parse ifc model use too many memory

[wen-bing]

When testing with the sample model:
BIMIT-Sample-Model

the model is 22MB , the parser used almost 1.2 GB memory.

Test with some real project 's model : a 93MB ifc model, the parser used almost 6GB memory, 220MB model used almose 10GB memory. according to antlr4's doc, it seems need to optimize STEP.g4 to improve the performance.

[ikeough]

How are you measuring this? I've got that sample model in the test set now, but I only ever run it through the unit testing framework. Might be good to add some memory tests.