[go: up one dir, main page]

Skip to content

Sphereserver/Source-X

Repository files navigation

SphereServer X

Ultima Online game server, developed in C++.

GitHub License     GitHub Repo size     GitHub Stars
Coverity Scan Build Status     GitHub Issues

Join the SphereServer Discord channel!
Discord Shield

Releases

Core

Branch: Master
(most stable pre-releases)
Branch: Dev
(most recent, potentially unstable)
GitHub last commit on Master branch   Changelog GitHub last commit on Dev branch   Changelog
Nightly builds: Nightly builds:
Build status: Windows x86_64 Build status: Windows x86_64
Build status: Windows x86 Build status: Windows x86
Build status: Linux x86_64 Build status: Linux x86_64
Build status: Linux x86 Build status: Linux x86
Build status: MacOS x86_64 Build status: MacOS x86_64
Build status: MacOS ARM Build status: MacOS ARM

Click the badges or follow the links:

ScriptPack

The official script pack is fully compatible with X new syntax, has all the new X features and preserves legacy/classic systems, which can be activated back in place of the new ones.
It is currently being revamped to add original OSI features.
Beware, it's still not 100% complete!

Resources

Coming from a different SphereServer version?

  • From 0.56d? Here a list of major scripting changes!
  • From an older 0.55 or 0.56 version? This might help resuming major changes until 0.56d.

Why the fork?

This branch started in 2016 from a slow and radical rework of SphereServer 0.56d, while trying to preserve script compatibility with the starting branch. Though, something has changed script-wise, so we suggest to take a look here. Most notable changes (right now) are:

  • Bug fixes and heavy changing of some internal behaviours, with the aim to achieve truly better speed and stability;
  • Support for x86_64 (64 bits) and ARM architectures, Mac OSX builds, Clang and GCC Linux builds, MinGW compiler for Windows;
  • CMake is now the standard way to generate updated build and project files;
  • Switched from MySQL 5.x to MariaDB client;
  • Added (and still adding) comments to the code to make it more understandable;
  • Reorganization of directories and files, avoiding big files with thousands of lines;
  • Code refactoring, updating to most recent programming standards and to the conventions described below.

Running

Required libraries (Windows)

  • libmariadb.dll (MariaDB Client v10.*package), found in lib/bin/cpu_architecture/mariadb/libmariadb.dll

Required libraries (Linux)

  • MariaDB Client library. Get it from the following sources.

From MariaDB website

See https://mariadb.com/docs/skysql/connect/clients/mariadb-client/

Ubuntu and Debian repositories

Ubuntu: Enable "universe" repository: sudo add-apt-repository universe Install MariaDB client: sudo apt-get install mariadb-client or sudo apt-get install libmariadb3 (depends on the OS version)

CentOS - Red Hat Enterprise Linux - Fedora repositories

Then install MariaDB client via yum (CentOS or RH) or dnf (Fedora): mariadb-connector-c

Required libraries (MacOS)

  • Install MariaDB Client library via brew install mariadb-connector-c

Building

Generating the project files

The compilation of the code is possible only using recent compilers, since C++20 features are used: the newer the compiler, the better. Oldest compiler versions supporting C++20: Visual Studio 2019 version 16.11, GCC 8, MinGW distributions using GCC 8, Clang version 10.
You need to build Makefiles or Ninja files (and project files if you wish) with CMake for both Linux (GCC) and Windows (MSVC and MinGW).
Both 32 and 64 bits compilation are supported.
No pre-built project files included.
Does CMake give you an error? Ensure that you have Git installed, and if you are on Windows ensure also that the Git executable was added to the PATH environmental variable (you'll need to add it manually if you are using Git Desktop, here's a quick guide).

Toolchains and custom CMake variables

When generating project files, if you don't specify a toolchain, the CMake script will pick the native one as default.
How to set a toolchain:

  • Via CMake GUI: when configuring for the first time the project, choose "Specify toolchain file for cross-compiling", then on the next step you'll be allowed to select the toolchain file
  • Via CMake CLI (command line interface): pass the parameter -DCMAKE_TOOLCHAIN_FILE="..." When using Makefiles or Ninja, you can specify a build type by setting (also this via GUI or CLI) CMAKE_BUILD_TYPE="build", where build is Nightly, Debug or Release. If the build type was not set, by default the makefiles for all of the three build types are generated.

You can also add other compiler flags, like optimization flags, with the custom variables C_FLAGS_EXTRA and CXX_FLAGS_EXTRA.

Example of CMake CLI additional parameters:

-DC_FLAGS_EXTRA="-mtune=native" -DCXX_FLAGS_EXTRA="-mtune=native"

(Use the -mtune=native flag only if you are compiling on the same machine on which you will execute Sphere!)

Example to build makefiles on Linux for a 64 bits Nightly version, inside the "build" directory (run it inside the project's root folder):

mkdir build
cmake -DCMAKE_TOOLCHAIN_FILE=cmake/toolchains/Linux-GNU-x86_64.cmake -G "Unix Makefiles" -DCMAKE_BUILD_TYPE="Nightly" -B ./build -S ./

Compiling

Installing the required packages on Linux

Building will require more packages than the ones needed to run Sphere.

Ubuntu and Debian

Install these additional packages:

  • sudo apt-get install git cmake
  • MariaDB client: sudo apt-get install libmariadb-dev and libmariadb3 or mariadb-client (depends on the OS version)

If you are on a 64 bits architecture but you want to compile (or execute) a 32 bits binary, you will need to install MariaDB packages adding the postfix :i386 to each package name.

CentOS - Red Hat Enterprise Linux - Fedora

Then install these additional packages via yum (CentOS or RH) or dnf (Fedora): git gcc-c++ glibc-devel mariadb-connector-c mariadb-connector-c-devel

If you are on a 64 bits architecture but you want to compile (or execute) a 32 bits binary, you will need to install the appropriate gcc package and to install the MySQL packages adding the postfix .i686 to each package name.

Compiling on Linux

Just run the make command inside the build folder. You can pass the -jX argument (make -jX, where X is a number) to speed up the compilation and split the work between X threads.

Address Sanitizer and Undefined Behaviour Sanitizer

You can enable Address Sanitizer (ASan) and Undefined Behaviour Sanitizer (UBSan) with the ENABLE_SANITIZERS checkbox via the GUI, or via the CLI flag -DENABLE_SANITIZERS=true.
This is easier with GCC and Clang on Linux.
Since ASan redirects the error output to stderr, you can retrieve its output by launching sphere from cmd (Command Prompt) or shell with the following command: SphereSvrX64_nightly > Sphere_ASan_log.txt 2>&1

Coding Notes (add as you wish to standardize the coding for new contributors)

  • Make sure you can compile and run the program before pushing a commit.
  • Rebasing instead of pulling the project is a better practice to avoid unnecessary "merge branch master" commits.
  • Removing/Changing/Adding anything that was working in one way for years should be followed by an ini setting if the changes cannot be replicated from script to keep some backwards compatibility.
  • Comment your code, add informations about its logic. It's very important since it helps others to understand your work.
  • Be sure to use Sphere's custom datatypes and the string formatting macros described in src/common/datatypes.h.
  • When casting numeric data types, always prefer C-style casts, like (int), to C++ static_cast<int>().
  • Be wary that in SphereScript unsigned values does not exist, all numbers are considered signed, and that 64 bits integers meant to be printed to or retrieved by scripts should always be signed.
  • Don't use "long" except if you know why do you actually need it. Always prefer "int" or "llong". Use fixed width variables only for values that need to fit a limited range.
  • For strings, use pointers:
    to "char" for strings that should always have ASCII encoding;
    to "tchar" for strings that may be ASCII or Unicode, depending from compilation settings (more info in "datatypes.h");
    to "wchar" for string that should always have Unicode encoding.

Naming Conventions

These are meant to be applied to new code and, if there's some old code not following them, it would be nice to update it.

  • Pointer variables should have as first prefix "p".
  • Unsigned variables should have as first (or second to "p") prefix "u".
  • Boolean variables should have the prefix "f" (it stands for flag).
  • Classes need to have the first letter uppercase and the prefix "C".
  • Private or protected methods (functions) and members (variables) of a class or struct need to have the prefix "_". This is a new convention, the old one used the "m_" prefix for the members.
  • Constants (static const class members, to be preferred to preprocessor macros) should have the prefix "k".
  • After the prefix, the descriptive name should begin with an upper letter.

Variables meant to hold numerical values:

  • For char, short, int, long, llong, use the prefix: "i" (stands for integer).
  • For byte, word and dword use respectively the prefixes: "b", "w", "dw". Do not add the unsigned prefix.
  • For float and double, use the prefix: "r" (stands for real number).

Variables meant to hold characters (also strings):

  • For char, wchar, tchar use respectively the prefixes "c", "wc", "tc".
  • When handling strings, "lpstr", "lpcstr", "lpwstr", "lpcwstr", "lptstr", "lpctstr" data types are preferred aliases.
    You'll find a lot of "psz" prefixes for strings: the reason is that in the past Sphere coders wanted to be consistent with Microsoft's Hungarian Notation.
    The correct and up to date notation is "pc" for lpstr/lpcstr (which are respectively charand const char), "pwc" (wcharand const wchar), "ptc" for lptstr/lpctstr (tcharand const tchar).
    Use the "s" or "ps" (if pointer) when using CString or std::string. Always prefer CString over std::string, unless in your case there are obvious advantages for using the latter.
    Examples:
  • Class or Struct: "CChar".
  • Class internal variable, signed integer: "_iAmount".
  • Tchar pointer: "ptcName".
  • Dword: "dwUID".

Coding Style Conventions

  • Indent with spaces of size 4.

  • Use the Allman indentation style:
    while (x == y)
    {
        something();
        somethingelse();
    }

  • Even if a single statement follows the if/else/while... clauses, use the brackets:
    if (fTrue)
    {
        g_Log.EventWarn("True!\n");
    }

Licensing

Copyright 2024 SphereServer development team.

Licensed under the Apache License, Version 2.0 (the "License").
You may not use any file of this project except in compliance with the License.
You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0