Skip to content

Latest commit

 

History

History
182 lines (142 loc) · 6.15 KB

RELNOTES.rst

File metadata and controls

182 lines (142 loc) · 6.15 KB

Picireny Release Notes

21.8

Summary of changes:

  • Dropped support for Python 2.
  • Upgraded dependency to Picire 21.8 to use new/improved argument logging, CLI argument processing; and adapted code to the updated API.
  • Heavily simplified the signatures of picireny.build_with_antlr4 and picireny.reduce.
  • Changed the API of several functions and methods, made numerous arguments keyword-only.
  • Added a new phase that applies the coarse filter to tree nodes and runs both pruning and hoisting on them.
  • Fixed HDDr to correctly traverse the tree in case of filtered nodes.
  • Fixed line-column calculations for tree nodes.
  • Fixed "skip unremovable" transformation to correctly determine the unparsed representation of nodes for all parametrizations.
  • Upgraded dependency ANTLeRinator to Epoch 1 (breaking away from ANTLR version numbering) and made use of its new feature to generate the lexer and parser from the ANTLRv4 grammar at build-time.
  • Added direct dependency on ANTLR and upgraded it to v4.9.2.
  • Made use of the inators package to unify CLI argument handling and logging.
  • Dropped runtime dependency on setuptools.
  • Moved to pyproject.toml & setup.cfg-based packaging.
  • Improved log output.
  • Improved documentation.
  • Improved the testing infrastructure (stabilized tests, improved resource handling, better output on failure, testing Windows & PyPy).
  • Various internal refactorings.

21.3

Summary of changes:

  • Introduced phases of reduction to allow executing the same HDD algorithm variant multiple times with different parametrizations (e.g., run Coarse HDDr and HDDr after each other).
  • Added a new transformation-based reduction technique called hoisting, as a new optional phase, to complement the existing pruning-based approaches.
  • Added support for "tokens" section (i.e., token names without an associated lexer rule) in grammars.
  • Added support for grammars with resource files that contain utility code or base classes of lexers and parsers.
  • Upgraded dependency to Picire 20.12 to utilize its new generalized split factor concept and updated API.
  • Upgraded dependency to ANTLR v4.9 (via ANTLeRinator).
  • Bumped minimum Python 3 requirement to 3.5.
  • Improved log output.
  • Adapted versioning to use setuptools_scm (included distance from latest release into non-released version strings).
  • Added classification metadata to project.
  • Improved documentation.
  • Improved the testing infrastructure (linting, faster test suite, testing Python 3.8 and 3.9, testing macOS, migrated testing from Travis CI to GitHub Actions).
  • Various internal refactorings and performance improvements.
  • Minor bug fixes.

19.3

Summary of changes:

  • Made code Python 2 compatible (with the help of upgraded dependencies Picire 19.3 and ANTLeRinator 4.7.1-1).
  • Improved the testing infrastructure (testing Python 2.7 and 3.7 on Travis CI; maintenance changes to various CI configurations).

18.10

Summary of changes:

  • Added implementation for the recursive variant of the HDD algorithm (a.k.a. HDDr).
  • Upgraded dependency to Picire 18.10 to utilize its new config ID and prefix concepts.
  • Minor improvements.

18.2

Summary of changes:

  • Added support for multiple tree builders, and added srcML as an experimental builder in addition to the existing ANTLRv4-based solution.
  • Generalized HDD implementation to be parametric to express classic HDD and Coarse HDD as well.
  • Upgraded dependency to Picire 18.1 to utilize custom initial granularity.
  • Upgraded dependency to ANTLR v4.7.1 (via ANTLeRinator).
  • Added support for building tokens from hidden ANTLR channels (whitespace, comments, etc.) into the tree but also hiding them from the reducer (for inputs where whitespace or other hidden tokens may matter during tree unparsing).
  • Added new module for gathering statistics on trees and improved the logging of the results of tree transformation algorithms.
  • Improved various algorithms (minimal replacement calculation from ANTLRv4 grammars, tree flattening for non-syntax-conforming inputs, unremovable node detection for rules in addition to tokens).
  • Improved Python-Java interworking (for Java-based ANTLRv4 parsers).
  • Improved API usability (for use-cases when Picireny is not called via its CLI).
  • Improved the testing infrastructure (by using the Coveralls online service).
  • Minor bug fixes and internal refactorings.

17.10

Summary of changes:

  • Improved the way how input format can be defined by enabling the use of a more consistent and well-defined config file.
  • Upgraded dependency to Picire 17.10 to utilize its Windows support.
  • Minor bug fixes.

17.7

Summary of changes:

  • Added implementation for the coarse variant of the HDD algorithm.
  • Implemented heuristical optimization to flatten left and right-recursive tree structures.
  • Improvements to the internal tree representation.
  • Simplified usage and ANTLR dependency installation via ANTLeRinator, and upgraded dependency to Picire 17.6.
  • Improved the testing infrastructure (support for Python 3.6 and code coverage measurement).

17.1

Summary of changes:

  • Updated dependency to Picire 17.1 and adopted its support for content-based result caching.
  • Added "squeeze tree" and "hide/skip unremovable tokens" HDD tree optimizations.
  • Improved handling of erroneous input.
  • Extended the HDD algorithm with testing of single-node tree levels to ensure 1-tree-minimality of output.
  • Minor bug fixes and improvements.

16.12

Summary of changes:

  • Added support for Java-based input parsing to improve performance.
  • Implemented HDD* (fixed-point iteration of hddmin).
  • Minor bug fixes and improvements.
  • Upgraded dependency to ANTLR v4.6.
  • Added Picireny to PyPI.

16.7

First public release of the Picireny Hierarchical Delta Debugging Framework.

Summary of main features:

  • ANTLRv4-based input parsing and Picire-based ddmin.
  • Automatic "smallest allowable syntactic fragment" computation for both parser and lexer rules.
  • Support for island grammars.
  • Python 3 API and out-of-the-box useful CLI.
  • py.test-based testing and tox support.