From: Peter Rabbitson <ribasushi@cpan.org>
Date: Fri, 13 May 2016 17:15:18 +0000 (+0200)
Subject: Fully separate parent and child resultsource metadata
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=9e36e3ec;p=dbsrgits%2FDBIx-Class.git

Fully separate parent and child resultsource metadata

Ensure that all attributes are shallow-copied on "clone". Currently this
means that the following are *NO LONGER* shared between rsrc clones:

_unique_constraints
_primaries
source_info

This seems just as cocky and reckless as the clusterfuck in 4006691d/RT#107462
and it most likely will have the same invisible-yet-dire consequences for
various downstreams. However there is a plan in place several commits ahead
allowing sidestepping the impossibility to debug a potential fallout.
---

diff --git a/Changes b/Changes
index c737569..2abdb6e 100644
--- a/Changes
+++ b/Changes
@@ -10,6 +10,9 @@ Revision history for DBIx::Class
           the maintainer believe this is safe, but this is a very complex
           area and reality may turn out to be different. If **ANYHTING** at
           all seems out of place, please file a report at once
+        - The unique constraint info (including the primary key columns) is no
+          longer shared between related (class and schema-level) ResultSource
+          instances
         - Neither exception_action() nor $SIG{__DIE__} handlers are invoked
           on recoverable errors. This ensures that the retry logic is fully
           insulated from changes in control flow, as the handlers are only
diff --git a/lib/DBIx/Class/ResultSource.pm b/lib/DBIx/Class/ResultSource.pm
index 053b398..4ebb4c0 100644
--- a/lib/DBIx/Class/ResultSource.pm
+++ b/lib/DBIx/Class/ResultSource.pm
@@ -16,12 +16,18 @@ use DBIx::Class::ResultSet;
 
 use namespace::clean;
 
-__PACKAGE__->mk_group_accessors(simple => qw/
-  source_name name source_info
-  _ordered_columns _columns _primaries _unique_constraints
-  _relationships resultset_attributes
-  column_info_from_storage sqlt_deploy_callback
-/);
+my @hashref_attributes = qw(
+  source_info resultset_attributes
+  _columns _unique_constraints _relationships
+);
+my @arrayref_attributes = qw(
+  _ordered_columns _primaries
+);
+__PACKAGE__->mk_group_accessors(simple =>
+  @hashref_attributes,
+  @arrayref_attributes,
+  qw( source_name name column_info_from_storage sqlt_deploy_callback ),
+);
 
 __PACKAGE__->mk_group_accessors(component_class => qw/
   resultset_class
@@ -149,9 +155,10 @@ Creates a new ResultSource object.  Not normally called directly by end users.
     $self->{sqlt_deploy_callback} ||= 'default_sqlt_deploy_hook';
 
     $self->{$_} = { %{ $self->{$_} || {} } }
-      for qw( _columns _relationships resultset_attributes );
+      for @hashref_attributes;
 
-    $self->{_ordered_columns} = [ @{ $self->{_ordered_columns} || [] } ];
+    $self->{$_} = [ @{ $self->{$_} || [] } ]
+      for @arrayref_attributes;
 
     $self;
   }