[confirmed]Drift of the dead (snow lands) vs land conversion

[travolter]

(I don´t know if snowcovered lands are implemented in manalink. This is a bug related to them)

Describe the Bug:
Drift of the Dead power and toughness is wrong.
Power: number of your lands + opponent´s lands in play
Toughness: number of opponent landsfixed

Which card did behave improperly ?
Drift of the Dead

Which update are you using?
dev version 250c78bd updated to 11-5-19 /DUEL

What exactly should be the correct behavior/interaction?
power and toughness equal to your snowcovered lands

[Aswan jaguar]

Normal basic lands are considered as snow lands for most snow cards (some don't). It was done like that as devs back then couldn't change the engine to accept basic lands both as basic and have snow type, too.
This is why you get so many bugs about snow cards.
Korath has fixed this in shandalar but I don't remember (and can't find this info) if he said if this can be ported with moderate effort in manalink, too.

[travolter]

Ok. Ill report more snowcovered bugs if I find, at least for maintain the database of present bugs.

[Korath]

The problem with using the actual snow-covered basics instead of considering all basic lands to be snow-covered isn't with Snow-Covered Plains etc.; it's that there's so many places, both in magic.exe and the manalink source, that implement a check which should be "Is this card a basic land?" by instead asking "Is this card specifically Plains, Island, Swamp, Mountain, or Forest?". Most have been found and fixed - or at least, we can assume they have, from the lack of people complaining that Wastes doesn't work - but nobody's done a systematic search. (Also, Wastes won't work as a canary for this sort of bug in cases where something calls for "a basic Forest", for example.)

[travolter]

OK Korath I undestand. Also I really doubt if some people will play with these snowcovered lands someday. Just add this bug report as "known issues" instead trying to fix it, cause I see its so much pain and time for cards that rarely will be used.

[Aswan jaguar]

The problem with using the actual snow-covered basics instead of considering all basic lands to be snow-covered isn't with Snow-Covered Plains etc.; it's that there's so many places, both in magic.exe and the manalink source, that implement a check which should be "Is this card a basic land?" by instead asking "Is this card specifically Plains, Island, Swamp, Mountain, or Forest?". Most have been found and fixed - or at least, we can assume they have, from the lack of people complaining that Wastes doesn't work - but nobody's done a systematic search.

I think it worth's the trouble to try it out in dev and see if we have problems.

(Also, Wastes won't work as a canary for this sort of bug in cases where something calls for "a basic Forest", for example.)

For this I made a search and there are only the panorama lands and 5 more cards that ask specifically ask for a basic something land.
Bant Panorama,
Flower // Flourish,
Gaea's Touch,
Nissa, Vastwood Seer // Nissa, Sage Animist,
Jund Panorama,
Grixis Panorama,
Goblin Caves,
Goblin Shrine,
Naya Panorama,
Esper Panorama.

[Korath]

I believe - without any testing whatsoever, since I don't currently have a compiler installed that can deal with Manalink - that all that needs to be done to enable this in dev is

1. Enable the five snow-covered lands in Manalink.csv
2. Remove the check for is_basic_land() from is_snow_permanent() in cards/ice_age.c
3. Add the five snow-covered csvids to is_basic_land_by_id() in functions/functions.c
4. Grep for each of CARD_ID_PLAINS, CARD_ID_ISLAND, etc., as well as CARD_ID_SNOW_COVERED_PLAINS etc., doublecheck that the checks for the non-snow ones really only mean those specific cards, and that any existing checks for the snow ones are enabled. I currently see:
   1. Myriad Landscape will only let you search for Swamps if you choose a snow-covered land as the first card. It should be looking at the five land subtypes anyway, though I don't think it's currently possible to change those for cards in your library.
   2. Gangrenous Zombies looks specifically for a card named Swamp that's snow (so it won't see a Snow-Covered Swamp, just a non-snow-covered Swamp that's been made to be snow). It doesn't honor Magical Hack, either, and again should be looking for the swamp subtype, not name, anyway.
   3. land_into_new_basic_land_type() in the_dark.c (why it's in that file is anyone's guess) will probably want to convert snow-covered lands into Snow-Covered Swamp etc. instead of Swamp. But land conversion in Manalink is already all kinds of broken anyway - it doesn't preserve the card name or legendary supertype, adds the basic supertype, and so on.
   4. Gaea's Touch looks like it won't currently allow anything; it should pass SUBTYPE_FOREST to get_hacked_subtype() instead.
   5. Maybe allow the snow-covered lands in is_beta_land() in functions/challenge_mode.c. Or maybe not. Judgment call.
   6. Enable the checks for the snow-covered lands in unlimited_allowed_in_deck() in functions/rules_engine.c
   In particular, none of the cards you list above except for Gaea's Touch look like they'll need fixing; all the others that are in Manalink already correctly look for basic land subtypes, not the five basic lands themselves.
5. Replace the basic lands with their snow-covered versions in all the gazillions of decklists that are floating around. This can be Somebody Else's Problem except for the ones distributed with the game itself.

If you were paying attention while changing land_into_new_basic_land_type(), you've seen the hard part - the internal_card_id's of the five original basic lands are 0, 1, 2, 3, and 4. Those are impossible to grep for in the Manalink source, and even harder to find in the magic.exe decompilation. For example, sub_47EF50(), which implements the check for TARGET_SPECIAL_BASIC_LAND in the targeting selection and validation functions, checks for basic lands this way (though this specific function is probably not a problem, since nothing directly in the Manalink source uses the TARGET_SPECIAL_BASIC_LAND flag anymore, and I don't think any cards Manalink calls into magic.exe for do either).

[Aswan jaguar]

I did 1, 2, 3, 4a, 4b, 4d(for this part to work but still Gaea's Touch can't be activated to put land in play - that is a different issue though) and 4f.
4e not my choice and 5 absolutely not.

4c: Left it as is. Thespian's Stage seems to work fine in my tests with snow basic lands. nothing to do with 4c.

I searched with various methods for the internal_card_id issue as far as Manalink source is concerned only and the few ones I found seemed not to need change. Even if they do need change the problem will be very limited and not worth leaving basic snow lands as they were.

Is something else in magic.exe that you think it can cause issues so I can run specific tests?

[Aswan jaguar]

Fixed in 1090c90d. I don't know if we have another post for land conversion issues so I will leave this open for that part as I didn't touch it and is still broken.