[[HLF]Southclaw]

Pawn Community Compiler



What

This is a modified version of the Pawn 3.2.3664 compiler with many bug fixes and enhancements.

This project was originally started by Zeex but on 2017-12-31, the project was taken over by some members of the SA:MP community. Zeex still contributes to the project, along with the Compiler Team.

The original readme is available here

Why

This project exists to:

• Fix known bugs with the original compiler
• Provide a better development experience for the SA:MP community

If you find problem, you can open an issue and contributors can work on a fix. This isn’t true with the original compiler that comes with the SA:MP server distribution.

There are also new features that enhance the development experience, such as telling you which tags are involved in a “tag mismatch” or showing you where that pesky “symbol is never used” is actually declared.

There are plenty of features and fixes that are documented, see below for links:

• Known compiler bugs contains a list of bugs that the team are aware of
  with their status.
  
• What’s new contains features and other notable changes.
  
• Release notes list of all official releases of the compiler
  binaries.

How to Use

Binary packages can be downloaded from Releases, see the below sections for platform-specific installation instructions.

Note: You will probably get warnings/errors/fatals when you first build with this compiler — you need to add the -Z flag to your build configuration or add #pragma compat. See this page for more information.

You can also get the latest development binaries for Windows on AppVeyor. This archive is built automatically on every Git commit and can be pretty unstable, so use at your own risk.

Windows

If you just use an editor configured to run pawncc such as Pawno, Sublime Text or VS Code you can simply delete your existing pawncc.exe and replace it with the new one.

Download the ZIP archive and extract pawnc.dll, pawncc.exe, pawndisasmsm.exe to your original pawncc directory. If you use Pawno, this will be in the pawno/ folder that’s inside your server directory.

openSUSE/SLES

There is an installation package available for openSUSE/SLES users so that you can easily install the compiler on your distribution. Please follow these steps:

1. Go to https://build.opensuse.org/package/s...hnitzer/pawncc
2. On the right side, select your distribution (only if it’s not disabled!)
3. Click “Go to download repository”
4. Copy the link and enter in sudo mode in your shell:
   zypper ar $COPIED_LINK home:mschnitzer
5. Again as root, type: zypper ref
6. Install the package with zypper in pawncc
7. Run pawncc in your shell to test if it’s working

With sampctl

If you are a sampctl user, you are already using this compiler!

Build From Source

If you are interested in contributing or just using a specific version, check out this page for instructions for compiling for your platform.

Background

The project was originally started as a set of patches aimed to create a compiler that would be compatible with the compiler used in SA-MP (San Andreas Multiplayer).

SA-MP uses a modified version of Pawn 3.2.3664 [1] with Windows-only executables, and the developers said that they lost the source code for it which means it can’t be ported to other platforms (e.g. Linux) and newly discovered bugs can’t be fixed. So the main goal of the project is to re-create changes that were previously made by the devs as well as fix all known compiler bugs.

[1] It’s worth noting that the version of the AMX embedded into the SA-MP server seems to be based on an older release of Pawn.

[SyS]

Ahh really love this

[Graber]

your code will compile faster with this, use it!

[coool]

WOW.

[FinStar]

nice, good luck

[Undef1ned]

Good

[CaptainBoi]

Good Job

[Freaksken]

Would be great if this would replace the old compiler in the next server package. Just switching from the old compiler to this one improved the compile time on one of my test modes from 7.2s to 1.9s!

[solstice_]

Neat!

[Variable™]

Great work!

[cuddlee]

10/10 would recommend to a friend

[aymel]

ah yes open source *rubs hands*

[iLearner]

Yep recommended.

[wallee]

Quote:

Originally Posted by [HLF]Southclaw SA-MP uses a modified version of Pawn 3.2.3664 [1] with Windows-only executables, and the developers said that they lost the source code for it which means it can’t be ported to other platforms (e.g. Linux) and newly discovered bugs can’t be fixed. So the main goal of the project is to re-create changes that were previously made by the devs as well as fix all known compiler bugs. [1] It’s worth noting that the version of the AMX embedded into the SA-MP server seems to be based on an older release of Pawn.

if kalcor lost the source code to it and then you guys recreated it from scratch how do you know this matches all the functionality of the old one?

i am using this for about a year and haven't encountered any issues but i'm just curious

[Y_Less]

Quote:

Originally Posted by wallee if kalcor lost the source code to it and then you guys recreated it from scratch how do you know this matches all the functionality of the old one? i am using this for about a year and haven’t encountered any issues but i’m just curious

Kalcor didn’t do the original modifications, I did. Even if that weren’t the case, it is very easy to confirm that the changes match by taking code that would only compile on the old one and compiling it on the new one.

[NoahF]

Simply fantastic. Great work, thank you!

[Calisthenics]

Is this a normal behavior of const-correctness? The compiler complains (as it should) about 1D array but not for 2D, 3D or 4D arrays

pawn Код:

#include <a_samp>

main()
{
    new arr[10];

    func(arr);

    printf("value: %d", arr[0]);
}

func(const arr[])
{
    arr[0] = 1; // error 022: must be lvalue (non-constant)
}

pawn Код:

#include <a_samp>

main()
{
    new arr[10][10];

    func(arr);

    printf("value: %d", arr[0][0]); // prints "value: 1"
}

func(const arr[][])
{
    arr[0][0] = 1;
}

No warnings. It modifies the value even though it is meant to be a constant array.

[Y_Less]

No, that's not correct. Could you open an issue on github please?