Test Details Overview

Split `block_rebroadcaster.cooldown` into suppression and expiry tests (#5253)

block_rebroadcaster.cooldown pushed a block, waited for the peer to
store it and pushed it again, expecting the second push to land inside
a 500 ms cooldown. The cooldown starts when the first push is
processed, so the wait for the peer counted against it: on a Windows
runner the peer took longer than that, the second push went out as a
second rebroadcast and the test timed out waiting for it to be
suppressed.

Each half is now its own test, set up so that a delay can only help.
cooldown keeps the suppression check with a 60 s cooldown that no delay
between the pushes can outlast. The new rebroadcasts_after_cooldown
covers expiry with a 100 ms cooldown and waits at least that long
before the second push, so a delay only moves the push further past
expiry.

With a 1 s delay injected before the second push, the old test failed
at the suppression check on both backends and both new tests passed.

__________

Type: commit

Hash: 8b3dff4da65458cd2a838f1e9f982d6ec65d49e4

Created: yesterday

__________

gr0v1ty/nano-node:8b3dff4da65458cd2a838f1e9f982d6ec65d49e4