OK. Applied. BTW, I added a call to waitpid() after the fork() one in the main process, so that it waits for the child death before resuming its operations. Thanks -- Aurelien Aurelien Foret wrote:
Christian Hamar wrote:
Hello there.
So i completed other entry which was in TODO.
I hacked pacman-lib to use chroot() system calls and not the binary one. Entry was in TODO.
The attached patch is for latest checked pacman-lib, which is:
2005-10-29 09:18 aurelien * src/pacman/: conf.c, conf.h: fixed parseconfig usage
Patch description: ------------------
Removed: -------- * Removed old chroot binary detection part * Removed global declaration for chrootbin.
Added: ------ * Added a complete new way to chroot and do things.
Modified: --------- * Modified cmdline snprintf's to right way. Eliminated | (pipe) and echo things, which looked terrible ugly :)
As you need to know, this patch using a fork() and chroot() and execl() instead of system() call. execl() looks better and safer. fork() is required for chrooting in child process, if i do not do a fork() then whole process stays in chrooted env. So fork is neccessary as i think.
That's exactly how I wanted it to be.
Please test it and see it, if it is ok, then commit it. (I tested it, worked for me)
I'll have a look at it...
-- Aurelien