Update docs

CHANGELOG.md

@@ -7,10 +7,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
-### Fixed
+### Added
 
+- `Regexp::Lexer.lex` now streams tokens when called with a block
+  - it can now take arbitrarily large input, just like `Regexp::Scanner`
+  - this also slightly improves `Regexp::Parser.parse` performance
+  - note: `Regexp::Parser.parse` still does not and will not support streaming
 - improved performance of `Subexpression#each_expression`
-- improved `Regexp::Parser.parse` performance by about 20%
+- minor improvements to `Regexp::Scanner` performance
+- overall improvement of parse performance: about 10% for large Regexps
 
 ## [2.6.2] - 2023-01-19 - [Janosch Müller](mailto:janosch84@gmail.com)
 

tasks/benchmarks/log

@@ -2,13 +2,13 @@ Results of rake:benchmark on ruby 3.1.0p0 (2021-12-25 revision fb4df44d16) [arm6
 
 Parsing a minimal Regexp
 
-       Scanner::scan:    32710.2 i/s
-          Lexer::lex:    31091.8 i/s - same-ish: difference falls within error
-       Parser::parse:    27097.3 i/s - 1.21x  (± 0.00) slower
+       Scanner::scan:    32069.4 i/s
+          Lexer::lex:    30700.6 i/s - same-ish: difference falls within error
+       Parser::parse:    26248.5 i/s - 1.22x  (± 0.00) slower
 
 Parsing a complex Regexp (URI.regexp)
 
-       Scanner::scan:      877.0 i/s
-          Lexer::lex:      545.2 i/s - 1.61x  (± 0.00) slower
-       Parser::parse:      294.7 i/s - 2.98x  (± 0.00) slower
+       Scanner::scan:      843.4 i/s
+          Lexer::lex:      546.3 i/s - 1.54x  (± 0.00) slower
+       Parser::parse:      332.5 i/s - 2.54x  (± 0.00) slower