Speed up do_classes_matches?

[preciz]

Introduced a guard check using String.contains?/2. If the class attribute does not contain the target class substring, we return false immediately, bypassing the expensive String.split/2 operation.

There is an average 1.5x speed up.

Bench script:

Mix.install([
  {:benchee, "~> 1.0"},
  {:floki, path: "."}
])

read_file = fn name ->
  __ENV__.file
  |> Path.dirname()
  |> Path.join(name)
  |> File.read!()
  |> Floki.parse_document!()
end

# Use big.html for a realistic large DOM
doc = read_file.("big.html")

IO.puts "Document parsed. Running benchmark..."

Benchee.run(
  %{
    "find .mw-redirect (common)" => fn -> Floki.find(doc, ".mw-redirect") end,
    "find .reference (frequent)" => fn -> Floki.find(doc, ".reference") end,
    "find .nonexistent (not found)" => fn -> Floki.find(doc, ".nonexistent12345") end,
    "find .citation (partial of .citation-comment)" => fn -> Floki.find(doc, ".citation") end
  },
  time: 5,
  memory_time: 2
)

[Copilot]

Pull request overview

This PR introduces a performance optimization for CSS class matching by adding a guard check using String.contains?/2 before performing the more expensive String.split/2 operation. If the class attribute doesn't contain the target class substring, the function returns false immediately, avoiding unnecessary string splitting and enumeration.

• Adds String.contains?/2 guard check before String.split/2 in single-class matching
• Maintains correctness by still performing full validation when substring is present
• Targets the do_classes_matches?/2 function for single-class selectors

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[philss]

@preciz good call. Thank you!