Issue 415: Use sh:memberShape in SHACL SHACL
#417
Conversation
jeswr
changed the title
sh:memberShape in SHACL SHACLsh:memberShape in SHACL SHACL
jeswr
force-pushed
the
wip/shacl-shacl-memberShape
branch
from
44c396a to
eaf5d95
Compare
| sh:node shsh:ListShape ; # xone-node | ||
| sh:memberShape shsh:ShapeShape ; # xone-memberShape |
There was a problem hiding this comment.
Interesting. The test has caught up with an uncaught corner case of the spec.
Section 4.6.4 of the 1.0 spec., consistent with today's draft state, reads:
sh:xonespecifies the condition that each value node conforms to exactly one of the provided shapes.
However, the remainder of that section doesn't talk about whether sh:xone can or can't be empty. This shape in SHACL-SHACL is consistent with that lack of detail, but the validation for a user with any targeted nodes would always fail, because there isn't exactly one match.
Thinking about it a bit, this seems like it would be a backwards-incompatible change to add a sh:minListLength 1 here that fails a SHACL-SHACL run, because the sh:xone subject could be defined with no targets and sneak in as valid against SHACL-SHACL of SHACL 1.0.
Thoughts on whether we try fixing this now, or adding a warning-severity shape in SHACL-SHACL?
There was a problem hiding this comment.
I would suggest adding text to the 1.2 spec requiring at least one shape - and reflecting that here with a minListLength of 1.
There was a problem hiding this comment.
Drafting an explanatory comment, I now think this is better left delegated to a follow-on Issue. Please leave this thread unresolved for reference.
jeswr
force-pushed
the
wip/shacl-shacl-memberShape
branch
2 times, most recently
from
2f88f1d to
2ad5655
Compare
jeswr
force-pushed
the
wip/shacl-shacl-memberShape
branch
from
2ad5655 to
8ffae40
Compare
…:xone
This patch revises part of the implementation in PR 417.
* `in-minListLength` was not added to the Core document, but it was
referenced in SHACL-SHACL. Verbiage is added to Core in line with the
text drafted for Issue 429.
* `xone-memberShape` was added as a reference in SHACL-SHACL, but
doesn't exist in the Core document. `xone-members-node` is the anchor
in the Core document. `xone-members-node` is now used.
* The `PropertyShape` for `sh:in` set a minimum required length of 1,
which after discussion in Issue 429 appears to be a backward-
incompatible change. A shape paralleling the original suggestion for
`sh:xone` is added to house that new restriction at the same warning-
level severity.
An implementation style point: The property shapes housing
`sh:minListLength` are given IRIs after sketching a test
(`.../in-003.ttl`) and assuming it would be more desirable to have an
IRI-identified shape in the results graph, rather than a blank node.
A testing style point: Rather than running SHACL-SHACL over the whole of
a test, shapes from SHACL-SHACL are copied into some focused tests
(`.../{in,xone}-003.ttl`).
References:
* #429
* #417
Signed-off-by: Alex Nelson <[email protected]>
2 participants
Closes #415