Closed ajnelson-nist closed 2 years ago
If I'm correct, the wget on
...rdf-toolkit-build/lastSuccessfulBuild/...
might pull a new version in the future. This demands the new rdf-toolkit version to be backwards compatible with the current one to prevent the tool to potentially crash.At the same time, constraining the toolkit with a SHA512 to one single version only, prevent new versions to be applied, and might make the tool to stall for different builds.
I assume this to be the intended behaviour.
@plbt5 , you're right that this was the behavior I intended to encode. This download strategy hasn't changed since CASE's initial implementation of it.
However, you've made me realize that it assumes stability in the download, and that assumption is incorrect. The hard-coded SHA-512 "Pins" the expected JAR file to today's release. So, if a new release happens, the download will fail to verify.
I'll push another patch to revise the strategy to use CASE's redistribution rather than retrieving the original file. I think a preferable remediation would be accessing a version somehow "Pinned" upstream, but that's out of scope of this PR.
The update changes all ordered lists into sets of IRIs. The application can only select one set element.
I'm not quite sure I follow this remark - the syntactic style of the ordered lists is changed, but they have not been converted into sets. ( ... )
is Turtle shorthand for an RDF List. Or is there something I'm missing?
Closing as superseded by PR 398.
This Pull Request will resolve all requirements of Issue 373.
This PR adjusts the strategy of v1, PR 386, due to issues having been found when running a normalized file through
rdf-toolkit
multiple times. Issue documentation is on CASE-Examples PR 75.pre-commit
usage is removed from v2 of resolving Issue 373.Review steps taken: