Your message dated Sat, 18 May 2013 10:13:15 +0300 (EEST) with message-id <[email protected]> and subject line Bug #399769 filed against the timeout package has caused the Debian Bug report #399769, regarding timeout: bad support for processes which use a process group to be marked as done.
This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact [email protected] immediately.) -- 399769: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=399769 Debian Bug Tracking System Contact [email protected] with problems
--- Begin Message ---Package: timeout Version: 1.11-6.2 Tags: patch The timeout command has a bug which occures when the spawned process sends a signal to process group. The bad output: $ timeout 5 sh -c 'kill 0'; echo $? Timeout: aborting command ``sh'' with signal 9 Killed 137 I would like to do a NMU with solution for this bug and Bug#187092. -- .''`. Piotr Roszatycki : :' : mailto:[email protected] `. `' `---- timeout.c.orig 2001-09-09 23:46:10.000000000 +0200 +++ timeout.c 2006-11-21 21:40:47.000000000 +0100 @@ -34,6 +34,7 @@ #include <stdlib.h> #include <unistd.h> #include <stdio.h> +#include <sys/wait.h> extern int optind; @@ -41,6 +42,7 @@ #define perrorexit(s) { perror(s); exit(1); } +static pid_t child_pid; static int kill_signal = SIGKILL; static char *progname; static char *commandname; @@ -55,9 +57,16 @@ int sig; { signal(kill_signal, SIG_DFL); - fprintf(stderr, "Timeout: aborting command ``%s'' with signal %d\n", - commandname, kill_signal); - kill(0, kill_signal); + if (sig == SIGALRM) { + fprintf(stderr, "Timeout: aborting command ``%s'' with signal %d\n", + commandname, kill_signal); + } + else { + fprintf(stderr, "Got signal %d: aborting command ``%s'' with signal %d\n", + sig, commandname, kill_signal); + } + kill(-child_pid, kill_signal); + exit(-1); } int main(argc, argv) @@ -66,7 +75,6 @@ { int time_to_run; pid_t pid; - pid_t child_pid; int status; progname = argv[0]; @@ -87,11 +95,11 @@ * Run the command and its watchdog in a separate process group so that * both can be killed off with one signal. */ - setsid(); switch (child_pid = fork()) { case -1: /* error */ perrorexit("timeout: fork"); case 00: /* run controlled command */ + setsid(); execvp(argv[1], argv + 1); perrorexit(argv[1]); default: /* become watchdog */ @@ -99,10 +107,14 @@ (void) signal(SIGINT, terminate); (void) signal(SIGQUIT, terminate); (void) signal(SIGTERM, terminate); + (void) signal(SIGSEGV, terminate); + (void) signal(SIGPIPE, terminate); + (void) signal(SIGXCPU, terminate); + (void) signal(SIGXFSZ, terminate); (void) signal(SIGALRM, terminate); alarm(time_to_run); while ((pid = wait(&status)) != -1 && pid != child_pid) /* void */ ; return (pid == child_pid ? status : -1); } }
--- End Message ---
--- Begin Message ---Version: 1.19-1 Some (long) time ago you filed this bug against the timeout package. In the meantime the package has been removed and a timeout command is included in the coreutils package. If the issue you encountered is also applicable to timeout from coreutils (as released with Debian 7.0 wheezy or later) the maintainers would very much appreciate a fresh report using reportbug /usr/bin/timeout (you should also mention the number of this bug if you think it contains valuable information). Please note that further information sent to this bug is likely to remain unread, because the bug is currently not assigned to any existing package. Thank you for taking time to report bugs found in Debian packages. Kind regards, Andrei (with no relation to timeout, tct or coreutils maintenance)
--- End Message ---
