Bug #947: Signal handler block is never called when the program is too short

• Status: Closed
• Assignee: ko1 (Koichi Sasada)
• Target version: 1.9.2
• ruby -v: ruby 1.9.2dev (2010-03-24 trunk 27033) [i686-linux]

[yugui (Yuki Sonoda)]

=begin
forked from #727.

% ruby-trunk -e 'Signal.trap(:INT){p :ok}; Process.kill(:INT, $$)'
=> displays nothing
=end

[yugui (Yuki Sonoda)]

• Target version changed from 1.9.1 RC2 to 1.9.2

=begin

=end

[mame (Yusuke Endoh)]

• ruby -v set to ruby 1.9.2dev (2010-03-24 trunk 27033) [i686-linux]

=begin
Hi,

2008/12/29 Yuki Sonoda [e-mail]:

Bug #947: Signal handler block is never called when the program is too short
http://redmine.ruby-lang.org/issues/show/947

Author: Yuki Sonoda
Status: Open, Priority: Normal
Assigned to: Koichi Sasada, Category: YARV, Target version: 1.9.1

forked from #727.

% ruby-trunk -e 'Signal.trap(:INT){p :ok}; Process.kill(:INT, $$)'
=> displays nothing

It would be good to check whether signal is delivered or not before
process termination:

diff --git a/eval.c b/eval.c
index a1deab6..2db160c 100644
--- a/eval.c
+++ b/eval.c
@@ -127,6 +127,12 @@ ruby_cleanup(volatile int ex)
volatile VALUE errs[2];
rb_thread_t *th = GET_THREAD();
int nerr;

• void rb_threadptr_interrupt(rb_thread_t *th);

• void rb_threadptr_check_signal(rb_thread_t *mth);

• rb_threadptr_interrupt(th);

• rb_threadptr_check_signal(th);

• RUBY_VM_CHECK_INTS();

  errs[1] = th->errinfo;
  th->safe_level = 0;
  diff --git a/thread.c b/thread.c
  index a3dfe00..5273dd1 100644
  --- a/thread.c
  +++ b/thread.c
  @@ -290,7 +290,7 @@ reset_unblock_function(rb_thread_t *th, const struct rb_unblock_callback *old)
  native_mutex_unlock(&th->interrupt_lock);
  }

-static void
+void
rb_threadptr_interrupt(rb_thread_t *th)
{
native_mutex_lock(&th->interrupt_lock);
@@ -2644,18 +2644,13 @@ rb_gc_save_machine_context(rb_thread_t *th)

int rb_get_next_signal(void);

-static void
-timer_thread_function(void *arg)
+void
+rb_threadptr_check_signal(rb_thread_t *mth)
{

• rb_vm_t vm = GET_VM(); / TODO: fix me for Multi-VM */
  int sig;

• rb_thread_t *mth;

• /* for time slice */

• RUBY_VM_SET_TIMER_INTERRUPT(vm->running_thread);

• /* mth must be main_thread */

• /* check signal */
• mth = vm->main_thread;
  if (!mth->exec_signal && (sig = rb_get_next_signal()) > 0) {
  enum rb_thread_status prev_status = mth->status;
  thread_debug("main_thread: %s, sig: %d\n",
  @@ -2665,6 +2660,19 @@ timer_thread_function(void *arg)
  rb_threadptr_interrupt(mth);
  mth->status = prev_status;
  }
  +}

+static void
+timer_thread_function(void *arg)
+{

• rb_vm_t vm = GET_VM(); / TODO: fix me for Multi-VM */
• rb_thread_t *mth;
• /* for time slice */
• RUBY_VM_SET_TIMER_INTERRUPT(vm->running_thread);
• /* check signal */
• rb_threadptr_check_signal(vm->main_thread);

#if 0
/* prove profiler */

--
Yusuke ENDOH [e-mail]
=end

[mame (Yusuke Endoh)]

• Status changed from Open to Closed
• % Done changed from 0 to 100

=begin
This issue was solved with changeset r27513.
Yuki, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.

=end