On 8/26/06, <b class="gmail_sendername">Paul Davis</b> &lt;<a href="mailto:pjdavis@engineering.uiowa.edu">pjdavis@engineering.uiowa.edu</a>&gt; wrote:<div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>(gdb) bt<br>#0&nbsp;&nbsp;0x000000000040a2f8 in boost::detail::atomic_exchange_and_add<br>(pw=0x98, dv=-1) at sp_counted_base_gcc_x86.hpp:50<br>#1&nbsp;&nbsp;0x000000000040a46c in boost::detail::sp_counted_base::release<br>(this=0x90) at sp_counted_base_gcc_x86.hpp:143
</blockquote><div><br>Ah, the this pointer is off in la-la land here (0x90).<br><br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">#2&nbsp;&nbsp;0x000000000040a4d0 in ~shared_count (this=0x2514218) at
<br>shared_count.hpp:159</blockquote><div><br>which means the pi_ pointer, an sp_counted_base, got messed up.<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
#3&nbsp;&nbsp;0x00002aaaaccb849b in ~shared_ptr (this=0x2514210) at shared_ptr.hpp:106<br>#4&nbsp;&nbsp;0x00002aaaaccb88e1 in<br>__gnu_cxx::new_allocator&lt;boost::shared_ptr&lt;min_heap&gt; &gt;::destroy<br>(this=0x7fffffc2b410, __p=0x2514210)
<br>&nbsp;&nbsp;&nbsp;&nbsp;at new_allocator.h:107<br>#5&nbsp;&nbsp;0x00002aaaaccb9f66 in std::vector&lt;boost::shared_ptr&lt;min_heap&gt;,<br>std::allocator&lt;boost::shared_ptr&lt;min_heap&gt; &gt; &gt;::erase (<br>&nbsp;&nbsp;&nbsp;&nbsp;this=0x7fffffc2b410, __position={_M_current = 0x2514220}) at
<br>vector.tcc:115<br></blockquote></div><br>Hmm, you're guess is as good as mine.&nbsp; I couldn't tell you if it is a boost bug or not.&nbsp; But I would probably spend some more time checking your own code too, to see if maybe you have memory corruption resulting from a buffer overflow or something like that.
<br><br>Phillip Hellewell<br>