test_lightningd.py: test_closing_different_fees must wait for txs to hit mempool

Careful log examination revealed that we were generating a block before one
of the mutual close txs had entered the mempool.  This is rare because it
means that both peers have to be too slow.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

tests/test_lightningd.py

@@ -1366,6 +1366,12 @@ class LightningDTests(BaseLightningDTests):
         for c in closes:
             c.result(72)
 
+        # close does *not* wait for the sendrawtransaction, so do that!
+        # Note that since they disagree on the ideal fee, they may conflict
+        # (first one in will win), so we cannot look at logs, we need to
+        # wait for mempool.
+        wait_for(lambda: bitcoind.rpc.getmempoolinfo()['size'] == num_peers)
+
         bitcoind.generate_block(1)
         for p in peers:
             p.daemon.wait_for_log(' to ONCHAIN')