pytest: fix flake in test_closing_anchorspend_htlc_tx_rbf

Turns out we resubmit two txs (the commitment tx, and the anchor spend), but only wait
for one of them: if we mine a block before the anchor spend, it doesn't go in:

```
    @unittest.skipIf(TEST_NETWORK != 'regtest', 'elementsd anchors unsupported')
    @pytest.mark.developer("needs dev_disconnect")
    def test_closing_anchorspend_htlc_tx_rbf(node_factory, bitcoind):
...
        l1.daemon.wait_for_log('Peer permanent failure in CHANNELD_NORMAL: Offered HTLC 0 SENT_ADD_ACK_REVOCATION cltv 116 hit deadline')
        l1.daemon.wait_for_log('Creating anchor spend for CPFP')
    
        wait_for(lambda: len(bitcoind.rpc.getrawmempool()) == 2)
    
        # But we don't mine it!  And fees go up again!
        l1.set_feerates((3000, 3000, 3000, 3000))
        bitcoind.generate_block(1, needfeerate=5000)
    
        l1.daemon.wait_for_log('RBF anchor spend')
        l1.daemon.wait_for_log('sendrawtx exit 0')
    
        # And now we'll get it in (there's some rounding, so feerate a bit lower!)
        bitcoind.generate_block(1, needfeerate=2990)
    
>       wait_for(lambda: 'ONCHAIN:Tracking our own unilateral close' in only_one(l1.rpc.listpeerchannels()['channels'])['status'])
```

tests/test_closing.py

@@ -3775,7 +3775,8 @@ def test_closing_anchorspend_htlc_tx_rbf(node_factory, bitcoind):
     bitcoind.generate_block(1, needfeerate=5000)
 
     l1.daemon.wait_for_log('RBF anchor spend')
-    l1.daemon.wait_for_log('sendrawtx exit 0')
+    # We actually resubmit the commit tx, then the RBF:
+    l1.daemon.wait_for_logs(['sendrawtx exit 0'] * 2)
 
     # And now we'll get it in (there's some rounding, so feerate a bit lower!)
     bitcoind.generate_block(1, needfeerate=2990)