More ligatures

lib/LaTeXML/Core/Document.pm

@@ -1243,6 +1243,7 @@ sub closeText_internal {
         next if ($fonttest = $$ligature{fontTest}) && !&$fonttest($font);
         $string = &{ $$ligature{code} }($string); } }
     $node->setData($string) unless $string eq $ostring;
+    Debug("LIGATURE $ostring => $string") if $LaTeXML::DEBUG{document} && ($string ne $ostring);
     $$self{node} = $parent;    # Effectively closed (->setNode, but don't recurse)
     return $parent; }
   else {

lib/LaTeXML/Package/TeX.pool.ltxml

@@ -7112,19 +7112,33 @@ DefMathI('\to', undef, "\x{2192}", role => 'ARROW'); # RIGHTWARDS ARROW??? a bit
 DefPrimitive('\@@endash', sub { Box("\x{2013}", undef, undef, T_CS('\@@endash')); });
 DefPrimitive('\@@emdash', sub { Box("\x{2014}", undef, undef, T_CS('\@@emdash')); });
 
-DefLigature(qr{--}, "\x{2013}",
-  fontTest => sub { $_[0]->getFamily ne 'typewriter'; }); # EN DASH (NOTE: With digits before & aft => \N{FIGURE DASH})
-DefLigature(qr{---}, "\x{2014}",
-  fontTest => sub { $_[0]->getFamily ne 'typewriter'; });    # EM DASH
+sub nonTypewriter {
+  my ($font) = @_;
+  return ($font->getFamily ne 'typewriter'); }
+
+sub nonTypewriterT1 {
+  my ($font) = @_;
+  return ($font->getFamily ne 'typewriter') && (($font->getEncoding || 'OT1') =~ /^(OT1|T1)$/); }
+
+# EN DASH (NOTE: With digits before & aft => \N{FIGURE DASH})
+DefLigature(qr{--},  "\x{2013}", fontTest => \&nonTypewriter);    # EN dash
+DefLigature(qr{---}, "\x{2014}", fontTest => \&nonTypewriter);    # EM dash
 
 # Ligatures for doubled single left & right quotes to convert to double quotes
 # [should ligatures be part of a font, in the first place? (it is in TeX!)
-DefLigature(qr{\x{2018}\x{2018}}, "\x{201C}",
-  fontTest => sub { ($_[0]->getFamily ne 'typewriter')
-      && (($_[0]->getEncoding || 'OT1') =~ /^(OT1|T1)$/); });    # is this needed?
-DefLigature(qr{\x{2019}\x{2019}}, "\x{201D}",
-  fontTest => sub { ($_[0]->getFamily ne 'typewriter')
-      && (($_[0]->getEncoding || 'OT1') =~ /^(OT1|T1)$/); });
+DefLigature(qr{\x{2018}\x{2018}}, "\x{201C}", fontTest => \&nonTypewriterT1);   # double left quote
+DefLigature(qr{\x{2019}\x{2019}}, "\x{201D}", fontTest => \&nonTypewriterT1);   # double right quote
+DefLigature(qr{\?\x{2018}},       UTF(0xBF),  fontTest => \&nonTypewriterT1);   # ? backquote
+DefLigature(qr{!\x{2018}},        UTF(0xA1),  fontTest => \&nonTypewriterT1);   # ! backquote
+# These ligatures are also handled by TeX.
+# However, it appears that decent modern fonts in modern browsers handle these at that level.
+# So it's likely not worth doing it at the conversion level, possibly adversely affecting search.
+# DefLigature(qr{ff},               "\x{FB00}", fontTest => \&nonTypewriterT1);
+# DefLigature(qr{fi},               "\x{FB01}", fontTest => \&nonTypewriterT1);
+# DefLigature(qr{fl},               "\x{FB02}", fontTest => \&nonTypewriterT1);
+# DefLigature(qr{ffi},              "\x{FB03}", fontTest => \&nonTypewriterT1);
+# DefLigature(qr{ffl},              "\x{FB04}", fontTest => \&nonTypewriterT1);
+
 DefConstructor('\TeX',
   "<ltx:text class='ltx_TeX_logo' cssstyle='letter-spacing:-0.2em; margin-right:0.2em'>"
     . "T<ltx:text yoffset='-0.4ex'>E</ltx:text>X</ltx:text>",

t/tokenize/ligatures.tex

@@ -8,6 +8,8 @@ \section{Text Ligatures}
 
 A ``quote'' like this.
 
+?`Espa\~nol? !`Hola!
+
 \section{Ignore comments}
 %foo
 `%bar

t/tokenize/ligatures.xml

@@ -21,6 +21,9 @@ In interjection — like this — gets em-dash.</p>
     <para xml:id="S1.p3">
       <p>A “quote” like this.</p>
     </para>
+    <para xml:id="S1.p4">
+      <p>¿Español? ¡Hola!</p>
+    </para>
   </section>
   <section inlist="toc" xml:id="S2">
     <tags>