Build is green

Thanks, the button with the links are really needed!

Add an anim that disappear below ground for a time so that it doesn't popup
Add global variable.

Build has FAILED

Build is green

Change the actor somehow to stop showing the animation after playing it once.
We either have to add support for local simulation timers or GUI timers with arguments. Both are easy and could be done in a separate patch.
Whatever timer is used can delete the animation some time after it played.

Grugnas found out that the session code doesn't work as advertized.

Just tested the uploaded version with the simulation timers. That indeed yields an OOS as expected. It's important to notice that the animation is not always played exactly twice, but sometimes 2.5 times, sometimes 3.5 times before the arrows disappear.
I don't expect that this can be solved with (local) simulation timers nor GUI timers, because it is tied to the renderer. Also it should be played exactly once.

With grep -Ri 'scriptInterface&' * | grep -v 'const ScriptInterface&' we can find few more candidates:

At least all those GUI init ones cannot be made const, since one of the functions they use isn't const, and cannot be made const.

perhaps this isn't the cleaner solution.

Build is green

Good patch.

In D818#32474, @temple wrote:

if ever prepare time was larger than repeat rate ... Maybe it would mess up the animation

Would, refs D258

Build is green

A couple other things I remembered.
The first is that there may or may not be problems if ever prepare time was larger than repeat rate. But that's actually not the case for any unit. (Maybe it would mess up the animation too, I didn't test.)
The second is that healers are probably susceptible to the same bug. However, currently Cleopatra is the only aura/tech that modifies the repeat time, and she doesn't affect healers, so it's not an issue at the moment. It's probably the same fix so I think I should include it in this patch.

imo the place isn't that bad, but this can probably happen with many more functions (not that I can give examples). So shouldn't we skip the functionName check and use this for every timer function with a target (or other entity related data)? (I didn't check or test whether this crashes elsewhere in code)

A more generic way might work, though I'm still not sure if that logic should be in cmpTimer.

So the correct feature description would be "don't ping people if their nick appears in a historic lobby chat message?".

(Actually would be nicer to have it listed correctly and find a way to close something, even if not accepted prior.)

In D760#32449, @leper wrote:

Has anyone considered setting up a VM?

If one isn't collecting VM images and has deduced that a one-line patch must be correct,
then spending that hour or two to install and configure BSD on a VM, installing the dependencies, checking out, patching, compiling and running 0ad is moderately unproportional.

Only question, why always const, but there are functions, which look like the changing of the script interface?

Because we can, and because it might point out some cases where we are likely to do stupid things in the future. Some functions are indeed strange as const, since they do seem to change internal state, but that is in SpiderMonkey and not the ScriptInterface, so at that point we're mostly doing it for consistency and because we can.

Agree, would appear nicer.

Has anyone considered setting up a VM, or even asking downstream if that works (since downstream reported that failure)?

Actually OSX does provide that, see the second comment on rP19801. They just didn't update the documentation.

In D606#32220, @fatherbushido wrote:

I don t feel myself of good advice for judging such a ui feature

Added @elexis 's suggestions.

Ah, I guess I didn't squash that commit into what became D739. In which case you might be interested in the unused GetScriptInterface in XmppClient.h.