Re: Bug Thread: Discussion
 

I've found a funny bug. Was playing around with the Eater of the Dead. I have cast GoR on him, gave pedant of luck and ring of regeneration, made him a prophet, fed him to the 3rd form and pushed him deep into enemy territory. He quickly made himself to the 1st place in the hall of fame, then, as usual, he became uncontrollable and started rampaging random provinces. The icon in the hall of fame was showing he was dead, so I had cast a Ritual of Rebirh, AND the SECOND INDEPANDANT Eater with items and prophet status has appeared in the province first one has gone berserk in. Now there are 2 Eaters stalking around. That's not all. There's supposed to be only one Eater in the world, and when you resummon it the independant one dissapears, so I was eager to know wich one would dissapear, or would they dissapear both... So I've cast Call the Eater again... To my surpise - none of them did, and I've received the third one.

Re: Bug Thread: Discussion
 

Are bvithg those Eaters of the Dead the same monster number? It has several forms, so that could be playing havoc with the normal unique mechanics. Once the second one reaches Unfettered form, it would be nice to see if it disappears or not...

Re: Bug Thread: Discussion
 

The two rampaging eaters are U997 - the Unfettered form, the third one I've summoned is U994 - the smallest. I'll try to feed him to the U996 form to watch the result.

Added: I just fed the 3rd one to the U996 form, few turns after he's becane U997 and gone out of control, captured the province and then just dissapeared the same turn. The previous two are still wreaking havoc in the world.

Re: Bug Thread: Discussion
 

I'd guess it has something to do with the one being a prophet?

Re: Bug Thread: Discussion
 

Must be. Prophetizing it changed it from a normal Eater of the Dead to type "prophet" even though the monster number is the same. It lists as dead because it went independent (much like charmed and enslaved units are counted dead). Then the second one went indie but is a regular Unfettered, so it stays too. The third one is fine as Eater, but once it reaches Unfettered status, Dominions does a check that says it shoudln't so it goes poof instead.

That's my guess as to how the game logic works with it. Just speculation though, but it fits the observed behavior.

Re: Bug Thread: Discussion
 

Well, I also noticed if the first Eater was growing in power and reached about 7000 hp, while the prophet one was still about 500 hp. And it seems like the game consideres the prophet Eater to be my prophet as I'm still unable to make a new one.
Then something else has happened: when the prophet Eater was killed - the same turn the Third Eater U997 has appeared in the same province... Well not third now since prophet is dead - but there are still two indie Eaters in the world, here's screenshot of hall of fame.

Re: Bug Thread: Discussion
 

Problem with .map files and #specstart - LA Atlantis and LA MArignon are nations that want coastal provinces. But if game has none valid coastal starts they can start on normal inland prov. Problem occurs when you want to place nations by #specstart, Mari/Atlantis either starts nowhere [and dies next turn] or replaces other nation [which dies]. Placing all nations by #specstart and leaving 1 free spot for Mari/Atlantis does not help, it makes it even worse [it still replaces one nation but mari/atlantis still dies, so 2 nations vanish]. Are there any solutions to that?

Re: Bug Thread: Discussion
 

I don't know if this a game bug or a mod bug.

I've been playing with the Vampiri mod race and three times now during turn generation Dominions has vanished. Loading it up and hitting End Turn always works, though.

The reason I suspect the game is that it's also when I upgraded to 3.21.

Re: Bug Thread: Discussion
 

"Vanished" is not a useful error report. The game gives an error message when it crashes, you should copy and paste that to determine the kind of problem there is.

Re: Bug Thread: Discussion
 

The Eater of the dead is no longer an entity with one body. For several patches now it has had the ability to split itself into neighboring provinces, slowly growing and turning every province into a field of animated flesh. Each manifestation of the otherworldly eater of the dead has a singular body, but the entity can manifest in several places.

The perceived bug might be a bug, but might also be a side effect of the multi-bodied entity mechanic.

Re: Bug Thread: Discussion
 

It's only not useful if it's not clarified in any way. But Dominions still has the "CTD during turn resolution" problem sometimes, and when that occurs, it just dumps you to desktop without any fanfare or explanation. It certainly sounds like that is what Loren is reporting (I think it was listed in the patch notes as fixed in 3.21, but I can confirm seeing it since then as well).

Re: Bug Thread: Discussion
 

I mean vanished. No error message. One instant it's there, the next it's gone. I've never seen exactly what phase of the turn generation it's on when it happens.

This probably means a stack overflow.

Re: Bug Thread: Discussion
 

This sounds exactly like what I'm seeing.

Re: Bug Thread: Discussion
 

In-game messages like "I'm setting taxes to 0% so you go bankrupt" can cause the turn to crash when they are viewed.

The game formats messages that are displayed on the screen using one of the printf calls. It should pass ( ..., "%s", message ) for user-entered messages. But instead it passes ( ..., message ). So if 'message' contains printf formatting, it will expect additional arguments. And if the printf code is '%s' (spaces between the '%' and 's' don't matter), it will treat whatever happens to be on the stack as an address to read a string from. If this address is invalid, that can cause a crash.

Re: Bug Thread: Discussion
 

This is a serious problem. This can cause much more than a simple crash. If someone makes a malicious message, it could take over your computer.

http://en.wikipedia.org/wiki/Format_string_attack

Re: Bug Thread: Discussion
 

Oh, wow. %n does not modify the output from printf but instead treats its arguments are a memory address and sets it to the number of characters printed so far. That raises the threat potential from printing out the contents of your Dom3 process to modifying memory, including the instruction pointer. http://julianor.tripod.com/bc/formatstring-1.2.pdf

It's interesting that vfb reports that this will cause crashes. Maybe Dom3 is compiled in a mode that does stricter checking of printf, and throws an exception if the wrong number of arguments is supplied. In that case it's not a security threat after all.

-Max

Re: Bug Thread: Discussion
 

No, I just said %s will cause crashes. I did not think of %n, I was not aware of that actually.

The printf call used does check for a null argument to %s on the stack and prints "(null)" in that case, but it's going to seg fault (crash) if there's something on the stack like a random integer value.

It's impossible to do a compile-time check of the printf arg count when the format string itself is variable. And that's the problem here, the format string should be "%s" instead of the user-entered message.

It's also impossible for a library function like printf to know how many arguments it was actually passed. Whatever is on the stack is just there, and it will try to use it according to the format string.

Re: Bug Thread: Discussion
 

Black Laurel, Ivy Crown and Crown of the Ivy King do not provide an armor to the wearer, bug or WAD? I remember that a couple of crowns were missing their armor, and KO said something like "remind me of any other crowns like this".

Re: Bug Thread: Discussion
 

Should they provide armor? They're leaves after all :)

Re: Bug Thread: Discussion
 

I don't know if this has been mentioned but:

If I use the random map generator once, quit that game and generate a new map I get the exact same map grahpics but with each province redefined (water provinces will often be land and vice versa).

Re: Bug Thread: Discussion
 

That happens if you save the map under the same name as before and is because of image caching in Dom3. You'll have to restart the game in-between or use another name. It's a bug that has screwed at least one MP game that wanted to start that I played in.

Re: Bug Thread: Discussion
 

Bug using gems against PD when mayusegems is initially 0:

I had a battle against PD, where my commander was loaded up with a couple nature gems (I was expecting something more).

So, he didn't cast the spells he was scripted for, I checked the debug log, and I see (mayusegems 0). No mystery there.

My army proceeds to trash the PD, the infantry runs away, and there's just archers left shooting at me still. My mage lets off some buffs, etc, and then ... Swarm! My army had taken very minimal losses, just engaged the archers, there were no new enemy summons on the field, but I check the debug log and for some reason I see for my casters (mayusegems 1). It toggled on in turn 8 of the battle.

Is this a known problem?

Re: Bug Thread: Discussion
 

Units with Forest Survival + Gluttony still starve inside of Forests? Is this WAD?

To clarify, I have an army as Lanka, in a Forest. Every unit present has Forest Survival. Yet, it still states "Supply Used 490", I got a Starvation report, and my Palankashas all have the Starving debuff on them.

Re: Bug Thread: Discussion
 

I believe I've seen this problem before as well. Battles where my mages don't use the gems they were scripted to use then use the gems later on, even when I'm in a winning battle. The worse part is that they are always spells I would never bother casting.

Also, I'm not sure how the opposition is weighed to determine when to use gems or when not to, but it appears that a lone black servant with almost no gear and scripted to retreat can cause all my mages to use gems. This is rather annoying as the unit presents no threat to my army, even if it were to attack rather then flee.

Re: Bug Thread: Discussion
 

I think the gem use starts because the mage's fatigue gets high and he doesn't want to drop unconscious when the battle takes a few turns, even when the opposition is outmatched. Still buggy, but that might be the root of the problem. Just a thought.

Re: Bug Thread: Discussion
 

They eat a lot more than the forest survival bonus provides in forests.

Re: Bug Thread: Discussion
 

Thanks, if I have the same situation I'll try scripting my casts, followed by "Stay Behind" instead of "Cast Spells", and see how that goes. I'm hoping the mage might just decide to do nothing instead of burn gems.

Re: Bug Thread: Discussion
 

I haven't used "cast spells" for months. I still suspect that this is a problem. Still a thought to consider.

Re: Bug Thread: Discussion
 

:bug: Hi, didn't saw this one in the bugs shortlist. When scripting cave drakes (conj 2 summon) to fire closest, they don't seem to fire. In fact, I never saw them firing anything - I would expect something like shards or stones... or maybe they do fire and this is a graphical glitch, but I don't think so, they just run into melee. Ranged attack seems broken for these poor creatures...

Re: Bug Thread: Discussion
 

Cave drakes do not have a ranged weapon unlike fire and ice drakes.

Re: Bug Thread: Discussion
 

Ouch. I thought they had...

Re: Bug Thread: Discussion
 

Then wouldnt it be a bug to even be able to choose "fire"?

Re: Bug Thread: Discussion
 

I suppose, but every one gets the "fire" command.
So it's more of a design decision than a bug.

Re: Bug Thread: Discussion
 

Be hard to restrict the fire command. You can mix archers and infantry and archers can lose their arms, in which case they lose their bows. If the availability of the fire command was dependent on the presence of missile weapons, it would lead to all kinds of micromanagement problems. It could also have all kinds of other related bugs attached if that were the case. So my vote is that it was a design decision as well and in my opinion a good one.

After all, a unit that has received mistaken fire orders but can't execute them will default to attack anyway, so no harm done on that score. And if the whole strategy depended on the non-missile units being missile units, well, that's the general's fault, isn't it?

Re: Bug Thread: Discussion
 

I've just suffered a very strange and somewhat disruptive bug...

As MA R'lyeh I've received the Visitor hero, and have sent him out conquering with an army in tow. However, this turn I got two messages:

Suddenly the Visitor disappeared in a puff of smoke.
Suddenly the Visitor appeared at the gates of your Dark Citadel and wanted to fight for your cause.

The problem is that all the aquatic troops are stranded in the province he disappeared from and all the amphibious troops are in squads under his command, but are not being counted (the army setup screen claims he is commanding "0 units in 4 squads")

Now, it's an MP game on the llamaserver, so I can probably just roll the turn back, unload the troops and then spend a couple of turns moving the Visitor back to them ('cause of the mapmove 10), but it would be a crippling bug for most nations to have a national hero suddenly snap back to the capital with a bugged army.

EDIT:

It appears that the troops under his command are actually still in the province he disappeared from (though they appear on his list in the capital). They can be selected individually and removed from the squad, at which point they appear in the province he disappeared from.

Re: Bug Thread: Discussion
 

And that's why mother warned you about tagging about with extradimensional entities! I mean, really, they don't even have decent bathrooms in Limbo...

Re: Bug Thread: Discussion
 

I don't know if this a bug or some sort of strange mishap, but I tried a EA Abysia nation, and somehow the misbreds acted like undead. They can only be placed under commanders with undead leading capacity. But they are not listed as undead beings.

But they are not listed as magical beings. At first I thought this was a mod issue. But after I removed the mods, it still happened.

Sadly, this makes them fairly useless. There are only two leaders who can lead undead. And only small amounts. Is this a bug? Or intentional?

Re: Bug Thread: Discussion
 

Misbreds are intentional and that's hardcoded.

Re: Bug Thread: Discussion
 

Aren't them demons? :)

Re: Bug Thread: Discussion
 

No, which is why they are such a surprise to many people. They would make EA Abysia nearly unstoppable if they were normal units, because massing them is trivial. It's finding leadership that is difficult.

Re: Bug Thread: Discussion
 

Yeah, but it is odd that they don't have the undead icon. So it is hard to see that only special units can lead them.

Re: Bug Thread: Discussion
 

They're not undead. Or demons either, which is actually what they'd be.

They can't be banished etc.

The only trace of their demonic heritage is the need for special leadership.

It would be nice if it were more obvious. Is there anything in the description that at least hints at it?

Re: Bug Thread: Discussion
 

Nope, nothing, that is why I thought it was a bug.

Re: Bug Thread: Discussion
 

If memory serves me correctly, the current setup is a fix from when they use to be regular demons. The change occurred because MA Abysia had demon priests, which had the unintended and unthematic consequence of allowing them to reanimate. The demon tag has since been removed and their undead leadership requirement is now hard-coded in.

I would suggest that these units be given a new tag to indicate their status. Perhaps a half-demon tag?

Honestly, undead leadership isn't as hard to find as you might think. It is trickier to find undead leadership on units that also fly and are stealthy, which is what you need to fully realize the misbred's potential. Take a god with blood and death. Start with Revive King and move onto Bind Succubus. Vampires and Arch Devils will be optimal from there on.

Re: Bug Thread: Discussion
 

Or give a Rod of the Leper King to a Demonbred.

-Max

Re: Bug Thread: Discussion
 

I'm not sure if the Demonbred are still immune to disease. Have you tested this?

Re: Bug Thread: Discussion
 

A new symbol for one unit only? Though not impossible, extremely unlikely to happen.

Re: Bug Thread: Discussion
 

Looks like this will be fixed in the next patch. From the Changelog:
* Creating new game with random map and same name without quitting resulted in wrong map graphics being displayed, fixed.

Re: Bug Thread: Discussion
 

Bug: Fortress not collecting resources from nearby provinces

This is on the Glory of the Gods MP map. The game is using the conceptual balance mod. I'm playing Atlantis, and I built a Fortified City on province 78, near the bottom right corner, adjacent to the two sea spaces that should be but aren't neighbours. I control all neighbouring provinces, one land and the rest sea. The province has only 32 resources, when it should have (by my estimate) over 100.

Re: Bug Thread: Discussion
 

Fortresses in land provinces don't collect resources from sea provinces. I believe this is mentioned in the manual.

Since seas have so many neighbours, it would be easy to accidentally deplete large seas out of resources if this were different.