In D4804#204522, @Stan wrote:

Maybe it prevents the background from being the same color as the rest aka gray?

Code introduced in patch https://github.com/0ad/0ad/commit/f63b762cf19a7613027f8cab8dca8b1f860217d0

In D4498#192151, @Stan wrote:

#6435

Probably part of this patch: #6435

In D3959#191620, @Stan wrote:

https://trac.wildfiregames.com/wiki/TranslatingMods

In D3959#191619, @bb wrote:

(sure, but not what I meant). With "the list" I meant to refer to the list of json's of all the mods. Those we have to get from mod.io or so.

In D4488#191538, @vladislavbelov wrote:

Should be fixed in the last update.

In D3959#191589, @bb wrote:

This would mean that we start providing translations for mods. Putting more stress on our translators.

The game and the atlas launch successfully. Messages appear in the console:

In D4475#190939, @vladislavbelov wrote:

Could you try the last version?

I confirm.
Constant flashing glitches on unit textures (about once per second)
GL ARB renderer

Linux x32/Clang 11
build, start game, save-load, atlas, js-console: all successfull, found no errors

In D4428#188938, @wraitii wrote:

Did you manage to fix that error? I didn't run into it.
The others I'm aware of, need to fix em.

@Silier Moved the mechanism for determining the name of the rank technology down two levels - to make it easier to replace the technology identification system in the future (if required)

added rankTechName() method in Identity.js (Encapsulating the technology name mechanism)

In D4340#188962, @Silier wrote:

If technology file does not follow the naming, this will not work.
It would be better if it could be avoided.

In D4428#188938, @wraitii wrote:

Did you manage to fix that error? I didn't run into it.
The others I'm aware of, need to fix em.

I have not build this version out of the box (linux/clang-11)

In file included from /home/nwtour/ff/0ad/libraries/source/spidermonkey/mozjs-91.5.0/modules/fdlibm/src/e_acos.cpp:44:
/home/nwtour/ff/0ad/libraries/source/spidermonkey/mozjs-91.5.0/modules/fdlibm/src/math_private.h:34:21: error: typedef redefinition with different types ('__double_t' (aka 'double') vs 'long double')
typedef __double_t  double_t;
                    ^
/usr/include/math.h:156:21: note: previous definition is here
typedef long double double_t;
                    ^
1 error generated.
ERROR: SpiderMonkey build failed

@wraitii Please look, this change correct?

In Mesa/Linux uses glDrawRangeElements() when called glDrawElements().
So in principle there should be no performance changes

smoke test passed in game and atlas. no critical change in performance was noticed (for me)

Yes, I forgot to mention - this is non-Windows build code-block

+ forum thread https://wildfiregames.com/forum/topic/37230-file_systemcpp122-function-call-failed-return-value-was-1

work

Look like part of this change:
Developer panel -> Click on "Hier pathfinding". Crash : "dataSize == width * height * bpp"

work

In D4416#188504, @Stan wrote:

@nwtour can you test this?

Look like part of this changes:
Open Developer panel in-game, click "Shadow frustrum" (build with make config=release)

In D4413#188357, @vladislavbelov wrote:

It seems dangerous since we use types like size_t.

After upgrading the system to
clang11.0 + boost-lockfree-devel-1.76 + libatomic1-10.3.1
error is not reproducible
workround is not required (for me)

In D3475#188182, @Stan wrote:

My main concern was the GPU thing, but

Keep keywords except: Realtime -> Real-Time, Mauryans -> Mauryas

In D3475#188175, @Stan wrote:

@nwtour: feel like commandeering this?

How it works
For this to work in the gnome3 (KDE did not study) - you must install a separate package https://packages.ubuntu.com/bionic/gnome/switcheroo-control

list of files

In D4410#188162, @Freagarach wrote:

Can pyrogenesis cope with a list of files?

"invert" -> "toggle"

In D4400#187915, @vladislavbelov wrote:

It may be the cause of deep freeze UI on low-spec PC

Any evidence?

Setter -> Invertor
Fix button "Shadows"
Really remove hardcode from AtlasUI/ScenarioEditor/Sections/Object/Object.cpp
Fix lint "Innacuracy Year"

In D4400#187883, @Stan wrote:

Thanks for the patch! Question, does changing the option in the actor viewer enable the shadows permanently?

In rP26117#54991, @vladislavbelov wrote:

Fixed in rP26127.

In rP26117#54972, @vladislavbelov wrote:

Does it crash or not?

Looks like Atlas stops considering rendering or shader settings
Minimal graphics settings, render "OpenGL ARB".

In D4362#187464, @Freagarach wrote:

If you want, there are more descriptions to be fixed. ;)

"health indicator" as "exempli gratia" of "status bars"

FYI Mesa update from 20.3 to 21.0.3 solves the issue with glGetQueryivARB

Change synonym of "status bars" : "health and capture points" -> "health bars". Freagarach request

In rP26093#54857, @Stan wrote:

Can you upload your hwdetect_userreporter.txt files? Check if they are different depending on the mesa and forcegl flags.. It seems to crash on GL_ARB_occlusion_query which means maybe your GPU is pretending to support something it does not.

In rP26093#54852, @vladislavbelov wrote:

Also could you try same options with forceglversion = "false"?

In rP26093#54849, @vladislavbelov wrote:

And this crash was introduced in that commit?

In rP26093#54844, @vladislavbelov wrote:

You might want to use rendererbackend = "glarb", see default.cfg.

In rP26093#54842, @vladislavbelov wrote:

The preferGLSL option is outdated, now it's called rendererbackend.

Executes the code for glsl with the option disabled (this did not happen before the commit)

In D4386#187197, @Freagarach wrote:

Fix serialisationtest.

In D721#187187, @Stan wrote:

Only occurs if you had old files lying around, make sure to delete libraries/source/glad before patching. It happens quite a lot with the CI.

@Stan Failed

In D4386#187168, @Stan wrote:

@nwtour can you try the rejointest (in the developper panel) option (with no AIs)

In D4386#187168, @Stan wrote:

@nwtour can you try the rejointest (in the developper panel) option (with no AIs)

Applied the patch, started a new game, save -> load - errors did not appear

I may be wrong, but look like missing changes of persians/infantry_archer_b.xml:

$ grep -RE 'scyth_tunic(_|_basic_)1_(\w)' binaries/data/mods/public/art/actors/* | awk '{print $1}' | uniq
binaries/data/mods/public/art/actors/units/athenians/infantry_archer_c.xml:
binaries/data/mods/public/art/actors/units/persians/infantry_archer_b.xml:
binaries/data/mods/public/art/actors/units/persians/cavalry_axeman_b_r.xml:

by the name of the files the error of this commit
After loading from the save file:

In D4383#187003, @Grapjas wrote:

Not sure how to make a patch that includes deleting a file, or if it's even possible.

Fix lint notice "Expected indentation of 4 tabs but found 5"

Removing the addition of a new translation line from selection_panels_helpers.js

In D4377#186882, @Silier wrote:

but for the sake of nice english string idc too much

In D4362#186760, @Freagarach wrote:

It'll also show e.g. the packing and resource bars.

fix lint warnings

added ternary operator for all three forms :
1 = fixed "1" (beautiful string in english)
2 = singular form (for languages with singular != "1")
3 = plural form

Removed block with translation of campaigns and dev panel

Now the behavior of the game strongly depends on whether the hotkey is in default.cfg

fix lint warnings

Test - patch really incorrect. Apparently, it plan to turn minimap

Silier requested changes to this revision.
This comment was removed by Silier.
removed a reviewer: Silier.