hyperion

In the past Mauryas were only able to train rams from captured fortress, the Mauryas specific model existed for this purpose only it seems. I wager this was deliberate.

That's very not obvious If a git client is installed just call git clone if not complain that the user needs to install git to use this feature. Mostly meant to make it easy for users to share mods via github for instance. You don't, the user will have to. Users can add keys they trust, you only ship the wfg one.

As long as it is an improvement The current mod selection screen is obviously one of those parts which were meant to have something working as soon as possible and make it easy to see if something goes wrong, a tiny step up from text files. In early development this is how it goes. That doesn't mean there isn't a lot of room for improvement or that it was built with the end user as first priority even. Some features I'd like to see as well: Allow adding mods via filebrowser Allow adding mods from a git repo uri Show if a mod was signed by wfg or other trusted party and successfully verified ---- The backend has issues as well, should force disable all reverse dependent mods on mod upgrade, guess we'd loose a lot less users that way.

The game tries to load mods that no longer work. Open your user.cfg file with a text editor and change the line starting with mod.enabledmods to only contain "mod public" and this hopefully gets you back on track. See https://trac.wildfiregames.com/wiki/GameDataPaths for where to find user.cfg

"You can't trust developers to do ui testing, because then they use text files", made my day

Not quite but similar. pyrogenesis would be like pyrogenesis -mod=mod but would only present main mods aka games in a fashion closer to the current map browser. Each game mod could have it's own (enabled) mod list modifiable by the current mod selector or a successor thereof which ofc would only present dependent mods then. Well, this is only an idea that came up at the spur of the moment, not well thought out, it appears there might be issues, though I don't know what @Stan` was referring to when mentioning troubles for Spring.

Sure, but would be neat if plain pyrogenesis would pop up a neat game selection screen, where game would be any of the full conversion mods. Also does something like alias exist on Windows?

Just like public for 0ad. A unique identifier if a number often is called ID, if a human interpret-able string a name or qualified name. So the use here is perfectly valid. Label or Display Name are rather common denominators for what is shown in UI as "name". Displaying the name property is sort of having debug info shown, same for Folder, so I agree with @Stan` to possibly drop those columns and rename Mod Label column but leaving the underling meaning untouched. This then ofc makes the Dependencies column less obvious to interpret and might require an adjustment as well. The current mod selection screen really isn't that great. @wowgetoffyourcellphone, you seem to have a rather clear idea what you want the mod selection screen to look like, beyond just column names, mind sharing a mock-up? PS: What I think would also be nice is having pyrogenesis be a multi-call binary, so if invoked as 0ad, it loads 0ad, if invoked as hyrule_conquest or delenda_est correspondingly.

A good game has simple rules which are reasonably easy to discover, this proposal has the taste of high complexity and a read the code feature. At least the implementation is the easy part here. We want some level of snowballing but not much, the current promotions are probably already boarder line. So I agree with @alre that this needs a lot of consideration. Most like this proposal will be OP or useless, close to impossible to find a sweet spot. Further this proposal builds on very shaky foundations. A prerequisite for this proposal is to fix the imbalance of for ex. ranged cav vs melee inf wrt. promotions.

First you might want to preload the mod mod and public mod as well, using the -mod= switch, as your mod might depend on them. Then the size says 1kb, so I have my doubt it actually worked at all. Also the message another process using the file already might mean the creation is still ongoing. Not seeing any output to the terminal suggest the actual worker thread detached somehow and pyrogenesis returned early? Not using Windows, so can only say that looks suspicious. Just renaming a .zip to .pyromod also works.

The first graph is by WFG, the latter two by Canonical. The 1k in the WFG graph is for v-16 (a25 I guess), while v-11 is only ~400. Still shockingly high I'd say. The covid spike is quite apparent in the graphs from Canonical, the one from WFG only covers the last month. Linux users don't download from WFG at all (at least hardly), distributions either mirror the sources or offer built binaries themself. The graphs from Canonical mostly show the shift from deb to snap. There isn't much other information to glean there other than estimating user base based on the plateau. The only thing I'm wondering is the exact definition of "weekly active devices"

It's rather recent that Ubuntu force snap down their users throat and that there was an issue with wfg ppa, so that conclusion can't be made from the above graphs at all even if it might be true. Also making some assumptions and doing some math it's not unlikely that the active (playing weekly) player base is between 300k and 500k.

@Stan`, loosely followed https://trac.wildfiregames.com/wiki/Basic3DImplementation Packed it all into a mod but the result doesn't seem to be available in the atlas, must (read-only / owned by root) public mod be edited for that?

Ok, so I followed a tutorial to create a procedural gold cube. Baking the diffuse gives me a black texture as one would expect, stetting metallic to 0 and baking diffuse gives something that looks to be albedo. Roughness should be grayscale, the inverse would still be grayscale while spec for gold should be closer to albedo for gold, right? Guess having an additional shader/material for use with blender baking for modders would be nice even if existing textures used the old one.

So there is no option to bake a spec map for 0ad any longer with recent blender? Why is there still diffuse then? Makes sense, thanks.

I don't see spec in bake type options Why not use the same UV for all maps?

Ask your infra guy Yeah, hardly possible to setup in a few minutes.

My idea of an announce mailing list is for it to be read-only except for wfg addresses. Traffic should be very low, announcing releases certainly qualifies. People subscribe to the list if they are interested instead of you directly reaching out. Target are primarily direct or indirect contributors, but anyone interested is free to subscribe. This would obviously replace you actively reaching out to distributions via mail.

Was meant as a joke but then again that's how it used to be. On that matter a 0ad-announce mailing list might make sense for packagers and modders alike. A dev mailing list maybe but for users I highly doubt it. hilarious, something is terribly wrong with discord then.

Just search public 0ad channels with your matrix client, I think 0ad central that channel was called. Don't have a client on this machine ... sorry

Are there any other team members willing to be available on discord? If you are the only one you could also just go with "stan's 0ad discord channel" and ditch subscription to all others and be done with it. In the end it's about what you and other team members want to offer and not what some folk want to consume. With the forum and the irc dev channel as well as the xmpp user channel (lobby) there is enough official means already to get in touch if people feel the need. And for real issues there is trac. Can we have a mailing list, please!

Just checked matrix and there is a 0ad channel with close to 70 people, sure there are channels for other mediums as well, why is this not good enough? How does make them official improve anything?

IIRC giant is 512 tiles a 5 m, so roughly 2.5 km

Which might be a pro Anyway how is clicking the chat link in the forum navbar a challenge for not technical people? On paper the best alternative to IRC for opensource, for plain chat perfectly fine for anything beyond you don't need a degree in CS but sure might be painful. The landscape is far more centralized which could be considered a con, however, as a consequence most everything works out of the box. --- For xmpp and matrix there are IRC bridges which could be setup if desirable --- Anyway, as for official channels, why are the current dev irc channel and a user irc and xmpp channel not sufficient?