And on further review, no, none of the cards redirecting or preventing damage to permanents target those permanents. They are:
Alabaster Potion,
Amulet of Kroog,
Argivian Blacksmith,
Guardian Angel,
Healing Salve,
Jade Monolith,
Oasis,
Rakalite,
Samite Healer. (And a handful of cards identical to one of those (mostly Sammy) except for cost/keyworded abilities/power/toughness:
Alabaster Wall,
Benevolent Ancestor,
Clergy en-Vec,
Femeref Healer,
Healing Leaves,
Kithkin Healer,
Militant Monk.) I'm currently leaning towards leaving them broken - since they don't reference their putative target at all, they can't be patched in-place like I did with most of the others; and if nobody's been up in arms about
Samite Healer's ability to prevent damage to
Black Knight, I can't see them objecting too strenuously to
Oasis being able to do the same for
Horizon Drake. At least for the time being.
Other than those, I've gotten through just about all of the others. The remainder are four astral cards - Orcish Catapult, Goblin Polka Band, Power Struggle, and some of Faerie Dragon's effects. While these all have a similar problem as the damage prevention cards, in that they don't (or don't always) fully target the cards they affect, they're all effects that I can feasibly rewrite at this point.
---
The reason I've been putting off cards that have both mana and non-mana activated abilities can more or less be summed up by the scenario: "You have nothing on the battlefield but a
Rishadan Port. Can you tap your opponent's land with it?" Obviously not, but the way this engine deals with mana costs is to compute the total possible mana that could be produced, then compare each card's activation costs against that. There's nowhere to store
which card is responsible for a given chunk of potential mana, and so no way to directly indicate a card like
Rishadan Port can't activate to tap a land with mana paid by tapping itself.
Manalink's solution is to pretend these cards' non-mana abilities have a higher cost when it's trying to determine whether they can be activated -
,
for
Rishadan Port,
,
for
Kjeldoran Outpost,
,
, sacrifice for
Dreamstone Hedron, and so on - and, when it
is being activated, temporarily tap it before charging mana.
The second part of this works ok in practice - there aren't, to my knowledge, any cards that change the amount of mana they produce depending on whether some other card you control (the one activating, in this case) is tapped - but the first is very fragile. Enchant that
Rishadan Port with a
Wild Growth or get a
Mana Flare onto the battlefield, and suddenly its first ability produces {C}
or {C} {C} and the game thinks that's enough to pay for the second ability again.
This is a problem that I've been irritated with for a couple years now. For a long time, I couldn't come up with any better solution than the fairly obvious one: to silently tap the Port, recompute potential mana to see if it had enough to pay the
for its second ability, then untap it and either re-recompute potential mana for the benefit of other cards or restore a stashed copy made at the start of the process. MPMTG's CPU requirements are modest by the standards of today's hardware, and I doubt that something like that would have a noticeable effect even when multiplied by thousands of AI branches per turn, but it's still pretty distasteful - it defeats the whole point of precomputing potential mana in the first place, and that's a quite elegant optimization most of the time.
Another fairly obvious method - make every card that registers itself as able to produce mana also respond to another event that makes them
remove their mana from the potential mana count instead of adding it, like EVENT_COUNT_MANA does. This was promising, but turned out to be impractical: in edge cases similar to the ones that break when you pretend
Rishadan Port costs
to tap a land, this ends up needing to be able to find which specific bit of potential mana was previously announced by the card. And if that's recorded anyway, there's no need for the new event.
My current thinking is to put this off for a while longer until I begin to tackle cards with restrictions on how their mana costs or produced mana may be paid or spent -
Imperiosaur,
Myr Superion,
Eldrazi Temple,
Cavern of Souls, and so on. I'm already going to have to mark each point of mana with what produced it both in the mana pool and in the potential mana pool to deal with the first two of those examples, and probably to deal with the latter two as well. I may do something like the "good enough" Manalink solution as a stopgap before then - I
really want
Kjeldoran Outpost - but it'd be error-prone to future-proof it for when I
do have a permanent solution in place.
On the other hand, looking at
Strip Mine and
Elephant Graveyard in my to-rewrite list for miscellaneous protection made me realize that none of that really matters if the non-mana ability has no mana in its cost, so I went through and programmed all the otherwise-blindingly-easy cards that looked like "
: Add [some mana] to your mana pool.
/
/ sacrifice something / etc: Do something else."
- and some other cards that caught my eye | Open
- Alaborn Veteran
Auriok Bladewarden
Kabuto Moth
Kitsune Loreweaver
Loxodon Mender
Pang Tong, "Young Phoenix"
Steelshaper Apprentice
Stern Marshal
Takeno's Cavalry
Baleful Stare
Blinkmoth Infusion
Filigree Sages
Neurok Prodigy
Oboro Breezecaller
Soratami Mirror-Guard
Soratami Rainshaper
Teardrop Kami
Withering Gaze
Acolyte of Xathrid
Cabal Torturer
Cackling Imp
Hunger of the Nim
Irradiate
Scuttling Death
Wall of Blood
Akki Raider
Blood Rites
Fireslinger
Ogre Recluse
Orcish Cannoneers
Screaming Fury
Shinka Gatekeeper
Steam Vines
Accelerated Mutation
Arbor Elf
Fyndhorn Brownie
Rebuking Ceremony
Roaring Slagwurm
Tel-Jilad Lifebreather
Thriss, Nantuko Primus
Traproot Kami
Trolls of Tel-Jilad
Viridian Lorebearers
Viridian Scout
Crenellated Wall
Galvanic Key
Honor-Worn Shaku
Imi Statue
Nine-Ringed Bo
Scale of Chiss-Goria
Scuttlemutt
Thermal Navigator
Voltaic Construct
Ghost Town
Heart of Yavimaya
High Market
Karakas
Maze of Shadows
Pendelhaven
Petrified Field
Quicksand
Swarmyard
Unstable Frontier
Wasteland
This keeps me from adding effects that increase activation costs, but there aren't many of those, and I couldn't have done them properly while any cards at all are left still using their MicroProse implementations in Shandalar.exe anyway.
Gloom is a theoretical problem, so I left out
Gaea's Touch;
Iceberg is the only other enchantment with both mana and non-mana activated abilities, and it has a mana cost. I can accept for now that these will incorrectly show up as activateable if there isn't enough other mana to pay for them
and there's a
Gloom on the battlefield
and the card's been
Sleight of MindPurelace'd to the right color
and either there's an
Enchanted Evening on the battlefield or it's a copy made with
Copy Artifact.
Posted: 05 Mar 2016, 04:59
