Changeset 27990 for other-projects

Timestamp:

2013-08-06T22:20:41+12:00 (11 years ago)

Author:

ak19

Message:

2 fixes: 1. The Tudor collections' html source documents have stray carriage returns (r) that are not cleaned up by html tidy and make it into the linux doc.xml which for the rest use only linefeed chars. In contrast, the windows doc.xml was explicitly processed to convert all carriage-return-line-feed (rn) into linefeed by removing the carriage returns. So there were stray carriage returns in the linux doc.xml but these had been removed in the windows doc.xml, resulting in differences. 2. Partly truncated ampersand entities in the xml report are now completed so that things don't break when the xslt is applied during the summarise command that generates the report.

File:

• other-projects/nightly-tasks/diffcol/trunk/diffcol/diffcol.pl (modified) (3 diffs)

other-projects/nightly-tasks/diffcol/trunk/diffcol/diffcol.pl

@@ -654 +654 @@
             
                 my $win_contents = $testIsWin ? \$test_contents : \$model_contents;
-            
+                my $lin_contents = $testIsWin ? \$model_contents : \$test_contents;
+                
                 # remove all carriage returns \r - introduced into doc.xml by multiread after pdf converted to html
                 $$win_contents =~ s@[\r]@@g;
@@ -665 +666 @@
                 #FOR MAC: old macs use CR carriage return (see http://www.perlmonks.org/?node_id=745018), so replace with \n?)
                 # $$win_contents =~ s@\r@\n@mg;
+                
+                # remove solitary, stray carriage returns \r in the linux doc.xml, as occurs in the tudor collection owing to the source material
+                # containing solitary carriage returns instead of linefeed
+                $$lin_contents =~ s@[\r]@@g; #$$lin_contents =~ s@[\r][^\n]@@g;
             }
             
@@ -862 +867 @@
 
         # make sure there are no stray ampersands/partial ampersands that need to be completed as < or >
-        if($strOutput =~ m/&(.{1,2})?$/) { # &lt => < or &g => >
+        if($strOutput =~ m/&(.{1,2})?$/ || $strOutput =~ m/&amp$/) { # &lt => < or &g => > or &a(m)=> & or &amp => &
         if(defined $1 && $1) {
             my $rest = $1;
-            if($rest eq "g" || $rest eq "l") {
+            if($rest =~ m/^a/) {
+                $strOutput =~ s@am?p?$@amp;@;
+            }
+            elsif($rest eq "g" || $rest eq "l") {
             $strOutput .= "t;"; # close the known tag
             }
-            elsif($1 eq "gt" || $1 eq "lt") {
+            elsif($rest eq "gt" || $rest eq "lt") {
             $strOutput .= ";";
-            }
+            }           
         } else { # & on its own
             #$strOutput = substr( $strOutput, 0, 977); # lop off the &