDocumentCommand \@@_Block: { O { } m D < > { } m }
+ { \@@_Block_i #2 \q_stop { #1 } { #3 } { #4 } }
% \end{macrocode}
%
% \medskip
-% The first argument of |\@@_Block:| (which is required) has a special syntax.
-% It must be of the form $i$|-|$j$ where $i$ and $j$ are the size (in rows and
-% columns) of the block.
+% The first mandatory argument of |\@@_Block:| has a special syntax. It must be
+% of the form $i$|-|$j$ where $i$ and $j$ are the size (in rows and columns) of
+% the block.
% \begin{macrocode}
-\cs_new:Npn \@@_Block_i #1-#2 \q_stop { \@@_Block_ii:nnnn { #1 } { #2 } }
+\cs_new:Npn \@@_Block_i #1-#2 \q_stop { \@@_Block_ii:nnnnn { #1 } { #2 } }
% \end{macrocode}
%
% \medskip
% Now, the arguments have been extracted: |#1| is $i$ (the number of rows of the
-% block), |#2| is $j$ (the number of columns of the block), |#3| are the tokens
-% to put before the math mode and |#4| is the label of the block. The following
-% command must \emph{not} be protected because it contains a command
-% |\multicolumn| (in the case of a block of only one row).
+% block), |#2| is $j$ (the number of columns of the block), |#3| is the list of
+% key-values, |#4| are the tokens to put before the math mode and |#5| is the
+% label of the block.
% \begin{macrocode}
-\cs_new:Npn \@@_Block_ii:nnnn #1 #2 #3 #4
+\cs_new_protected:Npn \@@_Block_ii:nnnnn #1 #2 #3 #4 #5
{
% \end{macrocode}
-%
-% \medskip
-% In the case of a block of only one row, we create a special node of shape
-% |coordinate| in order to remember the $y$-value of the baseline of the current
-% row.
-% \begin{macrocode}
- \int_compare:nNnT { #1 } = 1
- {
- \begin { tikzpicture } [ remember~picture , baseline ]
- \coordinate
- ( nm - \int_use:N \g_@@_env_int
- - Block
- - \int_use:N \c@iRow
- - \int_use:N \c@jCol ) ;
- \end { tikzpicture }
- }
- \bool_gset_true:N \g_@@_medium_nodes_bool
-% \end{macrocode}
% We write an instruction in the |code-after|. We write the instruction in the
% beginning of the |code-after| (the |left| in |\tl_gput_left:Nx|) because we
% want the Tikz nodes corresponding of the block created \emph{before} potential
@@ -6909,23 +7324,57 @@ c & c & b & a & e
% \begin{macrocode}
\tl_gput_left:Nx \g_@@_code_after_tl
{
- \@@_Block_iii:nnnnn
+ \@@_Block_iii:nnnnnn
{ \int_use:N \c@iRow }
{ \int_use:N \c@jCol }
{ \int_eval:n { \c@iRow + #1 - 1 } }
{ \int_eval:n { \c@jCol + #2 - 1 } }
- \exp_not:n { { #3 $ #4 $ } }
+ { #3 }
+ \exp_not:n { { #4 $ #5 $ } }
}
+% \end{macrocode}
+% It's not allowed to use the command |\Block| twice in the same cell of the
+% array. That's why, at the first use, we link the command |\Block|
+% to a special version. The scope of this link is the cell of the array.
+% \begin{macrocode}
+ \cs_set_eq:NN \Block \@@_Block_error:nn
}
% \end{macrocode}
%
+% \begin{macrocode}
+\cs_new:Npn \@@_Block_error:nn #1 #2
+ {
+ \@@_error:n { Second~Block }
+ \cs_set_eq:NN \Block \use:nn
+ }
+% \end{macrocode}
%
+% \medskip
+% \begin{macrocode}
+\keys_define:nn { NiceMatrix / Block }
+ {
+ tikz .tl_set:N = \l_@@_tikz_tl ,
+ tikz .value_required:n = true ,
+ white .bool_set:N = \l_@@_white_bool ,
+ white .default:n = true ,
+ white .value_forbidden:n = true ,
+ }
+% \end{macrocode}
+%
%
% \medskip
-% The following command |\@@_Block_iii:nnnnn| will be used in the |code-after|.
+% The following command |\@@_Block_iii:nnnnnn| will be used in the |code-after|.
% \begin{macrocode}
-\cs_new_protected:Npn \@@_Block_iii:nnnnn #1 #2 #3 #4 #5
+\cs_new_protected:Npn \@@_Block_iii:nnnnnn #1 #2 #3 #4 #5 #6
{
+% \end{macrocode}
+% The group is for the keys.
+% \begin{macrocode}
+ \group_begin:
+ \keys_set:nn { NiceMatrix / Block } { #5 }
+% \end{macrocode}
+%
+% \begin{macrocode}
\bool_if:nTF
{
\int_compare_p:nNn { #3 } > \c@iRow
@@ -6934,59 +7383,118 @@ c & c & b & a & e
{ \msg_error:nnnn { nicematrix } { Block~too~large } { #1 } { #2 } }
{
% \end{macrocode}
-% If the block has only one row, we have to do a special work in order to have
-% the contains of the node aligned with the contents of the other rows of the array.
-%
+% We put the contents of the cell in the box |\l_@@_cell_box| because we want
+% the command |\rotate| used in the content to be able to rotate the box.
% \begin{macrocode}
- \int_compare:nNnTF { #1 } = { #3 }
- {
- \begin { tikzpicture }
+ \hbox_set:Nn \l_@@_cell_box { #6 }
+% \end{macrocode}
+%
+% \medskip
+% The construction of the node corresponding to the merged cells. First, we
+% construct |\coordinate| corresponding to the upper-left and the lower-right
+% corners of the cell. They are called |one| and |two|.
+% \begin{macrocode}
+ \begin { tikzpicture }
+ \int_set:Nn \l_tmpa_int { #1 }
+ \int_decr:N \l_tmpa_int
+ \int_set:Nn \l_tmpb_int { #2 }
+ \int_decr:N \l_tmpb_int
+ \@@_extract_coords:w ( row - \int_eval:n { \l_tmpa_int + 1 } )
+ \dim_set_eq:NN \l_tmpa_dim \pgf@y
+ \@@_extract_coords:w ( col - \int_eval:n { \l_tmpb_int + 1 } )
+ \dim_set_eq:NN \l_tmpb_dim \pgf@x
+ \coordinate ( one ) at ( \l_tmpb_dim , \l_tmpa_dim ) ;
+ \@@_extract_coords:w ( row - \int_eval:n { #3 + 1 } )
+ \dim_set_eq:NN \l_tmpa_dim \pgf@y
+ \@@_extract_coords:w ( col - \int_eval:n { #4 + 1 } )
+ \dim_set_eq:NN \l_tmpb_dim \pgf@x
+ \coordinate ( two ) at ( \l_tmpb_dim , \l_tmpa_dim ) ;
+% \end{macrocode}
+%
+% \begin{macrocode}
+ \bool_if:NT \l_@@_white_bool
+ {
+ \fill [ white , line~width = 0 pt]
+ ( [ yshift = -\arrayrulewidth ] one )
+ rectangle
+ ( [ xshift = -\arrayrulewidth ] two ) ;
+ }
% \end{macrocode}
-% First, we compute in |\l_tmpa_dim| the $y$-value of the baseline of the row.
-% We have constructed a special node of shape |coordinate| in this order.
+%
+% Now, we can construct the real node fitting the nodes |one| and |two|.
+% We have to insert |\l_@@_tikz_tl| in the list of the options and that's why we
+% use |\use:x|.
% \begin{macrocode}
- \tikz@parse@node \pgfutil@firstofone (Block-#1-#2)
- \dim_set:Nn \l_tmpa_dim \pgf@y
- \node
+ \use:x
+ {
+ \exp_not:N \node
[
- fit = ( #1 - #2 - medium . north~west )
- ( #3 - #4 - medium . south~east ) ,
+ fit = ( one ) ( two ) ,
inner~sep = 0 pt ,
+ \l_@@_tikz_tl
]
- (#1-#2) { } ;
+ }
% \end{macrocode}
-% With the following instruction, we retrieve the $x$-value and the $y$-value of
-% the center of the block. We will only use the $x$-value, available in |\pgf@x|.
+% The node of the block has a name |(#1-#2-block)| and we will use it internally just
+% after to put the label of the block.
% \begin{macrocode}
- \tikz@parse@node \pgfutil@firstofone (#1-#2)
- \path (\pgf@x,\l_tmpa_dim) node [ anchor = base ] { #5 } ;
- \end { tikzpicture }
- }
+ (#1-#2-block) { } ;
% \end{macrocode}
-% If the number of rows is different of $1$, it's necessary to create two Tikz
-% nodes because we want the label |#5| really drawn in the \emph{center} of the node.
+%
+% \medskip
+% If the creation of the ``medium nodes'' is required, we create a ``medium
+% node'' for the block.
% \begin{macrocode}
+ \bool_if:NT \l_@@_medium_nodes_bool
{
- \begin { tikzpicture }
- \node
- [
- fit = ( #1 - #2 - medium . north~west )
+ \node
+ [
+ fit = ( #1 - #2 - medium . north~west )
( #3 - #4 - medium . south~east ) ,
inner~sep = 0 pt ,
- ]
+ ]
+ (#1-#2-block-medium) { } ;
+ }
% \end{macrocode}
-% We don't forget the name of the node because the user may wish to use it.
+% \medskip
+% Now, we will put the label of the block.
% \begin{macrocode}
- (#1-#2) { } ;
- \node at (#1-#2.center) { #5 } ;
- \end { tikzpicture }
- }
+ \int_compare:nNnTF { #1 } = { #3 }
+ {
+% \end{macrocode}
+% If the block has only one row, we want the label of the block perfectly
+% aligned on the baseline of the row. That's why we have constructed a
+% |\coordinate| on the baseline of the row, in the first column of the array.
+% Now, we retrieve the $y$-value of that node and we store it in |\l_tmpa_dim|.
+% \begin{macrocode}
+ \@@_extract_coords:w (row-#1-base)
+ \dim_set_eq:NN \l_tmpa_dim \pgf@y
+% \end{macrocode}
+% We retrieve (in |\pgf@x|) the $x$-value of the center of the block.
+% \begin{macrocode}
+ \@@_extract_coords:w (#1-#2-block)
+% \end{macrocode}
+% We put the label of the block which has been composed in |\l_@@_cell_box|.
+% \begin{macrocode}
+ \path (\pgf@x,\l_tmpa_dim) node [ anchor = base ]
+ { \box_use_drop:N \l_@@_cell_box } ;
+ }
+% \end{macrocode}
+%
+% \medskip
+% If the number of rows is different of $1$, we put the label of the block in
+% the center of the node (the label of the block has been composed in
+% |\l_@@_cell_box|).
+% \begin{macrocode}
+ { \node at (#1-#2-block.center) { \box_use_drop:N \l_@@_cell_box } ; }
+ \end { tikzpicture }
}
+ \group_end:
}
% \end{macrocode}
-%
%
-% \subsection{How to draw the dotted lines transparently}
+% \bigskip
+% \subsection*{How to draw the dotted lines transparently}
% \begin{macrocode}
\cs_set_protected:Npn \@@_renew_matrix:
{
@@ -7009,8 +7517,8 @@ c & c & b & a & e
% \end{macrocode}
%
%
-% \medskip
-% \subsection{Automatic arrays}
+% \bigskip
+% \subsection*{Automatic arrays}
%
%
% \begin{macrocode}
@@ -7063,12 +7571,12 @@ c & c & b & a & e
{
\cs_set_protected:cpn { #1 AutoNiceMatrix }
{
- \str_gset:Nx \g_@@_type_env_str
- { command ~ \c_backslash_str #1 AutoNiceMatrix }
+ \str_gset:Nx \g_@@_name_env_str { #1 AutoNiceMatrix }
\AutoNiceMatrixWithDelims { #2 } { #3 }
}
}
% \end{macrocode}
+%
% \begin{macrocode}
\@@_define_com:nnn p ( )
\@@_define_com:nnn b [ ]
@@ -7077,8 +7585,8 @@ c & c & b & a & e
\@@_define_com:nnn B \{ \}
% \end{macrocode}
%
-% \medskip
-% \subsection{We process the options}
+% \bigskip
+% \subsection*{We process the options}
%
% We process the options when the package is loaded (with |\usepackage|) but we
% recommend to use |\NiceMatrixOptions| instead.
@@ -7104,9 +7612,113 @@ c & c & b & a & e
\ProcessKeysOptions { NiceMatrix / Package }
% \end{macrocode}
%
-% \subsection{Error messages of the package}
+% \bigskip
+% \subsection*{Error messages of the package}
+%
+% The following command converts all the elements of a sequence (which are token
+% lists) into strings.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_convert_to_str_seq:N #1
+ {
+ \seq_clear:N \l_tmpa_seq
+ \seq_map_inline:Nn #1
+ {
+ \seq_put_left:Nx \l_tmpa_seq { \tl_to_str:n { ##1 } }
+ }
+ \seq_set_eq:NN #1 \l_tmpa_seq
+ }
+% \end{macrocode}
+%
+% The following command creates a sequence of strings (|str|) from a |clist|.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_set_seq_of_str_from_clist:Nn #1 #2
+ {
+ \seq_set_from_clist:Nn #1 { #2 }
+ \@@_convert_to_str_seq:N #1
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+\@@_set_seq_of_str_from_clist:Nn \c_@@_types_of_matrix_seq
+ {
+ NiceMatrix ,
+ pNiceMatrix , bNiceMatrix , vNiceMatrix, BNiceMatrix, VNiceMatrix
+ }
+% \end{macrocode}
+%
+% \bigskip
+% If the user uses too much columns, the command |\@@_error_too_much_cols:| is
+% executed. This command raises an error but try to give the best information to
+% the user in the error message. The command |\seq_if_in:NVTF| is not expandable
+% and that's why we can't put it in the error message itself. We have to do the
+% test before the |\@@_fatal:n|.
+% \begin{macrocode}
+\cs_new_protected:Npn \@@_error_too_much_cols:
+ {
+ \seq_if_in:NVTF \c_@@_types_of_matrix_seq \g_@@_name_env_str
+ {
+ \int_compare:nNnTF \l_@@_last_col_int = { -1 }
+ { \@@_fatal:n { too~much~cols~for~matrix } }
+ { \@@_fatal:n { too~much~cols~for~matrix~with~last~col } }
+ }
+ { \@@_fatal:n { too~much~cols~for~array } }
+ }
+% \end{macrocode}
+%
+%
+% \begin{macrocode}
+\@@_msg_new:nn { too~much~cols~for~matrix~with~last~col }
+ {
+ You~try~to~use~more~columns~than~allowed~by~your~
+ \@@_full_name_env:.~The~maximal~number~of~columns~is~
+ \int_eval:n { \l_@@_last_col_int - 1 }~(plus~the~potential~
+ exterior~ones).~This~error~is~fatal.
+ }
+% \end{macrocode}
+%
+%
+% \begin{macrocode}
+\@@_msg_new:nn { too~much~cols~for~matrix }
+ {
+ You~try~to~use~more~columns~than~allowed~by~your~
+ \@@_full_name_env:.~ Recall~that~the~maximal~number~of~columns~
+ for~a~matrix~is~fixed~by~the~LaTeX~counter~'MaxMatrixCols'.~
+ Its~actual~value~is~\int_use:N \c@MaxMatrixCols.~This~error~is~fatal.
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+\@@_msg_new:nn { too~much~cols~for~array }
+ {
+ You~try~to~use~more~columns~than~allowed~by~your~
+ \@@_full_name_env:.~The~maximal~number~of~columns~is~
+ \int_use:N \c@jCol\space (plus~the~potential~exterior~ones).~
+ This~error~is~fatal.
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+\@@_msg_new:nn { bad~value~for~baseline }
+ {
+ The~value~you~gave~to~'baseline'~(\int_use:N \l_tmpa_int)~is~not~
+ valid.\\
+ If~you~go~on,~a~value~of~1~will~be~used.
+ }
+% \end{macrocode}
%
%
+% \begin{macrocode}
+\@@_msg_new:nn { Second~Block }
+ {
+ You~can't~use~\token_to_str:N \Block\ twice~in~the~same~cell~of~the~array.\\
+ If~you~go~on,~this~command~(and~the~other)~will~be~ignored.
+ }
+% \end{macrocode}
+%
+% \begin{macrocode}
+\@@_msg_new:nn { empty~environment }
+ { Your~\@@_full_name_env:\ is~empty.~This~error~is~fatal. }
+% \end{macrocode}
%
% \begin{macrocode}
\@@_msg_new:nn { unknown~cell~for~line~in~code-after }
@@ -7121,7 +7733,7 @@ c & c & b & a & e
% \begin{macrocode}
\@@_msg_new:nn { last-col~non~empty~for~NiceArray }
{
- In~the~\g_@@_type_env_str,~you~must~use~the~option~
+ In~the~\@@_full_name_env:,~you~must~use~the~option~
'last-col'~without~value.\\
However,~you~can~go~on~for~this~time~
(the~value~'\l_keys_value_tl'~will~be~ignored).
@@ -7131,7 +7743,7 @@ c & c & b & a & e
% \begin{macrocode}
\@@_msg_new:nn { last-col~empty~for~NiceMatrix }
{
- In~the~\g_@@_type_env_str, you~can't~use~the~option~
+ In~the~\@@_full_name_env:, you~can't~use~the~option~
'last-col'~without~value.~You~must~give~the~number~of~that~last~column.\\
If~you~go~on~this~option~will~be~ignored.
}
@@ -7141,7 +7753,7 @@ c & c & b & a & e
\@@_msg_new:nn { Block~too~large }
{
You~try~to~draw~a~block~in~the~cell~#1-#2~of~your~matrix~but~the~matrix~is~
- too~small~for~that~block.\\
+ too~small~for~that~block. \\
If~you~go~on,~this~command~will~be~ignored.
}
% \end{macrocode}
@@ -7160,7 +7772,7 @@ c & c & b & a & e
\@@_msg_new:nn { Wrong~last~row }
{
You~have~used~'last-row=\int_use:N \l_@@_last_row_int'~but~your~
- \g_@@_type_env_str\ seems~to~have~\int_use:N \c@iRow \ rows.~
+ \@@_full_name_env:\ seems~to~have~\int_use:N \c@iRow \ rows.~
If~you~go~on,~the~value~of~\int_use:N \c@iRow \ will~be~used~for~
last~row.~You~can~avoid~this~problem~by~using~'last-row'~
without~value~(more~compilations~might~be~necessary).
@@ -7179,32 +7791,16 @@ c & c & b & a & e
% \begin{macrocode}
\@@_msg_new:nn { Outside~math~mode }
{
- The~\g_@@_type_env_str\ can~be~used~only~in~math~mode~
+ The~\@@_full_name_env:\ can~be~used~only~in~math~mode~
(and~not~in~\token_to_str:N \vcenter).\\
This~error~is~fatal.
}
% \end{macrocode}
%
% \begin{macrocode}
-\@@_msg_new:nn { Option~Transparent~suppressed }
- {
- The~option~'Transparent'~has~been~renamed~'transparent'.\\
- If~you~go~on~this~command~will~be~ignored.
- }
-% \end{macrocode}
-%
-% \begin{macrocode}
-\@@_msg_new:nn { Option~RenewMatrix~suppressed }
- {
- The~option~'RenewMatrix'~has~been~renamed~'renew-matrix'.\\
- If~you~go~on~this~command~will~be~ignored.
- }
-% \end{macrocode}
-%
-% \begin{macrocode}
\@@_msg_new:nn { Bad~value~for~letter~for~dotted~lines }
{
- The~value~of~key~'\tl_use:N\l_keys_key_tl'~must~be~of~length~1.\\
+ The~value~of~key~'\tl_use:N\l_keys_key_str'~must~be~of~length~1.\\
If~you~go~on,~it~will~be~ignored.
}
% \end{macrocode}
@@ -7213,7 +7809,7 @@ c & c & b & a & e
% \begin{macrocode}
\@@_msg_new:nnn { Unknown~key~for~NiceMatrixOptions }
{
- The~key~'\tl_use:N\l_keys_key_tl'~is~unknown~for~the~command~
+ The~key~'\tl_use:N\l_keys_key_str'~is~unknown~for~the~command~
\token_to_str:N \NiceMatrixOptions. \\
If~you~go~on,~it~will~be~ignored. \\
For~a~list~of~the~available~keys,~type~H~<return>.
@@ -7232,6 +7828,7 @@ c & c & b & a & e
end-of-row,~
exterior-arraycolsep,~
hlines,~
+ hvlines,~
left-margin,~
letter-for-dotted-lines,~
light-syntax,~
@@ -7241,14 +7838,15 @@ c & c & b & a & e
renew-matrix,~
right-margin,~
small,~
- and~transparent
+ transparent~
+ and~vlines.
}
% \end{macrocode}
%
% \begin{macrocode}
\@@_msg_new:nnn { Unknown~option~for~NiceArray }
{
- The~option~'\tl_use:N\l_keys_key_tl'~is~unknown~for~the~environment~
+ The~option~'\tl_use:N\l_keys_key_str'~is~unknown~for~the~environment~
\{NiceArray\}. \\
If~you~go~on,~it~will~be~ignored. \\
For~a~list~of~the~available~options,~type~H~<return>.
@@ -7256,6 +7854,7 @@ c & c & b & a & e
{
The~available~options~are~(in~alphabetic~order):~
b,~
+ baseline,~
c,~
code-after,~
code-for-first-col,~
@@ -7273,6 +7872,7 @@ c & c & b & a & e
first-col,~
first-row,~
hlines,~
+ hvlines,~
last-col,~
last-row,~
left-margin,~
@@ -7283,7 +7883,8 @@ c & c & b & a & e
renew-dots,~
right-margin,~
small,~
- and~t.
+ t,~
+ and~vlines.
}
% \end{macrocode}
%
@@ -7294,8 +7895,8 @@ c & c & b & a & e
% \begin{macrocode}
\@@_msg_new:nnn { Unknown~option~for~NiceMatrix }
{
- The~option~'\tl_use:N\l_keys_key_tl'~is~unknown~for~the~
- \g_@@_type_env_str. \\
+ The~option~'\tl_use:N\l_keys_key_str'~is~unknown~for~the~
+ \@@_full_name_env:. \\
If~you~go~on,~it~will~be~ignored. \\
For~a~list~of~the~available~options,~type~H~<return>.
}
@@ -7317,6 +7918,7 @@ c & c & b & a & e
first-col,~
first-row,~
hlines,~
+ hvlines,~
last-col,~
last-row,~
left-margin,~
@@ -7326,7 +7928,8 @@ c & c & b & a & e
parallelize-diags,~
renew-dots,~
right-margin,~
- and~small.
+ small,~
+ and~vlines.
}
% \end{macrocode}
%
@@ -7358,10 +7961,9 @@ c & c & b & a & e
% \begin{macrocode}
\@@_msg_new:nn { Zero~row }
{
- There~is~a~problem.~Maybe~your~\g_@@_type_env_str\ is~empty.~
- Maybe~you~have~used~l,~c~and~r~instead~of~L,~C~and~R~in~the~preamble~
- of~your~environment. \\
- If~you~go~on,~the~result~may~be~incorrect.
+ There~is~a~problem.~Maybe~you~have~used~l,~c~and~r~instead~of~L,~C~
+ and~R~in~the~preamble~of~your~environment. \\
+ This~error~is~fatal.
}
% \end{macrocode}
%
@@ -7369,29 +7971,28 @@ c & c & b & a & e
\@@_msg_new:nn { Use~of~:~in~first~position }
{
You~can't~use~the~column~specifier~'\l_@@_letter_for_dotted_lines_str'~in~the~
- first~position~of~the~preamble~of~the~\g_@@_type_env_str. \\
+ first~position~of~the~preamble~of~the~\@@_full_name_env:. \\
If~you~go~on,~this~dotted~line~will~be~ignored.
}
% \end{macrocode}
%
-%
-% \subsection{Obsolete environments}
+% \bigskip
+% \subsection*{Obsolete environments}
%
% \begin{macrocode}
\@@_msg_new:nn { Obsolete~environment }
{
- The~environment~\{\@currenvir\}~is~obsolete.~We~should~use~#1~instead.~
- However,~you~can~go~on~for~this~time.~
- If~you~don't~want~to~see~this~error~again,~you~should~load~'nicematrix'~
- with~the~option~'obsolete-environments'.
+ The~environment~\{\@currenvir\}~is~obsolete.~You~should~use~#1~instead.~
+ However,~it's~still~possible~to~use~the~environment~\{\@currenvir\}~(for~
+ a~few~months)~by~loading~'nicematrix'~with~the~option~
+ 'obsolete-environments'.
}
% \end{macrocode}
%
% \begin{macrocode}
\NewDocumentEnvironment { pNiceArrayC } { }
{
- \@@_error:nn { Obsolete~environment }
- { the~option~'last-col' }
+ \@@_fatal:nn { Obsolete~environment } { the~option~'last-col' }
\int_zero:N \l_@@_last_col_int
\pNiceArray
}
@@ -7401,8 +8002,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { bNiceArrayC } { }
{
- \@@_error:nn { Obsolete~environment }
- { the~option~'last-col' }
+ \@@_fatal:nn { Obsolete~environment } { the~option~'last-col' }
\int_zero:N \l_@@_last_col_int
\bNiceArray
}
@@ -7412,8 +8012,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { BNiceArrayC } { }
{
- \@@_error:nn { Obsolete~environment }
- { the~option~'last-col' }
+ \@@_fatal:nn { Obsolete~environment } { the~option~'last-col' }
\int_zero:N \l_@@_last_col_int
\BNiceArray
}
@@ -7423,8 +8022,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { vNiceArrayC } { }
{
- \@@_error:nn { Obsolete~environment }
- { the~option~'last-col' }
+ \@@_fatal:nn { Obsolete~environment } { the~option~'last-col' }
\int_zero:N \l_@@_last_col_int
\vNiceArray
}
@@ -7434,8 +8032,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { VNiceArrayC } { }
{
- \@@_error:nn { Obsolete~environment }
- { the~option~'last-col' }
+ \@@_fatal:nn { Obsolete~environment } { the~option~'last-col' }
\int_zero:N \l_@@_last_col_int
\VNiceArray
}
@@ -7445,7 +8042,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { pNiceArrayRC } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~options~'last-col'~and~'first-row' }
\int_zero:N \l_@@_last_col_int
\int_zero:N \l_@@_first_row_int
@@ -7457,7 +8054,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { bNiceArrayRC } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~options~'last-col'~and~'first-row' }
\int_zero:N \l_@@_last_col_int
\int_zero:N \l_@@_first_row_int
@@ -7469,7 +8066,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { BNiceArrayRC } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~options~'last-col'~and~'first-row' }
\int_zero:N \l_@@_last_col_int
\int_zero:N \l_@@_first_row_int
@@ -7481,7 +8078,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { vNiceArrayRC } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~options~'last-col'~and~'first-row' }
\int_zero:N \l_@@_last_col_int
\int_zero:N \l_@@_first_row_int
@@ -7493,7 +8090,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { VNiceArrayRC } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~options~'last-col'~and~'first-row' }
\int_zero:N \l_@@_last_col_int
\int_zero:N \l_@@_first_row_int
@@ -7505,7 +8102,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { NiceArrayCwithDelims } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~option~'last-col' }
\int_zero:N \l_@@_last_col_int
\NiceArrayWithDelims
@@ -7516,7 +8113,7 @@ c & c & b & a & e
% \begin{macrocode}
\NewDocumentEnvironment { NiceArrayRCwithDelims } { }
{
- \@@_error:nn { Obsolete~environment }
+ \@@_fatal:nn { Obsolete~environment }
{ the~options~'last-col'~and~'first-row' }
\int_zero:N \l_@@_last_col_int
\int_zero:N \l_@@_first_row_int
@@ -7665,9 +8262,9 @@ c & c & b & a & e
%
% Error message when the user gives an incorrect value for |last-row|.
%
-% A dotted line can no longer cross another dotted line (except the dotted lines
-% drawn by |\cdottedline|, the symbol |:| (in the preamble of the array) and
-% |\line| in |code-after|.
+% A dotted line can no longer cross another dotted line (excepted the dotted
+% lines drawn by |\cdottedline|, the symbol |:| (in the preamble of the array) and
+% |\line| in |code-after|).
%
% The starred versions of |\Cdots|, |\Ldots|, etc. are now deprecated because,
% with the new implementation, they become pointless. These starred versions
@@ -7763,6 +8360,25 @@ c & c & b & a & e
%
% Correction of a bug linked to |first-row| and |last-row|.
%
+% \subsection*{Changes between versions 3.11 and 3.12}
+%
+% Command |\rotate| in the cells of the array.
+%
+% Options |vlines|, |hlines| and |hvlines|.
+%
+% Option |baseline| pour |{NiceArray}| (uniquement).
+%
+% The name of the Tikz nodes created by the command |\Block| has changed: when
+% the command has been issued in the cell $i$|-|$j$, the name is
+% $i$|-|$j$|-block| and, if the creation of the ``medium nodes'' is required, a
+% node $i$|-|$j$|-block-medium| is created.
+%
+% If the user try to use more columns than allowed by its environment, an error
+% is raised by nicematrix (instead of a low-level error).
+%
+% The package must be loaded with the option |obsolete-environments| if we want
+% to use the deprecated environments.
+%
% \PrintIndex
%
% \tableofcontents
diff --git a/macros/latex/contrib/nicematrix/nicematrix.pdf b/macros/latex/contrib/nicematrix/nicematrix.pdf
index a9862fd6a3..f56d0ccfd1 100644
--- a/macros/latex/contrib/nicematrix/nicematrix.pdf
+++ b/macros/latex/contrib/nicematrix/nicematrix.pdf
Binary files differ