move comment to better place

some white spaces and comments

I can do separate patch for injured function and to unify this.hittpoints calling but I do not see reason for this diff to wait for it :)

move some code back to Reduce and rename function

I tried in game and one unit command works on minimap

fix tests

rebase

fix

found flaw

refactor

New implementation based on comments above.

builds, deprecated variants are replaced by valid equivalents

remove not needed clamp
mapdata is u8 (0 - 255), maxed value we can get here is 65280 what is in range of u16 (0 - 65535)

just for record: Corinthian Ismuth (4) 4 ai very hard still out of memory

wrong diff link, include should be replaced, gameboy reported broken build on forum ( I wonder how Harbormaster passed, I dont have time now to test build), from his report looks that uniform_real_distribution' : is not a member of 'boost'

I accept this because:

In D1813#74338, @elexis wrote:

15:34 < bb> just leaving it accepted for a short while (till tonight) in case someone wants to add something

• One may check for other unused vars in the same file (the compiler should already be checking for that)
• One may want to find the revision (for referencing, possible audit, revision of the underlying feature design, discovery of forgotton implementation ideas etc)
  
  may = may chose not to

Pictures are white-black or white-gray-black.
Maps looks good I think.

windows builds with no pch settings

I am on windows so as you want :)

Looking closer this does not work.

@Nescio would not be better to change vegetables to grain?

it is 5ms for all currently selected (200) units. I called this update from updateGUIObjects in sesson.js

Yes, one would assume that when computation takes from 0 to 1 ms for one unit, it will take around x ms for x units but ( i was surprised ) it does not

Looped 100 times and 200 times and did not got bigger than 5 ms if recall correctly. But when I tried to update this overlay for placement, it was 1 ms for that building so I think it would not be wise update ranged overlay for units if they move. Thats why given information would be incorrect for them and more misleading as it would depend on the position they were spawned.

Restricting this update to buildings. For units this to work it needs to enable recomputing range overlay on unit motion what could become expensive.

It is recomputed just if object changes owner, is created or some modification comes. Overlay for units is not updated white moving even when they get elevation bonus. Using let now = Date.now() and logging Date.now() - now at the end of GetElevationAdaptedRange returns 0 or 1 depending on range, bigger maximum range and elevationbonus longer it gets to compute. However with 1000 loop it freezes. But is it actually such problem if range overlay is not updated on movement ? We could restrict this computation just for buildings and make TODO to solve it for units.

added missing check for change in elevation bonus

Right, elevationBonus is number based on which is added range corresponding to how high is current terrain.
It is based on GuiInterface.js lines 410 - 426 which result is used to display (+x) in toolip for range attack.

sele ptol civ tag revert

Identity for player entity have been in game before this patch. See special/player/player.xml

should not be result of && operation boolean and not string ? But somehow (well javascript) it is string if first is true else it is nothing

simplify map function

just still dont get this javascript features somehow, there is return text && sprintf("..."), and aside from that the result of this expresion should be true, but it is replaced string

Uses of Formation.prototype.GetClosestMember
UnitAI.js
Line 184 - called without filter
Line 1920 - called with filter UnitAi.CanAttack(entity)
Line 4303 - called without filter
Line 4405 - called without fiter

I can confirm this is fixing assertion error in both cases described in ticket.

C4458 and C4456 not displayed while building project with --atlas. Project builds without errors. It is good for me.

periods

years

less lines

missing dir

2 variables -> one

ready to go

forgot 2 comments, sorry

fix translate

D1945 moves fullscreen crash avoid handling inside render function and removes from frame function,so there can be optional do not render as it applies to every situation window and fullscreen.

forgot to translate :)

remove wrong dots

oh sorry, I did not meant to do that :) I just used replacement

dots :)

rebase
@vladislavbelov chance to look into this? I would like to finish it :)

should be everything

@wraitii extend this diff or make another one?

We should realise that current behaviour returns only INVALID ENTITY if filter is passed what is cmpAttack. CanAttack and there is check if target is enemy esle returns false. So clearly unit cannot attack itself as is not enemy of itself.

In unitAi healer filters possible targets with injured mark using Health.prototype.IsUnhealable function (and another checks) so unit with maxed health is not healed and stays injured and checked in next search if still in range.

If unit is marked as injured, healers get it as unit to heal. So if unit with maxed health would be marked as as injured, healers would waist their time for unit that does not need it. I think, can someone think that this is not good change and disagree ?