Tag Info

New answers tagged


If I understand question 1 correctly, no, there's no stack or list of child processes. Somewhere in the structs that the kernel keeps around for a process, there's a PID field and a parent PID field. When a process exits, the parent PID of the exiting process gets a notification via a SIGCHLD (the default SIGCHLD handler is SIG_IGN, so nothing happens unless ...


To do this for the shell, use the source command (aliased to .). It will run the application in the context of the current shell rather than creating a new shell context. This is how you can set environment variables as well. If you notice all of the startup scripts in /etc/rc.d or /etc/init.d, you will see lots of scripts that are included by other scripts ...


You can also use the CDPATH environmental variable. As explained in man bash: CDPATH The search path for the cd command. This is a colon-separated list of directories in which the shell looks for destination directories specified by the cd command. A sample value is ".:~:/usr". So, for example, you could add ...


Your program does exactly what you tell it to do: it changes the working directory for itself to /home/enedil/projects/algo. But once it exits, the shell's working directory is restored. I guess what you want to achieve is to change the working directory of the parent process, i.e. the shell, without resorting to a simple cd. There is a method, but as its ...


The current working directory is local to the process. So, what you want is not possible.


You can refer to the official page: https://www.kernel.org/pub/linux/libs/uclibc/Glibc_vs_uClibc_Differences.txt


I don't know of any portable way to do this. I thought maybe ptrace(), but I can't see how from the manpage. Even if that works, "tracing" the other process in any way is probably unnecessarily invasive For Linux, your suggestion to use fopen("/proc/PID/status", "r") is about as clean and direct as you're going to get. It seems to be what gdb does.

Top 50 recent answers are included