Hooks

Mamba supports lifecycle hooks that run before and after the main run function. This matches cobra's hook chain exactly.

Hook types

Hook	When it runs	Scope
persistent_pre_run	Before pre_run	This command and all children
pre_run	Before run	This command only
run	Main command logic	This command only
post_run	After run	This command only
persistent_post_run	After post_run	This command and all children

Execution order

For a command with all hooks defined:

persistent_pre_run  →  pre_run  →  run  →  post_run  →  persistent_post_run

Example

const mamba = @import("mamba");

fn setupLogging(_: *mamba.Command, _: []const []const u8) !void {
    // Runs before every command in the tree
}

fn validateConfig(_: *mamba.Command, _: []const []const u8) !void {
    // Runs before this specific command
}

fn serve(cmd: *mamba.Command, _: []const []const u8) !void {
    cmd.writeOut("Server started\n");
}

fn cleanup(_: *mamba.Command, _: []const []const u8) !void {
    // Runs after this specific command
}

pub fn main() !void {
    var cmd = mamba.Command.init(.{
        .name = "serve",
        .persistent_pre_run = &setupLogging,
        .pre_run = &validateConfig,
        .run = &serve,
        .post_run = &cleanup,
    });
    try cmd.execute();
}

Persistent hooks

Persistent hooks cascade to all child commands. This is useful for cross-cutting concerns like logging or authentication:

var root = mamba.Command.init(.{
    .name = "app",
    .persistent_pre_run = &initLogger,      // runs for ALL commands
    .persistent_post_run = &flushMetrics,   // runs for ALL commands
});

var serve = mamba.Command.init(.{
    .name = "serve",
    .pre_run = &loadConfig,  // only for serve
    .run = &serveRun,
});

root.addCommand(&serve);

When running app serve: 1. initLogger (persistent, from root) 2. loadConfig (local, from serve) 3. serveRun (local, from serve) 4. flushMetrics (persistent, from root)

Error handling

If any hook returns an error, execution stops. The error propagates to the execute() caller.