# Competitor parsers for tools/bench_engines.py. Deliberately NOT in requirements-test.txt: that file
# is the correctness ORACLE toolchain, where a bump moves a verdict and has to be re-soaked. Nothing
# here oracles anything — these are the engines Frostwork is measured against.
#
# Pinned anyway, because a competitor bump moves two numbers this repo publishes: its speed row AND its
# value-parity-vs-lxml row. lexbor is an HTML5 parser and libxml2 is not, so the disagreement rate is a
# property of the exact pair of versions, not of "selectolax".
selectolax==0.4.11
beautifulsoup4==4.15.0
# bs4's CSS engine, and the reason a bs4 row is a CSS-selector row at all. A transitive bs4 dep, pinned
# explicitly because it decides which of the corpus's production selectors that row can even express.
soupsieve==2.9.1
