Single Player bugs and crashes v35 plus (SVN) - After the 18th of August 2014

[Talin2009]

They are expanding if you have the traits off.

Updated SVN today, 50-60 turns passed - they still don't expand (not a single city have been built).

So what do i have to turn off in new game so AI plays normally?

 

[dan1ell1ma]

I got a strange bug.

I make a rogue, all going normal when suddenly the rogue cant attack, just push the enemy.

Anothe issue is my archer, when i destroy a road of enemy, the archer desapear!

 

[Thunderbrd]

@thunderbrd

This crash happens because of a stack overflow in this pLoopUnit->jumpToNearestValidPlot() call in CvUnit::setXY. You changed the previous canCoexistWithEnemyUnit into canUnitCoexistWithEnemyUnit that must cause it.

					if (isEnemy(pLoopUnit->getTeam(), pNewPlot) || pLoopUnit->isEnemy(getTeam()))
					{
						if (![B][COLOR="Red"]canUnitCoexistWithEnemyUnit(pLoopUnit, pNewPlot)[/COLOR][/B])
						{
							if (NO_UNITCLASS == pLoopUnit->getUnitInfo().getUnitCaptureClassType() && pLoopUnit->canDefend(pNewPlot))
							{
								[B]pLoopUnit->jumpToNearestValidPlot()[/B]; // can kill unit
							}
							else
							{
								if (!m_pUnitInfo->isHiddenNationality() && !pLoopUnit->getUnitInfo().isHiddenNationality())
								{
									GET_TEAM(pLoopUnit->getTeam()).changeWarWeariness(getTeam(), *pNewPlot, GC.getDefineINT("WW_UNIT_CAPTURED"));
									GET_TEAM(getTeam()).changeWarWeariness(pLoopUnit->getTeam(), *pNewPlot, GC.getDefineINT("WW_CAPTURED_UNIT"));
									GET_TEAM(getTeam()).AI_changeWarSuccess(pLoopUnit->getTeam(), GC.getDefineINT("WAR_SUCCESS_UNIT_CAPTURING"));
								}

								if (!isNoCapture())
								{
									pLoopUnit->setCapturingPlayer(getOwnerINLINE());
								}

								pLoopUnit->kill(false, getOwnerINLINE(), true);
							}
						}
					}

Hmm... that's an important difference there. I'll have to evaluate the situation itself to see if there's a way around it causing a crash.

Updated SVN today, 50-60 turns passed - they still don't expand (not a single city have been built).

So what do i have to turn off in new game so AI plays normally?

Would be interesting to see a savegame. I don't suppose you're playing with a unit tile limit or something odd like that?

I got a strange bug.

I make a rogue, all going normal when suddenly the rogue cant attack, just push the enemy.

Anothe issue is my archer, when i destroy a road of enemy, the archer desapear!

Both are quite interesting. Please provide a save that can replicate the situation. The archer at least... I can probably solve the rogue issue with a test case here.

 

[Talin2009]

Would be interesting to see a savegame. I don't suppose you're playing with a unit tile limit or something odd like that?

Nope, never liked unit tile limit playstyle. Here's savegame.

Started another game with developing leaders on, no positive traits at start on, focused traits on, will play it for couple days and see how it goes.

p.s. Im using civs megapack from modmods section (not sure if it's included in SVN now, im unpacking it manually each time i update SVN).

 

[dan1ell1ma]

Both are quite interesting. Please provide a save that can replicate the situation. The archer at least... I can probably solve the rogue issue with a test case here.

Here, the save, just try to destroy the road in Europe.

 

[dan1ell1ma]

CTD when load.

Dump file empty.

 

[Thunderbrd]

Ok... I've fixed the problem that caused the crash due to the recent changes for tunnel dynamics and also the one regarding always hostile HN units being unable to attack. I'll have to look into each of those last three one at a time here.

 

[dan1ell1ma]

Ok... I've fixed the problem that caused the crash due to the recent changes for tunnel dynamics and also the one regarding always hostile HN units being unable to attack. I'll have to look into each of those last three one at a time here.

OK, Thanks!

Now, when i attack, i can't move after.

 

[Thunderbrd]

OK, Thanks!

Now, when i attack, i can't move after.

Please be more specific. I need to have something to test. In many cases, units aren't supposed to be able to move after attacking anyhow.

I've just tested both of your saves btw and it appears both are fine now - it's entirely possible I fixed both issues with the last update.

 

[dan1ell1ma]

Please be more specific. I need to have something to test. In many cases, units aren't supposed to be able to move after attacking anyhow.

I've just tested both of your saves btw and it appears both are fine now - it's entirely possible I fixed both issues with the last update.

Sure, sorry!

I provide a save here.

I have a rogue near egypt, This rogue (and all others hidden units) only walk when attack, but not if just try to move.

 

[Noriad2]

I got a strange bug.

I make a rogue, all going normal when suddenly the rogue cant attack, just push the enemy.

In my last game something happened that might be similar. On two occasions, I attack a unit, but both my unit and the target unit survive. The target unit is very low health (one shot away from dying) but cannot be attacked anymore (attempts to attack fail). However if I save/reload, the target unit disappears, making it very hard to replicate the situation. I have reloaded the autosave a few times and redid the attack to try to replicate the bug but failed (enemy unit died as it should).

 

[Thunderbrd]

Interesting - I found I do have a replicable situation here. Furthermore, it looks like we might, in this, have found one of the triggers for the ghost unit problem here. Very interesting indeed... looking into it now. Thanks for the report!

EDIT: Yeah, it's not the ghost unit issue. I think that was solved by a correction regarding HP and Damage recently. THIS is far more... interesting. Some catch that shouldn't be taking place in canmove.

EDIT: Ok, found the issue and have resolved it. Update pending.

 

[dan1ell1ma]

Interesting - I found I do have a replicable situation here. Furthermore, it looks like we might, in this, have found one of the triggers for the ghost unit problem here. Very interesting indeed... looking into it now. Thanks for the report!

EDIT: Yeah, it's not the ghost unit issue. I think that was solved by a correction regarding HP and Damage recently. THIS is far more... interesting. Some catch that shouldn't be taking place in canmove.

EDIT: Ok, found the issue and have resolved it. Update pending.

Thanks, i test and its ok.

Can you see the other problem of destroy the trail path? When i do, all the units desapear of tile.

 

[Noriad2]

crash no longer happens; thx for the quick fix.

 

[Thunderbrd]

Thanks, i test and its ok.

Can you see the other problem of destroy the trail path? When i do, all the units desapear of tile.

I'll test this again later but my test of your save last night showed no problem.

 

[JosEPh_II]

Just updated to 8191 and Composite Bowmen no longer update to Bellybow or early crossbow. Last night, prior to 8191 update, my composite would upgrade to Belly. I just got early crossbow after the update.

Is this intentional that Composite can't upgrade to either belly or early crossbow now?

JosEPh

 

[Noriad2]

SVN 8191 research values do not get reset upon regenerate map.

How to replicate:
-start a game (in my case large map/deity/nightmare)
-settle in place. You now get 11 research points/turn
-regenerate map. Settle in place. You now get 21 research points per turn.

Seems like the classic type of bug of not setting variables to zero before
starting a calculation.

 

[Stormwind]

Rev 8191:

I have 3 Hashishins and whenever they are in the top of a stack or alone in a plot their lower part is a red blob. See screenshot.

View attachment 387454

 

[Thunderbrd]

Just updated to 8191 and Composite Bowmen no longer update to Bellybow or early crossbow. Last night, prior to 8191 update, my composite would upgrade to Belly. I just got early crossbow after the update.

Is this intentional that Composite can't upgrade to either belly or early crossbow now?

JosEPh

Not to my knowledge. I haven't been fiddling with unit upgrade paths at all lately so that certainly warrants taking a look into further. Setting up a test on this would be difficult so if you can provide your save it may help.

 

[Thunderbrd]

SVN 8191 research values do not get reset upon regenerate map.

How to replicate:
-start a game (in my case large map/deity/nightmare)
-settle in place. You now get 11 research points/turn
-regenerate map. Settle in place. You now get 21 research points per turn.

Seems like the classic type of bug of not setting variables to zero before
starting a calculation.

This will be interesting... I've not found where regeneration of the map takes place in the code yet but it shouldn't be too tough.

You can always CTRL-SHIFT-T for a recalc - that would fix it in play.