Descriptive Problem Summary:
I have alot of stuff on my map, and also alot of open multi-z enviroments that are rendered using vis_contents per individual open space turf, when a turf below open space tile is also an open space they form this recursive chain that lets us see on the lowest possible turf, and i use this alot, half of the map is like that.
It seems that perhaps i somehow hit a limit of how many things client can render? Which leads to these very weird and random artifacts all over the place.
Here's how it looks in motion: https://cdn.discordapp.com/attachments/382276327614447647/ 1360570196129218710/2025-04-12_13-59-50.mov?ex=6800df3c&is=6 7ff8dbc&hm=beb4cf19431b100873bbae9de8cea2c7853e4c320291b6982 2d471b0932e375e&
My first instinct was to confirm whenever its a problem specific to my code or is it byond that fails to do this, so i switched to 514, here is the video:
https://cdn.discordapp.com/attachments/382276327614447647/ 1360583397050679436/2025-04-12_14-52-24.mov?ex=6800eb88&is=6 7ff9a08&hm=d61909bf4cc219d95c57e7ae1e67c540299dc9758b3cce328 fad746e25dfc7cd&
As i thought it wasn't occuring on 514 byond.
P.S. Videos look like that because i turned off all plane_master logic to ensure its not my plane_master behaviour doing that.
Numbered Steps to Reproduce Problem:
I wasn't able to reproduce these artifacts on a clean project with just turfs and open spaces unfortunately, I would gladly help to assemble it if there are ideas on how to.
Expected Results:
Here's the part of the map with all vis contents turned off on the map:
https://media.discordapp.net/attachments/ 1228611598554370080/1362045551826309121/ image.png?ex=6800f745&is=67ffa5c5&hm=3d12a6647e17ead6f431386 052250be25733418f0df2f0c14e5f3ecd36e8840c&=&format=webp&qual ity=lossless
Actual Results:
Same part of the map but with all vis contents on the map turned on:
https://media.discordapp.net/attachments/ 1228611598554370080/1362045434259964055/ image.png?ex=6800f729&is=67ffa5a9&hm=8e8dd6b80f7d2dbc3f389eb 7d574ecf5b6b8ddf659bbb8c6a95184d5c0e2450f&=&format=webp&qual ity=lossless
Does the problem occur:
Every time? Or how often?
Consistently, every time.
In other games?
Can't confirm
In other user accounts?
Yes.
On other computers?
Yes.
When does the problem NOT occur?
The less turfs on the map use vis contents the less artifacts i can find on the map.
Did the problem NOT occur in any earlier versions? If so, what was the last version that worked? (Visit http://www.byond.com/download/build to download old versions for testing.)
I searched for the specific version this issue starts to occur at, and i found that 515.1615 version is the first one where the issue starts to occur, with nothing of sorts occuring before it. And yes, it still occurs on any 516 version.
Workarounds:
Increase of x size of the map shifts all visual artifacts to the right.
Increase of y size of the map shifts all visuals artifacts vertically up.
Increase of z size of the map makes artifacts go up a z-level.
Turning off vis contents on the map makes artifacts disappear.
Cutting amount of vis contents on the map by 1/5 made artifacts cease to appear as well.
Is this enough information or testbed is still critical to solving this?
What else can i help with in order to figure this out? Provided i am not yet able to reproduce it on a clean project?
ID:2969915
![]() Apr 16, 6:01 am
|
|||||||||||||
| |||||||||||||
![]() Apr 16, 10:41 am
|
|
Lummox JR changed status to 'Test case needed'
|
Without a test case, even if it's a full project, there's nothing I can do to look into this. If the project is compile-ready for 516 and I can fire it up in the debugger, and you have instructions that can get me to a specific place the issues start to appear, I can take a look at that.
I will however once again bang the drum on two things: 1) Putting the turf below each open space in that space's visual contents is the wrong way to approach multi-Z views; it will miss turfs that are just off to the side and it's also wildly inefficient. This needs to be done with blocks of turfs. 2) The time to move your codebase to 516, at least for 516-readiness, was last summer during the alpha, or at latest during the beta. 515 is over. |
I have spent hours on end to try to replicate this in another project, but its not happening.
However, i have found that removing vis_hide on my lighting objects stops this bug from occuring in my project, i already found which version this starts at, namely 515.1615. I think the code that made "The server didn't send appearance updates for objects in visual contents whose VIS_HIDE flag was turned on. (Cadis E. Di Raizel)", is the code that makes what i experience happen. What i also found that running server on version that is lower than 1615, and client joining to the server with the version that is equal or higher to 1615, client does not experience the bug on his screen. Here's the demonstration: https://cdn.discordapp.com/attachments/1228611598554370080/ 1362151292751053021/2025-04-16_22-42-25.mov?ex=680159bf&is=6 800083f&hm=c33ccfe68ee0805adf933a341687a38a935aa4f0680807342 26cf009c9972dc7& I think this has something to do with objects in vis_contents that have vis_hide and that are displayed recursively, but that's my conjecture. I am still trying to get the bug to occur on the testbed though. |
Login to reply.