v0.8.3

Automa v0.8.3

Diff since v0.8.2

Closed issues:

• Generated code is highly redundant (#53)
• Pushdown automata (#59)
• Same char used in two different patterns of an alternation (#77)
• Automa.execute gives wrong result in some case (#78)
• Shift-based tables for small machines (#81)
• Allow construction of RegExp from invalid UTF-8 strings (#83)
• Optimize away dict lookup at machine exit (#84)
• Tokenisation fails to produce an error (#88)
• Make ambiguity errors easier to understand (#93)
• Verify keys in RE.actions are correct on setindex! (#100)
• Keyerror on remove_dead_notes (#104)
• Bug with regex set negation (#105)
• Make FSM generation deterministic by simply renaming nodes (#106)
• (Optionally) despecialize generated code on its IO (#113)
• Correctly finding matching double quotes when Unicode characters are present (#114)

Merged pull requests:

• Improve debugging of Automa (#64) (@jakobnissen)
• Fix execute (issue 78) (#79) (@jakobnissen)
• Optimize away dict lookup when possible (#85) (@jakobnissen)
• Add workflow for downstream tests (#86) (@CiaranOMara)
• Have Tokenizer check if in accept state when EOF (#89) (@jakobnissen)
• Small tweaks to Tokenizer internals (#90) (@jakobnissen)
• Small cleanup of Automa codegen (#92) (@jakobnissen)
• Fixup parse_class (#94) (@jakobnissen)
• Breaking changes for V1 (#95) (@jakobnissen)
• Remove ByteSet (#96) (@jakobnissen)
• Use smaller integer width for table generator (#97) (@jakobnissen)
• Disable unambiguity check again (#98) (@jakobnissen)
• Fix variables in generate_reader (#99) (@jakobnissen)
• Remove StableSet and StableDict (#107) (@jakobnissen)
• Better error on compiling null regex (#108) (@jakobnissen)
• Remove unused files (#109) (@jakobnissen)
• Remove dependency on Printf (#117) (@jakobnissen)
• Bump ScanByte to 0.4.0 (#121) (@jakobnissen)