[LON-CAPA-cvs] cvs: loncom /homework/cleanxml post_xml.pm xml_to_loncapa.pm

[damieng]

damieng		Fri Jan  8 20:32:42 2016 EDT

  Modified files:              
    /loncom/homework/cleanxml	post_xml.pm xml_to_loncapa.pm 
  Log:
  better space handling for style and pre elements
  
Index: loncom/homework/cleanxml/post_xml.pm
diff -u loncom/homework/cleanxml/post_xml.pm:1.5 loncom/homework/cleanxml/post_xml.pm:1.6
--- loncom/homework/cleanxml/post_xml.pm:1.5	Wed Jan  6 16:44:32 2016
+++ loncom/homework/cleanxml/post_xml.pm	Fri Jan  8 20:32:42 2016
@@ -1,7 +1,7 @@
 # The LearningOnline Network
 # Third step to clean a file.
 #
-# $Id: post_xml.pm,v 1.5 2016/01/06 16:44:32 damieng Exp $
+# $Id: post_xml.pm,v 1.6 2016/01/08 20:32:42 damieng Exp $
 #
 # Copyright Michigan State University Board of Trustees
 #
@@ -50,7 +50,7 @@
 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 @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 @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',
@@ -2394,8 +2394,29 @@
   my $type = $node->nodeType;
   if ($type == XML_ELEMENT_NODE) {
     my $name = $node->nodeName;
-    if ((string_in_array($all_block, $name) || string_in_array(\@inline_like_block, $name)) &&
-        !string_in_array(\@preserve_elements, $name)) {
+    if (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
       if (defined $node->firstChild && !string_in_array(\@no_newline_inside, $name)) {
         my $first = $node->firstChild;
@@ -2481,26 +2502,6 @@
           $node->lastChild->setData($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);
-        }
-      }
     }
   }
 }
Index: loncom/homework/cleanxml/xml_to_loncapa.pm
diff -u loncom/homework/cleanxml/xml_to_loncapa.pm:1.5 loncom/homework/cleanxml/xml_to_loncapa.pm:1.6
--- loncom/homework/cleanxml/xml_to_loncapa.pm:1.5	Wed Jan  6 16:44:32 2016
+++ loncom/homework/cleanxml/xml_to_loncapa.pm	Fri Jan  8 20:32:42 2016
@@ -1,7 +1,7 @@
 # The LearningOnline Network
 # convert_file takes a well-formed XML file content and converts it to LON-CAPA syntax.
 #
-# $Id: xml_to_loncapa.pm,v 1.5 2016/01/06 16:44:32 damieng Exp $
+# $Id: xml_to_loncapa.pm,v 1.6 2016/01/08 20:32:42 damieng Exp $
 #
 # Copyright Michigan State University Board of Trustees
 #
@@ -81,7 +81,7 @@
     if (defined $parent->parentNode) {
       $grandparent_name = $parent->parentNode->nodeName;
     }
-    my @no_escape = ('m', 'script', 'display', 'parse', 'answer');
+    my @no_escape = ('m', 'script', 'style', 'display', 'parse', 'answer');
     if (string_in_array(\@no_escape, $parent_name) &&
         ($parent_name ne 'answer' ||
         (defined $grandparent_name &&