Sandbox Fatal fixes 01.2024, round 2 (#2315)

* guard revertScript from inner Tokens

* allow deep recursion for computing box sizes in larger tables

* allow deep recursion related to digesting parameter reads

* fix infinite bibitem loop with OmniBus natbib loading

* pgfmath recognizes a single dot as 0

* avoid perl warning for regex on undef

* add a couple of latex.ltx internals

* guard listings infinite loop

* protect IEEEtran.cls.ltxml from renewed definitions of \sc and \it

* allow for optional arguments in IEEEtran lists, and rebind the default latex envs

* Revert "guard revertScript from inner Tokens"

This reverts commit 328a7df.

* avoid Fatals in vertical/align checks over document without a root node

* allow trailing space in rgb colorspec (see arXiv:1705.04144)

lib/LaTeXML/Common/Font.pm

@@ -22,7 +22,7 @@ use LaTeXML::Common::Font::Metric;
 use LaTeXML::Common::Font::StandardMetrics;
 use LaTeXML::Common::Color;
 use List::Util qw(min max sum);
-use base       qw(LaTeXML::Common::Object);
+use base qw(LaTeXML::Common::Object);
 
 # Note that this has evolved way beynond just "font",
 # but covers text properties (or even display properties) in general
@@ -263,6 +263,7 @@ sub toString {
 
 # Perhaps it is more useful to list only the non-default components?
 sub stringify {
+  no warnings 'recursion';
   my ($self) = @_;
   my ($fam, $ser, $shp, $siz, $col, $bkg, $opa, $enc, $lang, $mstyle, $flags) = @$self;
   $fam = 'serif' if $fam && ($fam eq 'math');

lib/LaTeXML/Common/Object.pm

@@ -28,6 +28,7 @@ my %NOBLESS = map { ($_ => 1) } qw( SCALAR HASH ARRAY CODE REF GLOB LVALUE);
 # Since the next two are used in debugging and error messages,
 # be careful to avoid recursive errors
 sub Stringify {
+  no warnings 'recursion';
   my ($object) = @_;
   my $string = eval {
     local $LaTeXML::IGNORE_ERRORS = 1;

lib/LaTeXML/Core/List.pm

@@ -73,6 +73,7 @@ sub toString {
 
 # Methods for overloaded operators
 sub stringify {
+  no warnings 'recursion';
   my ($self) = @_;
   my $type = ref $self;
   $type =~ s/^LaTeXML:://;

lib/LaTeXML/Core/Parameter.pm

@@ -126,6 +126,7 @@ sub reparse {
         return $value; }); } }
 
 sub digest {
+  no warnings 'recursion';
   my ($self, $stomach, $value, $fordefn) = @_;
   # If semiverbatim, Expand (before digest), so tokens can be neutralized; BLECH!!!!
   if ($$self{semiverbatim}) {

lib/LaTeXML/Core/Tokens.pm

@@ -81,6 +81,7 @@ sub stringify {
   return "Tokens[" . join(',', map { $_->toString } @$self) . "]"; }
 
 sub beDigested {
+  no warnings 'recursion';
   my ($self, $stomach) = @_;
   return $stomach->digest($self); }
 

lib/LaTeXML/Core/Whatsit.pm

@@ -175,6 +175,7 @@ sub getString {
 
 # Methods for overloaded operators
 sub stringify {
+  no warnings 'recursion';
   my ($self) = @_;
   my $hasbody = defined $$self{properties}{body};
   return "Whatsit[" . join(',', $self->getDefinition->getCS->getCSName,
@@ -243,6 +244,7 @@ sub computeSize {
         @boxes = $boxes[0]->unlist; } }
     else {
       push(@boxes, $sizer); }
+    no warnings 'recursion';
     return $$props{font}->computeBoxesSize([@boxes], %options); } }
 
 #======================================================================

lib/LaTeXML/Package/IEEEtran.cls.ltxml

@@ -180,8 +180,13 @@ RawTeX(<<'EoTeX');
 \def\theparagraph{\thesubsubsection\alph{paragraph}}         % I-A1a
 \fi
 EoTeX
-DefMacro('\format@title@font@section',    '\sc');
-DefMacro('\format@title@font@subsection', '\it');
+DefPrimitiveI('\ltx@ieeetran@it', undef, undef,
+  font => { shape => 'italic', family => 'serif', series => 'medium' }, locked => 1);
+DefPrimitiveI('\ltx@ieeetran@sc', undef, undef,
+  font => { shape => 'smallcaps', family => 'serif', series => 'medium' }, locked => 1);
+
+DefMacro('\format@title@font@section',    '\ltx@ieeetran@sc');
+DefMacro('\format@title@font@subsection', '\ltx@ieeetran@it');
 DefMacro('\figurename',                   'Fig.');
 DefMacro('\tablename',                    'TABLE');
 DefMacro('\thetable',                     '\Roman{table}');
@@ -340,23 +345,39 @@ DefMacro('\IEEEiedlabeljustifyc', '');
 DefMacro('\IEEEiedlabeljustifyl', '');
 DefMacro('\IEEEiedlabeljustifyr', '');
 
-DefEnvironment('{IEEEitemize}',
+# TODO: Use the optional argument.
+# also, we skip the internal @-named variants for now.
+DefEnvironment('{IEEEitemize}[]',
   "<ltx:itemize xml:id='#id'>#body</ltx:itemize>",
   properties      => sub { beginItemize('itemize', '@item'); },
   beforeDigestEnd => sub { Digest('\par'); },
   locked          => 1, mode => 'text');
-DefEnvironment('{IEEEenumerate}',
+DefEnvironment('{IEEEenumerate}[]',
   "<ltx:enumerate  xml:id='#id'>#body</ltx:enumerate>",
   properties      => sub { beginItemize('enumerate', 'enum'); },
   beforeDigestEnd => sub { Digest('\par'); },
   locked          => 1, mode => 'text');
-DefEnvironment('{IEEEdescription}',
+DefEnvironment('{IEEEdescription}[]',
   "<ltx:description  xml:id='#id'>#body</ltx:description>",
   beforeDigest    => sub { Let('\makelabel', '\descriptionlabel'); },
   properties      => sub { beginItemize('description', '@desc'); },
   beforeDigestEnd => sub { Digest('\par'); },
   locked          => 1, mode => 'text');
 
+# override LaTeX's default IED lists
+Let('\itemize',                  '\IEEEitemize');
+Let('\enditemize',               '\endIEEEitemize');
+Let('\enumerate',                '\IEEEenumerate');
+Let('\endenumerate',             '\endIEEEenumerate');
+Let('\description',              '\IEEEdescription');
+Let('\enddescription',           '\endIEEEdescription');
+Let(T_CS('\begin{itemize}'),     '\IEEEitemize');
+Let(T_CS('\end{itemize}'),       '\endIEEEitemize');
+Let(T_CS('\begin{enumerate}'),   '\IEEEenumerate');
+Let(T_CS('\end{enumerate}'),     '\endIEEEenumerate');
+Let(T_CS('\begin{description}'), '\IEEEdescription');
+Let(T_CS('\end{description}'),   '\endIEEEdescription');
+
 # V1.7 provide string macros as article.cls does
 DefMacro('\contentsname',     'Contents');
 DefMacro('\listfigurename',   'List of Figures');

lib/LaTeXML/Package/LaTeX.pool.ltxml

@@ -290,7 +290,7 @@ DefConstructor('\lx@newline OptionalMatch:* [Glue]', sub {
       $document->insertElement('ltx:XMHint', undef, name => 'newline'); }
     else {
       my $context = $document->getElement;
-      if    ($context->getAttribute('_vertical_mode_')) { }
+      if    (!$context || $context->getAttribute('_vertical_mode_')) { }
       elsif (($document->getNodeQName($context) eq 'ltx:p')
         && $context->parentNode->getAttribute('_vertical_mode_')) {
         $document->maybeCloseElement('ltx:p'); }
@@ -1372,7 +1372,7 @@ DefEnvironment('{flushright}', sub {
 sub setupAligningContext {
   my ($document) = @_;
   my $node = $document->getElement;
-  AssignValue(ALIGNING_NODE => [$node, $node->lastChild]);
+  AssignValue(ALIGNING_NODE => [$node, $node->lastChild]) if $node;
   return; }
 
 sub applyAligningContext {
@@ -4971,6 +4971,8 @@ DefConstructor('\raisebox{Dimension}[Dimension][Dimension]{}',
   beforeDigest => sub { reenterTextMode(); },
   sizer        => sub { raisedSizer($_[0]->getArg(4), $_[0]->getArg(1)); });
 
+DefMacro('\@finalstrut{}', '\unskip\ifhmode\nobreak\fi\vrule\@width\z@\@height\z@\@depth\dp#1');
+
 #**********************************************************************
 # C.14 Pictures and Color
 #**********************************************************************
@@ -5514,6 +5516,8 @@ DefPrimitiveI('\textregistered', undef, UTF(0xAE));     # REGISTERED SIGN
 DefPrimitiveI('\texttrademark',  undef, "\x{2122}");    # TRADE MARK SIGN
 DefConstructor('\textsuperscript{}', "<ltx:sup>#1</ltx:sup>",
   mode => 'text');
+DefConstructor('\@textsuperscript{}', "<ltx:sup>#1</ltx:sup>",
+  mode => 'text', locked => 1);
 # This is something coming from xetex/xelatex ? Why define this way?
 #DefConstructor('\realsuperscript{}', "<ltx:text yoffset='0.5em' _noautoclose='1'>#1</ltx:text>");
 DefConstructor('\realsuperscript{}', "<ltx:sup>#1</ltx:sup>",

lib/LaTeXML/Package/listings.sty.ltxml

@@ -1355,7 +1355,16 @@ sub lstProcess_internal {
   my $classes    = LookupValue('LST_CLASSES');
   my $literate   = LookupValue('LST_LITERATE');
   my $lit_re     = ($end_re ? $LaTeXML::LITERATE_INNER_RE : $LaTeXML::LITERATE_RE);
+  my $loop_guard = 'loop guard';
+
   while ($LaTeXML::listing ne '') {
+    if ($LaTeXML::listing eq $loop_guard) {
+      # we must make headway on every step, or we risk infinite loops
+      Error('listing', 'infinite_loop', undef,
+        "lstProcess_internal failed to execute correctly.", "content was: '$loop_guard'");
+      $LaTeXML::listing = '';
+      last; }
+    $loop_guard = $LaTeXML::listing;
     # Matched the ending regular expression? (typically a close delimiter)
     if ($end_re && $LaTeXML::listing =~ s/^($end_re)//s) {
       $LaTeXML::colnum += length($1);
@@ -1449,7 +1458,6 @@ sub lstProcess_internal {
     elsif ($LaTeXML::QUOTED_RE && $LaTeXML::listing =~ s/^($LaTeXML::QUOTED_RE)//) { # Something quoted.
           # Don't just past together, and watch for leading \ (a common quoter)
       lstProcessPush(map { ($_ eq '\\' ? T_CS('\textbackslash') : T_OTHER($_)) } split('', $1));
-
       $LaTeXML::colnum += length($1); }
     else {
       if ($LaTeXML::listing =~ s/^(.)//s) {    # Anything else, just pass through.

lib/LaTeXML/Package/natbib.sty.ltxml

@@ -485,10 +485,13 @@ DefMacro('\shortcites Semiverbatim', '');
 #   \bibitem[\protect\citename{Jones et al., }1990]{key}...
 #   \harvarditem[Jones et al.]{Jones, Baker, and Williams}{1990}{key}...
 
-DefMacro('\bibitem',
+DefMacro('\lx@nat@bibitem',
 ##         '\reset@natbib@cites\refstepcounter{@bibitem}\@ifnextchar[{\@lbibitem}{\@lbibitem[\the@bibitem]}',
   '\reset@natbib@cites\refstepcounter{@bibitem}\@ifnextchar[{\@lbibitem}{\@lbibitem[]}',
   locked => 1);
+# Careful: since OmniBus also defines \bibitem and LaTeX.pool has a save/restore mechanism,
+# it is safer to \let to a constant natbib-specific macro definition. Test with arXiv:2007.09909
+Let('\bibitem', '\lx@nat@bibitem');
 
 RawTeX(<<'EOTeX');
 %%%

lib/LaTeXML/Package/pgfmath.code.tex.ltxml

@@ -705,6 +705,7 @@ expr :
     | /0b[01]+/                  { oct($item[1]); } # !!!
     | /0x[0-9a-fA-F]+/           { hex($item[1]); }
     | /0[0-9]+/                  { oct($item[1]); }
+    | /\./                       { 0.0; } # pgf treats a single dot as a zero
 
     UNIT :
       /(?:ex|em|pt|pc|in|bp|cm|mm|dd|cc|sp)/

lib/LaTeXML/Package/xcolor.sty.ltxml

@@ -290,7 +290,7 @@ sub DecodeColor {
   my $color_re = qr/($color_expr_re)(($func_expr_re)*)/;
 
   my $color;
-  if ($expression =~ /^$color_re$/) {
+  if ($expression =~ /^$color_re\s*$/) {
     #DG: Dear reader, I present to you: maintenance hell:
     my $prefix     = $2 || $10;
     my $name       = $3 || $11;

lib/LaTeXML/Post/MathML.pm

@@ -693,7 +693,8 @@ sub stylizeContent {
     if (my $default = $role && $default_token_content{$role}) {
       $text = $default; }
     else {
-      $text = ($iselement ? $item->getAttribute('name') || $item->getAttribute('meaning') || $role : '?');
+      $text = ($iselement ? ($item->getAttribute('name') || $item->getAttribute('meaning')
+            || $role || '') : '?');
       $color = 'red'; } }
   elsif (($text eq '-') && $role && (($role eq 'ADDOP') || ($role eq 'OPERATOR'))) { # MathML Core prefers unicode minus
     $text = "\x{2212}"; }
@@ -720,7 +721,7 @@ sub stylizeContent {
     : ($plane1 ? $variant : undef));
   my $u_text = ($tag ne 'm:mtext') && $u_variant && unicode_convert($text, $u_variant);
   if ((defined $u_text) && ($u_text ne '')) {    # didn't remap the text ? Keep text & variant
-    $text    = $u_text;
+    $text    = $u_text || '';
     $variant = ($plane1hack && ($variant ne $u_variant) && ($variant =~ /^bold/)
       ? 'bold' : undef); }                       # Possibly keep variant bold
                                                  # Use class (css) to patchup some weak translations