Remaining pruneBlankNodeIdentifiers test options.

[davidlehn]

frame-manifest.jsonld still has a few tests with the "pruneBlankNodeIdentifiers": false option. This was removed in #629. How should this be handled? @dlongley suggested the flag may be useful in 1.0 mode. Should those tests stay around and be marked as 1.0? If so then it seems implementations would have to know how to handle a flag which has no spec text. If the tests are removed then how do implementations test that flag?

[gkellogg]

No, the flag’s gone, so we should make these depend on the processing mode, which can be set in the manifest.

Implementations May keep such a flag, but the CG manifest won’t reflect this.

[gkellogg]

These are removed in PR #643.