[PATCH 1 of 7 saem_ref V2] [seda] Configure the seda compound tree to stop considering Activity

Denis Laxalde denis.laxalde at logilab.fr
Mon Apr 10 15:33:40 CEST 2017

Sylvain Thenault a écrit :
> # HG changeset patch
> # User Sylvain Thénault <sylvain.thenault at logilab.fr>
> # Date 1490876699 -7200
> #      Thu Mar 30 14:24:59 2017 +0200
> # Node ID b8599a52fa6dd09ce1b4457840a9ecb3179a1eb5
> # Parent  fe766631d60a999d4d7338169a3404bacdb23cd1
> [seda] Configure the seda compound tree to stop considering Activity
> We've already fixed permission issues from the same problem (Activity being
> added by saem_ref and in several compound trees). Now, it appears that 'used'
> and 'generated' relations appear in the SEDA_PARENT_RTYPES structure of
> seda's hooks, leading to a container relation being set for activities even when
> they belong to a scheme, concept or authority record, even though this relation
> doesn't exists in the schema. This causes later problem, such has being unable
> to remove those entities (hence the non regression test).
> Instead of simply fixing SEDA_PARENT_RTYPES, make the seda compound tree
> configurable and skip activity once for all so we may expect no more problems
> because of this.
> A migration script is added to cleanup the 'container' relation by removing all
> records where subject of the relation is one of the AuthorityRecord, ConceptSchema
> and Concept types (only SEDA* types should have the container relation). This
> is done using SQL and not RQL since those records can't even exist according to
> the schema.

I still don't quite understand why things "can't even exist according to 
the schema". Not sure I actually want to understand this though...

More information about the saem-devel mailing list