List of issues/bugs in 1.102

[mystikx21]

Plus some fixes (or proposed fixes). I'm partly cataloguing these here so I can work on some of them in a checklist and post some fixes.

1) Banaue/Lalibela are not hill-limited on +1 / or territory pops. Doesn't appear to be any lua code to scan for hills anywhere. The effect also is not city limited. Any tile in the city radius is improved regardless of what city works it. Appears fixed

2) No artwork used for new wonders appears to cause some "ghost" cities. Easily fixed with available art. Fixed.

3) Panama doesn't appear to provide +1 using the tradition bonus as it's not listed on the tooltip. (Could just be a very busy wonder and it got shoved off but I doubt it as there are other busy wonders). Itaipu does this also. Must have to do with the tradition code. I suspect it doesn't detect them for lack of culture? I will check. Fixed.

4) Courthouse gold price is fixed at a 100 building cost price, while production price varies with size (+10 per?). It's therefore always more efficient to buy it even though it is intended to be far more expensive with gold. I'm testing this one as a fix by putting the per pop cost modifier ahead of the hurry gold price. My guess is it runs the one before the other and causes a mismatch in the gold cost when going to buy it. Fix didn't work but I'm pretty sure that is what is happening is gold cost is fixed. Could be cost is fixed too. Appears fixed, still testing.

5) Liberty specialist +2 is way too powerful and needs to be adjusted for balance. Ideally to a + bonus for using a specialist per city rather than per specialist. Early game finisher is as powerful as a late-game wonder right now and makes most mines and most villages essentially useless for most of the game. Fixed for now. Possible adjustments to AI still needed to make the policy changes work more effectively but better balanced.

6) City screen hangs when changing/locking tiles/specialists. Slight delays on selecting workers or cities late in game. Hard to track what the cause is in the mod but seems replicated by many others. Much improved.

7) Unit upkeep appears to adjust upward with game speed. I'm not sure if that's intended or not, but it's looking like ~1.5x or 2x the cost on slower game speeds and 2/3 on faster. When deleting a unit, change in net gold is more/less than the displayed cost and the change in unit maintenance costs is higher also. For an ironclad of 10 , it caused a difference of 22, Rifle of 13 , 28, on a marathon game in unit costs.

It displays in the hover text as though it's still at normal speed so I suspect it's not intended. Should be easily fixed if gamespeed or a related variable can be used in the formula. Potentially a ton of money on epic or marathon games especially late with high cost units. The formula runs after costs have been increased for units in the units directory as it should. I'm not sure if it runs after any game speed modifiers to that cost or not have taken effect. Nor am I sure when the upkeep cost as tooltip is generated for display.

8) "City happiness", ie excess local happiness, appears to cause a disparity in the top bar UI happiness, which is then often several points off by the late game. In a current game in 1920s it's off by 11. This added up roughly to the difference the last few times I checked within 1.

9) Some text issues on some tooltips. Faith in city screen works correctly when moused over but not on visual inspection. Some buildings/units in city build queue help text are incorrectly defined. Tooltips are working for each correctly but text once they appear in the queue is incorrect or out of date. That could be confusing. The tooltips themselves work properly however. Other than belief related effects on shrines.

10) Shrines appear to be providing belief bonuses associated before beliefs are selected. (working on this one now). Appears fixed.

11) Promotion tooltips in some cases need updated, mostly reflecting recent promotion changes like swords +25% vs cities. Fixed.

I had some conflicts with my own personal modifications which brought back some zombie promotions (sentry and coastal raider 2 most notably, as I think both are nixed). I'll test this one again when the next version comes out as I think I got it sorted out finally, but it's not as significant as a mouse over on some promotions conflicting with their bonuses in actual combat. (Again most of these are easy to fix). Fixed it. (Mostly, still glitchy on melee ships promos).

12) Diplomatic text piles up on one side during peace talks making it hard to sort out agreements of who gets what. Also terms cannot be adjusted. The apparent error is at least consistent, even if I could not identify the effective source in the actual code. Fixed.

 

[Lord Oden]

Air repair is also not an option for bombers at level 4 as the guide suggest. It is an option for fighters at level 3. Why the difference in levels?

 

[Lord Oden]

Japaniese Dojo does not give +1exp per turn in city as stated.

 

[mystikx21]

Air repair is also not an option for bombers at level 4 as the guide suggest. It is an option for fighters at level 3. Why the difference in levels?

Presumably the difference is to give fighters as defensive units for the AI more potency and slow down (our) bombing slightly.

I'm trying to figure out why it wouldn't be available now. It looks fine on paper. (ditto the Dojo, though that's a lua effect that I'm less sure about the debugging)

 

[Thalassicus]

The courthouse problem is likely a mistake in:
/civup/3_interface_branches/ProductionPopup.lua

The slight delay when moving tiles or specialists is the game recalculating all yields for all human-player cities. The code for this is written in lua, which is slower than the core c++ language. I think the only way to speed things up is change the main part of civup to a core mod, which I estimate will take about 100-150 hours, when I get time for it.

I'll see what I can do about the rest.

 

[mystikx21]

Banaue seemed to be working. Some of the tooltips were already fixed. I'll test the wonder fixes more and make sure they work as expected.

The wonders happiness issue looked related to the isvisible check in the traditions sql effects and the lack of artwork. Maybe?

 

[jacktannery]

The slight delay when moving tiles or specialists is the game recalculating all yields for all human-player cities. The code for this is written in lua, which is slower than the core c++ language. I think the only way to speed things up is change the main part of civup to a core mod, which I estimate will take about 100-150 hours, when I get time for it.

Thal, does this mean that the number of humanly-controlled specialists can impact on this lag? For example, if I played a specialist-heavy game with fifty cities would that theoretically have a longer lag than a specialist-light game with twenty cities? Or does the overall number of cities and specialists (AI & human) and thus map-size make more of an impact?

 

[mystikx21]

I notice the lag when adjusting tiles period. That does cause unemployed specialists temporarily but I am pretty sure Thal is right it has to do with all the yield calculations and not specialists, probably not map size either.

The speed fixes when yield tables were not updated as frequently (and as well) demonstrated this was a big issue.

 

[jacktannery]

Ok thanks Mystik - I misread Thal's post, reading 'specialists' for 'tiles'. One question though:

-Why does UP/GEM need to count yields in human cities so much more than vanilla? What exactly does this do, other than the little culture button that tells me how many turns my city will expand. All the yield counting in Up/GEM should be pretty much the same as the yield-counting in Vanilla, no?

 

[mystikx21]

Probably yes, but there's other sources of yields than in vanilla that have to be factored in the equations.

There aren't per pop faith, gold, or culture buildings for instance (I think), for which the totals are then modified by percentage buildings and policies and so on. I don't know if this changes how often it has to run the totals, but it changes what has to be checked for when you do it.

 

[Thalassicus]

Civup adds these sources of yields:

• First specialist yield change
• Culture per pop
• Trade deal modifier
• Golden age points
• Happiness from terrain
• Distributed citystate rewards
• Open borders gold modifier
• City gold per luxury percent
• Golden age surplus yield mod

To add things like this two years ago, I needed to recreate the entire part of the game which gives us yields. This is slightly slower because the lua programming language (which we had access to) is less efficient than c++ (the game core stuff). Each city calculates each source of base yields and modifiers, surplus yields and modifiers, total yields, modifiers, rate, storage, and yield needed. It does similar calculations for each player.

It's possible to translate all this into the faster game core c++ code. I'm going to work on that when I have time. It took hundreds of hours to do it the first time around, so there are no quick solutions.

"City happiness", ie excess local happiness, appears to cause a disparity in the top bar UI happiness, which is then often several points off by the late game. In a current game in 1920s it's off by 11. This added up roughly to the difference the last few times I checked within 1.

This is probably a forgotten source of happiness in either the game calculation or tooltips. It would help if someone can post a savegame when they encounter this bug, so I can total up all the sources of happiness, and figure out what might be missing.

 

[jacktannery]

Thank you Thal.

 

[mystikx21]

"The display of unit maintenance now scales to game speed (visual change only)."

Was unit upkeep cost intended to increase with game speed? (building upkeep does not)

Or is it easier to fix that it's providing correct information in the short term and look into how to shift it later?

 

[mystikx21]

This is probably a forgotten source of happiness in either the game calculation or tooltips. It would help if someone can post a savegame when they encounter this bug, so I can total up all the sources of happiness, and figure out what might be missing.

Here's what I've got at the moment. The second screen appears to be the total the game is using based on the change in culture from tradition, and any production penalties, growth penalties, or science bonuses when applicable.

I've been checking it every time the difference changes and it now seems to be the dummy "city happiness" buildings that appear after events (I think) that add up to the difference correctly each time.

The only other disparities between the two summary screens are terrain (caused by events and natural wonders in territory?) is not listed and the social policies total is different. But social policies includes a chunk of local happiness caused from wonders in cities, so I'm guessing that's not it. I am pretty sure I have 22 wonders in that game to explain the gap, plus a couple that weren't working with tradition correctly. But I didn't check the total against the older wonders glitch.

It's an slightly older version (1.102) that I'm finishing up a game on, so a savegame may be moot (might be fixed already?). I'll post one when I finish and resume updating, if it re-appears.

(the tile values disappear sometimes on mouse positions in the city screen too, so there aren't any in the last shot... but I have no trouble getting them back so it's not a big concern for me).

 

[Thalassicus]

The totals appear to add up correctly: 64 = 772-708. I thought you said it was off by 11? The topbar tooltip is missing a line for city happiness, which I'll add.

I thought I'd disabled those dummy city happiness buildings. That's weird! I'll fix that.

I see I also need to update the economy screen. It's easy to forget because I have not used the economy screen or non-tech left-side panels in two years... all the important information I need is available with a quick glance at the top bar. The other screens should be accurate though, for those who want to use it, and I will fix that.

Was unit upkeep cost intended to increase with game speed? (building upkeep does not)

Or is it easier to fix that it's providing correct information in the short term and look into how to shift it later?

The game core handles scaling of unit upkeep, so changing it is impractical at this time.

 

[mystikx21]

The actual bug is that the top bar adds up correctly but is not displaying my actual game happiness. The economy screen is however displaying whatever the game is using correctly on the resource/happiness screen. This is very important near break points. Not so important when I have +51 instead of +64.

The only thing that adds up to the difference (off by 13) is the dummy city happiness buildings total in my empire.

Hmm. Weren't there variables for unit upkeep in the mod? So what we are dealing with is the game has its own upkeep changes for speed? Even on top of unit upkeep changes the mod makes? Or that the unit upkeep values also scaled for speed and weren't being displayed before?

 

[Thalassicus]

In other words, the top bar is correct, and you're not getting what it says you should? This leads me to believe it's a problem with the code creating a stack of national happiness in the capital to modify happiness.

We can set the maintenance cost of units, but not (easily) change how the game uses that information.

 

[mystikx21]

Yeah that is basically the problem. It's routinely off by double digits by the late game. I think it has something to do with events adding happiness. But I would not swear to that.

Ahh. Having the correct information is a good start then. The game shouldn't really be scaling it. It makes tanks extremely expensive

 

[Drakir89]

Monarchy reduces unhappiness from capital like it does in vanilla, in addition to its other bonuses.

 

[mystikx21]

It does, and I'm not sure it was supposed to. But I don't think that's causing a disparity in the UI. It would be off 17-18.