Currently, the numbers of items sent by the second sender are determined
incorrectly in the mpsc_send test, resulting in the receiver seeing a
bunch of messages it doesn't expect and missing some that it does. This
doesn't break the test in loom, because we send substantially fewer
messages, and the multiplication happens to be the same as what we
intended (as the second tx only sends one message under loom). However,
the test fails incorrectly when running normal tests with real threads.
Currently, the numbers of items sent by the second sender are determined incorrectly in the
mpsc_sendtest, resulting in the receiver seeing a bunch of messages it doesn't expect and missing some that it does. This doesn't break the test in loom, because we send substantially fewer messages, and the multiplication happens to be the same as what we intended (as the second tx only sends one message under loom). However, the test fails incorrectly when running normal tests with real threads.This commit fixes that.