When analyzer fails, seeing the intermediate parse trees generated is useful.
1 - parse trees stop being generated before the error.
2 - parse trees can help debug the problem, as error often occurred a few passes earlier.
3 - you can look at the state of the parse just before the error occurred.
4 - you can see where a parse tree becomes corrupted.
@ddehilster
Well, going to the directory/file listing, the intermediate trees are there for examination.
So this is low priority, if necessary at all to put into the NLP++ side.
When analyzer fails, seeing the intermediate parse trees generated is useful. 1 - parse trees stop being generated before the error. 2 - parse trees can help debug the problem, as error often occurred a few passes earlier. 3 - you can look at the state of the parse just before the error occurred. 4 - you can see where a parse tree becomes corrupted. @ddehilster