...
- A. Soroka
- out sick
- Eric James
- Running into problems that should be documented
- Can update the config, but that doesn't change what's shown by the repo – need to refresh. Is there a simple way to refresh the repository?
- Will add errors to the wiki
- Content changes show up in the repo, but browser doesn't find it reliably because of caching issues, will write IT to demonstrate this
- Running into problems that should be documented
- Unknown User (escowles@ucsd.edu)
- Clustering worked once firewall was relaxed to allow nodes to talk to each other
- Working through triplestore use cases
- Most look like they are supported by what we have now
- Will expand SPARQL Recipes to cover more of the use cases
...
- Eric James
- Eclipse bugs yesterday, switching to IntelliJ. Trying to get IT to verify federated content updates when filesystem is updated.
- When changing repository.json and restarting repo, it's still using the old config.
- Adam: Mode may be storing projection info in repository and not updating it for some reason.
- A. Soroka
- We are running into the limitations of JCR content type system, may need to do something else (implement our own, adopt a different one)
- There are lots of interesting possibilities, but they aren't practical to implement quickly.
- SPARQL may be a better strategy: in a transaction, a user could create content, run SPARQL queries to validate, and then commit or rollback based on that validation. Could add a single-step process for doing that automatically.
- Will put on wiki/mailing list for feedback.
- Unknown User (escowles@ucsd.edu)
- Finished reviewing triplestore use cases and found them well-supported. Updated SPARQL Recipes and RESTful HTTP API - ObjectsContainers to demonstrate some functionality, like updating multiple objects with a single SPARQL Update, using more vocabularies, etc.
- Will investigate Mulgara-style recursion functionality. This is not available in standard SPARQL, but Jena and/or Sesame may have similar extensions.
- Adam: May be able to get hierarchy from repo, and then build a simple SPARQL query.
- Will try to replicate Eric's federation config reloading issue.
...
- Unknown User (escowles@ucsd.edu)
- Fixing SPARQL Recipes (some objects weren't getting indexing enabled), and getting the recursive collection membership query working correctly (and simplifying it too)
- Testing filesystem federation and data copying options (uploading to filesystem via REST API, copying between repo and federated filesystem)
- Uploading to federated filesystem works, but messes up properties
- Will file a bug on that and start trying to track it down
- Eric James
- Trying to get auth working with federated filesystem
- Docs may be out of date, auth delegate config (docs still call it PEP)
- Andrew: should look at auth.war for config, and email ff-tech/Mike Daines about the docs.
- Andrew Woods
- David Wilcox
- Working on feature docs messaging
- Will coordinate with Andrew acceptance testing for large file support and external triplestore features
- A. Soroka
- Snow day
3/18
- A. Soroka
- Just getting back and starting to fix issues that Andrew had with PR.
- Want to schedule special topics call for validation.
- Andrew Woods
- Talked to Kai Sternad about properties updates in transactions and merged in his update.
- Also talked to Kai about some other transaction issues.
- Looked at federation issues and agrees with Adam that there is a MODE issue here, but maybe we can figure out how to handle this or we are doing something wrong. Will continue to look at this.
- Merged in work from Mike Daines for user groups via HTTP headers and updated docs for how to wire that in.
- Working on 4.0 feature set docs.
- David Wilcox
- Will be contacting users for doing acceptance testing of large file handling and external datastreams.
- Eric James
- Working on getting auth up and running with federated filesystem.
- Using new config docs in the wik, which is running but not completely working. Trying to track that down.
- Andrew: make sure to ask for help if that isn't working
- Unknown User (escowles@ucsd.edu)
- Working on writing to filesystem federation using REST API. The repository is writing links to versions that don't exist and this is breaking reading the node metadata.
- We should probably have a supertype of fedora:resource that's not versionable, but this may be problematic. Working on more limited approach to just trap the errors and continue with RDF generation instead.
- Adam: We should be able to turn off the generation of those triples. Or we can remove the mix:versionable mixin from fedora:resource, which would require explicitly adding the versionable mixin if desired (which is how it worked previously). We may also want to ask about adding a method to the connector interface for querying whether a connector supports versioning.
3/19
- Andrew Woods
- Reviewing code updates and merging them.
- Working on 4.0 feature documentation.
- Trying to wrap up federation configuration updates.
- Lots of meetings today and tomorrow.
- Would like to get a release out before traveling on Sunday, so we need to get code changes finalized early enough to make that happen.
- David Wilcox
- Emailed people who submitted use cases for large files and external datastreams.
- Expect testing to go forward over next few weeks with new 4.0 alpha release.
- Will work with Andrew on messaging.
- Eric James
- Working on federation with authorization.
- Found issue with root node access not having any roles, and opened PR to fix that.
- Next step is setting up two federated directories and keeping one open and locking down the other one.
- Looking into how to add mixin to federated node using REST API to set access roles.
- A. Soroka
- Working on content modeling agenda for tomorrow, trying to engage with some people outside of the core group to broaden input.
- Will work with transform module.
- Unknown User (escowles@ucsd.edu)
- Updating to use FedoraFileSystemConnector instead of Modeshape's FileSystemConnector.
- Testing copying between repo and federated filesystem.
- Versioning documentation changes.
- Will continue with testing large files and clustering, and help with federation configuration issues.