Not his fault something is wrong with premake
Thu, Nov 24
Tested on GLES.
Can't test ARB (causes assertion)
Can't test high quality water (wrong qualifiers)
Normal water looks okay
Silhouettes seemed to work
MSAA FXAA and stuff don't seem to do anything there.
Wiki Alpha26 => Gameplay.
Maybe cause you ran it without a template with a different root than entity?
Also not sure what happen if you run 0 A.D. with -dumpSchema which allows to validate templates ?
There a few template validator that rely on the relaxng file generated by that?
Maybe because of the order of templates?
Wed, Nov 23
Oh, yeah indeed, nowadays I use another way. Could be fixed indeed.
Problem is in my case I have many VS versions, so I need to be specific.
Tue, Nov 22
Probably have to teach checkrefs about it.
Well that's my point it doesn't have to. The only constraint should be playerdefaults >= gaia + max num of players we can put an ensure for that would that allow me to commit it ?
Can you be more specific ? I can't think of case where it happens. max player count should always be 8 + gaia in all cases
Mon, Nov 21
Looks good to me might be split in two commits.
C# uses .First() or .FirstOrDefault() (former throws an exception if element is not found)
Correct patch, still will fail on VS2017
Indeed failure is not linked to you. Touching glad usually breaks the CI sadly.
Sun, Nov 20
Yeah that's what I meant :)
May be nice to update glad headers with 2.02 to reduce diff size i think glad unused was added in there. No strong feeling
Tue, Nov 15
Mon, Nov 14
@Mercury thoughts on this?
Hey, sorry for the very delayed answer I was on holidays.
Sun, Nov 6
Could have an ENSURE that checks that player defaults contains at least 9 (8 + gaia) players.
I don't see how iterating from 1 to 8 like it's done everywhere in the code means a 6 months FF?
We could just use an extra check there or iterate players instead in another patch?
Just like the atlas preview it just reveals a bug that's been there for a while.
The patch allows to select "Player 9" and place an entity in Atlas which triggers an error.
You can already do that (just use numpad). Not linked to this. Would be worrysome if you could add up to 16 players. Which is not the case.
It is hard because animation variants contain tools props. You could restrict mesh and textures nodes but then you can't have bloodied corpses. Which you should override I suppose.
Sat, Nov 5
Indeed but we have checkrefs for that case :)
Well yeah kinda the point, you group things to update at the same time.
How does this make it easier to add more variation? If anything it reduces duplication and makes it easier to ensure consistency in actors. Also makes diff easier to review since less files get touched
That's an interesting patch. Don't forget the XML headers in the group files.
Oct 27 2022
@Mercur, @wraitii maybe you should find a time to discuss this live, on irc, matriw discord or whatever. It seems that it might be more productive. What do you think
Oct 24 2022
This looks clean, but I'm not sure if it has performance implications or anything thereof ? Does it change anything? Or are you just fixing my code now that we can presumably use std::optional because we upgraded Xcode since I wrote my code
Seems to be the case indeed.
Oct 23 2022
Is there any advantage of having an index attribute rather than relying on the order of definition? I guess it would allow having similarly named tracks grouped while not having them played like that, but is that worth the tradeoff? An attribute also makes it abundantly clear to someone not familiar with this code.
All indexes seem to follow the array index anyway, except one which I presume is a test case.
Maybe you could fabricate a replay and use @mbusy network's savegame feature? Or just tweak the c++ code to do it for you. Wonder if it can be done through the RL interface.
Oct 22 2022
I think we should be a bit smarter than this. What your patch does is break the randomness because tracks are only shuffled once.
Ah right I didn't think about std::max for list. I suppose we have to parse the whole list anyway whether it's to sort it to be fast after, or to determine which element is the max. It would be nice to be able to stop after the first 3 has been detected though.
Can you fix the lint warnings as well?
Maybe we could use a function where we pass CTexture::** to reduce duplication?
IIRC there are some broken network tests that have been disabled maybe you could use them to test after fixing them. Since you are on Linux you could also open two instances of the game to test what @elexis said.
Oct 21 2022
Missed a rebase maybe?
Still one step too low
Your patch was generated in source instead of root :)
Oct 20 2022
It's supposed to look artificial to be "balanced"
Oct 19 2022
Oct 18 2022
BTW looks like you broke the tests.
Is #5614 related to this?
Oct 17 2022
Oct 16 2022
Hey thanks for the patch. Any idea why it was there in the first place? Maybe it prevents the background from being the same color as the rest aka gray?
Might notice it on the very slow hardware running the game.
Oct 15 2022
Hey thanks for the patch. @Samulis might have info about the actual order we need but it can be done at a later date.
Oct 14 2022
Can you ask Fatton how he'd like to be credited ?
Oct 13 2022
Oct 11 2022
Not sure why you need to remove the variant from the actors but not the actual variant?
Oct 8 2022
Shouldn't the description be use vec4 + vec2 instead of mat2?