Changeset 27969 for other-projects/nightly-tasks

Timestamp:

2013-08-05T15:28:32+12:00 (11 years ago)

Author:

ak19

Message:

When test-running diffcol on the Enhanced-PDF tutorial, it would choke on intermediate ~ files, the cached directory (see previous commit), and the FileSize metadata could differ slightly between OS. Further, the XSLT processing of the report would break if error messages were truncated at an incomplete entity (& symbol) since it becomes invalid xml.

Location:

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

Files:

• diffcol.pl (modified) (5 diffs)
• gdbdiff.pm (modified) (1 diff)

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

@@ -576 +576 @@
         # Files to be skipped because they get generated on one OS but not the other
         # On windows, files of the form col.invf.state.\d\d\d\d get generated (e.g. Small-HTML.invf.state.1228) that aren't there on linux
-        my $skipfiles_re = qr/\.invf\.state\.\d+$/; # Create a regex of all files to be skipped, see http://perldoc.perl.org/perlop.html
+        my $skipfiles_re = qr/(\.invf\.state\.\d+$)|~$/; # Create a regex of all files to be skipped, see http://perldoc.perl.org/perlop.html
         @aryInModel = grep { $_ !~ m/$skipfiles_re/ } @aryInModel; # http://stackoverflow.com/questions/174292/what-is-the-best-way-to-delete-a-value-from-an-array-in-perl
         @aryInTest = grep { $_ !~ m/$skipfiles_re/ } @aryInTest;
@@ -592 +592 @@
                 my $strNewTest = &FileUtils::filenameConcatenate($strTest,$strEachFile);
                 # now additionally ignoring the earliestDatestamp file and the index/idx/*.idh binary file when diffing file 
-                if(!($strEachFile eq "log" || $strEachFile eq "earliestDatestamp" || $strEachFile =~ m/\.cfg$/g || $strEachFile =~ m/\.((g|j|l|b)db|idh|i.*|wa|td|tsd|ti|t|tl|w|jpe?g|gif|png|wmf)$/g)) # wmf = windows meta file
+                if(!($strEachFile eq "log" || $strEachFile eq "earliestDatestamp" || $strEachFile =~ m/\.cfg$/g || $strEachFile =~ m/collect\.bak$/g || $strEachFile =~ m/\.((g|j|l|b)db|idh|i.*|wa|td|tsd|ti|t|tl|w|jpe?g|gif|png|wmf)$/g)) # wmf = windows meta file #  || $strEachFile =~ m/\~$/g to get rid of ~ files
                 {
                     push(@Errors,TestEach($strNewModel,$strNewTest,$intLevel,$strColName));
@@ -612 +612 @@
                 my $strNewModel = &FileUtils::filenameConcatenate($strModel,$strEachFile);
                 my $strNewTest = &FileUtils::filenameConcatenate($strTest,$strEachFile);
-                if(!($strEachFile eq "log" || $strEachFile eq "earliestDatestamp" || $strEachFile =~ m/\.cfg$/g || $strEachFile =~ m/\.((g|j|l|b)db|idh|i.*|wa|td|tsd|ti|t|tl|w|jpe?g|gif|png|wmf)$/g))
+                if(!($strEachFile eq "log" || $strEachFile eq "earliestDatestamp" || $strEachFile =~ m/\.cfg$/g || $strEachFile =~ m/collect\.bak$/g || $strEachFile =~ m/\.((g|j|l|b)db|idh|i.*|wa|td|tsd|ti|t|tl|w|jpe?g|gif|png|wmf)$/g)) #  || $strEachFile =~ m/\~$/g to get rid of ~ files
                 {
                     push(@Errors,TestEach($strNewModel,$strNewTest,$intLevel,$strColName));
@@ -626 +626 @@
     else
     {
-        my $ignore_line_re = "<Metadata name=\"(lastmodified|lastmodifieddate|oailastmodified|oailastmodifieddate|ex.File.FileModifyDate|ex.File.FilePermissions|ImageSize)\">.*</Metadata>\\s*\\n*";    
+        my $ignore_line_re = "<Metadata name=\"(lastmodified|lastmodifieddate|oailastmodified|oailastmodifieddate|ex.File.FileModifyDate|ex.File.FilePermissions|ImageSize)\">.*</Metadata>\\s*\\n*";
         
         my $strResult;
@@ -859 +859 @@
 
       if ( length( $strOutput ) > 1000 ) {
-            $strOutput = substr( $strOutput, 0, 978) . "... (output truncated)";
+            $strOutput = substr( $strOutput, 0, 978);
+
+        # make sure there are no stray ampersands/partial ampersands that need to be completed as &lt; or &gt;
+        if($strOutput =~ m/&(.{1,2})?$/) { # &lt => &lt; or &g => &gt;
+        if(defined $1 && $1) {
+            my $rest = $1;
+            if($rest eq "g" || $rest eq "l") {
+            $strOutput .= "t;"; # close the known tag
+            }
+            elsif($1 eq "gt" || $1 eq "lt") {
+            $strOutput .= ";";
+            }
+        } else { # & on its own
+            #$strOutput = substr( $strOutput, 0, 977); # lop off the &
+            $strOutput .= "gt;"; # 50% chance the closing tag is what was missing (else can make this &amp;)
+            # but even so, when the xslt is applied to report it doesn't break as long as & is not left dangling
+        }
+        }
+        $strOutput .= "... (output truncated)";
       }
 

other-projects/nightly-tasks/diffcol/trunk/diffcol/gdbdiff.pm

@@ -66 +66 @@
     # The total_numbytes field can vary depending on how many backslashes exist in the urls in the main body text, as each
     # of these windows slashes get escaped with another backslash, and the resulting string is used as key into rel link db
-    my $ignore_line_re = "\n<(lastmodified|lastmodifieddate|oailastmodified|oailastmodifieddate|ex.File.FileModifyDate|ex.File.FilePermissions|total_numbytes)>([^\n])*";
+    my $ignore_line_re = "\n<(FileSize|lastmodified|lastmodifieddate|oailastmodified|oailastmodifieddate|ex.File.FileModifyDate|ex.File.FilePermissions|total_numbytes)>([^\n])*";
     $model_text =~ s/$ignore_line_re//g;
     $test_text =~ s/$ignore_line_re//g;