elexis

-load=C:\Users\lg2015\Documents\My Games\0ad\replays\0.0.23\2017-11-16_0003\metadata.json NOT metadata.json but simulation.dat! The simulation.dat is inside the savegame of that replay. The savegame has the fileending .0adsave, must be renamed to .zip and unzipped as I mentioned above.

Which means that the simulation.dat wasn't loaded

Isn't a normal-sized map already filled by 2-3 civic centers if one starts with 8 players?

They are not supposed to connect. The replay should start from the savegame point. So the house should have been built already when the replay starts and the remaining 30 min should be replayable. I can test the patch later again to see if it works for me.

You can tell from the logo in the background that these barbarians must be big fans of 0 A.D..

Use the original commands.txt with the savegame where it started. No modification of the file needed. The replay you uploaded is 30minutes long. Are you sure you have compiled the game with the patch?

Why discourage people from expanding their territory aggresively towards the enemy?

There are some funny bugs. I recall placing a tower or barracks near an allied civic center does the job. (One can reach into that territory by building buildings towards the territory consecutively).

I didn't actually test the patch since a year ago, so I don't knowif it will be successful, but it should work because the relevant code didn't change. Apply the patch and compile the game. Rename your savegame to something.zip and you can use any zip program to unzip it. The zip folder contains a simulation.dat file. The replay contains a commands.txt file. You need both. If you are on Windows, you need to execute the following command: pyrogenesis.exe -load="C:\folder\simulation.dat" -visual-replay="C:\folder\commands.txt" load_simstate+visual_replay.diff

You are clearly talking about the Reflection map _kali never finished. #4760 He ran into terrain connectivity issues and the github code misses some library function he didn't upload. Agree it would be great to have it on every map as an option. I think it's quite fun to play non-symmetrical though as the element of surprise and the ncessity to scout everything really adds to the game.

You see those two commands.txt files right? The first one ends at turn 312, the second one should start at turn 312 where it now says turn 0. One could write a script to add 312 to every turn number in the second file. Then one can copy&paste the files together. Hm. Actually I have an old patch somewhere that plays replays from a savepoint. Digging. Found. Alpha 21. You are lucky, the patch still applies. You have to compile the game with the patch, unzip the savegame, then start pyrogenesis with -replay-visual="/path/to/second_replay/commands.txt" -load="/path/to/simulation.dat" and it should work in theory. I guess we could commit this part as a debugging feature if we don't priortize implementing it for the average user. Index: source/ps/Game.cpp =================================================================== --- source/ps/Game.cpp (revision 18875) +++ source/ps/Game.cpp (working copy) @@ -37,10 +37,11 @@ #include "ps/LoaderThunks.h" #include "ps/Profile.h" #include "ps/Replay.h" #include "ps/Shapes.h" #include "ps/World.h" +#include "ps/GameSetup/CmdLineArgs.h" #include "ps/GameSetup/GameSetup.h" #include "renderer/Renderer.h" #include "renderer/TimeManager.h" #include "renderer/WaterManager.h" #include "scriptinterface/ScriptInterface.h" @@ -52,10 +53,12 @@ #include "tools/atlas/GameInterface/GameLoop.h" extern bool g_GameRestarted; extern GameLoopState* g_AtlasGameLoop; +extern CmdLineArgs g_args; + /** * Globally accessible pointer to the CGame object. **/ CGame *g_Game=NULL; @@ -189,11 +192,21 @@ bool CGame::StartVisualReplay(const std: JSContext* cx = scriptInterface.GetContext(); JSAutoRequest rq(cx); JS::RootedValue attribs(cx); scriptInterface.ParseJSON(line, &attribs); - StartGame(&attribs, ""); + + CStr savegame = ""; + if (g_args.Has("load")) + { + std::ifstream file (g_args.Get("load").c_str(), std::ifstream::in); + std::stringstream buffer; + buffer << file.rdbuf(); + savegame = buffer.str(); + } + + StartGame(&attribs, savegame); return true; } /**

There is a difference between low FPS and 'simulation lag'. While too many units and a too big mapsize worsens the performance both in the simulation and rendering, it's also possible to have a crappy framerate even if the simulation runs smoothly. FPS is certainly getting worse with every tree and unit added to the screen. So you gotta resort to reducing the graphics settings and limiting the mapsize + population anyhow as Stan recommended. (There is also an FPS limiter in the options in case you want to get above 60 fps)

Ah, that was the reason why we should serialize the turnnumber in savegames as well. Yes, one can concatenate these two replays, but not before computing every turn number. Can implement that eventually depending on how old I get.

Replays can only be loaded for games that start at minute 0.

lyx is what-you-see-is-what-you-mean GUI to create LaTeX without having to write or even read code. By configuring some hotkeys one can write equations more quickly than with a pen. With copy & paste solving linear equation systems becomes trivial. Since one can only add/remove pairs of parentheses and select the entire term, one also never runs into issues with them that can easily occur otherwise

I did forget what the patch was about. The background was the intended change to the networking code to send only the changed setting instead of the entire gamesettings object, so that the server can check whether the modification was permitted, right? And this rewriting iteration didn't store the chosen settings in g_GameAttributes but in the GUI objects and constructed the gamesettings object from the GUI objects and the map data before launching the game. I recall. Maybe getSetting could be renamed to getGameAttribute to prevent confusion with the common settings.js code. 3 could become "ai:".length.

> Build traps around the owned territory I didn't hear that one yet, it would add a new dimension to the game and would be historically accurate too I'm thinking of the roman siege walls that had horrible spikes. We already have these spike models, but they don't deal any damage. About the gamesetup patch to allow players to select the gamesetup settings, it is a very important patch as it would also be one of the preconditions of dedicated servers (which are not as relevant as they used to be but still a nice toy). (I would avoid the word rewrite and have done so in the option unification patch because people often say we need a/one/the gamesetup rewrite in order to solve any problem in the gamesetup. Should give credit not too much and not too few credit to each rewrite. We for instance will need another major refactoring for the persist-match-settings crap and another one to allow maps to lock some settings, provide defaults for others while continuing to persist user settings by default. Also if everyone calls their thing gamesetup rewrite, people don't really know what's in it. Perhaps the term host-agnostic gamesetup would be more precise for https://github.com/0ad/0ad/compare/master...Imarok:just_another_gamesetup_rewrite?)

So this happens if the AI starts to trade and then saving the game? Always or only sometimes?

Sure you didn't try to load an old svn savegame and then clicked on save?

Perhaps one could also replace the small ones with big ones? Dunno. But it reminds me of the spartan headless units, some of these units had a different helmet depending on the animation.

If you upload a replay with unreachable trees, I'll take a look.

So upload it.

If I recall correctly, downloading as zip from github only download the files, but doesn't create a local repository folder. So it can't be updated then unless I'm wrong. Installing tortoise svn and downloading the source that way (see build instructions) is indeed the recommended way (no need to compile on windows either).

> I wish to know why have you entered a open-source project, your motivation and thoughts on the subject (open-source community, open-source at all, proprietary codes, etc) Free software is written in the interest of the user, while commercial software is written in the interest of the corporation. The latter seek consolidation until they become a monopoly and we frequently observe them working in diametrical opposition to the best interest of the user. > Do you think that open-source should be implemented as a default, and have no proprietary apps? Publishing software under a free license is right thing to do, but it doesn't mean proprietary software should be prohibited with force. > How do you feel about the project itself 0 A.D. feels like a triple A game to me and I'm proud to be part of a group that so many wise men were part of and that worked on it for more than a decade. > I mean anyone can go on and program 0 A.D in their free time Most are too exhaused in their free time to work on such stressful tasks then. > is it important that your free time is covered up in programing open-source? if yes, why? What do you think about the ethics in this? I don't understand the question.

Sounds like a map of an outdated version of 0ad is loaded, maybe an outdated, broken or otherwise changed copy of _default.xml in one of the http://trac.wildfiregames.com/wiki/GameDataPaths that are searched for maps?