openbsd-ports/lang/ghc/patches/patch-rts_posix_Signals_c
kili b2e79ee47e Update to ghc-7.4.2, but keep it marked as broken (and not only for amd64).
Most of the -nopie changes are from pascal@ (thanks); I only had to
merge them for ghc-7.4 and tweak configure and mk/config.mk.in.
2012-09-02 19:53:46 +00:00

55 lines
1.5 KiB
Plaintext

$OpenBSD: patch-rts_posix_Signals_c,v 1.2 2012/09/02 19:53:46 kili Exp $
There are several problems with the original code:
- It doesn't work with uthreads (instead of running the default
handler, the custom handler is called again, which causes an
endless cascade of signals once the process gets a SIGTSTP).
- It isn't posix compliant (sigprocmask(2) behaviour is unspecified for
multithreaded processes, where you would have to use
pthread_sigmask(3)).
- It's far too complicated compared to the simple kill(getpid(),
SIGSTOP).
--- rts/posix/Signals.c.orig Wed Feb 1 19:10:32 2012
+++ rts/posix/Signals.c Sat Jun 9 21:36:25 2012
@@ -519,7 +519,7 @@ empty_handler (int sig STG_UNUSED)
The trick we use is:
- catch SIGTSTP
- - in the handler, kill(getpid(),SIGTSTP)
+ - in the handler, kill(getpid(),SIGSTOP)
- when this returns, restore the TTY settings
This means we don't have to catch SIGCONT too.
@@ -541,17 +541,8 @@ sigtstp_handler (int sig)
}
}
- // de-install the SIGTSTP handler
- set_sigtstp_action(rtsFalse);
-
// really stop the process now
- {
- sigset_t mask;
- sigemptyset(&mask);
- sigaddset(&mask, sig);
- sigprocmask(SIG_UNBLOCK, &mask, NULL);
- kill(getpid(), sig);
- }
+ kill(getpid(), SIGSTOP);
// on return, restore the TTY state
for (fd = 0; fd <= 2; fd++) {
@@ -559,8 +550,6 @@ sigtstp_handler (int sig)
tcsetattr(0,TCSANOW,&ts[fd]);
}
}
-
- set_sigtstp_action(rtsTrue);
}
static void