Changeset 38071 for other-projects/nightly-tasks/diffcol

Timestamp:

2023-09-09T22:23:24+12:00 (8 months ago)

Author:

anupama

Message:

CDS-ISIS test collection diff issues on Mac: CAPS suffixes were generated for filenames listed in archiveinf-src and archiveinf-doc of Mac test CDS-ISIS collection, and didn't compare successfully against its Linux model collection. So some extra processing to normalise the Mac collection's archiveinf files.

File:

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

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

@@ -118 +118 @@
     
     if($testIsWin == $modelIsWin) {     
-    # both linux or both windows, do the basic test we did on linux machines:
+    # both unix or both windows, do the basic test we did on linux machines:
             # ignore absolute path prefixes in modelcol and testcol (necessary for archiveinf-doc and -src.gdb files)
 
@@ -133 +133 @@
             $test_text =~ s@^([^\\/]*(//)*).*(\\|/)(collect(\\|/)$strColName)(.*)$@$1$4$6@mg;
             #$model_text =~ s@^([^\\//]*).*(\\|/)(collect(\\|/)$strColName)(.*)$@$1$3$5@mg;         
-            #$test_text =~ s@^([^\\//]*).*(\\|/)(collect(\\|/)$strColName)(.*)$@$1$3$5@mg;                      
+            #$test_text =~ s@^([^\\//]*).*(\\|/)(collect(\\|/)$strColName)(.*)$@$1$3$5@mg;
+
+            # File extensions on darwin may be listed in archiveinf-src|doc as CAPS, make them lowercase. This problem popped up in CDS-ISIS collection
+            if($test_os =~ m/darwin/ && $dbname =~ m/archiveinf-/) {
+                
+                my $collection_path = $strTestCol; # full path to a windows collection
+                my $tmp = ""; # rebuild darwin test file's set of lines after processing them one by one
+
+                for my $line (split /^/, $test_text) { # split the string into newlines
+                if($line =~ m/\@THISCOLLECTPATH\@[^\.]*\.([^\.]+)/s) {
+
+                    my $file_suffix = $1;
+
+                    if(!defined $file_suffix || $file_suffix eq "") {
+                    next;
+                    }
+                    # there can be a mixture of lowercase and uppercase suffixes if test_os is darwin, so don't skip processing any line
+                    #if($file_suffix !~ m/[A-Z]+/) { # if there are all caps suffixes, it will be consistently so in the DB file, if not skip lowercasing them
+                    #last;
+                    #}
+                    
+                    my $lc_suffix = lc($file_suffix); # lowercase it
+
+                    $line =~ s/$file_suffix$/$lc_suffix/; #$line =~ s/(\@THISCOLLECTPATH\@[^\.]*\.)($file_suffix)/$1$lc_suffix/;
+                }
+                $tmp .= $line; # don't need to add a newline at end
+                }
+
+                if($tmp ne "") {
+                $test_text = $tmp; 
+                #print STDERR "NEW FILE:\n************\n$test_text\n***********\n";
+
+                # need to re- sort the keys, now that the absolute paths to tmp locations has been removed
+                # so that we get the files in the same order as the model collection                
+                
+                $test_text = text_to_db_to_text($test_text, "test.gdb");
+                }
+            }
     }
     
@@ -238 +275 @@
     
     # Windows or linux: if index is a flat db file, then ensure the docIDs listed in <contains> field of
-    # both test and model flat db file are alphabetically sorted. So to the numbers in <mdoffset> field.
+    # both test and model flat db file are alphabetically sorted. So too the numbers in <mdoffset> field.
     # Despite PERL_PERTURB_KEYS envvar being set to 0 on both machine when generating model collections
     # and when test collections were generated on test machine, still collections like Images-GPS and some