TriGParser: distinguish empty entities from no-entity being read #42
Conversation
|
See #37 (comment) What is your (@k00ni, @pietercolpaert ) preferred way to deal with that? I can either adjust the parser to consistently return empty subject/predicate/object on "diamonds" ( |
|
Adjusting the test to expect an exception takes my preference |
* removed the prefixed-only IRIs input line from the first test scenario as this does not belong to the testBlankNodes() tests and is tested aleady in testIssue37() * turned the empty prefixed IRIs test scenario into two - first, where and error is expected due to unknown document base IRI and second, where parsing succeeds thanks to `documentIRI` parser option being set
zozlak
changed the title
There was a problem hiding this comment.
👍 Thank you @zozlak, looks good to me, just a few remarks.
I also support @pietercolpaert's preference.
Off topic: The current coding style should be addressed in the future, because it makes code (especially the tests) very hard to read. Therefore only a few remarks in this regard.
Solves the #37 by throwing an informative exception describing why the parsing can not be done without the document base IRI being set.