The description strings are in need of improvement.

Can you make another patch for the renaming ? I agree it should be done but this patch is already big enough.

Here you go: D3331.
Once that's committed, another patch should enable the assembly and give it more meaningful values (costs etc.). This patch can then simply focus on introducing the carnyx unit, making things less complicated to review.

For city phase, the gains must be considered.
What kinds of gains could we make? what values do you have in mind?

How about +100% health, +50% capture points, and +5 population? And perhaps add a minor stone cost, e.g. 200 wood and 50 stone?

being able to train women without the need for technology?

That's a good idea! Or maybe add the basic infantry spearman? But not any cavalry or ranged units.

Yes, I see your point, and I'd like to see apartments used in game. However, in my opinion they should be a nice extra, in addition to ordinary houses, not instead of, and since I tend to build most my houses in the town phase, I prefer the city phase for the apartments.

Well spotted!
I agree it makes sense to make loot proportional to costs for these units too. 10% is fine for support units, that's also what (most?) soldiers have. (Don't forget the slave template, though.) Since support units can't fight back, I'd favour removing their <xp>.
As for ships, I don't think they must have 10%, because they're a bit different. Structures have loot 20% of their costs, and that's also the percentage I proposed for siege engines in D3319. Speaking of which, feel free to include siege engines loot standardization here in this patch too, they're a bit out of place in D3319.
On the other hand, ships sink, so I'd be fine with giving them no resource loot at all as well. And maybe make warship loot <xp> proportional to their health?

Feel free to suggest alternatives!

• also lower fortress health, per @borg-

With the resistance line removal, a 10% health reduction would work out at 5185 health, so 5200. Or do you prefer to round it down to 5000 (i.e. −13%)?

I still have doubts about the health increase to 6000. even if it is an adequate number due to the reduction of armor, it still seems high to me.

6000 merely looks a lot higher, the difference isn't that much in practice (about 4% harder to destroy with crush damage, although it's 15% easier with hack or pierce).

• rebased

• rebased

While I like the snapping possibility, I also like freely placing things at aesthetically pleasing angles and distances. Having to press Ctrl to activate snapping is less inconvenient than having to press Ctrl to deactivate it, especially when I'm already rotating foundations by pressing [ or ].

This patch could just only change the fortress cost, keeping their building time, health, resistance and the Iberians as is, if that's what you prefer.

Wouldn't values like 4600 be enough?

No. 4200 health + 3 resistance equals 5761 health. (And 6000 health − 3 resistance equals 4374 health.)

The extra armor removal is good but I don't like the increase to 6000, I would use a value of 4500.

Removing the resistance without compensating the health makes them much weaker.

Would like to keep the Iberian fort (as well as its towers) with more health.

Towers, sure, fortress, I beg to differ, but it's two against me, so I'll revert.
What to do with the iber fortress building time? +50% seems rather severe.

Wasn't the boosted health for iber fortresses made to depict their turtling nature?

The Iberian fortress received a higher cost and health in rP9277, the health was further raised in rP11148 and again in rP11390 (nine years ago), and the cost in rP17706 as well. None of the commit messages explain why. My guess is iber fortresses are stronger simply because iber towers are stronger.

Increasing the health instead of the crush damage means that it's now harder for all entities to destroy it, while before it was only harder for siege engines.

Fortresses receive not only +3 crush, but also +5 hack and pierce resistance, so this patch makes it actually somewhat easier for other entities to destroy them. However, given the very low effectiveness of those damage types, it doesn't really matter.

EDIT: As noted by @Freagarach on IRC this also nerfs Fire javelins. (Resistance of all buildings to fire is 0).

Good point! Is that bad? (I still don't understand those new status effects.)

Thanks for the explanation. What version is used by the Windows builds? I have:

Package libxml2-devel-2.9.10-8.fc33.x86_64 is already installed.

That shouldn't matter. The font files used by the game are images generated by script and located under binaries/data/mods/mod/fonts/.
For instance, my text editor and editor have no difficulty with displaying Chinese (e.g. 漢語拼音), but when pasting that to a name string in 0 A.D., it's displayed as unrecognized characters, because those glyphs are not present in the game fonts:

This needs to be rebased.

In D2856#144142, @ValihrAnt wrote:

Personally not a big fan of adding increased cost. What about getting rid of that and reducing the health increase by 5 or 10 percent?

@ValihrAnt, what do you think of the values proposed by @Palaxin and those of @borg-?

In D2856#145989, @borg- wrote:

@Nescio can add a new icon to this technology? Currently it has the same "Horse Racing" icon, and both is researching on stable, so it can be confusing.

Any icon you recommend?

Nice catch! I should have checked when updating D2987 the last time.

@Nescio do you see the name correctly without the diff ?

Yes, actually I do, this is how it looks on my end (Fedora 33) without the patch:

• increase cavalry health technology cost to 400 food, 150 metal, per @borg-

Can you make another patch for the renaming ? I agree it should be done but this patch is already big enough.

True. Perhaps you could commit the actor and icon already? Auras and templates can be done here afterwards.
I also recommend against introducing that technology.

400 food 150 metal.

Or wood instead of metal?

Since the assembly is not really a theatre, could you rename it to assembly.xml, change the parent to template_structure.xml, and update the file accordingly?
Perhaps the Gaul champion infantry could be trained at the assembly too, instead of at the barracks and fortress?
Shouldn't the carnyx-man be a support unit (without an attack)?

That's quick!

I would increase the cost of "Horse Breeding" tech to 400 food 200 metal.

Wouldn't that be a bit expensive? It's only 10%. And the city phase batch training cost technologies cost only 500 food.
How about 400 food and 100 metal, or 450 food and 150 metal?

Nisean tech should also, but there is another patch for that, so I don’t think it’s necessary.

Yes, D2856, which is why I didn't change it here.

Diplomacy tends to be a political affair, not a commercial business.

(seems this needs changes)

Yes, it does. However, it'll have to be rebased again if D3315 is committed, so I'll probably wait for that, if you don't mind.

I would probably like a spread bonus more, but it needs to be substantial (CS archer have spread 3 which is huge, vs 1 for champions which is still rather large all in all).

How about what I wrote earlier:

In D3246#145198, @Nescio wrote:

I like +10% damage with −15% spread, but in that case ranged champion damage ought to be reduced from 2× to 1.5× that of basic troops, and ranged heroes ought to inflict proportionally less damage too.

which would work out to:

          range, spread
basic   : 1    , 1
advanced: 1.1  , 0.85
elite   : 1.21 , 0.72
champion: 1.5  , 0.5

You can see everything your allies see, not just their traders.

• forgot svn:mime-type and svn:eol-style

• change icons per @borg-

• rephrase description spotted by @wraitii

• restore erroneously deleted line

• restore ram's 200 metal cost, per @borg-
• remove 0 loot lines, per @Freagarach
• standardize loot of other siege engines

• rebased and corrected

As I wrote earlier, I really like the concept, but disagree with the technologies raising both crush and pierce resistance.
Moreover, could you rename them properly?

cd binaries/data/mods/public/simulation/data/technologies/
svn mv armor_cav_01.json soldier_resistance_hack_01.json
svn mv armor_cav_02.json soldier_resistance_hack_02.json
svn mv armor_infantry_01.json soldier_resistance_pierce_01.json
svn mv armor_infantry_02.json soldier_resistance_pierce_02.json
svn propset svn:mime-type text/json soldier_resistance_*
svn propset svn:eol-style native soldier_resistance_*

Furthermore, update the template_structure_military_forge.xml production queue.

Summary says +25% and techs +20%.

Oops, apparently I uploaded the wrong version.

• bolt-shooter spread to 1, stone-thrower to 6, per @wraitii
• gave towers the same projectile values as their siege engine counterparts

Good points, thanks for the feedback!
So bolt-shooters (and towers) a spread of 1? And raise the spread of stone-throwers (and towers) and quinqueremes, but by how much? Currently they have 4; for comparison, basic archers have 3, champions 1 (and champions have an attack range of 70).

I compared the models with the values and everything seems appropriate.

Before you accept, do you think you could do at least one test game with e.g. @ValihrAnt? I'm a bit worried what effect it will have on brit and gaul.

How about increasing or decreasing the garrison capacity depending on the size of the models, what do you think?

Well, I actually thought about it when writing this patch, but I was unsure by how much, or whether it's desirable; being able to garrison rams with fewer swordsmen seems a clear disadvantage.

I'm not sure about metal reduction, I think it can make spam easier and we have some problems with rams vs spear / pike, so I would keep the value of 200 for metal.

I lowered the metal value because rams are a lot less effective in A24 than they were in A23, and because rams are much more expensive than the structures they're supposed to raze (walls and towers). But if you think keeping 200 is better, I can do that, of course; or 150?

Is there a replacing bonus?

You mean for brit and gaul? Their structures already have lower wood costs and lower building time, giving them a clear advantage in the early game. Having the population increase (a third bonus) effectively means having to spend even less wood and time on houses, making things even worse. Moreover, a historical justification why exactly these two civs and not others should get it is absent.

I would have liked to have retained the Wonder class in the civinfo page, but I won't change that when committing. We'll see what the community says (if they even notice) once A24 is out.

Wonders were not included in the civilization overviews in A23 or any previous alpha I remember, so that's why I left them out. Besides, all civs have them, and they're all statistically identical. If people really want to show wonders there, then one could simply insert a CivSpecific class into the temlate_structure_wonder.xml; there is no need for hard-coding the Wonder class there (you don't know what mods want to do).

The problem has been solved by D2733/rP24535, civ-specific structures are again properly displayed in the civilization overviews.

I prefer than removal. The changes are good, I would risk high values like 15% and cost like 200 food.

Due to the way batch training works, I'd rather not lower the batch time modifier below 0.7.
How about lowering the base training times of domestic animals instead? And if so, to how much?

• change existing Persian levy technologies from −20% time and −5% health into simply −10% time, per @borg-

Same reason it needs an actor. It just happens to be there. Would be interesting to dig the archive logs to see what was planned for it.

See rP6391 (twelve years ago), rP10965, rP11497, rP12183 (eight years ago), all of which very large and not mentioning anything about the files in question.
Anyway, it's unused, ugly, confusingly named, and not touched for the better part of a decade. Surely that's sufficient to warrant deletion? Keep in mind everything is preserved in the svn revision history, so it can be recovered later, if someone really wants to.

BTW for some reason images are buggy on Phab,

Yeah, after someone updated phabricator. I assumed it was to speed up loading speed. Anyway, it will show up if you click on the “it is a mystery”.

just add {image, width=256} to make them show up (IIRC giving them a title also works)

Thanks, I'll give it a try next time I want to display an image.

Why does it need a foundation? It's not used in any template, is it?
Also the plot.xml is really ugly:

Sorry, I'm not sure I understand what you mean. Did I delete a wrong file?
This is how the plot_orchard.xml looks without the patch:

What is the assembly currently used for ?

You mean the gaul theatre? It's not buildable at the moment.

Should we allow the carnyx player to be recruited from there? IIRC the files are committed, and the actor can be taken from DE. Remaining is template work.

I'm not exactly sure what you mean; feel free to make a patch.

Those files were made for the orchard. Maybe they should be renamed to depict that ?

The orchard has its own files and doesn't need the ones deleted here; try it yourself (the plot.dae mesh is kept).

I know that I'm usually the only one to review those, but it's nice to have the O12 group, makes it easier to search for such tickets.

I checked, my other open [art] patches have the O12: Art group as a reviewer; apparently I forgot adding it here.

If it doesn’t make historical sense to train at the temple then the barracks is the most correct place for that.

The temple is the location @genava55 more than once recommended, hence why I didn't opt for the barracks.

We have in mind a patch to change champions from fortress to barracks, perhaps this should also apply to those who are trained in the temples.

Some civs already have specific structures in the city phase at which they can train their champions (e.g. athen gymnasium, pers apadana, spart syssition). Ideally all would, but new structures require new art, which is not trivial.

Does it make historical sense to keep fanatics in the temples?

Not really that much (see forum post), but the barracks makes even less sense.
I wouldn't mind simply deprecating the fanatic as well, but people are already complaining civs are rather similar (they're right).

• replace technology with one that lowers corral batch training time

• increase rome army camp attack range from 60 to 70, per @borg-

Yes, you're right, there is an empty character between the s and the <, which shouldn't be there, and is correctly removed by this patch. I do wonder how it got introduced, though.

Both strings really look the same to me, I wonder what's changed.

Having suggested it myself, it's hardly surprising I'm in favour of having separate technologies for hack and pierce resistance. However, I'm unsure whether it's better to have them applicable to all soldiers (similar to Age of Mythology); have separate versions for infantry and cavalry; or for melee and ranged; or make them apply to melee troops only, excluding ranged troops from receiving resistance increases.
I would highly appreciate feedback on this patch from more players.
Moreover, I don't really understand why a technology increases both crush and pierce resistance: elephants and archers have very different attacks.

While I like the idea of giving cart something different, I dislike it apartments are upgrades. Moreover, I'd recommend limiting them to the city phase and omitting the unlock technologies.

Perhaps you could make a forum post (with spoilers), showing the proposed new icons next to the current, old ones, to get more feedback from a wider audience (and avoid phabricator's quirks)?

@wraitii?

It looks good to me.

Good enough to accept?

The proposed icon might work as a placeholder but is in my opinion not good enough for a release.
What I did in my mod is take the art/textures/ui/session/portraits/technologies/grain.png, give it a transparent background, and use that as a placeholder: https://github.com/0abc/0abc-a23/tree/master/art/textures/ui/session/icons/resources :

I think we need to rethink this kind of thing. Just adding more information to tooltips isn't a solution IMO.

Yes, I agree, in general. However, there is infantry in game that can gather and infantry that can't, just looking at their name and icon won't necessarily give you this information, and not everyone knows what classes to look for, hence this patch to make it more explicit.

Thank you, this is a great improvement and long overdue!
Is just displaying an icon without any explanatory text clear enough, though?
Just so you know, arc patch D3310 has trouble properly adding the new images.

Maybe we would need nicer fish icon.

And maybe we need a better food icon, one that's more representative.

• −1 <MaxGatherers>
• goats grant 80 food, per @borg-

The idea is nice, but I'm unsure about the proposed implementation. I don't think treating visible classes differently than (non-visible) classes (or ranks) is desirable.
Moreover, having a file for each visible class seems file inflation to me. I'd prefer having one file listing all classes, with one line per entry (cf. credits files).

I guess you're planning for this but this calls for a "Gatherer" class and the removal of "Worker".

Perhaps. I agree it would be more consistent and admit I've toyed with the idea more than once. However, it's less straightforward, units are not necessarily able to gather all resources. Female citizens and citizen infantry can, but citizen cavalry can only gather meat, no other resources, so basically both the Citizen and the Worker classes have gatherer functions, as does the FishingBoat.
Moreover, a mod might want to introduce separate units (a butcher, farmer, miner, stone mason, etc.), so it might be better to introduce a class for each resource subtype (e.g. MeatGatherer, GrainGatherer, WoodGatherer). However, displaying that many classes makes things less readable. Maybe we should not make those classes visible and simple display the gather rates instead. Unfortunately those numbers are wrong, gather rates vary by resource subtype, e.g. citizen infantry has:

<ResourceGatherer>
  <Rates>
    <food.fruit>0.5</food.fruit>
    <food.grain>0.25</food.grain>
    <food.meat>1</food.meat>
  </Rates>
</ResourceGatherer>

which is displayed in game as “food: 0.58” (the average). But that's a different discussion.

• fix typo spotted by @wraitii
• rename worker elephant aura

This kind of seems like extremely poor design given how micro-heavy that is, but that's a different conversation.

It is. Personally I really dislike how it currently works in 0 A.D. and would strongly prefer corrals to be disabled by default. However, I'm not proposing that. Let's focus the discussion here on the actual patch.

This patch was more intended as a clean-up of elephant-related things missed in other commits than one that drastically alters gameplay balance. @Freagarach, since you're uncomfortable with committing it as is (with hero elephant twice the attack damage of champions), how about giving hero elephants merely exactly the same attack as their champion counterparts?
@Feldfeld hasn't been on phabricator in the past six monts and I'm not sure if and when he'll be back.

• change class name to CivSpecific
• fix identifying classes in the StructuresSubsection.js so these structures are actually displayed (and not Wonders)
• replace “special” in subsection headers with “specific” for consistency