Consider splitting errors into tokenizer and generator errors

[josecelano]

After splitting the parser into Tokenizer and Generator,, I think it also makes sense to split the errors in error module.

We can create two new error modules inside generators and tokenizer modules, and use an enum the main error module.