parsimonious

This project publishes no release notes. Between 0.9.0 and 0.10.0 there were 21 commits, 19 of them substantive:

• Faster node cache
• Ensure all LazyReference objects are resolved
• Remove unnecessary if-statement
• Pythonic code: itervalues(rule_map) -> rule_map.values()
• Binary grammars and string literal precedence fix (#202)
• Fix RST formatting in README
• Add a failing test for #134.
• Add release note.
• Fix star token grammar (#203)
• Simplify reference resolution. (#205)
• Add note to release log.
• remove some redundant code in Sequence expression (#176)
• Modernize Syntax for python 3 (#206)
• Use --tb=native to stack traces less verbose.
• Add ability to run particular tests.
• Use raise from to improve exception messages.
• Infinite recursion left associative (#211)
• fixed recursion errors of Expressions.__eq__ (#212)
• \0.10.0 Release (#214)

This project publishes no release notes. Between 0.8.1 and 0.9.0 there were 50 commits, 48 of them substantive:

• Fix Grammar.repr backslash escaping
• Add ascii flag
• Remove py3.3 in travis-ci
• Remove Python 3.3 from the trove classifiers.
• Merge dropping of 3.3 support, which should fix Travis. Close #147. Fix #146.
• Add ASCII flag to regex expressions.
• Update README.rst
• Update README.rst
• Merge in an example for the readme. Close #150.
• Added two ticks for the NodeVisitor class and key, values as variables in visit_entry(). Close #151.
• Support functions, class methods, and descriptors for custom rules
• Migrate callability tests & remove redundant grammar tests
• Move to function-style print in README
• Travis CI: Remove py34 add py37, py38
• Fixup: ensure method-descriptor-binding is used during illustrative descriptor test case (h/t @evanh)
• Fixup: update callability test case
• Nit: rename dictionary key from 'descriptor' to 'descriptor_rule' for readability, particularly in callability unit test
• Merge removal of Python 3.4 from CI and addition of 3.7 and 3.8. Close #165.
• Merge function-style print in readme. Close #158.
• Consistency: use single-quoted key names during dictionary access

…and 28 more.

This project publishes no release notes. Between 0.8.0 and 0.8.1 there were 2 commits, 1 of them substantive:

• use print function for Python 3 compatibility

This project publishes no release notes. Between 0.7.0 and 0.8.0 there were 40 commits, 39 of them substantive:

• Remove 3.2 from the trove classifiers, since we can't get it to run under tox anymore.
• Type check when comparing nodes
• Better nose based test
• No need for the None check anymore
• Merge fix for crash when comparing Nodes with instances of other classes. Close #98.
• Remove duplicate VisitationError import.
• Remove duplicate VisitationError import. Close #100,
• Expose BadGrammar and VisitationError as top-level imports. Fix #101.
• Drop python 2.6 support and add python 3.6 to the build matrix.
• Remove StrAndRepr from classes which define their own repr
• Fix a bug in unicode encoding of reprs.
• Add failing test that rule ordering is preserved.
• Make visitor parse rules in order.
• Remove direct calls to dict()
• Add failing test that rule ordering is preserved on shallow copies.
• Fix ordering of shallow copy.
• Have Grammar inherit directly from OrderedDict.
• Remove unused imports.
• pin six>=1.9.0 for python_2_unicode_compatible
• Merge six pinning at 1.9.0. Close #109.

…and 19 more.

This project publishes no release notes. Between 0.6.2 and 0.7.0 there were 35 commits, 29 of them substantive:

• Added the ability of parsimonious to parse lists of tokens.
• Merge master in to get up to date.
• Clean up Token.
• Simplify design of token-based parsing.
• Slightly improve tests, and fix a TokenMatcher bug that revealed. Move Token to utils.
• Merge experimental token-based parsing, via TokenGrammar class, for those operating on pre-lexed streams of tokens.
• Remove what turns out to be a useless slots definition from TokenMatcher, add a useful one to Token, and confirm with a test that slots does not behave in CPython according to the language docs.
• Fix #53: Make Grammar an immutable mapping.
• Drop Python 3.1 support.
• Add .travis.yml file for travis support.
• Fix #16: Keep parenthesis when printing expressions.
• Use io.open to read README.rst
• Fix spelling of a few comments
• README: Syntax highlight Python samples
• Add Py3 compatibility.
• Update some glue code and fix some bugs
• Add a test for repr py3k incompatibility
• Fix string_escape bug
• Update travis.yml to use tox-travis (hopefully).
• Drop support for python 3.2 due to pip incompatibility

…and 9 more.

This project publishes no release notes. Between 0.6.1 and 0.6.2 there were 1 commits, 1 of them substantive:

• Speed up grammar compilation by 100x. Close #62. Bump version to 0.6.2.

This project publishes no release notes. Between 0.6 and 0.6.1 there were 1 commits, 1 of them substantive:

• Fix bug which made the default rule of a grammar invalid when it contained a forward reference. Bump version to 0.6.1. Fix #61.

This project publishes no release notes. Between 0.5 and 0.6 there were 24 commits, 23 of them substantive:

• Refactor RuleVisitor._resolve_refs
• Improve exception message when you forget to declare a visitor method.
• Make clear that the grammar is a string, not some magical pile of operator overloading.
• Show the production visit_or_term unpacks.
• Correct visit_or_term example.
• Improve NodeVisitor documentation.
• Add parse() and match() convenience methods to NodeVisitor, and add the concept of a default grammar for a visitor.
• Implement grammar-from-docstrings visitors. Close #46. Consolidate public API in init.
• Improve some Grammar docs.
• Tersen just a little bit. Make members consistently a tuple, not sometimes None.
• Drastically simplify _resolve_refs(). Close #34.
• Stop relying on the object identity of multiple instantiations of ().
• Make Grammars extensible through custom Python code using "custom rules". Close #47.
• Replace GrammarFromDocstrings with @rule. Everything works fine except subclassing; I haven't figured that out yet.
• Bring readme up to date.
• Move custom rules to a custom kwarg.
• Make _match() public. Close #52.
• Make Compound public, in case anyone wants to make custom Expressions based on it.
• Replace default_rule kwarg with .default(). Refs #30.
• Make UndefinedLabel no longer a kind of VisitationError.

…and 3 more.

This project publishes no release notes. Between 0.4 and 0.5 there were 12 commits, 11 of them substantive:

• Work around tox bug where it sees that whole comment as part of the dir name.
• Support multi-line rules, comments in the middle of rules, and all other sorts of horrific behavior. Fix #19.
• Add error reporting. Fix #28.
• Factor out some duplication in parse(), and tweak some comments and docs.
• Add pos kwarg to Grammar.parse as well.
• Add line and column reporting to ParseError.
• Tweak a few things in the readme to be more encouraging.
• Come up with a benchmarking framework, and add a basic speed test.
• Tolerate whitespace after opening parens.
• Add support for single-quoted literals. Ref #20.
• Besmirch my error reporting, which isn't great, but I want to release for the other fixes.