Posts

Just a few weeks after the last update here I’m again to write the next one in which I’m going to show you the performance of my new hash table implementation.

After all the work carried out during the last months and culminated in the previous update, I took some time to experiment with some different things. I wanted to leave the rendering side for a bit and take a look at how to optimize other parts of the engine.

During those months two very important features appeared in the nCine. The first one is the integration of Lua for scripting, a language which is very easy to integrate and runs very fast. With Lua the user can quickly prototype ideas or actually write the entire game with just scripts, in a way similar to other engines.

Plenty of work has been done during these few last months as I decided once again to update the engine renderer. This time it went from using OpenGL 2 and OpenGL ES 2 to 3.3 and 3.0 respectively.

Are you curious about the time I spend to compile the nCine? 😄 In this post I’m going to show how much time is needed to compile the engine on different platforms.

A lot of work has been carried out during the last three and a half months. First of all I have added some macros to allow asserts in the code, think about checks like this: ASSERT_MSG_X(index < size_, "Index %u is out of bounds (size: %u)", index, size_); The macro also invokes a breakpoint if a debugger is connected, allowing to inspect the context around the failure. It was a long needed feature that I finally had the time to implement. 😄

During those three months I have been working on two big features. The first one has been the support of SDL2 GameController mapping format. Initially my plan was to build a layer on top of my joystick input functions and leave the mapping code outside, as helper functions in a file distributed along the source of my tests and only linked by them. Later on I decided to refactor everything and bring the code inside the engine, in an effort to make it easier for an application to transparently use it.

During June and July 2017 I have been working as usual, in my spare time, on the project. 😉 The first big June addition has been the automatic screen culling of sprites, a very important feature needed in order to support games extending on multiple screens. The culling works on sprites of any kind (regular ones, particles, text nodes) and regardless of their scaling or rotation parameters. If those sprites are completely outside of the screen they will just not be rendered, saving draw calls from being issued.

I am writing this to present my latest and biggest project to date, the nCine. I was originally going to publish an article the day that I was going to release the source code but for various reasons you will have to wait a bit more. 😢