Changes to satisfy Stan, Lint and Jenkins. ;)

The = defaultwas ok but not in the header file.

This is a good fix.

I did not test that sending reports still work but i have saw Report Submission on thread 2. So i think it works.
GetShutdown on L257 should be done bevore the task is created(appended).
More things should be changed: code-coments thread -> task and condition-variable-> ?. for code changes i wrote inline coments.

There is a counting_range in boost which could be used, if we are not going to implement iota.

• I'm not convinced that for-range iterating ::iota, or using std::transform is more readable than a regular for loop, nor much simpler to read. It's going to be an uncommon sight in the codebase, and that makes it meh to use in my opinion.

More includes

Triger rebuild

In D4614#198203, @vladislavbelov wrote:

Ideally I'd prefer to not have back relation CFont > CFontManager (via callback). I don't think CFontManager being friend of CFont is worse than that, because it makes CFont a pure data holder without much logic.

Heavy Warships also have this behavior.

The Flag dont disapear on halt and on attack the spears dont move (this is good).
Just for understanding what is the Footprint stuff about?

include and type change

CFont now only has const public member functions.
CFont can only be created using CFont::Create.

Test results from Mentula on #0ad-dev with the patch: https://paste.gnome.org/KNeEYmHZZ
and bevore the patch:

With SDL_VIDEODRIVER=wayland(or disabeling xwayland) sdl_video_backend is Wayland and without x11.
After using atlas sdl_video_backend is none.
Screenshots work in the game and atlas.

It looks fine. I don't notice a performance difference.

In D4636#197087, @vladislavbelov wrote:

In D4636#197050, @phosit wrote:

Why are the constructors private?

To emphasize that they should be created only by their device.

In D4636#197040, @vladislavbelov wrote:

In D4636#197031, @phosit wrote:

Why not use variant? (instead of a pointer to base class)

A bit longer usage, having GL (and other backends) includes (direct or indirect) for regular code to know the size

Why not use variant? (instead of a pointer to base class)

Why are so many CMatrix3D fonctions defined in the header?

You don't know what they do:

• They might just declare a constant
• They might be a "function"
• They might modify local variables, without me passing them as argument.

• rebase
• using instead of typedef

As always macros should be avoided but we can do that in another diff.

test if the error/wargnig is logged.

I encountered no error.

CFonts are const outside of CFontManager
Improved spelling in comments

Pascal case for CFont::SetTexture

This solves #6476

Sory, i missonderstood Stans comment.
Go ahead

In D4603#195828, @Stan wrote:

Lib update is post FF material so committing this sounds good.

ther are more spots were i would use reinterpret_cast

In D4602#195807, @vladislavbelov wrote:

Did you update to the last svn?

It does not compile on my machine:

In D4279#195318, @vladislavbelov wrote:

In D4279#195314, @phosit wrote:

I think using the sea level (as of now) is a nice mindset. So the solution would be to raise the wather level in the affected maps.

That doesn't work when a map is on a side of a hill.

This is becoming to complex and we will never draw the exact "field of vision".
In my opinnion there should be one "intersection plain" across the map.

In D4573#194982, @vladislavbelov wrote:

In D4573#194980, @phosit wrote:

If it might be the same bug isn't the patch just hiding it? Could you tell the reason of the crash from the ticket?

yes it is. no i don't

It'd be useful to understand the reason. Can you reproduce the crash from the ticket?

In D4573#194979, @vladislavbelov wrote:

In D4573#194978, @phosit wrote:

The specivic defect from the trac issue is solved/unreachable.
But ther are other bugs in Atlas, or it is the same bug from a different angle.

If it might be the same bug isn't the patch just hiding it? Could you tell the reason of the crash from the ticket?

The specivic defect from the trac issue is solved/unreachable.
But ther are other bugs in Atlas, or it is the same bug from a different angle.

I get sometimes a crash when doing the following:

• Start Atlas
• switch to object tab
• click on Switch to Actor View
• click on Switch to Game View
• switch to map settings
• click on generate map
• switch to object tab
• click on Switch to Actor View

In D4573#194928, @Langbart wrote:

I am getting a crash when doing the following:
...

I followd your instructions but don't get a crash.

Applied suggestions from @vladislavbelov.

In D4570#194847, @Vulcan wrote:

Build is unstable, some tests have failed - The Moirai have given mortals hearts that can endure.

builderr-release-gcc7.txt
In file included from ../../../source/pch/atlas/precompiled.h:26:
../../../source/tools/atlas/GameInterface/Messages.h: In function 'void AtlasMessage::fGetTerrainGroupPreviews(AtlasMessage::qGetTerrainGroupPreviews*)':
../../../source/tools/atlas/GameInterface/Messages.h:305:8: warning: '#'target_mem_ref' not supported by dump_expr#<expression error>' may be used uninitialized in this function [-Wmaybe-uninitialized]
 struct sTerrainTexturePreview
        ^~~~~~~~~~~~~~~~~~~~~~
../../../source/tools/atlas/GameInterface/Messages.h:305:8: warning: '#'target_mem_ref' not supported by dump_expr#<expression error>' may be used uninitialized in this function [-Wmaybe-uninitialized]
../../../source/tools/atlas/GameInterface/Messages.h:305:8: warning: '#'target_mem_ref' not supported by dump_expr#<expression error>' may be used uninitialized in this function [-Wmaybe-uninitialized]
../../../source/tools/atlas/GameInterface/Messages.h:305:8: warning: '#'target_mem_ref' not supported by dump_expr#<expression error>' may be used uninitialized in this function [-Wmaybe-uninitialized]
 struct sTerrainTexturePreview
        ^~~~~~~~~~~~~~~~~~~~~~
../../../source/tools/atlas/GameInterface/Messages.h:305:8: warning: '#'target_mem_ref' not supported by dump_expr#<expression error>' may be used uninitialized in this function [-Wmaybe-uninitialized]
../../../source/tools/atlas/GameInterface/Messages.h:305:8: warning: '#'target_mem_ref' not supported by dump_expr#<expression error>' may be used uninitialized in this function [-Wmaybe-uninitialized]

Link to build: https://jenkins.wildfiregames.com/job/docker-differential/7564/display/redirect

Updated Copyright Year

last patch could not be applied

git diff --no-prefix -U300 --output=diff

• Direct initialization of two floats
• More const variables
• Folowing the brace-style