GitLab

- syntaxtree.py: transformation_factory decorator for processing table functions; further type annotations

- syntaxtree.py: reversed parameter order in transformer-functions: node is now always the first paramter again.

- can use operators "+" (Sequence) and "|" (Alternative) for constructing parsers in Python; better testing suppoer

- example MLW extended by adding rudimentary testcases

syntaxtree, class Node: XML-output now puts tags and content of the innermost node on one line for better readability

- virtualized Parser.__class__.__name__ by introducing Parser.ptype (which by default is set to Parser.__class__.__name__)
- removed all isinstance(node, ...) statements from ebnf.EBNFCompiler, because this does not work with mock trees which are be needed for testing.

split compileEBNF into two functions: compileEBNF and compile_parser; adjustments for compatibility with pypy3 / python 3.5

- better syntaxtree.mock_syntax_tree() generation: Can now read actual ASTs in S-expression form. Useful for testing of AST-transformation or -validation as long as these do not call isinstance() on any parser.

tests/test_syntaxtree.py to syntaxtree.py

- potential infinite loops are now caught when running parser and an error is reported; bug fixes, tests

- DSLsupport.run_compiler now produces callable DSL compiler script stubs; code reorganization all DHParser modules moved to a subdirectory containing the DHParser package; dhparser.py script remains in the main directory

- renamed parser.py to parsercombinators.py to avoid possible name conflicts with the parser module from the Python standard library