GitLab

That may occurs if it's run several time in the same transaction (e.g. when
EnsureArchiveUnitHasAContentOp is run after it has been called).

Closes #16684239

use module prefixed style notation.

with additional rules so that SEDA 0.2 is disabled when there are more than one
item in the history or if SEDAWhen is linked to an item, because in SEDA 0.2
custodial history is saved as texte, making those meaning less (well, even the
notion of 'history item' actually...).

Related to #16269136

Move xsd_* methods all together, with the first one being the structural ones (i.e archive, archive unit, data object).

to avoid too much switches depending on SEDA version (more to come...).

Extract most of its content to separated methods so differences between the two implementations
is clear.

Related to #16269136

Related to #16269136

[profile gen] Consider keyword reference entity to control presence of KeywordReference tag in SEDA 1.0 / 0.2 export

Related to #16269136

by e.g. the saem cube

Ensuring we're able to validate a sample SEDA 0.2 profile.

Functional testing will be introduced by a later cset.

Related to #16269136

Functional testing will be introduced by a later cset.

Related to #16269136

* ContentDescription should always appears for first level archive unit but not
  necessarily for sub-levels

* LatestDate should appear before OldestDate in content description

* but language should occur first anyway

Merge xsd_content_description implementations along the way, not worth for a
single changing line.

to control appearance of the mimeCode attribute which was previously always
forbidden.

additionaly to Identification. This comes from a sample profile used in SAEM
tests.

by looking for the closes rule and climbing up to the transfer node.

Closes #15927695

(i.e. if data_object_reference can't be found).

This avoid clash with `simplified_profile` predicate breaking the data object
creation form because `is_typed_reference` had unexpectidly the highest score.

cw 3.24.1 has been released

They should only appear in the main tab.

This is not necessary to extend cw primary/tab views'selector, the specific one
is enough

Closes extranet #12350128

At this point we expect a container as in its sibling '_render_triggers' method.

In such case, there are no 'seda_*' relations in the rql since we're not in the
context of an archive transfer which hold them.

Find a suitable concept scheme for encoding/mime type/algorithm relations of a component archive unit

Former constraint was relying on a 'code list version' relation which was setup
on the archive transfer, which a component archive unit doesn't have. To find a
suitable scheme, we may still find the one that is linked to the correct entity
and relation type in the schema (at least until we've several of those, which is
not the case for now).

To do so, add a new `SCHEME_FROM_CONTAINER` structure in xsd2yams (close to
where constraints are defined) so one may easily get the correct query depending
on the container type (ie. wether we are in an archive transfer or an archive
unit).

Also, extend tests to check both case for the 3 aforementionned relations, since
the rql constraint deserves to be tested...

Well, technically an archive unit sequence.