enable overflow for accu keys, to_analysis_urls, and submission_cmd_args

modules/Bio/EnsEMBL/Hive/DBSQL/AccumulatorAdaptor.pm

@@ -48,6 +48,32 @@ sub default_table_name {
 }
 
 
+sub slicer {    # take a slice of the hashref (if only we could inline in Perl!)
+    my ($self, $hashref, $fields) = @_;
+
+    my $overflow_limit = $self->overflow_limit();
+
+    return [ map { eval { my $value = $hashref->{$_};
+                          my $ol = $overflow_limit->{$_};
+                          if (defined($ol) and defined($value) and  (length($value) > $ol)) {
+                              $self->db->get_AnalysisDataAdaptor()->store_if_needed($value);
+                          } else {
+                              $value;
+                          }
+                   }
+
+
+             } @$fields ];
+}
+
+sub overflow_limit {
+    return {
+        'key_signature' => 255,
+        'struct_name'   => 255,
+    };
+}
+
+
 sub fetch_structures_for_job_ids {
     my ($self, $job_ids_csv, $id_scale, $id_offset) = @_;
     $id_scale   ||= 1;
@@ -63,6 +89,9 @@ sub fetch_structures_for_job_ids {
 
         ROW: while(my ($receiving_job_id, $struct_name, $key_signature, $stringified_value) = $sth->fetchrow_array() ) {
 
+            ($key_signature, $struct_name) = map {$self->check_and_dereference_analysis_data($_)}
+                ($key_signature, $struct_name);
+
             my $value = destringify($stringified_value);
 
             my $sptr = \$structures{$receiving_job_id * $id_scale + $id_offset}{$struct_name};

modules/Bio/EnsEMBL/Hive/DBSQL/BaseAdaptor.pm

@@ -530,6 +530,31 @@ sub store {
 }
 
 
+=head2 check_and_dereference_analysis_data
+
+   Arg [1]     : string data that may reference an analysis_data_id
+   Usage       : my $value = $self->check_and_dereference_analysis_data($$fetched_row[0])
+   Description : Checks to see if the passed value matches the regular expression
+               : /^_ext(?:\w+)_data_id (\d+)$/ (e.g. "external_data_id 3", pointer to an analysis_data_id).
+               : If so, it returns the entry from the "data" column in analysis_data from the
+               : row containing the given analysis_data_id.
+               : If not, it returns the passed parameter unchanged.
+   Returntype  : string
+
+=cut
+
+sub check_and_dereference_analysis_data {
+    my $self = shift @_;
+    my $possible_analysis_data_id_ref = shift @_;
+
+    if ($possible_analysis_data_id_ref  =~ /^_ext(?:\w+)_data_id (\d+)$/) {
+        return $self->db->get_AnalysisDataAdaptor->fetch_by_analysis_data_id_TO_data($1);
+    } else {
+        return $possible_analysis_data_id_ref;
+    }
+}
+
+
 sub DESTROY { }   # to simplify AUTOLOAD
 
 sub AUTOLOAD {

modules/Bio/EnsEMBL/Hive/DBSQL/DataflowRuleAdaptor.pm

@@ -55,6 +55,13 @@ sub default_insertion_method {
 }
 
 
+sub overflow_limit {
+    return {
+        'to_analysis_url' => 255,
+    };
+}
+
+
 sub object_class {
     return 'Bio::EnsEMBL::Hive::DataflowRule';
 }

modules/Bio/EnsEMBL/Hive/DBSQL/ObjectAdaptor.pm

@@ -58,7 +58,7 @@ sub slicer {    # take a slice of the object (if only we could inline in Perl!)
                     ? $object->dbID()
                     : eval { my $value  = $object->$_();
                              my $ol     = $overflow_limit->{$_};
-                             (defined($ol) and length($value)>$ol)
+                             (defined($ol) and defined($value) and length($value)>$ol)
                                 ? $self->db->get_AnalysisDataAdaptor()->store_if_needed( $value )
                                 : $value
                       }

modules/Bio/EnsEMBL/Hive/DBSQL/ResourceDescriptionAdaptor.pm

@@ -55,6 +55,13 @@ sub default_insertion_method {
 }
 
 
+sub overflow_limit {
+    return {
+        'submission_cmd_args' => 255,
+    };
+}
+
+
 sub object_class {
     return 'Bio::EnsEMBL::Hive::ResourceDescription';
 }