--- loncom/homework/cleanxml/post_xml.pm 2016/01/08 20:32:42 1.6 +++ loncom/homework/cleanxml/post_xml.pm 2016/01/20 00:40:39 1.8 @@ -1,7 +1,7 @@ # The LearningOnline Network # Third step to clean a file. # -# $Id: post_xml.pm,v 1.6 2016/01/08 20:32:42 damieng Exp $ +# $Id: post_xml.pm,v 1.8 2016/01/20 00:40:39 damieng Exp $ # # Copyright Michigan State University Board of Trustees # @@ -41,6 +41,8 @@ use Cwd 'abs_path'; use XML::LibXML; use HTML::TokeParser; # used to parse sty files 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 @@ -552,11 +554,11 @@ sub replace_m { # Returns the HTML equivalent of LaTeX input, using tth sub tth { my ($text) = @_; - my ($fh, $tmp_path) = tempfile(); - binmode($fh, ':utf8'); - print $fh $text; - close $fh; - my $output = `tth -r -w2 -u -y0 < $tmp_path 2>/dev/null`; + my $output = &tth::tth($text); + my $errorstring = &tth::ttherror(); + if ($errorstring) { + die $errorstring; + } # hopefully the temp file will not be removed before this point (otherwise we should use unlink_on_destroy 0) $output =~ s/^\s*|\s*$//; $output =~ s/
<\/div>//; # why is tth using such ugly markup for \newline ? @@ -567,7 +569,7 @@ sub tth { sub html_to_dom { my ($text) = @_; $text = ''.$text.''; - 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   into a character my $dom_doc = XML::LibXML->load_xml(string => $textref); my $root = $dom_doc->documentElement; @@ -1812,8 +1814,8 @@ sub remove_useless_notsolved { sub fix_paragraphs_inside { my ($node, $all_block) = @_; # blocks in which paragrahs will be added: - my @blocks_with_p = ('loncapa','library','problem','part','problemtype','window','block','while','postanswerdate','preduedate','solved','notsolved','languageblock','instructorcomment','togglebox','standalone','body','form'); - my @fix_p_if_br_or_p = (@responses,'foil','item','text','label','hintgroup','hintpart','hint','web','windowlink','div','li','dd','td','th','blockquote'); + my @blocks_with_p = ('loncapa','library','problem','part','problemtype','window','block','while','postanswerdate','preduedate','languageblock','instructorcomment','togglebox','standalone','body','form'); + my @fix_p_if_br_or_p = (@responses,'foil','item','text','label','hintgroup','hintpart','hint','web','windowlink','div','li','dd','td','th','blockquote','solved','notsolved'); if ((string_in_array(\@blocks_with_p, $node->nodeName) && paragraph_needed($node)) || (string_in_array(\@fix_p_if_br_or_p, $node->nodeName) && paragraph_inside($node))) { # if non-empty, add paragraphs where needed between all br and remove br