Friday, 9 August 2019

Last & Furious

In November 2017, Ivan Mogilko and Jim Reed created Last & Furious for a MAGS competition.  Unusually for AGS, it was a top-down racer in the mould of SuperCars II.  It went on to win a handful of annual AGS awards as it spoke to the versatility of that engine.  

Fortunately, they also shared the source code on github.  A previous attempt at porting it to XAGE hadn't gotten very far, as it relied heavily on functionality I hadn't yet implemented (DynamicSprites, new audio mechanism, scripted keyboard handling etc). 

As work continues on [REDACTED], I periodically return to old ports as a palette-cleanser and also to see how far the conversion tools and engine have progressed.  After a few weeks of late nights (with some additional pointers from Ivan), I'm happy to say the port is complete and now available on itch.io.



Download link:  https://clarvalon.itch.io/lastandfurious  (40MB, Win x64)

This is using the latest version of the XAGE engine, running on a standalone deployment of .NET Core 3.  It is not using a CoreRT build as the .NET foundation have recently decided to instead focus on a different AOT approach for .NET 5 once it launches in 2020.  

Performance seems decent - on my 6 year old laptop it averages at about 5% utilisation for both CPU and GPU while in the middle of a race.  By comparison, the AGS version averages about 20% for CPU.  This is likely because it is using a software renderer whereas XAGE is able to offload more work onto the GPU.  I also optimised a handful of things along the way, so not exactly a fair comparison.  Where AGS still trumps XAGE comprehensively is memory usage.

The conversion process was 95% automated.  The C# generated should look very familiar to anyone comfortable with AGS Script:


There were a few manual tweaks to:
  • Correct some AGS variable type inconsistencies (as with every port).
  • Simplify some of the walkable areas, here used determine obstacles and how the car controls on various surfaces.
  • Suppress pathfinding (as everything is handled manually).
  • Change the scope of a handful of arrays to reduce the amount of garbage being generated (the game still generates a lot, but mainly Gen0, and not enough for it to affect performance).
  • Prevent new DynamicSprites being created for every single rotation - instead let the GPU handle it.
  • Force a custom texture to be stored locally (to speed up GetPixel calls).
I left in the debug hooks so pressing the usual console key (`) will open the ImGui debug console.  It should otherwise look and feel identical to the original AGS version.

As I've been working primarily on something behind the scenes, it's nice to be able to show how general engine development is progressing!

Sunday, 3 March 2019

Experimenting with CoreRT

I hope you like acronyms.

As work on porting [REDACTED] continues, I've been putting more thought into distribution.  Now that XAGE uses .NET Core, we have the option to publish 'standalone' executables, which means there is no requirement to have a version of the .NET framework installed on the end user's machine.  This is a good thing, though results in a version of .NET Core being bundled in with the final executable, which at the moment means lots of individual files and a much larger overall size.

There is a mechanism being developed to allow these individual files to be essentially be bundled together and extracted to a temporary location at runtime, though I'm not convinced that this is the best approach.

A more interesting prospect is leveraging CoreRT in order to compile everything ahead-of-time (AOT) into a native executable.  This comes with some benefits:

  • The final standalone executable is much smaller and completely self-contained.
  • The startup time is much quicker as no just-in-time (JIT) compilation is required.
  • General performance is better.
  • Decompilation of the code is much harder with native executables compared to .NET. 

At the cost of a few downsides:

  • The process of the AOT compilation & linking to remove unused code means that you need to provide some details on items not to remove in the form of an XML file.
  • One of FNA's selling points is that you can package your game with MonoKickstart in such a way that it will work for x86 and x64 Windows, Mac and Linux.  Using CoreRT will mean that separate packages need to be prepared per platform. 

The CoreRT project is still in early preview, but has made a lot of progress in the last few years.  In the process of trying to get it work with XAGE, I came across a major stumbling block:  CoreRT does not currently support XmlSerialization and Protobuf-net - both of which are serialization techniques used by XAGE.  The reason for this is that CoreRT does not support Reflection.Emit, which generates Intermediate Language (IL) which is JITed into native code.

There are some changes on the way that should make both viable with CoreRT, but in the meantime I wanted to explore some different serialization techniques, which is when I discovered Biser.  This little-known library allows you to generate serialization C# classes AOT that you can include in your final executables or libraries, meaning no run-time reflection is required (not dissimilar to something I'd first tried back in 2010).  I was able to rewrite this and include it in the XAGE workflow such that it could be used as the single serialization method within the engine runtime:


And voilĂ , we have our minimum viable game runtime - here for Windows x64:


Where the executable is under 24MB and there are only 11 files in total:

  • Four content files specifically for the XAGE game (game.zip, graphics.zip, audio.zip and voice.zip - the latter two of which are optional)
  • Six native .dlls in the x64 directory - five required for FNA (FnaLibs) and one for Dear ImGui, which you probably wouldn't need as part of the Release build anyway.
  • Strictly speaking I should also include FNA.dll.config, but it seems to work fine without.

I like this a lot, as we get the bleeding edge functionality & performance from .NET Core while also producing neat and tidy native distributables.

I'm not sure whether I'll stick with my fork of Biser (unimaginatively named Xiser) or return to Protobuf-net once it matures further and is able to support CoreRT.  I have a lot more confidence in Marc Gravell's ability to maintain a robust & feature rich serializer than my own.  But it's good to have options.

Tuesday, 1 January 2019

The state of things

Last march I committed to releasing XAGE publicly in 2018.  This hasn't happened, but for a good reason.

For most of 2018 I've been working on a port of an upcoming commercial game, details of which are under NDA.  The game itself represents a scale & complexity I've not worked with previously, and has helped focus attention on the areas that most needed work.  At a high level:

  • Several months were spent improving the AGS code parsing so that the converter could, in this case, automatically produce 36,000 lines of equivalent C#.
  • Time was spent further optimising the Editor and build tools to handle the large volume of game items and assets, in such a way that re-ports can still be iteratively performed quickly.
  • Font handling was rewritten to convert .ttf, .sci and .wfn formats to XNA-style spritefonts.
  • Support for 'new-style' AudioClips & AudioChannels was introduced.  DynamicSprites and DrawingSurfaces were also partially implemented.
  • AGS scripting coverage jumped from 39% to 64% for properties and 50% to 62% for methods.
  • Engine script-threading was re-written to more closely align it with AGS (i.e. Global & Room threads) and with some events being queued rather than run immediately.
  • The AGS Exporter plugin now optionally uses Potrace (for vectorising walkable areas) and ffmpeg (for converting all audio to .ogg).  
  • The runtime now uses the latest version of FNA with FAudio handling the .ogg files.
  • A runtime debug console was added using Dear ImGui & ImGui.Net, with a custom inspector and logger added to help with debugging.

There have been 745 commits to the XAGE source code repository in 2018.  Counting commits is not the best metric, but it is a metric, and hopefully one that gives some indication that this project is very much a going concern.

Seeing the port of [REDACTED] progress from thousands of compile errors to a close facsimile of the AGS version has been very rewarding.  It's not quite there yet - there is still plenty to do - but there have been times recently when debugging where I've forgotten whether I'm playing the AGS or XAGE version of the game.  If the same rate of change can be maintained then with luck you'll be able to play [REDACTED] on Xbox One and other fancy platforms sometime in 2019. 

Wednesday, 14 March 2018

AGS Script to C#

I've never been out on New Year's Eve.  It always seemed noisy and expensive, and I've never been around people who were inclined to celebrate, so it stands to reason that I was working on XAGE on the evening of December 31st, 2017.

I'd posted some progress pics and got chatting to Scavenger, who kindly sent me the source code to his AGS game Terror of the Vampire.  After poking around for a few hours I was able to convert the bulk of the game to XAGE, albeit with a worryingly large 10K+ of C# scripting errors.

At high volume the number becomes somewhat meaningless - a missing bracket can create hundreds of syntax errors in a single class - but it was clear this was a complex game that came with a lot of challenges:

  • Massive AGS Modules like Tween were used which were not being parsed well.
  • AGS Engine plugins were used and unsupported.
  • AGS-style structs were used throughout, which were not being parsed and converted into equivalent C# classes.
  • There were gaps in the scripting API, mainly due to newer AGS functionality as part of version 3.2 - things like AudioClip and AudioChannel.

By looking at the error patterns in Visual Studio, I was able to start whittling the error count down by improving the parsing, handling new data structures, implementing API placeholders and some workarounds for various AGS quirks and oddities (especially around scope).  Two and a half months later and we're down to 81 errors following a fresh conversion.

The law of diminishing returns dictates that trying to automatically resolve these would involve a big time investment that is better spent elsewhere, given they are generally simple to solve by hand, and can often be fixed beforehand in the AGS source.  They mostly revolve around C# being somewhat stricter than AGS code:

  • Casting.  C# doesn't like to implicitly cast certain types, like a bool or char into an integer.  AGS is rather lax about typing.  The solution is to add an explicit cast. 
  • Function integrity.  AGS allows you to define a function with a return type and not actually return anything.  C# does not ("not all code paths return a value").  Solution is to fix the return values.
  • Method variable scope.  AGS allows you to define variables within inner blocks that share the same name as a variable in the outer block.  C# does not.  Solution is to fix the scope or rename the variables.

These are all simple things to fix manually, meaning we are now able to fully compile the C# solution and get in-game:



As can be seen from the screenshots, there is a lot of engine functionality that has yet to be implemented or isn't quite right, but this is where the real work begins.  I will most likely start with DynamicSprite and DrawingSurface as these are all throwing null exceptions currently and are used throughout this game.  This is where I finally start to flesh out the Engine API and turn some of those table cells green.

A big caveat is that AGS Engine plugins still remain unsupported - for the time being I added a few placeholder classes to act as a dummy implementation to allow the code to compile.  It may be that we can autogenerate C# bindings to the original C++ plugin .dll, but the simpler option may be to simply re-implement each required plugin in C# and keep everything in managed code.

Other items

  • I've foolishly committed to releasing XAGE in some form or other this calendar year.
  • There has been some promising experimentation around embedding VS Code directly into XAGE Editor for a more seamless (if diluted) user experience.
  • PUBG is still causing me an enormous amount of pain.

Saturday, 14 October 2017

The Ever-Changing Horizon

As of a few days ago, XAGE runs on .NET core.  What does that mean?  Let's have a quick and probably inaccurate history lesson.

.NET Core

In 2014 Microsoft announced they were open-sourcing .NET.  Their new direction involved splitting the framework, from the old monolithic windows-only .NET Frameworks and introducing the new fancy, cross-platform .NET Core.  While Microsoft committed to updating both, it was clear that they saw .NET Core as the future.

For a while everything was confusing.  The command line tools didn't work very well and were constantly being changed.  Microsoft toyed with using a new project.json file to replace the old xml-based *.csproj files, only to revert back again.  The documentation was initially poor and often out of date with each iteration.

With the release of NET Standard 2.0, things are looking considerably better.  The framework API now covers a much larger set of the existing functionality.  All key platforms can target this standard i.e. Windows, Mac, Linux, iOS & Android (via Xamarin) and UWP.  The tooling is now better, with support built directly into Visual Studio, by far the best IDE for development.

XNA, MonoGame & FNA

In 2006 Microsoft first released XNA, a simple framework built on top of .NET for making games.  The main draw was the ability to write games in C# and to target the Xbox 360 console.

A few years later, an open source implementation of XNA specifically for iOS appeared called XNATouch.  This later became MonoGame as the number of supported platforms grew and more developers became involved.

By 2013 Microsoft finally retired XNA.  The writing had been on the fall for a while, though its abandonment upset a great many hobbyist developers who had invested a lot of time and energy building their games upon it.  MonoGame and FNA stepped up to fill the gap.

FNA was a fork of MonoGame by Ethan Lee, using an SDL2 renderer instead of the unsupported OpenTK as used (then) by MonoGame.  One impressive aspect on FNA was its laser-focus on accuracy, which shows in the quality of the Linux and Mac ports it has produced (e.g. Fez, Roque Legacy).

XAGE

At various points, XAGE has been built upon each of the above, including other implementations not mentioned here (e.g. Silversprite).  At the time of writing, MonoGame is used for its content building tools and handheld platforms.  FNA is used for the main desktop platforms.

With the recent support for .NET Core and potential plans for additional platform support (given the ubiquity of SDL2), the case for using FNA is strengthened further, though MonoGame remains an important and useful project.

Being able to use .NET Core solves distribution problems (no need to install the .NET framework, no complicated or brittle Mono binaries).  It also feels good to be back on the bleeding edge, which is a far cry from all the anxiety around XNA's protracted demise.

There remain some blockers for making the switch fully.  One being that XAGE uses Protobuild for creating all the relevant Visual Studio solutions and projects for the user.  Protobuild does not currently support .NET Core.

Regardless, we're in a pretty good position right now.  The big graphics rewrite started a year ago turned into a larger refactoring project that is resulting in a saner, more manageable codebase.  Once a few more reliability and usability milestones have been hit, we'll be looking to put out a release into the wild.

Wednesday, 15 March 2017

Xbox One Support

Since starting work on the big refactoring piece I haven't paid too much attention to broadening platform support. However, since Microsoft announced the Xbox Live Creators Program (essentially a replacement for the Xbox 360's Xbox Live Indie Games), I was curious to see how much effort it would be to get XAGE up and running on the Universal Windows Platform. It turned out to be this: two evenings, thanks to MonoGame v3.6.  Using the latest version of Protobuild, XAGE Editor will now automatically create a Visual Studio solution for UWP:


Some of the changes to the engine could be done without breaking anything, though some required introducing more preprocessor soup (#if #else #endif) due to breaking framework changes. Ideally once dotnet core matures then we can move towards NETStandard libraries for all platforms.
  • UWP forces all file IO to be asynchronous, so the simplest solution was to use a synchronous wrapper specifically for the WindowsUniversal libraries. 
  • Minor API changes were needed (e.g. replacing Delegate.CreateDelegate with MethodInfo.CreateDelegate, generic use of SystemException with System.Exception).
  • Replace all Stream.Close() with Stream.Dispose() where a using block is not used for automatic cleanup. 
  • Remove all errant use of System.Console.Write() for debugging.
  • Identifying the main UI thread using Task.CurrentId instead of Thread.CurrentThread.ManagedThreadId (access to threads is not exposed in UWP).
  • Inserting GetTypeInfo() into various reflection method calls and properties due to a very UWP specific workaround.
The result of this work is that XAGE now tentatively supports the Windows 10 Store and the Xbox One family of consoles. I've updated the list of supported platforms accordingly.


Not everything is quite implemented yet (serialization, audio etc) but I'll return to finish this up once the main refactoring is complete.

Sunday, 20 November 2016

Interfaces vs Concrete Wrapper Classes

Software development can often be very rewarding, with interesting problems to solve and little pockets of endorphin releases as a reward.  It can also be a pretty tedious drudgery of box ticking with little to no fanfare.  The last few weeks have mostly been the latter.

Previously, when working on game scripts, a developer would be using C# interfaces defined within a standard interface library (XAGE.Interface.dll), which is how the scripts are de-coupled from the engine runtime.  The .NET naming convention for interfaces is to use an 'I' prefix e.g. ICharacter.

However, it's not possible to define a static method or property as part of an interface (which, when you think about it, wouldn't make sense anyway), so these were instead implemented as part of a second class e.g. Character.  This leads to some confusing and inconsistent typing in game scripts, e.g.:

ICharacter myCharacter = Character.GetAtScreenXY(mouse.x, mouse.y);

To work around this, I've created a concrete wrapper class for each interface:

// Instance of interface injected into concrete wrapper class
private ICharacter engineCharacter;

// Properties
public int x
{
    get
    {
        return engineCharacter.x;
    }
    set
    {
        engineCharacter.x = value;
    }
}

// Methods
public void SayAt(int x, int y, int width, string message)
{
    engineCharacter.SayAt(x, y, width, message);
}

Creating these wrapper classes has not been fun, even though it was partially automated, as there were of hundreds and methods and properties to wrap.  It does however have the following benefits:

  • The developer only needs to know about the single concrete class instead of the interface.  The newly styled documentation is also more straightforward as a result.  This will be fleshed out in future with examples and annotations.
  • Any discrepancies between the interface class and AGS scripts can be hidden as part of the wrapper class.  The main one being how AGS primarily uses integer IDs compared to XAGE's string IDs.
  • Automatic AGS conversions now no longer need clumsy string swaps for certain types.

There are some downsides however.  There is a tiny performance overhead in some instances, and some additional complexity when maintaining lists of objects in both the script and the engine.  These are outweighed by the benefits however.

So the last few weeks of spare evenings have been pretty dull, much like this blog post, but it's another small step in the right direction.  Next up:  Re-wiring and profiling.