MoM bugs reports (to do list)

[I like Serena]

Finally, I have a case for Gorgon Vs. Phantom Warrior. The bug is that Gorgon got too few damages. (Phantom warrior has stone immunity, and Gorgon does not have true sight, and thus 5+ damage is expect to Gorgon per warrior.)

Enter the attached saved file, and you will got the only one movable unit (magic spirit) at current turn. Just move it to the adjacent nature node, you will find some Gorgons. Just cast phantom warriors a few times, you will see what I mean.

I haven't checked your save file yet, but here's what my combat simulator predicts:

Unit Green Gorgons 4x 10-0-9-11-9 (36) +2/+0 Stoning Gaze/-2,Flying
attacks unit Blue Phantom Warriors 6x 3-0-0-6-1 (6) +0/+0 Non Corporeal,Poison Imm,Stoning Imm,Death Imm,Illusion,Combat only

Illusion reduces defense of Green Gorgons to 0
Unit of 4 Green Gorgons attacks for 20 damage
Unit of 6 Blue Phantom Warriors attacks for 5.4 damage
Unit of Green Gorgons has 4 units left with a total of 30.6 hp
Unit of Blue Phantom Warriors has -14 units left with a total of -14 hp

So expected is that all phantom warriors together deal 5.4 damage, not 5+ per warrior.
Does this match your case or not?

[Sajin]

1.Land unit upon sea with fly ability by spell.
2.Chaos Channel create it winged.
3.Remove Fly spell from it.
4.Unit suddenly disappear by dropping sea.

Oh no!,Where is my hero!?

[Asfex]

This is savegame for Unit is immobilized bug, when no combat button work properly, except Auto button. And pressing the "auto" is the only way to finish combat.
Very high probability to reproduce. Happens on vanilla and the 1.40m patch.

[kyrub]

Thanks for the save!

[kyrub]

Can you get me the latest your IDA database? To save my time, and answer a question "compatibility between mods"(due to extra difficulty level).My own IDA database is incompatible with your version.
And, of couse, this function have to have arg_0.

Believe it, or not, I don't have any "latest" IDA database. I still use the original one (see download section of ILSe). Plus I have a separate IDA database on 90% of changes I have done with small notes inside. (I admit it's not practical at all... I never thought the patch would grow to this scale). That's about 160 IDA db files on MoM alone and counting.

When there is a bug, I make an interpolation with various versions, until I find where the bug is "created". Then I find the appropriate IDA database, search for bugs...

[Ariel]

Has anyone hit this bug?:

Sharee starts casting Spell of Mastery, so I go kill her. My MP goes negative, but it self-corrects (goes to 30,000) after I proceed to the next turn. (That is a bug, but it's not the one I'm concerned about.)

Even after the MP goes to 30,000 things are still a bit broken because I can't cast spells in battle, but apparently this is a well known issue and you just have to reduce MP so I alchemy a bunch of MP to GP and the problem goes away. (So, that's another bug, but still not the one I am concerned about.)

Now when I killed Sharee, she had almost all of Myrror to herself, and had lots of cities/armies/mana/casting-skill, so I was a bit surprised because she didn't cast Spell of Return. Instead, her armies went away, her cities reverted to neutral cities, and she was apparently gone for good. (Three cheers! The Armageddon/Great-Wasting/Meteor-Storms casting witch is dead!)

But then some turns later, I get a notification that Sharee has started casting Spell of Mastery again. I go check the Magic screen -- she's still banished. I go check her old cities -- they're still neutral.

What do I do now? I don't have sorcery so Spell Blast isn't an option (and even if I did have it, my spell skill is puny, plus I don't know that I could even use it on her since the game kind of thinks she's banished). Normally I would put together a big stack (or two) and take her out again -- but she's still wherever wizards go when they're banished, so I can't get to her to kill her.

If a wizard casts Spell of Mastery while banished like this, does it actually work? I.e., can I just ignore it and play on knowing that Sharee can't hurt me? Or am I in trouble?

Is my only option to try to perma-kill both of the other wizards before the SoM finishes? Because I doubt that's possible. (Banish them both, maybe. But they'll just cast Spell of Return -- they have lots of cities and there's no way I can take out all of those cities in a short time.)

(I am using 1.40m; just load the attached save game and hit end turn and the bug happens right away)

edit:

I started writing a program to dump the save-game contents and verify that the contents are valid at the same time. The idea is to try to figure out where the game went corrupt and what went corrupt first, then maybe hack the save game to fix it.

It looks like it will take Sharee another 70 turns to finish casting Spell Of Mastery, so maybe I could take out the other two wizards in that time:

0x01840: Wizard #3
0x01840: portrait: 2
0x01841: name: Sharee
0x01855: race: Dark Elf
0x01856: banner color: 3
0x01858: personality: Chaotic
0x0185a: objective: Militarist
0x01864: fame: 24
0x01866: power base: 70
0x0186a: research: 0
0x0186b: mana: 100
0x0186c: skill: 0
0x0186e: summoning circle location: 16, 28, Myrror
0x01874: research candidates:
0x0188e: current spell being cast:
0x0188e: MP left: 4930
0x01890: full MP: 5000
0x01892: Spell Of Mastery

But I'm not sure if the game would still consider Sharee "alive" enough that it still wouldn't end in victory at that point.

[Ariel]

OK, I am looking at save games right before and after I killed Sharee, and I have a hypothesis. First, what I am seeing in the save games:

I already mentioned my MP went negative. The specific #s I see are that it was 298 before the battle and -16248 after the battle.

What is more interesting is that Sharee's MP also went negative. She had 30,000 before the battle, and -32768 after the battle. (Her power base [save-game file offset 0x1866] also went from 1809 to 0.)

Before battle she was casting Spell of Mastery. Now when I banished Sharee the game acted like she was completely dead (cities revert to neutral, no notification that she has started Spell of Return), but the after-battle save-game seems to indicate that she did in fact start casting Spell of Return.

Also, the after-battle save game (and all save games after that) are showing her status as "eliminated" (byte at save-game file offset 0x6607 has value of 0).

Now for my hypothesis:

When she lost the battle, the game decided whether or not to give a notification that she is casting Spell of Return. Because her MP was negative and she had no power base, the game decided she couldn't cast it. However, because her MP was at -32768 (and not 0), it's possible some other part of the code is either treating it as unsigned or is doing some kind of "MP - X > 0?" test which would result in an overflow even for small X, and that is how the Spell of Return is working. Some turns later Spell of Return "finishes". Even though Sharee was already marked "eliminated" and stays that way, since Spell of Return has finished she picks a new spell to cast -- Spell of Mastery. (I.e., I am guessing that when one spell finishes, the game logic doesn't even bother to check if the wizard is still alive before picking the next spell to cast. And that logic might be fine. It's the logic that allowed her to become "dead and casting" in the first place that's broken.)

In the save game I attached to my previous post (which is some turns after the battle), Sharee is still casting Spell of Return, and the MP left for casting that spell (save-game offset 0x188e) is right at 0. That's why hitting "end turn" immediately results in notification that she's casting Spell of Mastery. Her MP seems to have auto-corrected somehow, as it is no longer at -32768 and is instead at 0. (Though with her power base stuck at 0, I'm not sure how she's managing to cast anything.)

So, how am I going to fix this mess? I have already tried going back to the save game right before the battle and refighting the battle. But I still got the unexpected result of negative MP and Sharee apearing to perma-die rather than announcing Spell of Return. I didn't save that attempt though so I can't verify if it has the same negative MP for Sharee, but I'm guessing it did. And I think the problem might be that Sharee just has too much MP when I kill her -- if I hack her MP to be around 15,000 instead of 30,000 before going to battle, maybe the bug won't happen.

I'll let you know how it goes...

BTW, this is the very first game I have ever played on the "Impossible" setting (and only my sixth game of MoM ever). I don't think I was fully informed on exactly what "Impossible" entailed!

Also, if my above hypothesis is correct, then I think that the save-game I uploaded is worthless to you -- the game data consistency was likely already broken before that point. The save game right before the battle is the one that might let you see the bug as it happens.

[Dunhere]

Looks like I've bumped into the same bug Ariel did; after attacking Sss'ra's tower to stop a SoM, he was defeated instead of banished and my mana dropped to -15k. I have not yet played further to see if he attempts to cast again. Attached are saves just before and just after the battle.



Also, just for fun, a shot of one of Sss'ra's front line commando units valiantly assaulting one of my road building crews.

[Ariel]

Success! I changed Sharee's MP to 15,000, replayed the battle, and this time I got the notification that she's casting Spell of Return and my MP didn't go negative.

Dunhere, you've definitely hit the same bug. In your save games I see the same changes happening:

Rjak:[INDENT]MP: 1482 -> -15067[/INDENT]

Sss'ra:[INDENT]current spell: Spell of Mastery -> Spell of Return
MP: 30,000 -> -32,768
status: active -> eliminated[/INDENT]

I've taken the liberty of making the same patch to your pre-battle save-game that I made to mine. (I.e., change Sss'ra's MP to be 15,000 instead of 30,000.) You'll have to refight the battle, but then you should be good after that -- you should get the notice that Sss'ra is casting Spell of Return and your MP shouldn't go negative.

edit:

BTW, I think that the AI's MP is going to -32,768 first, and then because we banished the AI wizard the game is trying to give us 1/2 of the AI's remaining MP. Since we both had only small amounts of MP before the battle, adding -16,384 MP drove our MP negative.

I think I also figured out the auto-correct that happened to my MP: The code for preventing MP from going over 30,000 must be using unsigned compares. So the first time it bumps MP for any reason it does an unsigned compare against 30,000 and sees that it's "bigger" (e.g., 16 bit -15,067 signed is 48,289 unsigned), and then "decreases" it to 30,000.

I still don't have any detailed reason why the AI MP would go to -32,768 in the first place though. Even if there was some kind of overflow, it seems really odd that any code would be saturating at -32,768. (I can understand saturating at 0, or 30,000, but why -0x8000?)

[Malefice]

Hi, I started the game with 11 Life spellbooks and cast Incarnation to summon Torin as soon as possible. I discovered that even if I didn't have the mana to pay for his upkeep, I only got a message stating "Unit Dissipated: (some random unit name)" and Torin was still around. I tried for a few more turns and it's the same, so basically I had a free Torin rampaging everyone around me.

I played the game in the past and vaguely remembered Torin left if I didn't pay his mana upkeep, so I tried playing without the 1.40m patch and just the 1.31 one, and got the same result.

Does this happen on everyone's version of MoM?