Ok, im fine with 15%.

Up new patch.

Restricting Organic already changes the gameplay balance significantly, so I don't see why Unit is suddenly problematic.

In D2782#118216, @Nescio wrote:

A few more things:

• The current vision range seems rather excessive; perhaps it should be lowered to 40 or less. What do you think?
• If rams can no longer attack (organic) units, that makes them significantly less effective, so should its cost be changed?
• Could you change the <PreferredClasses> to "Gates Defensive Structure", in that order (cf. D2684), so rams prioritize structures capable of shooting arrows over others?

In D2682#117489, @Stan wrote:

@borg- thoughts?

In D2533#118092, @Nescio wrote:

No, actually not: although kennels are buildable in town phase, war dogs themselves don't have a phase restriction. So if a player starts with or captures a kennel, he can train war dogs immediately in the village phase. I think that makes sense, humans had dogs long before the first towns existed.
While I like the idea of moving them to the house (in fact, I suggested it myself), I'm not sure that's indeed the best option. Dogs require 0 population, hence the current training restrictions. If each house allows training two, then there would be no upper limit; e.g. 100 houses allow an additional force of 200 war dogs. Moreover, houses are usually one of the first structures to be built, and earlier and in much larger numbers than the barracks.

IIRC they reason why rams can attack units is that otherwise one can surround a ram by a few units, making it useless. Currently in that situation (at least most of the time), it will attack the surrounding units, trying to break free. Though it already prefers attacking structures over units (so it only attacks units if there are no structures around).

I've ran a 1vAI and the quick-hashes matched, as expected (full hashes don't match as serialisation is different).
Profiling on the following test map, no obvious difference in LOS computation despite LOS taking time.
A debug-hash rejoin test passed.

A few more things:

• The current vision range seems rather excessive; perhaps it should be lowered to 40 or less. What do you think?
• If rams can no longer attack (organic) units, that makes them significantly less effective, so should its cost be changed?
• Could you change the <PreferredClasses> to "Gates Defensive Structure", in that order (cf. D2684), so rams prioritize structures capable of shooting arrows over others?

For civilizations using stone launchers or slingers, it may be unviable.

D2497
As stated in the summary, I doubled the stone cost to keep the resource cost total unchanged, and because wonders dwarf fortresses, which cost 1000 stone. If you think it's better, I could instead just remove the food cost, keeping the other resources unchanged, thus making the wonder cheaper. I don't really care, what matters to me is food is for units, stone for structures.
D2660 is another wonder patch.

Successful build - Chance fights ever on the side of the prudent.

15%

Successful build - Chance fights ever on the side of the prudent.

Fix tests, I stopped halfway through.

Build failure - The Moirai have given mortals hearts that can endure.

Build failure - The Moirai have given mortals hearts that can endure.

Handle entity renaming explicitly (couldn't actually break it in my tests, but it seems prudent).

I don't see myself finding time to review/test this soon ;( Sorry.

Build failure - The Moirai have given mortals hearts that can endure.

Build failure - The Moirai have given mortals hearts that can endure.

Rebased & don't rebase on top of D2763 but of master.

In D2702#118170, @Freagarach wrote:

How much different is this from resetting finished orders in CallMemberFunction and setting finished order in FinishedOrder? (Apart from the (un)set in position.)
Reason for asking is the shift of responsibility from the entity to the formation(controller).

Build failure - The Moirai have given mortals hearts that can endure.

Build failure - The Moirai have given mortals hearts that can endure.

Turns out removing the hack wasn't so hard.

How much different is this from resetting finished orders in CallMemberFunction and setting finished order in FinishedOrder? (Apart from the (un)set in position.)
Reason for asking is the shift of responsibility from the entity to the formation(controller).

If it is historically a problem, then I must agree with @genava55.
That it is unique tech to Gaul civilization, the gain could be greater, perhaps 15%? The cost seems good to 15%.

I agree with 15%.

I agree with the changes in the pyramid.
About the wonder I agree in parts. Yes I think the cost of food should be removed, but I think a lot of 2000 stone resources. For civilizations using stone launchers or slingers, it may be unviable.
What about 1500w/1000s/1000m? Seems to me to be a fairer value in cost/benefit.

I will create a <Tooltip> as soon as I get home.
It seems that we agree with the three proposed changes.
We cannot change to "UNIT", it would change the balance of the game a lot. Players use ram to destroy other ram or siege weapons. Some civilizations would be very restricted to melee units, to combat siege weapons, this certainly would not work.

Successful build - Chance fights ever on the side of the prudent.

Successful build - Chance fights ever on the side of the prudent.

Successful build - Chance fights ever on the side of the prudent.

I also agree with this change, but it probably needs to be rebased and merged quickly so that it doesn't end up in merge conflicts everywhere. Might be a good candidate for a post-A24 commit.

Successful build - Chance fights ever on the side of the prudent.

Fix issue noted by Angen.

Stan's remarks.

Fix issues & hack.
Ran a rejointest, & visual/non visual-replays, all appears in order

I personally somewhat dislike that we rely so much on garrisoning for defence, so I would keep it limited, but I like how the effect looks, so it's good to have the possibility. The first concern is more related to balancing ofc.

Successful build - Chance fights ever on the side of the prudent.

• aura
• cost

Successful build - Chance fights ever on the side of the prudent.

Any opinions on the number of visible garrison slots on gates? In principle they could accomodate more units.

It doesn't seem to have been changed recently though, so something is weird.

Presumably wowgetoffyourcellphone has an outdated version of the UnitAI file in his mod.

I had indeed removed FinishOrder in Order.Stop by accident.

The Ram AI still attempts to attack units around them if there aren't any structures in vision range.

In rP23731#42614, @Angen wrote:

I believe garrison holder is still setting garrisoned flag for variants and art replays animations on variant change if it has active animation.

hmm, that seems like it's trying to close again when already closed, which shouldn't happen. Perhaps it's an art issue as we end up resetting some animation timer. I'll check.

I can probably do that myself quickly.

Great!

ramming a ship, if not historically accurate, is probably a good way to make a hole in the hull and sink it.

Ships ramming ships is historically accurate, it was the primary mode of naval warfare in classical times. However, ship rams were static extensions of the keel below the water line, whereas battering rams sling above ground. Moreover, there is the question how battering rams would reach the ships on the water. It's probably better if rams won't attempt attacking ships.

In rP23731#42611, @Nescio wrote:

One minor thing is that locked gates still play an opening-and-closing animation when garrisoning or ungarrisoning units; I suppose that's an art problem?

Thanks. I rebuilt the game and tried it out in combination with D2760. Everything seems to work fine.
One minor thing is that locked gates still play an opening-and-closing animation when garrisoning or ungarrisoning units; I suppose that's an art problem?

Now I don't know how the UnitAI works, but if adding this restriction means rams still chase and try attacking units, then I think a much better solution would be to simply reduce their damage vs units sharply; see D2684.

I guess the code is "AttackEntitiesByPreference" and that does seem like it's missing checks for restricted classes. But that just sounds like a UnitAI bug, and I would suggest fixing it instead of going for the above hack. I can probably do that myself quickly.

As listed in the summary, this patch does three different things, so let's discuss them in that order.

• I agree with the speed change. Though if the purpose is to allow pikemen to catch rams, it should be lowered even further.
• I have no objections to the fields restriction. In fact, I also included that in D2508. Also, animals can't attack structures, ships, or siege engines, so it's not unprecendented.
• I'm unsure about the organic restriction. The fact that it was present at some point and subsequently removed, as wraitii pointed out, suggests it may not have been an entirely good idea. Furthermore, wowgetoffyourcellphone wrote elsewhere:

In DE, Rams are prevented from attacking units. After years of experience with this change, I can see it's not pathfinding that needs adjusting, it's UnitAI. The Ram AI still attempts to attack units around them if there aren't any structures in vision range. What they need to do probably is just go idle, or attempt to flee (haven't given much thought on what they should do, probably just idle; they aren't support units).

Now I don't know how the UnitAI works, but if adding this restriction means rams still chase and try attacking units, then I think a much better solution would be to simply reduce their damage vs units sharply; see D2684.
Besides, I think it's better to restrict the attack vs Unit instead of Organic; battering rams attacking artillery and ships makes historically as little sense as targetting e.g. cavalry.
Furthermore, correct the indentation and add a <Tooltip> under <Identity>.

did not test it but looks good by checking the code

No, actually not: although kennels are buildable in town phase, war dogs themselves don't have a phase restriction. So if a player starts with or captures a kennel, he can train war dogs immediately in the village phase. I think that makes sense, humans had dogs long before the first towns existed.
While I like the idea of moving them to the house (in fact, I suggested it myself), I'm not sure that's indeed the best option. Dogs require 0 population, hence the current training restrictions. If each house allows training two, then there would be no upper limit; e.g. 100 houses allow an additional force of 200 war dogs. Moreover, houses are usually one of the first structures to be built, and earlier and in much larger numbers than the barracks.

The patch changes ram speed from 8.1 to 7.2, so from the same speed as an unpacked bolt shooter to the same speed as an unpacked catapult, making it easier for infantry and elephants to catch up to them.
Overall, I'm not sure if rams seem OP due to the ranged unit focused meta and some civs simply not getting good counter units to siege weapons or if they're actually OP. Still, I'm for this change as rams can be buffed again when some other meta changing balancing is done.

In D2781#118072, @wraitii wrote:

It's a fairly standard c pattern -> with unsigned types, "-1" is the max positive value.
The correct change is actually to use std::numeric_limits<size_t>::max() though, so I could change it to that.

As noted by stan -> See D576 for a situation that relies on local entities not sending messages to globally subscribed non-local entities.

In D2781#118060, @Stan wrote:

I assume -1 is to make sure they are always at the top of the list? I would have recommended static_cast but it seems consistent with the lines below. Maybe a comment?

It's a fairly standard c pattern -> with unsigned types, "-1" is the max positive value.
The correct change is actually to use std::numeric_limits<size_t>::max() though, so I could change it to that.

Any reason why it wasn't made a priority at first ?

You mean in past diffs? This comes straight from rP13167, I assume it wasn't deemed necessary or was simply forgotten.

I assume -1 is to make sure they are always at the top of the list? I would have recommended static_cast but it seems consistent with the lines below. Maybe a comment?
Any reason why it wasn't made a priority at first ?

Seems like a change for the better.

I think I'll proceed with this over the WE perhaps, need to remind myself to update the wiki pages and do the few useful changes pointed out.

garrisoned aura

In D2782#118046, @Lionkanzen wrote:

If we garrison units inside, it should increase the speed of the battering ram.

If we garrison units inside, it should increase the speed of the battering ram.

ok I will test this.

The best idea is train dogs in the houses. This does not the game unbalanced because they are training only in era 2. Can also put a limit of 1 or 2 dogs per house.

(Strictly speaking it's a tower, not an howdah.)

Successful build - Chance fights ever on the side of the prudent.

(commandeering, this isn't just a rebase).