What I see on the screenshot looks acceptable.

(Also for reference, some brief discussions with wraitii on http://irclogs.wildfiregames.com/2019-08/2019-08-12-QuakeNet-%230ad-dev.log)

• The flag was introduced in rP10534, all files touched in that commit are addressed.
• source/tools/dist/0ad.nsi and source/tools/dist/build.sh don't need a change
• extern_libs5.lua, clean-workspaces.sh, build-osx-libs.sh had been updated correctly
• found source/nvtt/patches/arm64-fix.patch having an empty hunk for libraries/source/spidermonkey/mozjs-31.2.0.rc0.tar.bz2 from rP16872, but it looks like it's ignored when applying. NVTT needs some attention anyhow I suppose.
• still had old mozjs shared libraries around: source/spidermonkey/lib/libmozjs38-ps-debug.so, libmozjs38-ps-release.so.
• Checked all folders that contains one of the files touched in that commit or mentioned here. In particular build/, grep for spider -i and mozjs.
• Found an unrevisioned symlink-libs.py in my osx/ folder, it seems to come from D1483 which I don't recollect applying. As it's not a revisioned file, it shouldn't be an issue. The header is the same as in build-osx-libs.sh.
• dehydra/printf-type-check.js also has something hardcoded about SpiderMonkey, but doesn't seem to need an update:

// SpiderMonkey:
if (name == 'JS_ReportError')
    return ['a', 'printf', 2, 3];
else if (name == 'JS_ReportWarning')
    return ['a', 'printf', 2, 3];

but should be remembered in case those two functions change.

Tested on VS2015.

Could we make it a factor instead? Like if it is 0.5 you've got a 50% chance of correctly estimating the position of a moving object and one can modify that number for different ballistics?
With that factor one can vary either the ETA or estimated velocity (or both)? (Normally distributed perhaps.)
Just throwing in some ideas ;)

Successful build - Chance fights ever on the side of the prudent.

Build failure - The Moirai have given mortals hearts that can endure.

the antivirus solution "G Data Internet Security" is activated and the cache completely cleared

This thing again :| okay, good way to fuzz test the VFS I suppose.

As an FYI: I confirm this is working on my machine, but don't expect all 4 cores to be 100% CPU, pathfinding is only a small part of it.

In D2140#90396, @ffffffff wrote:

didnt test but why not putting this option to have it global and dont get confused maybe additional to know wut general auto attack behaviour is like in options -> ingame -> unit auto attack -> capture/attack. Maybe thats good to add. is it good?

Yeah, I thought about that, but that means hardcoding "capture". Not sure whether that is wanted though.

Was just wondering what commit to mention in the commit message that this is a fix to. It seems it was rP16022 which also has some leftover concern apparently.

Thanks for the patch.

i like that, apply?

didnt test but why not putting this option to have it global and dont get confused maybe additional to know wut general auto attack behaviour is like in options -> ingame -> unit auto attack -> capture/attack. Maybe thats good to add. is it good?

Hey Angen,

Successful build - Chance fights ever on the side of the prudent.

Bring back the early returns;

Successful build - Chance fights ever on the side of the prudent.

In D1858#82303, @Stan wrote:

Haven't checked performance, but maybe that would be better to do data structures in C++ ? I'm sure boost or the stl have already such containers ?

Left a change in D274 incorrectly.

Successful build - Chance fights ever on the side of the prudent.

Un-nuke the commas.

Looks like you nuked all the semicolons...

Build failure - The Moirai have given mortals hearts that can endure.

Successful build - Chance fights ever on the side of the prudent.

Ran some more profiling, this seems comparable with svn. I think it's a tad slower as there's slightly more indirection, but that ought to be irrelevant in cost. MultiKeyMap could be switched to C++ I guess for best performance.

Fix formatting

When compiling with clang 8.0.1:

How is while (someconst) a violation but while (true) isn't?

Idea by minohaka
http://irclogs.wildfiregames.com/2019-07/2019-07-27-QuakeNet-%230ad-dev.log

20:44 < minohaka> Is there a shortcut for opening the diplomacy modal/window?

Successful build - Chance fights ever on the side of the prudent.

Rebased for commit.

Ran a few rejointests, all passing, which indicates that the serialisation keeps the deterministic non-linear order correctly, so this is good to go.

In D1252#90261, @Stan wrote:

@Mate-86 Sad to hear that but I hope you'll have fun with your new project. What is it called ?

Thanks for the fixes!

ok its working now but not as expected but i havent look deeply into code now just can say that one thread (main thread) is going when testing pathfinding with 6-16 threads with 1000-2000 units moving in a map very fast to 100% but the other threads are showing only small up usually 1-3% cpu usage some times 20% but very rare and though because main thread is at 100% exhausting and seem to not get more cpu time above 100% it starts lagging ofc and fps dropping even if non visual window (window hidden to taskbar). So i was expecting that the pathfinder work would be go well across all threads and especially not touching the main thread any more because there is also then graphics computation and other game logic stuff hitted which results again in fps droppage and lag game which should be avoided. So i think that would be the first step. then we have to look.

Not sure what was missing or too much, but it's fixed now.

Both concerns fixed.

Successful build - Chance fights ever on the side of the prudent.

Save the failing unit motion map in Atlas, so it probably can't be wrong now.

Successful build - Chance fights ever on the side of the prudent.

Successful build - Chance fights ever on the side of the prudent.