[Ns-developers] python bindings build failure

[Mathieu Lacage]

On Thu, 2010-04-15 at 09:53 +0200, Mathieu Lacage wrote:
> assert failed. file=../src/core/type-id.cc, line=138, cond="uid <=
> m_information.size () && uid != 0"
> Segmentation fault
> 
> Which looks like some weird TypeId registration problem
> (ns3::YansWifiPhy does not appear to be defined). Anyone had similar
> problems in the past ?

The corresponding backtrace looks like this:

#5  0x00007ffff1cc7ea1 in ns3::YansWifiPhyHelper::Create
(this=0x7fffffffd920, node=
      {m_ptr = 0x80d260}, device=Cannot access memory at address 0x0
) at ../src/helper/yans-wifi-helper.cc:233
#6  0x00007ffff1cfeff6 in ns3::PcapHelperForDevice::EnablePcap
(this=0x805ab0, prefix=
        {static npos = 18446744073709551615, _M_dataplus =
{<std::allocator<char>> = {<__gnu_cxx::new_allocator<char>> = {<No data
fields>}, <No data fields>}, _M_p = 0xab4468 "mixed-wireless"}}, nd=
      {m_ptr = 0x80d260}, promiscuous=false, explicitFilename=false)
    at ../src/helper/trace-helper.cc:417

But it's an impossible backtrace. 

But, if I look at the declaration of YansWifiPhyHelper, I see multiple
inheritance used:

class YansWifiPhyHelper : public WifiPhyHelper,
                          public PcapHelperForDevice, public
AsciiTraceHelperForDevice


which hints at something going horribly wrong with typecasting and
multiple inheritance and python wrappers. How come I am the only person
to see this ?

Mathieu