mapAlreadyAskFor doesn't have any effect unless a peer is sending invs for the same invalid transactions more than every 2 minutes, and even in that case it just slows down the unnecessary re-requesting.
Bitcoind never sends invs for the same tx that often (wallet rebroadcasts are at least 5 minutes apart), and if a slightly malicious node wants to cause a lot of unnecessary but low-cost network messages there are plenty of other ways to do that anyway.
Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/p4544_34ee579ea02daab5613e7078170e67f83501b5ab/ for binaries and test log. This test script verifies pulls every time they are updated. It, however, dies sometimes and fails to test properly. If you are waiting on a test, please check timestamps to verify that the test.log is moving at http://jenkins.bluematt.me/pull-tester/current/ Contact BlueMatt on freenode if something looks broken.
mapAlreadyAskFor doesn't have any effect unless a peer is sending invs for the
same invalid transactions more than every 2 minutes, and even in that case it
just slows down the unnecessary re-requesting.
Bitcoind never sends invs for the same tx that often (wallet rebroadcasts are at
least 5 minutes apart), and if a slightly malicious node wants to cause a lot of
unnecessary but low-cost network messages there are plenty of other ways to do
that anyway.
I always thought this map prevents us from requesting the same tx from multiple peers to not waste bandwidth, in case multiple peers announce the same tx to us at about the same time.
Oh, I see that! it's purpose makes sense now, thanks.
Thanks for trying to improve the network code and learning :)