implement verbatim*

brucemiller · Aug 5, 2023 · b569ec3 · b569ec3
1 parent 207eb53
commit b569ec3
Show file tree

Hide file tree

Showing 4 changed files with 74 additions and 38 deletions.
diff --git a/lib/LaTeXML/Package/LaTeX.pool.ltxml b/lib/LaTeXML/Package/LaTeX.pool.ltxml
@@ -1863,8 +1863,8 @@ DefRegister('\@itemdepth' => Number(0));
 #======================================================================
 
 # NOTE: how's the best way to get verbatim material through?
-DefEnvironment('{verbatim}',  '<ltx:verbatim>#body</ltx:verbatim>');
-DefEnvironment('{verbatim*}', '<ltx:verbatim>#body</ltx:verbatim>');
+#DefEnvironment('{verbatim}',  '<ltx:verbatim>#body</ltx:verbatim>');
+#DefEnvironment('{verbatim*}', '<ltx:verbatim>#body</ltx:verbatim>');
 
 DefMacroI('\@verbatim', undef,
   '\par\aftergroup\lx@end@verbatim\lx@@verbatim');    # Close enough?
@@ -1888,40 +1888,53 @@ DefConstructorI('\lx@end@verbatim', undef,
 # and also the usual environment capture.
 DefConstructorI(T_CS('\begin{verbatim}'), undef,
   "<ltx:verbatim font='#font'>#body</ltx:verbatim>",
-  beforeDigest => [sub { $_[0]->bgroup;
-      my @stuff = ();
-      if (my $b = LookupValue('@environment@verbatim@atbegin')) {
-        push(@stuff, Digest(@$b)); }
-      AssignValue(current_environment => 'verbatim');
-      DefMacroI('\@currenvir', undef, 'verbatim');
-      MergeFont(family => 'typewriter');
-      # Digest(T_CS('\par')); # NO! See beforeConstruct!
-      @stuff; }],
-  afterDigest => [sub {
-      my ($stomach, $whatsit) = @_;
-      #      $stomach->egroup;
-      my $font   = $whatsit->getFont;
-      my $loc    = $whatsit->getLocator;
-      my $end    = "\\end{verbatim}";
-      my @lines  = ();
-      my $gullet = $stomach->getGullet;
-      while (defined(my $line = $gullet->readRawLine)) {
-        # The raw chars will still have to be decoded (but not space!!)
-        $line = join('', map { ($_ eq ' ' ? ' ' : FontDecodeString($_, 'OT1_typewriter')) }
-            split(//, $line));
-        if ($line =~ /^(.*?)\\end\{verbatim\}(.*?)$/) {
-          push(@lines, $1 . "\n"); $gullet->unread(Tokenize($2), T_CR);
-          last; }
-        push(@lines, $line . "\n"); }
-      pop(@lines) if $lines[-1] eq "\n";
-      # Note last line ends up as Whatsit's "trailer"
-      if (my $b = LookupValue('@environment@verbatim@atend')) {
-        push(@lines, ToString(Digest(@$b))); }
-      $stomach->egroup;
-      $whatsit->setBody(map { Box($_, $font, $loc, T_OTHER($_)) } @lines, $end);
-      return; }],
+  beforeDigest    => [sub { beforeDigestVerbatim(0, @_); }],
+  afterDigest     => [sub { afterDigestVerbatim(0, @_); }],
+  beforeConstruct => sub { $_[0]->maybeCloseElement('ltx:p'); });
+
+DefConstructorI(T_CS('\begin{verbatim*}'), undef,
+  "<ltx:verbatim font='#font'>#body</ltx:verbatim>",
+  beforeDigest    => [sub { beforeDigestVerbatim(1, @_); }],
+  afterDigest     => [sub { afterDigestVerbatim(1, @_); }],
   beforeConstruct => sub { $_[0]->maybeCloseElement('ltx:p'); });
 
+sub beforeDigestVerbatim {
+  my ($starred, $stomach) = @_;
+  $stomach->bgroup;
+  my @stuff = ();
+  if (my $b = LookupValue('@environment@verbatim@atbegin')) {
+    push(@stuff, Digest(@$b)); }
+  AssignValue(current_environment => 'verbatim');
+  DefMacroI('\@currenvir', undef, 'verbatim');
+  MergeFont(family => 'typewriter');
+  # Digest(T_CS('\par')); # NO! See beforeConstruct!
+  @stuff; }
+
+sub afterDigestVerbatim {
+  my ($starred, $stomach, $whatsit) = @_;
+  #      $stomach->egroup;
+  my $font   = $whatsit->getFont;
+  my $loc    = $whatsit->getLocator;
+  my $end    = $starred ? '\end{verbatim*}' : '\end{verbatim}';
+  my $space  = $starred ? "\x{2423}"        : ' ';
+  my @lines  = ();
+  my $gullet = $stomach->getGullet;
+  while (defined(my $line = $gullet->readRawLine)) {
+    # The raw chars will still have to be decoded (but not space!!)
+    $line = join('', map { ($_ eq ' ' ? $space : FontDecodeString($_, 'OT1_typewriter')) }
+        split(//, $line));
+    if ($line =~ /^(.*?)\Q$end\E(.*?)$/) {
+      push(@lines, $1 . "\n"); $gullet->unread(Tokenize($2), T_CR);
+      last; }
+    push(@lines, $line . "\n"); }
+  pop(@lines) if $lines[-1] eq "\n";
+  # Note last line ends up as Whatsit's "trailer"
+  if (my $b = LookupValue('@environment@verbatim@atend')) {
+    push(@lines, ToString(Digest(@$b))); }
+  $stomach->egroup;
+  $whatsit->setBody(map { Box($_, $font, $loc, T_OTHER($_)) } @lines, $end);
+  return; }
+
 DefPrimitiveI('\@vobeyspaces', undef, sub {
     AssignCatcode(" " => 13);
     Let(T_ACTIVE(" "), '\nobreakspace');
@@ -1935,11 +1948,14 @@ DefMacroI('\verb', undef, sub {
     StartSemiverbatim('%', '\\', '{', '}');
     $STATE->assignCatcode(' ', CC_ACTIVE);
     my $init;
+    my $skippedSpace = 0;
     # As of texlive 2021, DO skip spaces before delimiter (even tho we've changed catcodes)
+    # but if we do skip spaces, * can be the delimiter
     do { $init = $gullet->readToken();
+      $skippedSpace = 1 if (defined $init && $init->getString eq ' ');
     } while (defined $init && $init->getString eq ' ');
     my $starred = 0;
-    if (T_OTHER('*')->equals($init)) {
+    if (T_OTHER('*')->equals($init) && !$skippedSpace) {
       $starred = 1;
       do { $init = $gullet->readToken();
       } while (defined $init && $init->getString eq ' '); }

diff --git a/t/tokenize/verb.pdf b/t/tokenize/verb.pdf
diff --git a/t/tokenize/verb.tex b/t/tokenize/verb.tex
@@ -21,6 +21,12 @@ \section{Verbatim Environment}
 keep this too    \end{verbatim} outside verbatim
 More Normal {\bf Bold} stuff.
 
+Verbatim 4:
+\begin{verbatim*} keep this
+AbNormal {\bf NonBold} stuff.
+keep this too    \end{verbatim*} outside verbatim
+More Normal {\bf Bold} stuff.
+
 % NOT allowed
 %\section{Inline \verb|\verb| verbatim.}
 \section{Inline verb command}
@@ -47,7 +53,7 @@ \section{Inline verb command}
 Note that spaces after a control-sequence are skipped on the \emph{next} read,
 by which time catcodes may have been changed.
 Prior to April, 2020, a space following \verb|\verb| would -- surprisingly -- have been treated as
-the delimiter! Since then, spaces are skipped and the following char is used as delimiter.
+the delimiter! Since then, spaces are skipped and the following char (even *) is used as delimiter.
 We'll adopt the newer approach.
 
 \verb |a}b#c^d_e$f| stuff
@@ -56,6 +62,8 @@ \section{Inline verb command}
 
 {\verb* Trick-roll} T }
 
+\verb * + i n * o u t +
+
 % Another tricky case
 \makeatletter
 \def\verbatimlisting#1{%

diff --git a/t/tokenize/verb.xml b/t/tokenize/verb.xml
@@ -35,6 +35,15 @@ AbNormal {\bf NonBold} stuff.
 keep this too    
 </verbatim>
       <p>outside verbatim
+More Normal <text font="bold">Bold</text> stuff.</p>
+    </para>
+    <para xml:id="S1.p5">
+      <p>Verbatim 4:</p>
+      <verbatim font="typewriter">␣keep␣this
+AbNormal␣{\bf␣NonBold}␣stuff.
+keep␣this␣too␣␣␣␣
+</verbatim>
+      <p>␣outside␣verbatim
 More Normal <text font="bold">Bold</text> stuff.</p>
     </para>
   </section>
@@ -78,7 +87,7 @@ More Normal <text font="bold">Bold</text> stuff.</p>
       <p>Note that spaces after a control-sequence are skipped on the <emph font="italic">next</emph> read,
 by which time catcodes may have been changed.
 Prior to April, 2020, a space following <verbatim font="typewriter">\verb</verbatim> would – surprisingly – have been treated as
-the delimiter! Since then, spaces are skipped and the following char is used as delimiter.
+the delimiter! Since then, spaces are skipped and the following char (even *) is used as delimiter.
 We’ll adopt the newer approach.</p>
     </para>
     <para xml:id="S2.p9">
@@ -91,9 +100,12 @@ We’ll adopt the newer approach.</p>
       <p><verbatim font="typewriter">rick-roll}␣</verbatim></p>
     </para>
     <para xml:id="S2.p12">
-      <p>Input snippet</p>
+      <p><verbatim font="typewriter">+ i n </verbatim> o u t +</p>
     </para>
     <para xml:id="S2.p13">
+      <p>Input snippet</p>
+    </para>
+    <para xml:id="S2.p14">
       <verbatim font="typewriter">|a}b#c^d_e$f|
 </verbatim>
       <p>after.</p>