TestItemRunner

Measure a tagged whole test item, including its setup and assertions. Run this selection from examples/bibliography. Prepare its item environment, then start Julia in that environment before evaluating the code:

julia setup.jl items
julia --project=.controller/items

using PerfChecker, TestItemRunner
listing = discover_testitems(pwd())
selected = filter(item -> item["name"] == "Bibliography export workload",
    listing["items"])
@assert length(selected) == 1
result = run_testitems(pwd(); ids=[only(selected)["id"]], samples=1,
    threads=1, timeout=180, reports="results/export-item")
@assert result["passed"]

The recorded native execution receipt records real extension activation, shared/performance item discovery, the actual Bibliography export-item profile handler and 1993.05 ms delivered to the native Testing API. That is a whole-item observation, including setup and assertions; it is not the export-operation timing curve. The Bibliography tutorial explains the item selection.

For a DataStructures item, run from examples/kitchen-sink:

julia setup.jl core
julia --project=. items.jl list
julia --project=. items.jl run queue

This second recipe is runnable but has no separately published queue-item receipt; the container operation charts cannot establish its whole-item cost.