-Never play if would suicide, unless we are losing the game next turn anyway AND it results in a double KO
-If AI has the cards that work better on small hand, do play Price of Progress even if it doesn't deal good damage to the enemy. (important synergy within Tempest block with major tournament tier cards)
TODO : Ideally, the AI should hold 4 mana untapped as soon as it has 4 lands in play, and play Sudden Impact ASAP if there are 4+ cards, before the human has a chance to use up any.
Note : This might be slightly suboptimal for decks where the AI is forcing the other player to draw, and thus might deal more damage if waiting, but considering cards that force draw are usually played before this due to their mana cost and timing, I don't think this will be a problem (if such AI decks even exist right now) - especially as using one now and another later is always better than only using one and those decks should have many copies.
Will now target players if granted Lifelink or equivalent even if otherwise would not - self damage is not a concern
Are there any other cards that need AILogic "SelfDamage" or Svar "LikeLifeLink" added? If you are aware of any, please add!
Do prefer targeting the player if has a "when this deals damage to an opponent" trigger
Important : The AI should be able to recognize the ability stacks - if there are multiple cards that deal damage, they should be used to kill the target together. (for example, tap two prodigal sorcerers to kill a 2/2 creature!)
Unfortunately, I don't think I can implement this with my current knowledge.
The existing IP selection logic was less than optimal on Linux hosts
where it would frequently select and report 127.0.0.1 as the IP address
to share. The new logic will create a datagram socket, use it to locate
an interface with a default route, and further backtrack to the
interface's IPv4 address.
Signed-off-by: Jamin W. Collins <jamin.collins@gmail.com>
Marked card AI playable.
Note : This also adds a new function to ComputerUtil,that checks if the AI is in danger of being killed, or if it has a same amount of life remaning - useful for future "pay life" AI decisions.
