zoot

Can't hurt, I guess: http://trac.wildfiregames.com/ticket/1719

I don't recall. The sound just felt superfluous to me when you also have tons of attack sounds at the same time.

Also, I've been thinking that we should probably suppress the sound if the camera is already trained on the location where the attack is happening. This would require an engine function exposed to the GUI context that returns the current camera position. But if that is too difficult to implement, we'll just wait and see what others say.

If you just push once you've removed it from the C++ side, I'll do some other stuff for a while. We should still fix this: Either you can have a crack at it, or I will do it when I get back.

https://github.com/zootzoot/0ad/commit/6a36d3e514e91151ad6065c7116d462f3ff9d04a

Seems good. About the 'time' parameter, we can remove that if you aren't using it (since no one else are). Should we?

The black dot issue is fixed. Is still get ping times up to 15-20 secs, though. Is that intentional? Or is it my frame rate that is being weird, perhaps?

I seem to be getting an issue where the 'dots' turn black for a while before they begin blinking: (Three frames depicted; match start, dot turns black, dot begins blinking.) Any idea what that might be?

I don't know, actually. But it is based on turns, I believe (not real-time).

If x and z are unused, I guess we can drop them from the message: https://github.com/zootzoot/0ad/commit/d7b90aee4de7c60d530c8277173e5eb481ce1439

Looks awesome! Only tiny problem I can find is that the ping time seems way too long. The blinking should at most continue a few secs after the entity is no longer being hit. Hmm. I'm not sure, but I think they are that high. These are 'pixel' or sub-tile positions, not coarse terrain tile positions.

https://github.com/zootzoot/0ad/commit/0f2aa3e2a9efcb057493dba4e627250059b54740

Yes, that can be done. I'll get on it in a sec (pending coffee ).

It's defined here: https://github.com/0ad/0ad/blob/master/source/gui/scripting/ScriptFunctions.cpp#L167

ICmpPlayerManager is also in the sim, so it will be player neutral too. In messages.js, I used the Engine.GetPlayerID() call. Perhaps you could track down how that works.

No, the sim needs to stay 'player neutral', I believe. It is used for all sorts of validation (anti-cheat, network sync, etc.), so the engine needs to be able to rely on two different hosts reporting the same values.

The PingMinimap call was sent from messages.js where a check on the player is done. (messages.js is in the GUI.)

The sim has no concept of who the current player is, any such check has to be done on the GUI side.

I guess what matters is whether you can sell that argument to a sufficient number of developers. Right now the pace of development doesn't seem huge. Will the employment of a manager affect it positively or negatively? I guess time will tell.

The error went away when I did this: https://github.com/zootzoot/0ad/commit/1a1641487a8d1570d93a45ee70c1d30296801e84 (If you pull you may have to recompile, since I've also merged in the latest 0ad/master.)

Actually, no, that's for component interfaces, not messages You seem to have added the message to MessageTypeConversions.cpp properly.

Hmm. The docs says you need to add a "COMPONENT" line to TypeList.h. Without that, the error could be due to the "MT_EntityAttacked" message type constant not being properly registered on the JS side.

Seems there may be a mismatch between your definition in MessageTypes.h and the format of the 'event' variable. If you can't make the definition in MessageTypes.h match, you should just leave the event variable out of the call (since it's used elsewhere and needs to retain its format) and pass a literal, like: Engine.PostMessage(target, MT_EntityAttacked, {target: target, posx: pos.x, posz: pos.z, time: cmpTimer.GetTime()}); I haven't used MessageTypes.h before, so I am not completely sure what is expected there, but hopefully something in that direction.

There is nothing unusual in your profiler output, so it would have to be something unprofiled. No one else seem to have encountered it over the ~6 month development cycle, though. (Edit: more like 3 month, but still.)

Hmm. What are you referring to? Entities do indeed have multiple components. But hopefully only one minimap component?