What I did today

Miscellaneous Topics
#5881

I have been working on this game template still. I don’t want to put it in WIP just yet because it’s just a template and not a game yet, but here’s what I’ve done. This update is for the particle emitter (which is supposed to look like fire), tile collisions with the player, the day / night cycle, and a half assed attempt at lighting (which is just a picture at the moment (sorry @orangepascal for using your light picture from a tutorial you made, but I wont use it as the final)). And as always if anyone wants to look at my code and tell me what I’m doing wrong I would be happy for that :slight_smile: https://github.com/IanFell/GameTemplate

Here’s the video:

NrU-sRw0xKI

#5882

So… today I successfully ported my first game to Light Engine, took a while and some stuff is broken(physics) but it works. The Light Engine base code was from that game, although the api remain mostly the same since my last commit which made it a very straightforward process (mostly package renaming) but I had to made some changes to the engine to make it work with the old rendering system. I found out that having a somewhat large abstraction layer made it easier because the game specific code is not affected by changes to the engine, even with the changes to the threading system (in my other reply) it work without issues.

#5883

Today I’ve posted my first WIP :smiley:

Just after adding first boss to my game

SNdl7ngevng

For now I have to refactor ai-tree tasks because are not modular enough which causes a lot of duplication. Next thing that I want to do is taking a look at some lighting system.

#5884

It was easy :smiley:

qiGblDDQXD4

#5885

Wow! That really looks great dude. Good job.

#5886

Thank you! It means a lot for me :slight_smile:

#5887

Been really into machine learning lately. Cooked up a little library that works (mostly) but still needs a lot of work. Here are some of the results.

edit: also if anybody could suggest some other interesting projects I could do I would appreciate it because I’m kinda out of ideas

Classic classification problem. The network will separate the two blocks no matter their position.

1276×718 13 KB

A more complicated problem but the network handles it well I think.

1276×718 23.9 KB

Another complicated example. The network actually freaks out after a while on this one if left alone. Not quite sure why that happens. ???

1276×718 44 KB

And finally, a program that tells asks whether or not an image is of a “6” or a “g”. An image that is a g has a value of -1 while an image of a 6 has a value of 1.

1276×718 5.19 KB

#5888

I started work on my magnum opus: a 100-player Battle Royal game. Brace yourself for its awesomeness:-

CI76GNMWJG4

#5889

Today I’ve added first NPC to my game along with talking with him. I am not alone in this world anymore :smiley:

j_567iE_eNg

#5890

Today I’ve added jump mechanics to my game. I wanted to make it more interesting with the abyss floors where you can just die by falling down when you lose your focus. I think that it went well but it was occupied with a lot of pain with box2d collision detection. I am glad that it works now :slight_smile: Later I will think about better “falling down” animation. It was an exhausting day.

d4xHcWSGkHE

#5891

Today I was thinking about dungeon instances in my game. To avoid players overflow- my first solution was to scale dungeon size along with number of players starting it. But I’ve found it hard to implement in my game case without causing game experience drop.

So I’ve decided to create dungeon per team instances. Now main question was: how to do it?

I was thinking about tickets (items that teleports player to the specific instance on use) distributed by some NPC. But for now I’ve dropped this solution.

What I did is: special room with few fields where player can stand and if there are any players on these fields, they will be teleported to the dedicated, newly generated dungeon instance. Teleportation is cyclic based (npc is telling when is the next departure), so players has to form a queue to play in their own teams.

I didn’t want to disallow player joining the dungeon if he doesn’t have enough party members, so I’ll need to scale difficulty level along with number of players.

It was a little tricky one, but I am fully satisfied with the result :slight_smile:

Video:

Ikh-4BxH16s

#5892

Implemented a basic SVG parser in to my basic vector engine which has now become more mature in itself. The SVG parser only parses the path elements so far but I can scale these and color them in independently. One thing I have learnt from this is the ludicrous design of the path element. With out doing some complicated maths its really hard to know whats a hole and what just another shape near by. If the shapes are correctly done using CW / CCW ordering I can parse them correctly but there is no rule that say you need to do this. Instead of building something nice in to the specification you just get the below which means a lot of calculation

https://www.w3.org/TR/SVG2/painting.html#SpecifyingFillPaint

#5893

@abcdef

I wrote an SVG parser that parses the specification and the paths classes are (I think) correctly implemented to handle all SVG path elements.

#5894

Thanks, I couldn’t see any rendering code (I can parse the data ok)? You have done a lot more parsing of the SVG spec than I have :smiley: I’m currently in search of an algorithm for easily doing triangulation based on the non-zero and even-odd windings (my current method just does non intersecting simple shape triangulation, I might need to look in to re writing what I have and adapting it)

#5895

I added the calculations for rotation along a path. All that’s left is incorporating my SVG font API to calculate the bounds of text elements and my parser should fully implement ~90% of the SVG 1.1 specification :slight_smile:

#5896

I finally got a basic phong shader working with a hard coded light position. Yea, i did this 20 years ago and it was easy. This time round it took weeks! So much for older wiser smarter. Why im doing this is a very very long story.


[/url]

#5897

Moved SilenceEngine forum to Flarum, and it is so beautiful!

https://silenceengine.goharsha.com/forum/

However I decided for it to be a fresh start because the original SMF version attracted a ton of spam.

#5898

Replacing my voxel cone tracing gi with a spherical harmonics based one, first working version uses ambient cubes though :slight_smile: currently only ambient based

800×600 74.7 KB

Edit: made smaller, sorry

#5899

thats literality a screenshot ;D

i can see every sneese :persecutioncomplex:

#5900

Haha, that may be my cat’s sneeze, she doesn’t like it when i work and that’s her way to tell me ;D