🚀 go-pugleaf

It came to my attention that the simple longjmp program:

#include <setjmp.h>

main()
{
   jmp_buf env;
   int err;
   
   if (err = setjmp(env))
   {
      printf("World (%d)!\n", err);
   } else {
      printf("Hello, ");
      longjmp(env, 1);
   }
}

was not working when stepped/nexted, but worked fine when run normally
or continued through.  Looking at some breakpoints, I saw:

(gdb) mai in br
Num Type           Disp Enb Address    What
-1  longjmp        keep n   0x00002404  <longjmp>
-2  longjmp        keep n   0x00002404  <longjmp>
-3  longjmp resume keep n   0x00000000  <_DYNAMIC>
1   breakpoint     keep y   0x0000227c in main at jmptest.c:5
(gdb)

Multiple breakpoints on the same address, hmmm.  Well, turns out this
was compiled using the sys5 libraries (on sunos 4.1.2) which has:

% nm /usr/5lib/libc.a | grep jmp
_setjmp.o:
0000002c T __longjmp
00000000 T __setjmp
0000002c T _longjmp
00000000 T _setjmp

Yes, folks, longjmp and _longjmp are the *same*.  Gdb sets breakpoints
for longjmp, _longjmp, and siglongjmp, and never checks for duplicates
believing that they will be always be different.  Ooops?

This small little patch seems to get me by for the time being.  I didn't
put a lot of thought into the correctness of the patch.  Someone with
more know how than I should take a look at the longjmp breakpoint code,
and see how it should make sense with the above in mind.

----------------------------------------------------------------------
*** breakpoint.c	Wed Aug 12 04:30:29 1992
--- breakpoint.c-dist	Mon Jul  6 10:10:37 1992
***************
*** 1290,1299 ****
  
    ALL_BREAKPOINTS (b)
      if (b->type == bp_longjmp)
!     {
!        b->enable = enabled;
!        check_duplicates (b->address);
!     }
  }
  
  void
--- 1290,1296 ----
  
    ALL_BREAKPOINTS (b)
      if (b->type == bp_longjmp)
!       b->enable = enabled;
  }
  
  void