No problem (yet), just a method of figuring out what the heck an user supplied to the system? :)

I imagine a mod that either consists of a single file (new assumption) or of an archive (zip/rar/jar)

We make sure we have a list of files (extracted from archive)

and we have a list of supported programming languages, and then we check what extension is used exclusively (if not exclusively, throw an exception?)

And then we know in what language the mod is written, and then we need separate strategies for handling the "file"

There has to be an 'entry point' to the mod though, doesn't it? One mod shouldn't be able to have two entry points (or should it?)

@SimonAndréForsberg That's a good point yes, and I'm still unsure how to deal with it, by convention might work best

I'm now more concerned to figure out in what language something is written in

Would Scala or Groovy also get packaged in a JAR?

Or well... they're still .class files I think

if a mod should only have one entry point, then there's not really a problem.

@skiwi I don't consider neither Scala or Groovy very relevant at this point in time.

@SimonAndréForsberg It's just to show that a JAR doesn't exclusively mean the source is Java

but on the other hand, a JAR contains .class files

Exactly. All JARs should be treated the same, no matter what they are.

On the other hand, a JAR is just an archive, nothing more (apart from manifest for example)

correct

If someone packaged Lua files into a JAR, is it fine if we can run it?

It would be foolish to package Lua files into a JAR

A JAR without Java class files should not be submitted as a mod

@SimonAndréForsberg Some people... are idiots

We shouldn't support idiots as modders.

I'm wondering if we should specify the language of a mod, or if it should try to figure it out by itself?

step 1. make it work, any way it wants to work. step 2. refactor.

i.e. IMO it doesn't matter for now.

So we'll go with supplying the language as well

well @skiwi the lambda thing did fix the problem, thank you! the only event related import that the class has is javafx.event.ActionEvent, maybe the EventHandler is imported in the actual scene?

also I had to give the command to the stage before switching to that stage or it wouldn't work right

@bazola Not sure abut that, it sounds like you were using the wrong class though

I am not totally sure.. anyway now I am getting this exception, but all the threads are properly closing now

java.net.SocketException: Socket closed
	at java.net.SocketInputStream.socketRead0(Native Method)
	at java.net.SocketInputStream.read(SocketInputStream.java:150)
	at java.net.SocketInputStream.read(SocketInputStream.java:121)
	at com.fasterxml.jackson.core.json.UTF8StreamJsonParser.loadMore(UTF8StreamJsonParser.java:176)
	at com.fasterxml.jackson.core.json.UTF8StreamJsonParser._skipWSOrEnd(UTF8StreamJsonParser.java:2606)
	at com.fasterxml.jackson.core.json.UTF8StreamJsonParser.nextToken(UTF8StreamJsonParser.java:645)

[Cardshifter/Cardshifter] skiwi2 pushed commit 1c55e9af to ecs-module-support

> The mod integration tests now delete their temp files.

@bazola When exactly do you get that exception?

as soon as I click the X to close the window

this is what I call when I click the X:

	public void closeGame() {
		try {
			this.in.close();
			this.out.close();
		} catch (Exception e) {
			System.out.println("Failed to break connection");
		}
		this.listenThread.interrupt();
		this.playThread.interrupt();
	}

@bazola That exception is expected.

You are closing a socket, but your other thread is still trying to read from the socket

[Cardshifter/Cardshifter] build #222 for commit 1c55e9af on branch ecs-module-support passed

You first need to interrupt that specific thread and then friendly close it

Interrupting the thread first sounds like a good idea though. Move the .interrupt calls to the top of the closeGame method @bazola

Not sure if it will work though. I think there will still be some exception.

trying to do this inside the controller is definitely not working, I have to assign it before switching to it:

		Node source = this.rootPane;
		Stage stage = (Stage)source.getScene().getWindow();
		stage.setOnCloseRequest(windowEvent -> {this.closeGame();});

> If this thread is blocked in an I/O operation upon an interruptible channel then the channel will be closed, the thread's interrupt status will be set, and the thread will receive a ClosedByInterruptException.

@bazola assign what before switching to what?

this works when I put it in the Launcher that ends up creating the actual Client:

				Scene scene = new Scene(root);
				Stage gameStage = new Stage();
				gameStage.setScene(scene);
				gameStage.setOnCloseRequest(windowEvent -> {controller.closeGame();});
				gameStage.show();

@bazola You can do just gameStage.setOnCloseRequest(windowEvent -> controller.closeGame()); here

putting the .interrupt calls at the top did not cure the exception

@skiwi that is what I am doing?

@bazola Well I more meant that you don't need the { ;} around it

ah okay I see now :)

@bazola let the exception be there for now, it doesn't do any harm, really

it is an expected exception

well I can live with that, but it does seem strange that an exception is guaranteed to happen when the program closes like that?

by doing this I did manage to make the exception stop though:

		this.listenThread.stop();
		this.playThread.stop();

Yes, but stopping threads is not recommended.

@skiwi or I (or @Marc-Andre?) can have a look at it sometime though. Just create a github issue for it.

[Cardshifter/Cardshifter] bazola pushed commit 1a2d774a to javafx

> Fixed issue of all threads not closing on exiting client

[Cardshifter/Cardshifter] bazola added label next release to issue #42: FX Client does not close all the threads

[Cardshifter/Cardshifter] bazola removed label in progress from issue #42: FX Client does not close all the threads

[Cardshifter/Cardshifter] build #223 for commit 1a2d774a on branch javafx passed

[Cardshifter/Cardshifter] bazola added label bug to issue #45: Exception caused when client closes

[Cardshifter/Cardshifter] bazola commented on issue #45: Exception caused when client closes

@Duga That was a very fast build, wow

I hope that clean building doesn't take too long for you guys with FindBugs added

Depends on how long 'too long' is

[Cardshifter/Cardshifter] skiwi2 pushed commit 3a63a4f7 to ecs-module-support

> Removed auto detection of mod language, added manual input for mod language.

@SimonAndréForsberg It takes about one minute on my PC, so I would say 5 minutes qualifies as too long

[Cardshifter/Cardshifter] Zomis commented on commit 3a63a4f7

> May I recommend an enum for the mod languages?

my builds usually take just a few seconds, but maybe I don't have findbugs running properly?

[Cardshifter/Cardshifter] skiwi2 commented on commit 3a63a4f7

> I don't think that would be the best solution, as mod languages might be saved as strings on disk/database, I imagine other people to only have a string available.

@bazola I don't think you build with the mvn command

Also depends on what we are talking about... I am talking about a project-wide clean and build

I'm not sure if I have FindBugs running or not, I was pretty sure I installed it

@SimonAndréForsberg About the enum thinigie

I don't really see an usecase for it

[Cardshifter/Cardshifter] build #224 for commit 3a63a4f7 on branch ecs-module-support passed

@skiwi I just spotted some code string duplication, that's all (for now)

@SimonAndréForsberg Well yes, but apart from the tests I do not see a need to hard-code the language in code

@skiwi Doesn't the actual mods need to specify it or something?

I don't think so

I imagine a server having a list of mods stored as disk locations and language attached to it

Alright

@SimonAndréForsberg it doesn't seem like the server is sending the ClientDisconnected message

@bazola I can't imagine why it wouldn't. Any details for how to reproduce that behavior?

[2014-10-01 13:00:14,426]  INFO   ClientIO [    Conn-3] (         ClientIO.java: 31) - Send to Player24: {"command":"resetActions"}
[2014-10-01 13:00:14,429]  INFO   ClientIO [    Conn-3] (         ClientIO.java: 31) - Send to Player24: {"command":"useable","id":1,"action":"End Turn","targetRequired":false,"targetId":0}
[2014-10-01 13:00:14,444]  INFO   ClientIO [    Conn-3] (         ClientIO.java: 31) - Send to Player24: {"command":"useable","id":5,"action":"Play","targetRequired":false,"targetId":0}

I can see the client disconnecting, but no message being sent to the other client

@bazola Might it to be because you close the socket and thus not receive data anymore?

there are two clients open

when I close one, the last message on the other one is the ResetAvailableActions message

@bazola You're right, problem found.

[Cardshifter/Cardshifter] Zomis pushed 5 commits to client-server

[Cardshifter/Cardshifter] Zomis pushed commit 739f9e9d to client-server

> Sending disconnected message also if there are 'no more bytes to read'

feel free to pull those changes @bazola (git pull origin client-server)

[Cardshifter/Cardshifter] build #225 for commit 739f9e9d on branch client-server passed

[Cardshifter/Cardshifter] bazola pushed 1 commit to javafx

[Cardshifter/Cardshifter] bazola pushed commit 418773a7 to javafx

> client now gets visual update when opponent leaves

[Cardshifter/Cardshifter] Zomis pushed commit de4dc87f to ecs-utils

> Added EventExecution interface, moved the event executing to the EventExecutor

one question is, should I be creating issues for problems I am about to solve? or does it not matter that all of those issues make it into the Done column?

@bazola I personally don't think it's any use in creating new issues just to solve them a few minutes later

[Cardshifter/Cardshifter] build #226 for commit 418773a7 on branch javafx passed

Although I have noticed @skiwi tends to do that for some reason I can't really understand...

[Cardshifter/Cardshifter] build #227 for commit de4dc87f on branch ecs-utils passed

I was thinking the same thing, just wanted to make sure

the only logical reason to do so would be to have a record in the Done column, but there is also a record in the commit history anyway

@SimonAndréForsberg I haven't been doing it much lately

@skiwi github.com/skiwi2/GithubHookSEChatService/issues/39

about a minute between opening the issue and closing it :)

@SimonAndréForsberg That's just fast implementation!

Well, perhaps I make exceptions to my own rules, but I do not know on which grounds

[Cardshifter/Cardshifter] Zomis created branch ais

[Cardshifter/Cardshifter] Zomis pushed commit dc60a50a to ais

> parameter on registering events for identification / easier debugging

[Cardshifter/Cardshifter] build #228 for commit dc60a50a on branch ais passed

@Duga That's the message I was waiting for!

[Cardshifter/Cardshifter] Zomis pushed commit 144f355c to ais

> Added logging of all possible actions to AI

[Cardshifter/Cardshifter] Zomis pushed commit 48fc0da2 to ais

> Added Scheduler and Guava to Server

[Cardshifter/Cardshifter] build #229 for commit 48fc0da2 on branch ais passed

@SimonAndréForsberg Nice!

I want to start working on a Java mod

what is the best way to create a minimalistic testing mod? (I only care for now that it starts up)

@skiwi How minimalistic is 'minimalistic' ?

@SimonAndréForsberg As bare as it can be

new ECSGame(); <--- done!

It doesn't have any players or anything in the game, but it is a game. It doesn't support player turns or anything else, but it is a game!

@SimonAndréForsberg Ok cool, I do think I'll need to make a whole Maven module for that

Ah maybe you misunderstood me

I want to produce a JAR that contains the most minimalistic game

Again I ask: How minimalistic is minimalistic?

Same constraints

:)

I'll write something up for you real quick

My idea is to create a Maven dependency to that JAR, then attempt to move it into the test resources folder, and then the big work can start to try to make it runnable

	public static ECSGame bareGame() {
		ECSGame game = new ECSGame();

		PhaseController phaseController = new PhaseController();
		game.newEntity().addComponent(phaseController);

		for (int i = 1; i <= 2; i++) {
			PlayerComponent playerComponent = new PlayerComponent(i - 1, "Player" + i);
			Entity player = game.newEntity().addComponent(playerComponent);

			Phase playerPhase = new Phase(player, "Main");
			phaseController.addPhase(playerPhase);

			ActionComponent actions = new ActionComponent();

Note that PhrancisResources.HEALTH can be replaced by any enum that implements ECSResource

@SimonAndréForsberg Was just about to ask ;)

Only you give a method that returns an ECSGame, I think mods work with a class that implements an ECSGame?

It's perfectly playable in ConsoleControllerECS, although I doubt that the JavaFX client would support it

@skiwi No, mods should not implement ECSGame.

ECSGame should not be extended. (although I have apparently not marked it as final yet)

So what is the entry method of a mod?

You just create a new ECSGame() and add stuff to it

A mod should be like a Supplier<ECSGame>

it can use an API to create the objects, of course

So the PhrancisGame has a method createGame I see

cardshifter-testmod-java, here I come!

@SimonAndréForsberg Can I tell maven to download dependencies and only that?

I cannot compile currently because it cannot resolve dependencies, I cannot add dependencies as it cannot build, because it cannot compile.

@skiwi Wait, what?

Yeah... that's the price I pay for copy pasting code (your code here in chat)

mvn clean ?

I commented out the offending code now and it worked, but it's not ideal

It's ECSResource btw, not IResource

The END_TURN_ACTION also happens to be defined in PhrancisGame @SimonAndréForsberg

@skiwi that's what I said:

@skiwi just use any string you want, it doesn't matter here

@SimonAndréForsberg Ah you ninja edited it since I last read it ;)

Okay, I think I've got this

@skiwi I edited in just a few seconds

[Cardshifter/Cardshifter] skiwi2 pushed commit d970b67a to ecs-module-support

> Added the cardsfhiter-testmod-java module which is a very basic Java mod.

[Cardshifter/Cardshifter] build #230 for commit d970b67a on branch ecs-module-support passed

[Cardshifter/Cardshifter] skiwi2 pushed commit f079098e to ecs-module-support

> Renamed the startGame method to createGame in the Mod interface.

Figuring out how to get that JAR into the test resources now...

@SimonAndréForsberg When in one of the children of a pom, you can do this:

<outputDirectory>${project.parent.build.directory}</outputDirectory>

Do you know if it is possible to go up to the parent, and then select another child, and then go down to build.directory?

[Cardshifter/Cardshifter] build #231 for commit f079098e on branch ecs-module-support passed

@skiwi I have no idea

Argh, I spent several hours figuring out why one thing didn't work... turns out I didn't do it when I thought I did it.

@SimonAndréForsberg The most but one annoying issue!

Only bugs inside programming languages are worse.

I think I need to make cardshifter-core a parent pom or something, such that cardshifter-testmod-java can live under it

Don't know how exactly I go about that though

Is it possible to "replace" cardshifter-core with cardshifter-core-parent and under that place cardshifter-core and cardshifter-testmod-java?

> The projects in the reactor contain a cyclic reference: Edge between 'Vertex{label='com.cardshifter:cardshifter-testmod-java:0.3-SNAPSHOT'}' and 'Vertex{label='com.cardshifter:cardshifter-core:0.3-SNAPSHOT'}' introduces to cycle in the graph com.cardshifter:cardshifter-core:0.3-SNAPSHOT --> com.cardshifter:cardshifter-testmod-java:0.3-SNAPSHOT --> com.cardshifter:cardshifter-core:0.3-SNAPSHOT -> [Help 1]

I'll check later this week onto how to mess this up more.

It might work if the mod copies the jar to the parent project, and then the core references that location

Massive spam! oh the humanity!

@SimonAndréForsberg we just got hit by the spam monster again

I thought I had secured the Q&A enough already, but apparently I missed some settings

It should be secured now. If it isn't, then there's something horribly wrong with it...

@skiwi the core shouldn't have a dependency on the testmod. At all.

[Cardshifter/Cardshifter] Zomis assigned Zomis to issue #22: GUI continues when game is over

[Cardshifter/Cardshifter] Zomis removed label ready from issue #22: GUI continues when game is over

[Cardshifter/Cardshifter] Zomis added label in progress to issue #22: GUI continues when game is over

[Cardshifter/Cardshifter] Zomis added label ready to issue #45: Exception caused when client closes

[Cardshifter/Cardshifter] Zomis added label ready to issue #43: Test for client/server connection

The core needs a built test jar in the test resources for the integration tests

Now really going to bed

If you need a jar for it, make a jar once and put it into the test/resources folder. Don't do it through Maven.

night @skiwi