Arch is using it

• https://github.com/archlinux/svntogit-community/commits/packages/0ad/trunk
• https://github.com/archlinux/svntogit-community/commit/5b5c066a9e9d04ebcdaa2c91223d10ca20115c3c

We merged the other GL loader.

@vladislavbelov asked me to clarify more things before committing this, because it might lead to a 6 months Feature Freeze or a painful revert.
We need to make sure it doesn't cause more spikes in other cases.

In D4758#203957, @Stan wrote:

I the meantime can you keep testing it with different loads and replays to make sure there isn't something unstable ? :)

EG. The custom combat huge demo or a 4v4 with ais on big maps etc.

I the meantime can you keep testing it with different loads and replays to make sure there isn't something unstable ? :)

Nah, this looks alright, will see about committing it after the release comes out.

Sorry about that, ZFS keeps failing these days, and the VM gets stuck.

Looks good to go. Wonder if they should have a copyright header. According to https://trac.wildfiregames.com/browser/ps/trunk/source/tools/LICENSE.txt it's MIT.

Just realised I'm an idiot and you need a way to specify the location because you don't want to put the rng in the mod folder (since you're probably gonna zip it afterwards) and it won't be there on the CI either.
Sorry, that's what happens when I do too much things at once.

Sadly I'm not sure I can set you as author. There seem to be a way but I dunno if trac and phab support it https://stackoverflow.com/questions/2107719/how-to-set-author-for-svn#2107898

Fix year and LGTM.

Maybe try this flag https://github.com/7wells/0ad/blob/master/binaries/system/readme.txt#L75

Can't you make it the responsibility of the receiver to destroy objects instead of adding the extra overhead of a unique ptr?

Windows failed because you broke the debug tests. Possibly a memory leak.

Updated version for A26.

Ok then go ahead the sooner the better.

I am in a bit of a pickle. Should this be committed regardless of string freeze? @Freagarach, @lyv, @wraitii, @bb

Ah nvm the last inline then ^^

In D4777#203582, @andy5995 wrote:

@Stan Well... who gives a shit what the ticket was about... if there's a better way to implement the ultimate goal of the ticket.

In D4777#203574, @andy5995 wrote:

I was thinking something like

wx-config --libs
if [ $? = 0 ]; then
  echo message
  exit 1
fi

@andy5995 It's not possible, premake does not build FCOLLADA. FCOLLADA, SPIDERMONKEY, and NVTT are built first then you build PREMAKE, then you run the script above, and then you run MAKE. You don't need wx-config for those.

Update. Example output:

Thanks. Looks okay.

Make sure to fix the lint inlines too

Does it by any chance fixes the bug reported here https://wildfiregames.com/forum/topic/85888-german-article-about-0-ad/?do=findComment&comment=508304 by Gurken Khan?

cc @Langbart @Mercury @jprahman this should hopefully be the last c++ patch for A26

There are some disabled network tests. They are disabled because they are broken but maybe they could be fixed. Easiest way to test is running multiple instances of the game on Linux, or on separate machines.

Ah indeed data() works, I tested so many things and forgot that one ^^

Changes look sane and were tested as part of the latest RC.

@phosit it seems one cannot cast from const_iterator aka _Array_const_iterator to T* any clue?

Before the patch

Don't forget #include <string_view>. :)

Maybe it would make more sense to do it in the ResourceSupply component, like the damage levels used for the Carthaginian buildings.

Congrats on uploading your first patch! I have added you to the Contributors group so the next time you upload your patch, the continuous integration will run. Please provide context for your patch by adding the -U5000 command line argument to your git diff command.

Mmmh won't that break all the balancing by changing DPS?
Also gotta be pretty confident to do random in the core of the sim should be tested in MP

numberOfWorkers < MIN_THREADS is perfectly fine.

Should be:

I think I've found all that a normal analysis can find now.

Could you answer @vladislavbelov inline @edoput? Also could you provide a test mod that uses that feature?

@Emperior said he found those strong too.

Perfect, thank you guys, sorry for the waste of time :)

Gotta have to check those includes, lots are missing. Can we do without boost? We're trying to reduce the dependency on it where possible.

Some unsigned / signed mismatches. There might be more. Sorry for all the nitpicks.