[BACK]Return to post_xml.pm CVS log [TXT] [DIR] Up to [LON-CAPA] / loncom / homework / cleanxml

Diff for /loncom/homework/cleanxml/post_xml.pm between versions 1.5 and 1.7

version 1.5, 2016/01/06 16:44:32 version 1.7, 2016/01/14 16:28:06
Line 41  use Cwd 'abs_path'; Line 41  use Cwd 'abs_path';
 use XML::LibXML;  use XML::LibXML;
 use HTML::TokeParser; # used to parse sty files  use HTML::TokeParser; # used to parse sty files
 use Tie::IxHash; # for ordered hashes  use Tie::IxHash; # for ordered hashes
   use tth;
   use Apache::html_to_xml;
   
 no warnings 'recursion'; # yes, fix_paragraph is using heavy recursion, I know  no warnings 'recursion'; # yes, fix_paragraph is using heavy recursion, I know
   
Line 50  my @inline_like_block = ('stringresponse Line 52  my @inline_like_block = ('stringresponse
 my @responses = ('stringresponse','optionresponse','numericalresponse','formularesponse','mathresponse','organicresponse','reactionresponse','customresponse','externalresponse','essayresponse','radiobuttonresponse','matchresponse','rankresponse','imageresponse','functionplotresponse');  my @responses = ('stringresponse','optionresponse','numericalresponse','formularesponse','mathresponse','organicresponse','reactionresponse','customresponse','externalresponse','essayresponse','radiobuttonresponse','matchresponse','rankresponse','imageresponse','functionplotresponse');
 my @block_html = ('html','head','body','section','h1','h2','h3','h4','h5','h6','div','p','ul','ol','li','table','tbody','tr','td','th','dl','dt','dd','pre','noscript','hr','address','blockquote','object','applet','embed','map','form','fieldset','iframe','center','frameset');  my @block_html = ('html','head','body','section','h1','h2','h3','h4','h5','h6','div','p','ul','ol','li','table','tbody','tr','td','th','dl','dt','dd','pre','noscript','hr','address','blockquote','object','applet','embed','map','form','fieldset','iframe','center','frameset');
 my @no_newline_inside = ('import','parserlib','scriptlib','data','function','label','xlabel','ylabel','tic','text','rectangle','image','title','h1','h2','h3','h4','h5','h6','li','td','p');  my @no_newline_inside = ('import','parserlib','scriptlib','data','function','label','xlabel','ylabel','tic','text','rectangle','image','title','h1','h2','h3','h4','h5','h6','li','td','p');
 my @preserve_elements = ('script','answer','pre');  my @preserve_elements = ('script','answer','pre','style');
 my @accepting_style = ('section','h1','h2','h3','h4','h5','h6','div','p','li','td','th','dt','dd','pre','blockquote');  my @accepting_style = ('section','h1','h2','h3','h4','h5','h6','div','p','li','td','th','dt','dd','pre','blockquote');
 my @latex_math = ('\alpha', '\theta', '\omicron', '\tau', '\beta', '\vartheta', '\pi', '\upsilon', '\gamma', '\gamma', '\varpi', '\phi', '\delta', '\kappa', '\rho', '\varphi', '\epsilon', '\lambda', '\varrho', '\chi', '\varepsilon', '\mu', '\sigma', '\psi', '\zeta', '\nu', '\varsigma', '\omega', '\eta', '\xi',  my @latex_math = ('\alpha', '\theta', '\omicron', '\tau', '\beta', '\vartheta', '\pi', '\upsilon', '\gamma', '\gamma', '\varpi', '\phi', '\delta', '\kappa', '\rho', '\varphi', '\epsilon', '\lambda', '\varrho', '\chi', '\varepsilon', '\mu', '\sigma', '\psi', '\zeta', '\nu', '\varsigma', '\omega', '\eta', '\xi',
   '\Gamma', '\Lambda', '\Sigma', '\Psi', '\Delta', '\Xi', '\Upsilon', '\Omega', '\Theta', '\Pi', '\Phi',    '\Gamma', '\Lambda', '\Sigma', '\Psi', '\Delta', '\Xi', '\Upsilon', '\Omega', '\Theta', '\Pi', '\Phi',
Line 552  sub replace_m { Line 554  sub replace_m {
 # Returns the HTML equivalent of LaTeX input, using tth  # Returns the HTML equivalent of LaTeX input, using tth
 sub tth {  sub tth {
   my ($text) = @_;    my ($text) = @_;
   my ($fh, $tmp_path) = tempfile();    my $output = &tth::tth($text);
   binmode($fh, ':utf8');    my $errorstring = &tth::ttherror();
   print $fh $text;    if ($errorstring) {
   close $fh;      die $errorstring;
   my $output = `tth -r -w2 -u -y0 < $tmp_path 2>/dev/null`;    }
   # hopefully the temp file will not be removed before this point (otherwise we should use unlink_on_destroy 0)    # hopefully the temp file will not be removed before this point (otherwise we should use unlink_on_destroy 0)
   $output =~ s/^\s*|\s*$//;    $output =~ s/^\s*|\s*$//;
   $output =~ s/<div class="p"><!----><\/div>/<br\/>/; # why is tth using such ugly markup for \newline ?    $output =~ s/<div class="p"><!----><\/div>/<br\/>/; # why is tth using such ugly markup for \newline ?
Line 567  sub tth { Line 569  sub tth {
 sub html_to_dom {  sub html_to_dom {
   my ($text) = @_;    my ($text) = @_;
   $text = '<root>'.$text.'</root>';    $text = '<root>'.$text.'</root>';
   my $textref = html_to_xml::html_to_xml(\$text);    my $textref = Apache::html_to_xml::html_to_xml(\$text);
   utf8::upgrade($$textref); # otherwise the XML parser fails when the HTML parser turns &nbsp; into a character    utf8::upgrade($$textref); # otherwise the XML parser fails when the HTML parser turns &nbsp; into a character
   my $dom_doc = XML::LibXML->load_xml(string => $textref);    my $dom_doc = XML::LibXML->load_xml(string => $textref);
   my $root = $dom_doc->documentElement;    my $root = $dom_doc->documentElement;
Line 2394  sub pretty { Line 2396  sub pretty {
   my $type = $node->nodeType;    my $type = $node->nodeType;
   if ($type == XML_ELEMENT_NODE) {    if ($type == XML_ELEMENT_NODE) {
     my $name = $node->nodeName;      my $name = $node->nodeName;
     if ((string_in_array($all_block, $name) || string_in_array(\@inline_like_block, $name)) &&      if (string_in_array(\@preserve_elements, $name)) {
         !string_in_array(\@preserve_elements, $name)) {        # remove newlines at the beginning and the end of preserve elements
         if (defined $node->firstChild && ($node->firstChild->nodeType == XML_TEXT_NODE ||
             $node->firstChild->nodeType == XML_CDATA_SECTION_NODE)) {
           my $text = $node->firstChild->nodeValue;
           $text =~ s/^\n+//;
           if ($text eq '') {
             $node->removeChild($node->firstChild);
           } else {
             $node->firstChild->setData($text);
           }
         }
         if (defined $node->lastChild && ($node->lastChild->nodeType == XML_TEXT_NODE ||
             $node->lastChild->nodeType == XML_CDATA_SECTION_NODE)) {
           my $text = $node->lastChild->nodeValue;
           $text =~ s/\n+$//;
           if ($text eq '') {
             $node->removeChild($node->lastChild);
           } else {
             $node->lastChild->setData($text);
           }
         }
       } elsif (string_in_array($all_block, $name) || string_in_array(\@inline_like_block, $name)) {
       # make sure there is a newline at the beginning and at the end if there is anything inside        # make sure there is a newline at the beginning and at the end if there is anything inside
       if (defined $node->firstChild && !string_in_array(\@no_newline_inside, $name)) {        if (defined $node->firstChild && !string_in_array(\@no_newline_inside, $name)) {
         my $first = $node->firstChild;          my $first = $node->firstChild;
Line 2478  sub pretty { Line 2501  sub pretty {
         if ($text eq '') {          if ($text eq '') {
           $node->removeChild($node->lastChild);            $node->removeChild($node->lastChild);
         } else {          } else {
           $node->lastChild->setData($text);  
         }  
       }  
     } elsif (string_in_array(\@preserve_elements, $name)) {  
       # collapse newlines at the beginning and the end of scripts  
       if (defined $node->firstChild && $node->firstChild->nodeType == XML_TEXT_NODE) {  
         my $text = $node->firstChild->nodeValue;  
         $text =~ s/^\n( *\n)+/\n/;  
         if ($text eq '') {  
           $node->removeChild($node->firstChild);  
         } else {  
           $node->firstChild->setData($text);  
         }  
       }  
       if (defined $node->lastChild && $node->lastChild->nodeType == XML_TEXT_NODE) {  
         my $text = $node->lastChild->nodeValue;  
         $text =~ s/\n( *\n)+$/\n/;  
         if ($text eq '') {  
           $node->removeChild($node->lastChild);  
         } else {  
           $node->lastChild->setData($text);            $node->lastChild->setData($text);
         }          }
       }        }
Removed from v.1.5  
changed lines
  Added in v.1.7

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>