João Salavisa--repeat-each parity fixture
A real blob report from npx playwright test --repeat-each=4 --retries=1 over a
single test that fails its first attempt on two of the four repetitions, plus
the stats block Playwright produced from that same blob.
It exists because --repeat-each is the flag used to verify de-flaking work, and
the summary used to be wrong for exactly that case. All four repetitions share
one project/file/title, so grouping on key collapsed them into a single test
whose only surviving attempt was retry 0 of the last repetition: the summary
reported 0 flaky of 1 executed where Playwright reported 2 flaky of 4.
Grouping on Playwright's test id, which is distinct per repetition, is what makes
the two agree.
Regenerate with the recipe in ../playwright-parity/README.md, replacing the
spec with a single test that fails the first attempt of some repetitions, and
passing --repeat-each=4.