[ruby-core:17342] Re: Release Plan: Ruby 1.9.0-2

From: "Bill Kelly" <[e-mail]>
>
> I ran `nmake test` again, but it did not crash this time.
> 
> I will try a few more times to see if it might happen again.

$ while [ true ] ; do nmake test ; done

...crashed after a few repetitions.

However, it's in a slightly different place this time, but again a field in
the [th] structure is zero:

   1140  rb_thread_execute_interrupts   native_mutex_lock       Normal  0
   3344  timer_thread_func              timer_thread_func       Normal  0
>  3588  thread_start_func_1            ubf_handle              Normal  0


EAX = 00000000 EBX = 00B3A838 ECX = 00000000 EDX = 009E4118 ESI = 009E2EE8 
EDI = 7C911538 EIP = 004E1303 ESP = 0101FEC8 EBP = 0101FECC EFL = 00000246 
CS = 001B DS = 0023 ES = 0023 SS = 0023 FS = 003B GS = 0000 
OV = 0 UP = 0 EI = 1 PL = 0 ZR = 1 AC = 0 PE = 1 CY = 0 


--- p:\code\ruby-svn\trunk\thread_win32.c --------------------------------------

  static void
  ubf_handle(void *ptr)
  {
  004E12E0  push        ebp  
  004E12E1  mov         ebp,esp 
  004E12E3  push        ecx  
      rb_thread_t *th = (rb_thread_t *)ptr;
  004E12E4  mov         eax,dword ptr [ptr] 
  004E12E7  mov         dword ptr [th],eax 
      thread_debug("ubf_handle: %p\n", th);
  004E12EA  xor         ecx,ecx 
  004E12EC  je          ubf_handle+20h (4E1300h) 
  004E12EE  mov         edx,dword ptr [th] 
  004E12F1  push        edx  
  004E12F2  push        offset string "ubf_handle: %p\n" (50CD54h) 
  004E12F7  call        dword ptr [__imp__printf (4F12B8h)] 
  004E12FD  add         esp,8 
      w32_set_event(th->native_thread_data.interrupt_event);
  004E1300  mov         eax,dword ptr [th] 
> 004E1303  mov         ecx,dword ptr [eax+44h] 
  004E1306  push        ecx  
  004E1307  call        w32_set_event (4E1200h) 
  004E130C  add         esp,4 
  }
  004E130F  mov         esp,ebp 
  004E1311  pop         ebp  
  004E1312  ret


If nobody else is getting these crashes, I would begin to worry my build
is corrupted or misconfigured somehow...

I'll try a fresh checkout from svn and start from scratch...


Regards,

Bill