…g connections to hide tx time

d45b344 Bucket for inbound when scheduling invs to hide tx time (Gleb)

Pull request description:

  It has been brought up to my attention that current random delays mechanism (originally intended to obfuscate transaction metadata) allows to easily estimate the time a transaction was received by a node.

  It may be done by connecting multiple observer nodes to the same node. Each of those nodes will generate its own schedule of delays. Combined metadata regarding those events from different sources allows an observer to estimate transaction time.

  After this patch a spy won't gain additional information by just creating multiple connections to a target.

Tree-SHA512: c71dae5ff350b614cb40a8e201fd0562d3e03e3e72a5099718cd451f0d84c66d5e52bbaf0d5b4b75137514c8efdedcc6ef4df90142b360153f04ad0721545ab1

…kable time

1a8f0d5 [tools] update nNextInvSend to use mockable time (Amiti Uttarwar)
4de6303 [tools] add PoissonNextSend method that returns mockable time (Amiti Uttarwar)

Pull request description:

  Introduce a Poisson helper method that wraps the existing method to return `std::chrono::duration` type, which is mockable.

  Needed for bitcoin#16698.

ACKs for top commit:
  ajtowns:
    ACK 1a8f0d5
  MarcoFalke:
    re-ACK 1a8f0d5
  naumenkogs:
    ACK 1a8f0d5, and let's merge it and come back to it later.

Tree-SHA512: 7e2325d7c55fc0b4357cb86b83e0c218ba269f678c1786342d8bc380bfd9696373bc24ff124b9ff17a6e761c62b2b44ff5247c3911e2afdc7cc5c20417e8290b

This seems to have been wrong already in the backported PR.

And remove manual calls accordingly.

fae169c test: Make bloom tests deterministic (MarcoFalke)

Pull request description:

  non-deterministic tests are useless, since a failing test could not be reproduced unless the seed is known.

Tree-SHA512: 4f634ff0c6adf663444f1ac504f6dbceaa46b78d697b840531977ba30006453ac559d5c21cc3eaef6d92b87d46008a34b0db6331ea3318001987fcfaec634acf

When we bumped before the call, it sometimes happened that in the following
loop one of the early nodes which got the network reactivated tried to
connect to one of the latter nodes with still disabled network. Such
connection attempts were dropped by the latter nodes and the retry
throttling then prevented the connection to be retried later. Doing the
bumping after setnetworkactive(True) should fix this.