Re: make check (linux,2.1.19)

[John W. Eaton]

On 24-Oct-1999, Etienne Grossmann <address@hidden> wrote:

| 
|   here is where the seg fault happens :
| 
| 
| (gdb) bt
| #0  0x4011ea7c in chunk_free ()
| #1  0x4011ea2a in __cfree ()
| #2  0x400b18a4 in __builtin_vec_delete ()
| #3  0x8260820 in octave_time::octave_time (this=0xbfffee38, address@hidden)
|     at oct-time.cc:58
| #4  0x8129a7f in Fmktime (address@hidden) at ./DLD-FUNCTIONS/time.cc:204
| #5  0x81db5f8 in octave_builtin::do_index_op (this=0x84c9780, nargout=0, 
|     address@hidden) at ov-builtin.cc:65
| #6  0x81d4613 in octave_value::do_index_op (this=0xbffff098, nargout=0, 
|     address@hidden) at ov.cc:547
| #7  0x81f95ea in tree_index_expression::rvalue (this=0x8532348, nargout=0)
|     at pt-idx.cc:88
| #8  0x820afe4 in tree_statement::eval (this=0x8532178, silent=false, 
|     nargout=0, in_function_body=false) at pt-stmt.cc:123
| #9  0x820b28a in tree_statement_list::eval (this=0x8532448, silent=false, 
|     nargout=0) at pt-stmt.cc:156
| #10 0x81b2d2b in main_loop () at toplev.cc:126
| #11 0x80bf911 in main (argc=1, argv=0xbffff9d0) at octave.cc:576
| #12 0x400dd18f in __libc_start_main ()
| (gdb) l
| 53     ot_unix_time = mktime (&t);
| 54
| 55     ot_usec = tm.usec ();
| 56
| 57   #if defined (HAVE_TM_ZONE)
| 58     delete [] t.tm_zone;
| 59   #endif
| 60   }
| 61
| 62   string

I don't see how this could be a problem, but I also don't know why it
was necesary to do

  #if defined (HAVE_TM_ZONE)
    string s = tm.zone ();
    t.tm_zone = strsave (s.c_str ());
  #endif

just before the call to mktime.  So, I've removed the strsave() and
the `delete[]', along with similar code in octave_base_tm::strftime.
Can you please tell me if it still fails for you after you get the
next source update?

Thanks,

jwe