Siege of Avalon lifted

- or an example of migrating old source code to newer grounds.

Compiled with latest Delphi (10.3.1 Rio) - Nice inventory graphics.

This post is just about the few steps I did on migrating the Delphi 4 source of Siege of Avalon to Delphi 10.3.1, and the intentions and road-map I have proposed to the players and modders of this nice game.


Intentions and road-map for my fork of the GitHub repository - as proposed on a German forum

1. Ensure that everyone can compile the original source with the newest free Delphi CE - which actually make it bearable to work with the source. DONE
2. Clean up the source - remove rot. And different separation, to help move code to cross-platform. WIP
3. Add the fixes, enhancements done by the great people who have gritted their teeth on the code previously.
4. Add documentation on the assets file-formats used.
5. Evaluate the tools - maybe redo/merge them - and include them in the repository.
6. Replacing DirectX - with what makes better sense (SDL2 would be easiest) - might include the DDraw fixes in code as first step (getting rid of DDraw.dll).
7. Cross-platform - that could be Windows, macOS, (Linux) and tablets (Android/iOS) - would need control overlays. Lowest priority since - clean code makes the rest easier.

Under the enhancements I refer to:

• HD/FullHD support - but I think that should be optional (also keeping the original 800x600) - so either add a "launcher" or re-organizing the "Options"
• Add a character/sex selector - again probably re-organizing the "Character Creator" dialog - Character in the center - male/female choices.
• Mod selection - more DLC style - extendable instead of destructive copy/paste exercises.
• Better multi-language support in the released source - right now it is crippled, like other parts of the UI code.

Backstory

Siege was open-sourced back in mid-2003, and at that time I grabbed the source with the intention to look into it some day.

Fast forward almost 16 years - and when Diablo was released on GOG.com (in corporation with Blizzard) two weeks ago - I remembered that I had the source laying about somewhere, found it and threw it on my private git server - where it should have been.

I then started to google about what had happened in the years gone by, finding that the proponent of the open-sourcing had put some of the original source, including an unfinished port to SDL and FPC on GitHub.

That had very recently been forked and picked up - compiled and working in Delphi 4 - with some else then adding support for higher resolutions screens, all based on the non-SDL code.

Process

So I thought that I could contribute to this project, I the way I would have liked it done - with focus on the code - so I did a fork of the Delphi 4 working repo - always building on the great shoulders of the people that came prior to us :)

Found my old CD of my Delphi 4 Enterprise - which in my opinion was the worst version of Delphi ever.

Ensured I could compile and play the game from that source - and the started to move the code to Delphi 7 and again Delphi 2007 until at last Delphi 10.3.1 Rio.

Apart from a bit of pointer fun and code clean-ups - it was fairly painless - the biggest pain was the changes that had happened in the Windows OS and DirectX area, breaking things and adding glitches - 2 days lost on that account :(.

So first goal achieved, the code is now able to compile in a free Delphi 10.3.1 Community Edition, to produce a fully working executable. Link to source code here.

There is still a lot of code cleanup to be done. The project is not trivial, and might take a bit of time - but it is just like eating an elephant - bit by bit - pun intended.

Why I did not try to fix the non-working FPC/SDL code

Order matters - both in the sense of keeping things tidy, but also in what order things are done to make it easier

• Code that compiles and works is funnier to work with.
• Based on working code you can refactor - otherwise it is bug fixing.
• Based on working code you can enhance and add features - otherwise it is useless.
• Getting clean working and accessible code invites others to contribute.
• By having more to contribute everyone wins - that is what it is about.
• Adding more platforms at this stages just adds more complexity - and more platforms to test and debug on.
• The working code was originally written in Delphi 4, so moving it to the latest "open/free" Community Edition is to me the natural choice - that is what the Community Edition is supposed to be used for.
• To me it is a very big difference to work in a Delphi 4/7/2007-like environment, and the latest modern version of Delphi.
• Spending my time effectively.

And why not just port the code to C#/Go/Java/Python/Unity?

• Well feel free, but to what purpose? Using the same assets and story, does not make it a better game - it could make it worse if the execution of the port is badly done. This game has sold worldwide - the original source is where to start.
• The game does not benefit from it - unless you rewrite the engine, modernize the assets, but then you end up with a different game.
• The engine is open-sourced, not the assets - so you can add new assets.

And should you not be aware; the Community Edition of Delphi - is free and feature-wise identical to the Professional edition which includes support for Windows, macOS, iOS and Android - you just need to keep your earnings with it below $5000 USD :) Get it here.

/enjoy