zoot

None that I can think of. There are no consumers of MT_AttackDetected yet. If there is any affected code, we should find out soon enough.

Lots of people with full-time jobs are not committed or even outright slackers, while, on the other hand, most of the work on this project has been done by unpaid "hobbyists". So nominal work hours doesn't really seem like a good measure of quality. But if all unpaid contributors on the team are behind it, and don't feel like they are being robbed, I suppose it could work, depending on how many more years they are willing to volunteer

Right, my bad

How does that relate to having a paid manager? Anyway, we'll see how things turn out.

A word of warning from Karl Fogel (a co-creator of SVN): I'm quite skeptical what a paid 'manager' would be able to add to the project, but we'll see Good to see that feneur stays around in some capacity, though.

To pull my changes into the current branch in your own repo, you would do something along the lines of: git remote add zootzoot https://github.com/zootzoot/0ad.git git pull zootzoot/attack-notification The first line adds my repo as a remote, and the second one does the pull. There must be some kind of interface between CCmpMinimap and the actual minimap already, but yeah, I don't know if it suffices for this task.

Here is what I have now, you'll decide whether it seems workable: https://github.com/zootzoot/0ad/commit/1d5db6f8d4511915266d859c25950b98fc72e238 Instead of calling Engine.PingMinimap(), this posts a message of type MT_EntityAttacked to the attacked entity. You should be able to intercept this message in the CCmpMinimap.cpp component, like it already intercepts MT_PositionChanged and MT_OwnershipChanged messages: https://github.com/0ad/0ad/blob/master/source/simulation2/components/CCmpMinimap.cpp#L139 The message type would have to be defined on the C++ side first, though.

Yes, you can always go back to the earlier commit. See e.g.: http://ariejan.net/2011/09/08/git-remove-reset-and-rollback-commits/ http://davsblog.me/how-to-do-rollbacks-in-git-in-other-words-oop Another option would be to create a new branch from your existing one (Git branches are designed specifically to be cheap and light). If your work in the new branch works out, you can merge it back into the old one. If it doesn't, you simply switch back to the old one. http://git-scm.com/book/en/Git-Branching-Basic-Branching-and-Merging

What OS are you using?

Yes, basically. We should still be able to see that units are under attack, even if the initially pinged fellow died.

Suggestion: What do you think about moving the ping functionality into the CCmpMinimap.cpp component (much like you originally suggested), but then instead of triggering it from messages.js once, together with the sound, it would be triggered directly by the AttackDetection.js component every time a unit is hit? That way units still under attack would keep blinking even after the one initially pinged has died, and it would make more units in a skirmish blink, thus making it easier to see.

Yes, definitely strive to add them to parent templates only. Otherwise you will waste a lot of your own time as well as make the templates harder to maintain.

It will, yes.

So something like: git remote add 0ad git://github.com/0ad/0ad.git This would add k776's repo as a remote under the name '0ad'. Then you should be able to do: git pull 0ad master to pull in the latest commits from his master branch.

Assuming you mean the master branch on k776's repo, you need to add his repo as a remote first: https://help.github....adding-a-remote

Yes. Do you get the error on your fully updated SVN working copy?

Hmm. Seems you need to pull from the 0ad repo and rebuild. That "Engine.GetSimRate" function which is referenced in your error output was only committed yesterday. Somehow your JS and compiled source has come out of sync.

Whoever recorded it has a good computer and/or only plays with at most one AI.

I pull with "git pull". To temporarily test a given remote branch, I just use "git checkout".

Actually it does, since it hasn't been optimized yet.

Great, it works now I'll just grab something to eat, and then I'll have a look at it.

(I guess it needs to be "MiniMap" instead of "Minimap". Unix systems are case-sensitive.)

So far so good, though now I get:

Yeah, I guess ISO C++ 2011 only has experimental support in GCC.

Great. I get errors when building, though: