*** empty log message ***

Author: bert-github

tools/htmlmathml-f-ent-to-awk.pl

@@ -0,0 +1,48 @@
+#!/usr/bin/env perl
+#
+# Tool to create htmlmathml.awk from
+# https://www.w3.org/2003/entities/2007/htmlmathml-f.ent
+#
+# Expects lines such as
+#
+#   <!ENTITY npart "&#x02202;&#x00338;" ><!--PARTIAL DIFFERENTIAL with slash -->
+#
+# The quoted replacement text contains text and numeric entities. If
+# the replacement text contains a character that would be an XML
+# delimiter, that delimiter is not written as "&#nnn;" but as
+# "&#38#nnn;". E.g., the replacement for "&lt;" would be "&#60;"
+# (which stands for "<"), but is written as "&#38;#60;".
+#
+# Each entity line creates an Awk array entry like
+#
+#    htmlmathml_ent["part"] = "∂̸"
+
+use warnings;
+use strict;
+use feature 'unicode_strings';
+
+binmode(STDOUT, ":utf8");
+print "# Defines the array htmlmathml::ent with replacements for\n";
+print "# all named character entities of HTML.\n";
+print "# This file is in UTF-8.\n\n";
+print "\@namespace \"htmlmathml\"\n\n";
+print "BEGIN {\n";
+
+while (<>) {
+  if (/<!ENTITY\s+([a-zA-Z][a-zA-Z0-9.-]*)\s*"([^"]*)"/) {
+    my ($n, $v) = ($1, $2);	# Entity name and value
+
+    # Replace hexadecimal and decimal character entities. Some
+    # replacements replace a name by another entity: (e.g., "lt" ->
+    # "&#38;#60") because the actual character would not be legal XML.
+    # We replace such entities as well.
+    $v =~ s/&#(?:38;#)?(?:(x)([0-9a-f]+)|([0-9]+));/$1 ? chr(hex($2)) : chr($3)/gie;
+
+    # Replace control characters, quotes and backslashes by hexadecimal escapes.
+    $v =~ s/[\x00-\x1F]|"|\\/sprintf("\\x%02x", ord($&))/ge;
+
+    print "ent[\"$n\"] = \"$v\"\n";
+  }
+}
+
+print "}\n";