Add onClientCommand event to cancel client-side commands

[ColombuxMaximus]

Adds onClientCommand event to intercept and cancel client-side commands before execution.

#3160

Before you go ahead and create a pull request, please make sure:

• your code follows the coding guidelines
• your commit messages are informative ("update file.cpp" is unhelpful. "fix missing model in getVehicleNameFromModel" is helpful)

If your work is incomplete, do not prefix your pull request with "WIP", instead
create a draft pull request: https://github.blog/2019-02-14-introducing-draft-pull-requests/

Thank you!

[ColombuxMaximus]

Let me know if there's something else. Not sure why some of my comments/answers are not showing.
( @Dutchman101 )

[ColombuxMaximus]

Hello @FileEX just in case you missed my comment to your review because Github says it's 'Pending'

[FileEX]

nullptr

[ColombuxMaximus]

Thank you @FileEX
Can you check here if everything is correct?

I'll test it tomorrow in game

[MohabCodeX]

Good job , i have made some changes that will enhance command handling by adding argument support to the event , but i don't have changing permission to your branch , anyway the final usage would be like this :

function onClientCommand(command, arguments)
    outputChatBox("Command: " .. command, 255, 255, 0)
    outputChatBox("Arguments: " .. arguments, 0, 255, 255)

    if command == "blocked" then
        outputChatBox("Command blocked!", 255, 0, 0)
        return false
    end

    return true
end

addEventHandler("onClientCommand", root, onClientCommand)

[FileEX]

Thank you @FileEX Can you check here if everything is correct?

I'll test it tomorrow in game

Okay, I didn’t realize that CCommands::Execute is part of the Core project. So let's keep everything within Deathmatch.

My suggestion is that COMMAND_Executed should return two booleans: something like wasCanceled and wasExecuted. Then, in CCommands::Execute, we display Unknown command or cvar: only if wasCanceled is false. The event can stay inside COMMAND_Executed.

Also, I noticed a bug. COMMAND_Executed should return false if:

g_pClientGame->GetRegisteredCommands()->ProcessCommand(szCommandBufferPointer, szArguments);

returns false – because that means the command doesn’t exist. Right now, it always returns true, even if the command doesn’t exist. As a result, when you type something like /foo, there’s no “unknown command” message – it just silently fails. You could fix this as well while you're at it.

Lastly, regarding executeCommandHandler – I think this function should also trigger onClientCommand, but with an additional parameter like executedByFunction to indicate whether the command was triggered manually (by typing) or programmatically (via function).

You could also consider having the CRegisteredCommands::ProcessCommand function trigger the event instead of COMMAND_Executed, and have it return two booleans – whether the command was canceled and whether it was executed (i.e., if it exists). That way, executeCommandHandler, which directly calls CRegisteredCommands::ProcessCommand, would benefit from the same logic. This way, we avoid triggering the event in two different places.