test_that("seedMatch runs on a tiny matching problem", {
  set.seed(0)
  dat0 <- data.frame(x = rnorm(8), z = c(rep(1, 3), rep(0, 5)), prc = rep(1L, 8))
  bd <- basicDistance(dat0)
  dat <- bd$dat
  cost <- bd$cost
  id1 <- dat$id[dat$z == 1]
  id2 <- dat$id[dat$z == 0]
  m <- blockDesign:::seedMatch(id1, id2, dat, cost, msetAdd = 0, ncontrols = 1, solver = "rlemon")
  expect_s3_class(m, "data.frame")
  expect_true("mset" %in% names(m))
  expect_true(all(m$z %in% c(0, 1)))
  expect_equal(nrow(m), length(id1) + length(id1) * 1L)
  expect_equal(sum(m$z == 1), length(id1))
  expect_equal(sum(m$z == 0), length(id1))
  expect_equal(length(unique(m$mset)), length(id1))
})

test_that("seedMatch requires z and id in dat", {
  cost <- matrix(1, 1, 2, dimnames = list("1", c("2", "3")))
  dat <- data.frame(id = 1:3, foo = 1:3)
  expect_error(
    blockDesign:::seedMatch(1, c(2, 3), dat[, "foo", drop = FALSE], cost, 0, 1),
    "z must be a column"
  )
  dat_z <- data.frame(z = c(1, 0, 0), foo = 1:3)
  expect_error(
    blockDesign:::seedMatch(1, c(2, 3), dat_z, cost, 0, 1),
    "id must be a column"
  )
})