minishell/src/executing/group_cmd/group_cmd_execute.h

/* ************************************************************************** */
/*                                                                            */
/*                                                        :::      ::::::::   */
/*   group_cmd_execute.h                                :+:      :+:    :+:   */
/*                                                    +:+ +:+         +:+     */
/*   By: khais <marvin@42.fr>                       +#+  +:+       +#+        */
/*                                                +#+#+#+#+#+   +#+           */
/*   Created: 2025/04/04 19:42:36 by khais             #+#    #+#             */
/*   Updated: 2025/04/29 14:40:03 by khais            ###   ########.fr       */
/*                                                                            */
/* ************************************************************************** */

#ifndef GROUP_CMD_EXECUTE_H
# define GROUP_CMD_EXECUTE_H

# include "../../minishell.h"

t_subprocess	group_cmd_execute(t_group_cmd *cmd, t_minishell *app);

#endif // GROUP_CMD_EXECUTE_H
group_cmd_execute: fork and passthrough to cmd_execute 2025-04-04 19:52:29 +02:00			`/* ************************************************************************** */`
			`/* */`
			`/* ::: :::::::: */`
			`/* group_cmd_execute.h :+: :+: :+: */`
			`/* +:+ +:+ +:+ */`
			`/* By: khais <marvin@42.fr> +#+ +:+ +#+ */`
			`/* +#+#+#+#+#+ +#+ */`
			`/* Created: 2025/04/04 19:42:36 by khais #+# #+# */`
fix(exec): prevent leak when calling exit() in a subprocess This was because at the point at which exit is called, we can only free a t_simple_cmd, but not the whole t_cmd tree. This commit introduces a convention of returning a t_subprocess enum from each function in exec. If the current thread is a subprocess, SUBPROCESS is returned, else PARENTPROCESS. We also no longer call exit in subproceses. This way, all processes bubble up to main, where if SUBPROCESS is returned, all memory is freed and the program exits. This also removes the previous should_exit convention. Still TODO: handling of the exit builtin, which should also bubble up in the same way. 2025-04-29 12:58:26 +02:00			`/* Updated: 2025/04/29 14:40:03 by khais ### ########.fr */`
group_cmd_execute: fork and passthrough to cmd_execute 2025-04-04 19:52:29 +02:00			`/* */`
			`/* ************************************************************************** */`

			`#ifndef GROUP_CMD_EXECUTE_H`
			`# define GROUP_CMD_EXECUTE_H`

			`# include "../../minishell.h"`

fix(exec): prevent leak when calling exit() in a subprocess This was because at the point at which exit is called, we can only free a t_simple_cmd, but not the whole t_cmd tree. This commit introduces a convention of returning a t_subprocess enum from each function in exec. If the current thread is a subprocess, SUBPROCESS is returned, else PARENTPROCESS. We also no longer call exit in subproceses. This way, all processes bubble up to main, where if SUBPROCESS is returned, all memory is freed and the program exits. This also removes the previous should_exit convention. Still TODO: handling of the exit builtin, which should also bubble up in the same way. 2025-04-29 12:58:26 +02:00			`t_subprocess group_cmd_execute(t_group_cmd cmd, t_minishell app);`
group_cmd_execute: fork and passthrough to cmd_execute 2025-04-04 19:52:29 +02:00
			`#endif // GROUP_CMD_EXECUTE_H`