Legend of the Invincibles

[arobinson]

I am getting a hang with a demon battle. Tried taking out the furious trait, but that does not seem to be the cause this time. I have to force quit Wesnoth when this is happening. It is happening during Azezel's turn on the save provided usually after Surcyn (the duke) is attacked (he still has his weapon raised like leadership when it has happened to me).

Turns out it was the "swarm" attacks (used the process of elimination) of the Azaag and Zis units in that save gave, so I must have incorrectly modified my save game the first time.

[mattsc]

arobinson, all: The cause of these hangups is the Fast Micro AI from Wesnoth 1.13 which Dugi imported into LotI. That AI was created for a somewhat different purpose: battles with very large numbers of mainline or mainline-like units, rather than battles with the type of "super units" appearing here. I have been meaning to change that anyway, so that the AI is able to deal with both of these types of "non-standard battles", and have used the hang-ups reported here recently as a motivation to actually do so.

I am pretty much done with that, just need some more clean-up and testing. The modified AI works without any long delays with the three replays posted here recently that I am aware of to cause hangups (LotI2-United_Turn_4.gz, LotI2-Halls_of_Miscreation_Turn_6_2.gz, LotI2-An_Old_F(r)iend_Turn_15.gz). If there are other test cases, please let me know about them. The more the better, in this case.

I will send an updated version to Dugi within the next couple days. Of course, then we will have to see what other bugs I introduced.

[arobinson]

Here is another save that will cause a hang. It is not the first attack, but after one or two it hung

[arobinson]

Another error seen

[Dugi]

The hang problem was fixed, I have seen no issue when using my version. Same for the error message bug. I have notified mattsc about it anyway.

_________________
Btw, I have passed physics PhD admission exams, so it is very likely that my activity here will remain quite constant for the next 4 years.

[xuanquang]

Hi Dugi,
Some problems about the newest version:
1. The AI: enemy leaders like to attack anything in their rank. Sometimes they go out from their castle and attack my units, get retaliation and die. This causes a lot of exp from units they recruit go away.
2. The Outcast bow: one modify it gives is "10% more attacks". It does not make sense! To get benefit from this modify the equipped unit have to have up to 10 bow base attacks, and until now I have never seen any unit could achieve that. I think "1 more attack" is better.
3. The comment in every item: it is interesting, but the item list is longer. Now I distinguish every item in the list by its grey comment line, but some items do not have their comments. It takes more time to find these ones.

[arobinson]

Some small things noticed with my Duelist Wizard when I gained his Ice dragon legacy spell:
The hood went away in place of hair when casting the "icy breath" (female attack image)
The explosion from Koschei's journal did not get applied to the "icy breath" attack

If I manually move the "advance" tags in the save file for KJ1 and KJ2 and then select "Alter advancement" and re-choose the KJ2 advancement, the explosive trait then gets added to the "icy breath". So perhaps you have logic that could ensure that the KJ* advancements are moved towards the end on level up?

[arobinson]

I've also seen that the blackguard from the gladiator duals always shows a message on advancement or changing items. Looks like it is from stats.cfg:

Code: Select all

		[if]
			[variable]
				name=has_charge
				equals=false
			[/variable]
			[then]
				{VARIABLE x $advanced.attack[$i].specials.damage.length}
				{VARIABLE advanced.attack[$i].specials.damage[$x].id "latent_charge"}
				{VARIABLE advanced.attack[$i].specials.damage[$x].multiply 1.5}
				{VARIABLE advanced.attack[$i].specials.damage[$x].apply_to both}
				{VARIABLE advanced.attack[$i].specials.damage[$x].active_on offense}
				{VARIABLE advanced.attack[$i].specials.damage[$x].filter_self.filter_adjacent.ability "charge_leadership"}
				{VARIABLE advanced.attack[$i].specials.damage[$x].filter_self.filter_adjacent.is_enemy "no"}
				[if]
					[variable]
						name=has_charging_backstab
						equals=true
					[/variable]
					[then] # Undo this if the unit would be also bastabbing
						{NARRATOR_MESSAGE "has_charging backstab $has_charging_backstab"}
...

Also, I'm curious, when do you think you'll release the next version (days, weeks, etc.)?

[Dugi]

@xuanquang
1. AI so you don't get exp for cheap kills. Problem?
2. It can make a change if you combine it with an item that adds some percentage to attacks.
3. I wanted the comments to be invisible on all lists of items, but it seems that I have forgotten it somewhere. I have comments written for most items now, so they should not be missing anymore. One of exception is the item named after you, what comment do you want it to have?

@arobinson
The hood went away - strange, I can't find what is causing it. Does it always happen or only sometimes?
The explosion from Koschei's journal - it does not apply to attacks added after getting the explosive advancement. I know that. In theory, I might push it on the top of the list, but exceptions to rules are usually not the way how to deal with things. I might, however, create a group of advancements that will always be pushed on the top of the list. I'll think about it.

@arobinson
I have fixed that bug long ago, I just haven't uploaded it yet.

I know that it needs an update. I was waiting for a guy to send something, I am not very sure when he'll do it tomorrow or in a week or later... I am also relatively busy till next Friday. We might make a deal that if I don't get it till the end of July I'll upload the update.

[mattsc]

The hang problem was fixed, I have seen no issue when using my version. Same for the error message bug. I have notified mattsc about it anyway.

Hi all -- just confirming that I checked this too and that both of those issues are fixed in the new version of the AI that I sent to Dugi a couple days ago (the latter was, in fact, already fixed in a previous version). So as soon as he gets the next version out, these should not be a problem any more.

[arobinson]

The hood went away - strange, I can't find what is causing it. Does it always happen or only sometimes?

It always happens. Here is a save game. The unit is Rabry and is near the top of the map. Use the 'icy breath' attack.

Also, it seems to be that there is no item type to match fists (troll rocklobber/boulderlobber for example). Do you think you could make the metal claws be usable with fists or a new craft-able type? Thanks.

[xuanquang]

@Dugi:
1. Yes. A leader normally recruit 4 - 5 lv3 units per turn (at chapter 6), so a big exp mount is wasted if he die at his first turn. I can keep my armies far from him, but in some scenarios my allies attack his castle, he goes out, fights, gets slowed, cannot return his throne, then die after 1 - 2 turns later. I am playing at hard difficulty, so I think an enemy leader should not die so easily and stupidly like that. In previous version, enemies leaders do their job better (just stand safely in the castle, call a lot of units to protect).
2. The "Outcast" is a bow, so do you mean there are some armors those add percentages to attack? Uhm, it sounds powerful. I'll look for some.
3. I am not good in writing, but the idea is somehow like that: "Don't afraid when seeing a bloody road. Keep calm and turn back, now you face the one made that road."

[Dugi]

@arobinson
The animation bug seems to be quite large and old. It seems to affect all animations added by AMLA from advancements that are not unit specific. It was usually not seen because most units are male-only and the gender difference is usually not very visible. I have fixed it.
And by the way, as I expected before, it did not always happen - the cap disappeared only in a half of the attacks.

@xuanquang
1. But these units would be cheap kills if you wanted to kill the leader as the last thing on map. This is not intended, I will think if I want to maintain the status quo or ask mattsc to change it.
2. I was telling bulldust, sorry. I have increased it to 20% now.
3. I have slightly edited it, but you got it there.

[woseshaman]

1. Something seems the be wrong with the flying boots. They set movement for Caves to 1 instead of mountains
2. Charging backstab doesn't seem to fit it's description (defenders damage not being doubled)
3. Also i thought about why Efraim and Lethalia are immune to poison: Since demigods created as such using debug mode CAN be poisoned it must be because of their undead past. Maybe the lack of legacies undead Units have is also considered some sort of legacy which offers no advancements but makes the unit immune to poison/drain and is not removed when they level up to demigods. Anyway creating a demigod efraim with the exact same items instead of having him sort of levelling up should fix it (same for lethalia of course).

LotI2-Odd_World_Turn_1.gz

Savefile of a poisoned Demigod

(954.87 KiB) Downloaded 226 times

LotI2-Nightmare_Cellar_Turn_2.gz

Lethalia and Efraim can't be poisoned

(1.83 MiB) Downloaded 246 times

4. Fun fact: I think i broke the Storyline in ch. 4

LotI1-Bloodbath_Turn_10.gz

Why are we fleeing again?

(893 KiB) Downloaded 219 times

[Dugi]

1. No, that's intentional. Gameplay reasons overrode logic.
2. This looks like a bug in wesnoth, I am considering reporting it.
3. I have found something that looked like a cause. If it's not the cause then the cause of the problem will be a real trouble because it was the only undead-related thing that remained on them. Undead immunities persisting long after the removal of the undead trait is already quite strange.

Thanks for reporting.