Fix for issue #677: Set the back reference

[danez]

When calling add...() on an BaseObject to add a crossreference, the generated code now sets the backreference on the Object that is given to the add function.
This ensures that all references are already set when the object is still not saved.

On the other hand I add a clone in the hasCode function to ensure that the the hasCode function never ever changes the original request as it is not necessary.

[willdurand]

looks good to me. @staabm?

[staabm]

@danez did the tests fail without the patch, otherwise LGTM.

You should always create the test first. This eases handling on our side and also makes sure that the fix works correctly.

[danez]

Okay sorry, next time i commit the tests separately first.
Without the fixes in generated object and BaseObject::hashCode:

1) GeneratedObjectRelTest::testManyToManySetterIsNotLoosingAnyReference
setRelCol is losing backreference on set objects
Failed asserting that actual size 0 matches expected size 1.

/Users/user/Propel/test/testsuite/generator/builder/om/GeneratedObjectRelTest.php:205

2) BaseObjectHashCodeTest::testUnsavedObjectCallingHashCodeIsNotChangingObject
Failed asserting that two objects are equal.
--- Expected
+++ Actual
@@ @@
     'age' => null
-    'collBooks' => null
+    'collBooks' => PropelObjectCollection Object (...)
     'collBooksPartial' => true
     'collEssaysRelatedByFirstAuthor' => null
     'collEssaysRelatedByFirstAuthorPartial' => null
     'collEssaysRelatedBySecondAuthor' => null
     'collEssaysRelatedBySecondAuthorPartial' => null
     'alreadyInSave' => false
     'alreadyInValidation' => false
     'alreadyInClearAllReferencesDeep' => false
     'booksScheduledForDeletion' => null
     'essaysRelatedByFirstAuthorScheduledForDeletion' => null
     'essaysRelatedBySecondAuthorScheduledForDeletion' => null
     'validationFailures' => Array ()
     '_new' => true
     '_deleted' => false
     'modifiedColumns' => Array (...)
     'virtualColumns' => Array ()
 )

/Users/user/Propel/test/testsuite/runtime/om/BaseObjectHashCodeTest.php:37

[danez]

I changed the order of the parameters so in the test the actual and expected values are correct.

[willdurand]

thanks!

[staabm]

@danez no problem, just a hint to make things better next time. thanks for your contribution!