Alpha 24 SVN Teamgame

[Stan`]

For A25 I plan more realistic impact sounds depending on what is hit

[Silier]

18 hours ago, nani said:

8) Sometimes untis seem to attack houses instead of capture if they are left idle

can you point to time in the replay pls?

Edit:
because only instance I found was when they had WalkAndFight order and that is not using Capture even in a23b, not saying thats cant be issue, but it is not broken by a24 so would wait for a25.

[nani]

49 minutes ago, Angen said:

can you point to time in the replay pls?

Edit:
because only instance I found was when they had WalkAndFight order and that is not using Capture even in a23b, not saying thats cant be issue, but it is not broken by a24 so would wait for a25.

Okey, I tried testing what I saw in that last game but I cant reproduce so what I think it was is just me clicking too fast while having lag/low fps, only reason I can think of. So false positive by my part.

Also the camera stuttering might be just lag that feels diferent from a23 lag so is more noticeable first time, so im not sure now.

Edited February 7, 2021 by nani

[faction02]

3 hours ago, Angen said:

can you point to time in the replay pls?

Edit:
because only instance I found was when they had WalkAndFight order and that is not using Capture even in a23b, not saying thats cant be issue, but it is not broken by a24 so would wait for a25.

• I think I saw iddle units attacking houses/barracks instead of capturing them too. Maybe I am not getting your point but in a23, this happens only for palisades/walls/fields that can't be captured, no?

• I think houses, corals, dropsites and market (????) for ptolemies now have the same construction time but they are cheaper. They do have less hp too but not sure it compensate for the economic bonus it is giving, especially for dropsites. Did I miss something there?

• The main change to the game experience seems to be unit movements for me (rather than sounds  ). It feels less fluid in overall. Units change frequently speed between walk, run, stop, rotation, formation. Maybe it is something that I have to get used to but currently I  find difficult to predict units movements and therefore more difficult to control them accurately. Maybe there could be some way to play with the numbers (rotation speeds etc...) to improve it if it is effectively a problem. 

• I love many of the graphism changes. 

[wraitii]

14 hours ago, faction02 said:

• I think houses, corals, dropsites and market (????) for ptolemies now have the same construction time but they are cheaper. They do have less hp too but not sure it compensate for the economic bonus it is giving, especially for dropsites. Did I miss something there?

The bonus was changed.

14 hours ago, faction02 said:

• The main change to the game experience seems to be unit movements for me. It feels less fluid in overall. Units change frequently speed between walk, run, stop, rotation, formation. Maybe it is something that I have to get used to but currently I  find difficult to predict units movements and therefore more difficult to control them accurately. Maybe there could be some way to play with the numbers (rotation speeds etc...) to improve it if it is effectively a problem. 

The main change is that rotation takes time - this makes unit movement less fluid. The unitMotion rewrite has changed things on the surface but I doubt it'd be very noticeable. Then formations-by-default has changed movement further, but that IMO increases fluidity.

The primary objective was to make 'dancing' much, much harder (which we've succeeded at).

 

[Silier]

On 07/02/2021 at 8:18 PM, faction02 said:

I think I saw iddle units attacking houses/barracks instead of capturing them too. Maybe I am not getting your point but in a23, this happens only for palisades/walls/fields that can't be captured, no?

yes, that's correct for palisades/walls/fields. 

I ll double check the replay, that should not happen and can't see reason from the code.

I couldn't reproduce it, but maybe I am missing some variable to get this result. 

I was saying that if you order units to walk and fight (attack move) they will not capture by default when they find target on the way.

Edit: we found something, that may be cause of the issue

[faction02]

11 hours ago, wraitii said:

The main change is that rotation takes time - this makes unit movement less fluid. The unitMotion rewrite has changed things on the surface but I doubt it'd be very noticeable. Then formations-by-default has changed movement further, but that IMO increases fluidity.

The primary objective was to make 'dancing' much, much harder (which we've succeeded at).

In most cases, the visible impact is rather small indeed, it becomes more visible when the number of rotation per unit of time becomes more important. For example woodcutting activities. I also remember a large group of units that wouldn't get into formation (I guess because they were surrounding a building), they were ordered to move on the other side of the building but because the first unit who should move was facing the wrong direction and needed a full rotation before moving. All the other units behind couldn't start moving and it became a bit chaotic.  

I understood the aim of the change (thanks for the effort!), this is why my question was more about optimizing the change rather than the change in itself. There is a trade-off here between efficiency in limiting dancing vs fluidity and I was hoping other players could share their own preference.

It might also just become normal to me after more hours of gameplay once I get used to it. 

[faction02]

8 hours ago, Angen said:

yes, that's correct for palisades/walls/fields. 

I ll double check the replay, that should not happen and can't see reason from the code.

I couldn't reproduce it, but maybe I am missing some variable to get this result. 

I was saying that if you order units to walk and fight (attack move) they will not capture by default when they find target on the way.

Edit: we found something, that may be cause of the issue

I have spent some time trying to reproduce it but unfortunately I failed. I had started to believe in some missclick too, thanks for the edit.  

I don't know if it can help but in this replay around minute 24, some archers start shooting at a barrack (the one north west to the temple), then a bit later I think it happened again with some other buildings.

commands.txt metadata.json

[faction02]

I got a few lost camels:

- Around 2:40, the first camel was requested to do an half circle exploration round around the territory. If I remember correctly, I shift-click that order, but the camels arrived late with respect to an a23 camel. I noticed that it stops for a while on the border of the map (probably because I clicked outside the map) 

 - Around min 4, 2 polite camels meet at the farmstead but can't decide who should go first. They get stuck for a while until a group of nice ladies come to their rescue

commands.txt metadata.json

[wraitii]

18 minutes ago, faction02 said:

- Around 2:40, the first camel was requested to do an half circle exploration round around the territory. If I remember correctly, I shift-click that order, but the camels arrived late with respect to an a23 camel. I noticed that it stops for a while on the border of the map (probably because I clicked outside the map) 

Yeah you ordered it to go somewhere unreachable and it took a few turns to realise it was indeed unreachable. Compared to A23, units will probably lose a little more time, but you'll also have fewer cases of units that fail to reach points that were actually passable. It's a tradeoff and for A24 it's in favour of 'correctness' over 'efficiency'. 

18 minutes ago, faction02 said:

 - Around min 4, 2 polite camels meet at the farmstead but can't decide who should go first. They get stuck for a while until a group of nice ladies come to their rescue

That's an unlucky situation, but it's not new, it could also happen in A23 and even earlier. Pretty rare that it happens with 2 moving units, and unfortunately not a _lot_ that can be done until unit pushing is introduced.

[faction02]

On 06/02/2021 at 9:44 PM, nani said:

List of possible problems and complains from game played a24 4vs4 random map, take into account that performance is comapred with a23. (last svn revision as for 2/6/2021 windows build)

1) Very heavy lag when using the scrollbar in gamestup. All players looks like they are about to disconnect. (Ex. Relic count, Relic duration scrolls) [this issue I think is very important to fix]
2) Chat with too much interline space (in a23 is much less space)
3) Map browser doesn't select first result if Entrer is pressed
4) Construction countdown  time texts and number of number of units building it text are too vertically close
5) Camera stutters hard when moving it with keyboard (also happens with middle click mouse)
6) Buildings icons too small/hard to see
7) Some orders/actions seem to be lost/ignored when in late game/lag/low fps
8) Sometimes untis seem to attack houses instead of capture if they are left idle
9) When ending game (session) and load -> stats page loading freezes and even Windows suggest to close 0ad as it doesn't seem to be responding.
10) Multiple occurrences of random 1-2 seconds for when the game drops to 4 fps on my case (looks random) and then returns to normal 17-30 fps

I also have the impression that some actions might be lost in game, as nani noticed in point 7). The reason might just be different than the one suggested by nani.

I have noticed that I suspiciously misclick more often than what I am used to do to assign units to collect resources. Since I often Shift-click and queue these types of orders, I am wondering if there could be a conflict between some type of orders + formation+Shift-click for example (which is also more likely to occur in late game because there are more units and could also be more easily noticed in a lagging environment)?

Some knowledge about the code would help to formulate a correct theory and help reject the wrong one directly. . If someone could tell me if something like this could make sense, I would test more. I am still trying to understand why units control feels different. Thanks ! 

 

[wraitii]

8 minutes ago, faction02 said:

Since I often Shift-click and queue these types of orders, I am wondering if there could be a conflict between some type of orders + formation+Shift-click for example (which is also more likely to occur in late game because there are more units and could also be more easily noticed in a lagging environment)?

Oh man you're right, thanks for investigating further. So having units in formation, ordering them to go somewhere in formation then shift-clicking to gather works, but giving 2 separate movement orders first doesn't. In that case, the gather order is lost. I totally missed that when 'implementing' queued orders for formations (which didn't work _at all_ in A23)

 

[seeh]

little bug

 

Installed in ubunto like so:

sudo apt install git
git clone https://github.com/0ad/0ad.git
#### get updates:
cd ~/0ad; git fetch https://github.com/0ad/0ad.git
cd ~/0ad/build/workspaces ; ./update-workspaces.sh -j13

you could find the little bug at the very end of the replay and some minutes before:

[13:45] <seeh> endless building

 seeh: Strange do you have the replay

[13:46] <seeh> its running

[13:46] <seeh> if i resign ...

[13:49] <seeh> has todo if walls splitted in parts and you reget it later

[13:49] ... seeh: Don't have to you can upload it later

[13:49] <seeh> where to send the replay?

...

[13:50] .... Could be on the forums

 

 

Edited August 26, 2021 by seeh

[seeh]

8 hours ago, seeh said:

little bug

 

2021-02-10_0002.zip

Installed in ubunto like so:

sudo apt install git
git clone https://github.com/0ad/0ad.git
#### get updates:
cd ~/0ad; git fetch https://github.com/0ad/0ad.git
cd ~/0ad/build/workspaces ; ./update-workspaces.sh -j13

you could find the little bug at the very end of the replay and some minutes before:

[13:45] <seeh> endless building

 seeh: Strange do you have the replay

[13:46] <seeh> its running

[13:46] <seeh> if i resign ...

[13:49] <seeh> has todo if walls splitted in parts and you reget it later

[13:49] ... seeh: Don't have to you can upload it later

[13:49] <seeh> where to send the replay?

...

[13:50] .... Could be on the forums

 

in a normal game the mistake almost never happens. not bad

known bug. https://code.wildfiregames.com/D1190

[22:01] <WildfireBot> D1190: Allow restoring lost wall turrets [Needs Review] – https://code.wildfiregames.com/D1190

Edited February 10, 2021 by seeh

[Dizaka]

Where should balance questions be fielded?

I'm concerned about two civs.  Macedonians and Romans.

For Macedonians:

• They effectively became "little Seleucid."  I believe they have everything Seleucids have but Seleucids have way more.  No real reason to play Macedonians.
• They were the only civ able to mass-spam siege really fast.  This was due to building multiple siege workshops and the ability to produce hero (siege dmg hero) simultaneously from fort.  Now all civs can do that and more-so Gauls (dmg hero in new building) with rams and Seleucids with eles (ele dmg hero from CC).

For Romans:

• They were unique in that they could force where battles are fought.  
• They were also unique in that they could create and turn games by enemy failure to scout.
• Both of the above points, currently, are difficult to achieve as the encampments do NOT produce siege weapons.

 

 

Which civs are the winners this patch?  I feel as Gauls and Seleucids won the lottery for A24.  Gauls because now they can take the place of Macedonians in early mass-spam siege (dmg hero comes from separate building, siege comes from separate buildings, have spear/skrim).  Previously, the bottleneck for Gauls was fort producing everything from hero, siege, and siege upgrades.

[Stan`]

We're building the first A24 release candidate revision: rP24875 Will upload links when it's done.

@Dizaka here is a good place for balance questions  @borg- should be the person you should talk to.

[borg-]

25 minutes ago, Dizaka said:

Where should balance questions be fielded?

I'm concerned about two civs.  Macedonians and Romans.

For Macedonians:

• They effectively became "little Seleucid."  I believe they have everything Seleucids have but Seleucids have way more.  No real reason to play Macedonians.
• They were the only civ able to mass-spam siege really fast.  This was due to building multiple siege workshops and the ability to produce hero (siege dmg hero) simultaneously from fort.  Now all civs can do that and more-so Gauls (dmg hero in new building) with rams and Seleucids with eles (ele dmg hero from CC).

For Romans:

• They were unique in that they could force where battles are fought.  
• They were also unique in that they could create and turn games by enemy failure to scout.
• Both of the above points, currently, are difficult to achieve as the encampments do NOT produce siege weapons.

 

 

Which civs are the winners this patch?  I feel as Gauls and Seleucids won the lottery for A24.  Gauls because now they can take the place of Macedonians in early mass-spam siege (dmg hero comes from separate building, siege comes from separate buildings, have spear/skrim).  Previously, the bottleneck for Gauls was fort producing everything from hero, siege, and siege upgrades.

Thanks for the feedback.
Well, what advantage does gaul have to have siege weapon spam faster than any other civ?
Our vision of gameplay is not limited to thinking that a civilization like Macedonia had only a fast siege as a game alternative. We added two new units.
The balance is under construction, and we plan specific changes for each civilization, but only for a25.
One of the proposals for Romans is that they can build immobile catapults for example.

Edited February 11, 2021 by borg-

[Dizaka]

2 hours ago, borg- said:

Thanks for the feedback.
Well, what advantage does gaul have to have siege weapon spam faster than any other civ?
Our vision of gameplay is not limited to thinking that a civilization like Macedonia had only a fast siege as a game alternative. We added two new units.
The balance is under construction, and we plan specific changes for each civilization, but only for a25.
One of the proposals for Romans is that they can create real immobile catapults for example.

Basically, Gauls went from "serial" production of Heroes/Siege to "parallel" production (see below) by spending more wood.  Wood in p3 is plentiful or can be planned out to be plentiful in p3.  Anyway, "parallel" production is similar to what Macedonians (and Romans) had. 

Additionally, a Gaul hero adds +20% dmg to siege rams.  This means with siege dmg upgrade rams have 44% more dmg (1.2*1.2-1).  Any civ being attacked by this will need to target rams first (Think 6-7 rams 14/15 min ... 7*1.44=10 ram dmg ...).  This is an issue as most players will be majority skirmishers, archers, or slingers with spear/pike meatshield.  Of the 3 ranged only slingers are effective against siege.  Additionally, with Carnucos units .... you can debuff dmg now ... huge imo when paired with parallel production for an extremely effective siege push.  A ton of synergy here for a siege push. 

The siege push was previously unique to Macedonians (mace have +20% dmg siege hero).  It's really effective b/c enemy needs to target siege to save base (the 7 rams equal to 10 rams of dmg) to worry about ... .  However, they have other units to worry about (Gaul general attack units), so enemy units get picked off because the enemy units (the skrimishers/archers/spear/etc (even sword in many instances b/c noone goes 40+ swordsmen)) prioritize the rams.  Otherwise, base goes bye bye.

Previously for Gauls (Serial Production):

• Castle costing 1,000 stone produced:

1. Heroes
2. Siege Rams
3. Siege upgrades
4. Will to fight

Currently (Parallel Production):

• Assembly of princes (400 wood)

1. Heroes (specifically, +20% champ)
2. Carnucos (-10% dmg for enemy units)

• Siege Workshop (300 wood)

1. Siege Rams
2. Siege Upgrades

• Castle (300 wood, 600 stone)

1. Will to Fight

 

Edit:  Previously only Romans/Macedonians could do the above strategy with rams.  Both needed a fort, however gauls don't even need a fort now.  Sele could somewhat do it with elephants (not as fast) and two forts.

Currently, using elephants, Seleucids/Carthagians can (sele ele hero, carth dmg hero) do the strategy.  For rams now you have Gauls(dmg hero, way more synergy)/Mace (siege hero)/Romans (dmg hero)/Carthagians (dmg hero).  

Would be interesting if Maury ele hero gave a elephant bonus (ranged and melee elephants).

 

Edit2:  One thing that *maybe* could be done for Macedonians is make siege workshop a p2 building that counts for p3.  This would be similar to Mauryas ele stable being a p2 building counting for p3.

Edited February 11, 2021 by Dizaka

[Langbart]

8 hours ago, seeh said:

known bug. https://code.wildfiregames.com/D1190

[22:01] <WildfireBot> D1190: Allow restoring lost wall turrets [Needs Review] – https://code.wildfiregames.com/D1190

Hey @seeh. As already noted  by @Stan` via IRC, the D1190 is a different issue. I just created a Ticket (#6006) in trac, I think you found an unknown bug. But so far I haven't found a clear rule as to why walls can't be built. If you find a way to consistently reproduce this bug, post it in the ticket or here in this forum thread.

[Stan`]

The release candidate is here!

Revision: rP24875

https://releases.wildfiregames.com/rc/0ad-0.0.24rc1-24875-alpha-win32.exe

https://releases.wildfiregames.com/rc/0ad-0.0.24rc1-24875-alpha-osx64.dmg

PLEASE READ THIS BEFORE POSTING & TESTING

The RC will overwrite the current installed version.

• The RC will use your current config including mods, so make sure to disable EVERYTHING (fgod, autociv, and badosu's mod are not supported)
• You might get warnings about hotkeys, we redesigned completely the system, so you might have to fix some of them (you may want to backup your config file if your intend to play A23B again
• It goes without saying but even though we are really close to a release, this is an experimental version.

If you want to help more you can also perform those steps

1. Launch a random game
2. Test your language's translations and fix them on Transifex.
3. Launch a normal skirmish.
4. Connect to the lobby
5. Play on the lobby with someone
6. Launch Atlas 
7. See if everything works there.
8. Open Unit tests demo (To see if there any breakage in displaying entity's) (It's in scenarios)
9. Try mods through modio only. (A23B MODS WILL NOT WORK)
10. If you need any help ping me
11. Enable feedback and see if it works (Main menu)
12. Connect to and use mod.io
13. Test replaying new games
14. Test Saving and loading a game.

Changelog

Spoiler

Angen committed:

    rP24859 Reuse the same formation controller when formationtemplate does not change
    rP24850 [gameplay-a24] Tweak elephant stable building time
    rP24849 [gameplay-a24] Make unlock_spies technology cheaper
    rP24848 Fix wrong selecting of the match when updating list introduced in rP24728

Imarok committed:

    rP24868 Fix positioning of the research progress icons below the counters overlay
    rP24843 Fix Phoenician Trireme

Stan committed:

    rP24874 New sling sounds
    rP24873 Reduced metallic impact for all melee sounds, blended existing sword sounds…
    rP24872 Improve training sounds for merchants and priests
    rP24871 Increase the difference between the chopping and building sounds
    rP24856 Make pages relying on hotkeys more resilient
    rP24844 Only parse querystring in RL Interface if it exists.

vladislavbelov committed:

    rP24870 Switches decals to VBChunk to reduce the number of state changes.
    rP24867 Batches decals and reduces the number of state changes.
    rP24861 Changes CVertexBufferManager container from std::list to std::vector.
    rP24860 Separates allocated vertex buffers into groups for data locality.

wraitii committed:

    rP24869 Fix no-formation orders after several formation orders.
    rP24866 Allow capture when falling back to "WalkAndFight" in COMBAT.APPROACHING
    rP24865 Better fix for formation waltzing, revert rP24831.
    rP24863 Ignore MoveToChasingPoint orders when not chasing.
    rP24858 Get the public IP from the lobby if not using STUN.
    rP24857 Lobbybots: update readme to reflect mod_ipstamp being optional
    rP24855 Fix units restarting orders when tasked to attack/gather/repair/heal same unit.
    rP24854 Fix a crash in the mod selector when using the mod downloadder.
    rP24852 Debounce gamesetup network messages & slider changes to avoid lag.
    rP24851 [gameplay] Reduce iberian fireship attack
    rP24847 Fix blurry textures in Bundle by bundling with mod mod in CI
    rP24846 Reduce interline space for lobby/gamesetup chats.

 

 

[Stan`]

Known bugs.

• Typing the /clear command will trigger errors.

[sarcoma]

@Stan`: In build instructions: in arch dependencies: the command needs sudo in front and wxgtk doesn't exist, it's wxgtk3 maybe

[Stan`]

1 minute ago, sarcoma said:

@Stan`: In build instructions: in arch dependencies: the command needs sudo in front and wxgtk doesn't exist, it's wxgtk3 maybe

Usually if you build with sudo then the game will throw you a tantrum

https://trac.wildfiregames.com/browser/ps/trunk/build/workspaces/update-workspaces.sh#L3

For wx-config you need an env variable

https://trac.wildfiregames.com/browser/ps/trunk/build/premake/extern_libs5.lua#L668

[sarcoma]

Not there, here https://trac.wildfiregames.com/wiki/BuildInstructions#ArchLinuxManjaro

sudo pacman ...  wxgtk3 ...

^missing                         ^missing

Edited February 11, 2021 by sarcoma

[Stan`]

My bad