But why push a valid unconfirmed tx that shall not be accepted into any valid block ever?
It shall.
For example the transaction
https://tradeblock.com/bitcoin/tx/2fb06fec2d549f1d05158f4a7a0812b50e7bb7cadf1697584e194ba01397886dwas discovered on the network 8 Oct 201
5 and confirmed 25 oct 201
6It was pushed by somebody over and over until confirmation succeded.
And there are a lot of such transactions
Isn't that a memory usage attack vector?
May be. But nodes can limit the size of mempool today
Here is the question then:
Can Bitcoin attacks and attack vectors be a malicious act or a misbehavior in your eyes?
If there is no code in sources which treat this as misbehavior - this is not misbehavior, but
a consensus of network usage. Sorry, my English is poor, I am not sure that I got a question
correctly and my answer is clear.