Report - Current Bugs and Issues

[sven]

It seems that if an allied empire merges with yours, any civillian population transports in motion will remain as a "ghost" of the old empire. This is the second game where I have one (I think both Phidi, even) that just bounces between two systems.

As best as I can tell, "ghost" transports were being left over when an empire was eliminated by conquest, but, from the tests I've done, transports should be e properly transferred over when an empire is absorbed. The "ghost" issue I was able to find and fix should be resolved by patch r38550, so if you see this happening in any games started after that patch, please upload the save and let me know.

[gaerzi]

This is a harder one to fix properly. The issue, more or less, is that the tool I'm using to draw the icons shown in the combat estimates only supports drawing single icons, but, of course, many different unit/population types may potentially be destroyed by ground combat. To solve the problem properly, I should really switch to a clustered icon drawing tool (which is a thing the game does have), and use that to form the suffix, rather than just kludging in a single static image. But... it's a minor, largely cosmetic issue, and making sure the clustered icons looked good and where positioned correctly would be a bit of a job to really get right.

I don't think we're talking about the same thing. I'm referring to the text blurb that says:

Our fleet is bombing Grumium I. We have destroyed 1,500,000 Gremak, 1,500,000 Human Slaves, and 1 battle mech brigade.

But the enemy, a minor marauder faction, does not have battle mech brigades, it only has marauder raiders. IIRC, when the combat started, they had 8 militia, 5 raiders, and 4 800 000 civilians.

Yeah, the pattern "if lua_load_complete and X" is sufficient to stop test code from running when the program first boots up for players, regardless of whether X is true or false. I'll usually try to remember to set X to false when I'm done with a test set anyways, just so I don't trigger the code accidentally if I'm editing the file for some other reason while the game is running, but, I don't always remember to. (In short, yes, it should be false, but there are probably many other incompletely disabled pieces of test code scattered around the codebase, and while they might surprise modders, they shouldn't cause bugs for players.)

Okay. I guessed it wasn't that much of a problem after I noticed the one in Drawers\ImportDrawer\quick_pop_icon_import.lua that changes the population on Dzibix and saw it didn't actually affect a Tinkers game. There's also one in AI\peace_war_trades.lua, one in Galaxy\stargates.lua, and one in debuggee\debugtest.lua.

This looks like a fairly deep bug in the tactical combat data structures. If you see a negative fighter count appearing again, please upload the save, I'd like to take a closer look at it.

Will do. Though we can't save during tactical combat, can we? So it'd be a save from after the fight.

[sven]

Will do. Though we can't save during tactical combat, can we? So it'd be a save from after the fight.

You actually should be able to save during a tactical combat -- though I'm sure many people don't notice the feature

[zolobolo]

You actually should be able to save during a tactical combat -- though I'm sure many people don't notice the feature

This was a very cool feature when I noticed it first time: having the ability to close the battle and lurking around on the galaxy map

[bjg]

game_8529
Fargone, can't build a planet from the Gas Giant because of the outpost.

[PrivateHudson]

You actually should be able to save during a tactical combat -- though I'm sure many people don't notice the feature

Bravo! This level of engineering does impress! I hadn't even tried it because was certain no one sane indie dev would spend the effort.

[Captainspire]

On a Windows 10 laptop, this is what I have. Im unsure if its the dual drivers, or something, but loading takes about 2 minutes, and each in game click takes 20-30 seconds to respond. With all the threads on graphics, I wasn't sure which one specifically dealt with this, or if my Laptop was just incapable of properly handling the game?

** EDIT** So apparently you can select which card/driver runs the program. I switched to the Nvidia one and it flys. ALL GOOD!

[gaerzi]

Being told that marauders are tanks annoyed me, so I went into the lua files, and eventually found:

@@util\format_lists.lua

Code: Select all

    if GroundCombatUnits[race]
      -- this is wrong!
      pushf('%d %s brigade%s',damage,singular | planet.tank_name:lower(),damage==1 and '' or 's')
    elseif RACE_INFO[race]
      pushf('%s %s',pop_string(damage), race_string_Plural | race  )
    end

Yes, this is wrong indeed, it just assumes all ground combat units are tanks. Now I'm not especially comfortable with Lua yet, so it's not a very elegant structure to just chain in a bunch of elseifs and I'd rather use a string array, but I made this change and it seems to work. I've gotta test it with attacking some planet where there actually are tanks, though.

Code: Select all

    if GroundCombatUnits[race]
		if race == 'tanks'
			pushf('%d %s brigade%s',damage,singular | planet.tank_name:lower(),damage==1 and '' or 's')
		elseif race == 'armored_raiders'
			pushf('%d space marine brigade%s',damage, damage==1 and '' or 's')
		elseif race == 'marauder_raiders'
			pushf('%d marauder part%s',damage, damage==1 and 'y' or 'ies')
		elseif race == 'pirate_raiders' -- those don't seem to be in the game currently but it doesn't hurt to include them
			pushf('%d pirate crew%s',damage, damage==1 and '' or 's')
		end
    elseif RACE_INFO[race]
      pushf('%s %s',pop_string(damage), race_string_Plural | race  )
    end

Spoiler: show

After testing on a planet with tanks, I somehow broke the singular, which befuddles me as the string construction line wasn't changed, but it otherwise works.

Spoiler: show

[gaerzi]

When annexing an empire, if it doesn't unlock any new technology (for instance, the race-exclusive ones were already researched), you get told that "the following technologies are now available for research: ." In this situation, the entire line should probably be withheld.

[nathanebht]

Draw occurred due to ally in battle. Upload game_8622.

Near the top, right of the map at the planet Izden. I wanted to capture Izden with its shipyard intact. Easily done with the fleet I have there.

The Phidi are allies and one of their troopships showed up for the battle. Twice I've played the battle and had this happen. Move up my ships and eliminate the planet's weapons. At that point the battle ends in a Draw.

Even the autocombat causes a Draw.

If you destroy the shipyard first and then take out the planet defenses. Then your given a win.

[gaerzi]

Did you board and capture the shipyard? Since it doesn't move, every ship can do it.

If you want to capture the shipyard intact, you need to capture the shipyard during space combat, and then capture the planet during the same turn. Any orbital station captured in combat is destroyed if you do not capture the planet also; but if the station isn't captured or destroyed then the fight is not finished.

[nathanebht]

Did you board and capture the shipyard? Since it doesn't move, every ship can do it.

If you want to capture the shipyard intact, you need to capture the shipyard during space combat, and then capture the planet during the same turn. Any orbital station captured in combat is destroyed if you do not capture the planet also; but if the station isn't captured or destroyed then the fight is not finished.

I take out the planet's defenses before capturing the shipyard. This works fine if an ally's troopship is not there. It definitely is a bug.

I could have probably worked around the bug if I really wanted to but capturing a shipyard is tedious anyway. Unless I really have a use for capturing one, I normally don't bother.

[sven]

Being told that marauders are tanks annoyed me, so I went into the lua files,

Hah. And well done -- I'd entirely forgotten about that wart in the code.

After testing on a planet with tanks, I somehow broke the singular, which befuddles me as the string construction line wasn't changed, but it otherwise works.

This is actually my fault, not yours. I have a noun form lookup for 'hover tanks' but not 'hovertanks'. So that logic was always broken in the hovertank case.

In any case, I've done an official version of this fix, and it should be posted as part of today's "in_development" patch.

[sven]

When annexing an empire, if it doesn't unlock any new technology (for instance, the race-exclusive ones were already researched), you get told that "the following technologies are now available for research: ." In this situation, the entire line should probably be withheld.

Yup. Should be fixed in the next patch. Thanks again for the help.

[sven]

I take out the planet's defenses before capturing the shipyard. This works fine if an ally's troopship is not there. It definitely is a bug.

Yeah, this one was connected to some fairly deep flaws in the logic around determining whether battles are won, lost, or stalemated in 3+ sided conflicts. I believe it should be fixed as of r38564.