Don't use if constexpr in a lambda. I think that lead to the error.
Again change variable names and comments.

Change naming and comments a bit.

In D4770#223669, @phosit wrote:

It should be visible by non-host-clients if it's a savegame. (It's not)

The information if a savegame is loaded is with this implementation only available to the controler.
The easiest solution would be to make the loadsaved-ness an init-attribute / setting like wraitii suggested. That requires me to rewrite and test much again.

In D5256#223809, @Vantha wrote:

In D5256#223808, @phosit wrote:

Before this diff the "forat translation" is allways done. That's almost certainly an issue.

Not sure what you mean by this, though. What should I change?

In D5256#223805, @Vantha wrote:

What do you suggest? Something like this?

GameSettings.prototype.Attributes.PlayerName.prototype.CountLabel =
	translate("%(name)s (%(count)s)");

Yes, almost. I assume the "s" is for strings. count is an integer so it should be translate("%(name)s (%(count)i)").

In D5189#223719, @phosit wrote:

In D5189#221449, @vladislavbelov wrote:

Have you measured performance for those changes?

It's nearly the same:

In D5256#223792, @Vantha wrote:

Do you want me to change it? I thought ' (${count})' looked better than " (" + count + ")".

No, I'm ok with that.

This inline went missing

I never saw Template Literals in this codepase. I don't object it. :)

Reduce "test_RecoverableError.js"
It should pass the tests now.

revert

It's not a big problem :)

Upload the earlier patch (that would be 11164) again. I could do that tomorow.

Sometimes there is a long "tooltip"

In D4770#223669, @phosit wrote:

It should be visible by non-host-clients if it's a savegame.

Does anybody have an idea where this could be shown?
An idea was to put it near the "cheats enabled" or "rated game" warnings but there is litle space.

I think i'll commit it with the deprecated path. We can decide later in the release cycle if we want to keep it. Propably dependant on how long it takes. And propably the will be breaking rmgen changens.

MakeExportable hasn't to be called from js

Remove #undef FLOODFILL

In D5189#221449, @vladislavbelov wrote:

Have you measured performance for those changes?

Revert embarrassing things.

Use promises for GUIpages (see summary)
Remove things related to multithreading.
Use std::queue instead of std::vector because of iterator invalidation.

Don't look at it it's embarrassing ^^

Hide the button in single player
Don't unlock civ-selection when changing assigement

rebase

We noticed bugs while testing:
When loading a game and changing starting position sometimes the civ column get's unlocked. (It should stay looked)
It should be visible by non-host-clients if it's a savegame. (It's not)
Loaded game should not be ranket. (We didn't acctually test that. I don't know the current behavior)

Thank you for testing.
I testet "2 humans" localy with two clients.
I could use your help to test it using the lobby. I'll have time on sunday morning (untill 12:00 UTC+2) or evening (from 17:00 UTC+2)

Rebase

How does this relate to #305?

In D5254#223622, @Stan wrote:

Related ? https://trac.wildfiregames.com/ticket/5989

In D4770#223608, @Stan wrote:

Well, can you break it? (Corrupted savegames, forged ones, spamming buttons, accessing it from somewhere you shouldn't) ? Possible bugs with AIs, rejoin check etc

In D5195#223604, @Stan wrote:

Well I would assume this changes the generated assembly which could be for the worst compared to macros. But I suppose it could / should be profiled.

Can somebody take another look. pls.

@Freagarach are you ok with the current wording?

Use std::any_of

In D5195#223594, @Stan wrote:

Dumb question, but doesn't that make the code slower?

Copyright year

• fix typo
• use exceptions
• ignore some warnings from json_spirit

Rebase

In D5240#223558, @Boston_Brooks wrote:

In D5240#223529, @phosit wrote:

If you zoom out (much), toggle the view and rotate the actor. There are artifacts.

This is a hard problem, I don't know if the orthogonal projection is interfering with rendering of "texture cubes" and "sky managers" (in source/renderer/SkyManager.cpp), so the screen isn't being cleared before the next frame is drawn.

In D5222#223533, @sera wrote:

Thinking a bit more about player spoofing, it probably is a bad idea in the first place and extending the command protocol to have both a "user" and "player" id would probably make implementing co-op play a lot cleaner. Sure it's not the intended scope of this patch ...

Use json_spirit

If you zoom out (much), toggle the view and rotate the actor. There are artifacts.

In D5222#223495, @sera wrote:

In D5222#223494, @phosit wrote:

IMO The behavior of the server shouldn't dependent on the simulation's init-atributes. Instead the controller should explicitly control the server.

allowCheats is an init-attribute, at least that's my take.

Is that also the case if we call it allowPlayerSpoofing?

IMO The behavior of the server shouldn't dependent on the simulation's init-atributes. Instead the controller should explicitly control the server.

rebase

Move the realTimeSinceLastFrame calculation back in to Frame.

In D5240#223444, @Boston_Brooks wrote:

Deduplicated default value vs initialised value.
Values are placed in the constructor,
in line with other code in the project.

elexis argues that i should change all the maps and remove the old way in this diff.

Readd the test

In D5220#223421, @sera wrote:

In D5220#223420, @phosit wrote:

In D5220#223418, @sera wrote:

Currently you add another way to do the same thing but using the old way doesn't complain loudly, how are map authors supposed to know that they are supposed to migrate. When will the old way be removed?

My plan is to commit this before the release of a27; add a warning after the release of a27 and remove the old way after the release of a28.

Did some api break happen since A26? If so some more doesn't matter at all anymore and you could just go ahead with dropping the old one. If not why only warn after a27 is released, the earlier the better me thinks? If you plan to break the API next cycle, you could just commit this one without the old way with the rest of the anticipated breakage, right?

All breaking changes should be scheduled the same. e.g. Introduce the new thing in a27, deprecate the old thing in a28, remove the old thing after a28.
@marder @wraitii @Stan

In D5220#223418, @sera wrote:

So what is the plan?

My plan is to

1. Learn rmgen
2. Document the flaws
3. Come up with a solution
4. Convince other devs

xD

In D5220#223414, @sera wrote:

Also now that you drive generation from cpp why the need to call export map from js?

In D5220#223413, @sera wrote:

global is as good a name as rootedvalue or habakuk.

If run in a browser you also have "document" and "window", so multiple ones is not the issue. "document" and "window" are well crafted names. We have "Engine" and now you add "global", the latter sounds like something one comes up on a spur.

So the question is why does "Engine" not work here and if not can we come up with a better name than "global"?

In D5220#223411, @sera wrote:

global.g_Map

why not Engine.map?

In D5240#223398, @Stan wrote:

@phosit is there a convention for floats : 1.0f, 1.f, 1f ?

Adopt empire as an example.

Add points at the end of comments

flatMap instead of map(...).flat()
{} instead of =

copyright year
missing include
typo

Now you can ;)

Rebase

Got more usefull with D5226.

Don't store the xmppClient in CGUIManager.

Use an Error.
Make it less branchy.
Move RunGenerator to ScriptFunction.
Prevent the "printing" from rmlib.

Add a test.
Use KiB more and include it's header.

Copyright year in NetServer.cpp