Various fixes for parsing #306

andreas-zeller · 2025-02-06T08:15:19Z

This pull request

fixes a number of (mostly minor) issues with parsing complex inputs (bits, regexes, alternatives...).
brings a new --format=repr output option, printing derivation trees using the Fandango internal representation (useful for assertions).
reworked the repr() and str() functions for derivation trees and symbols: repr() now returns the internal representation, and str() returns the (user-facing) associated string (as per Python conventions).
DerivationTree objects still have several to_xxx() conversion methods for producing "official" output formats.

All tests pass, and the documentation builds. Enjoy!

…return the internal representation

andreas-zeller added 17 commits February 5, 2025 18:12

New: have str(x) return a user-friendly representation; repr(x) will …

bb225e3

…return the internal representation

Adapted tests to new syntax

dc2c963

Fix: value to be parsed might be bytes

8f37843

New: use shlex.split() for easier reproduction of failing CLI tests

47dc4b7

Added tests for currently failing inputs

cd22d20

Fix: shorter repr() for ParseState

4037c44

Fix: did not check all parse alternatives

42a7802

New: ensure str() always returns a string

e77217c

Extended parse tests

3259f62

Refactoring

1d75ad7

(abc)*.txt: Added inputs as permanent resources

9948468

Moved CLI parsing tests into test_parsing

c96b079

Made tests executable from command line

2c092ff

Fix: regex scanning would accept extra bytes

84259a2

Code cleanup

1313df3

Added extra (hard) parsing tests

ecca309

Fix: reduced population size for test_infinity

9f88e28

joszamama merged commit 8e450b6 into main Feb 6, 2025
3 of 4 checks passed

Provide feedback