Re: [Boost-users] [multi_index] SIGABRT inmulti_index_containerconstructor

From: Manuel Jung (gzahl_at_[hidden])
Date: 2007-02-25 08:13:25

"JOAQUIN LOPEZ MU?Z" wrote:

> Hello Manual, thank you for not abandoning this
> problem!
> 
> ----- Mensaje original -----
> De: Manuel Jung <gzahl_at_[hidden]>
> Fecha: Domingo, Febrero 25, 2007 11:43 am
> Asunto: Re: [Boost-users] [multi_index] SIGABRT in
> multi_index_container constructor
> Para: boost-users_at_[hidden]
> 
>> Hi,
>> 
>> After a working version with std::queue's im now back to
>> multi_index_container. And new try new idea! I found out something:
>> 
>> This "#define" block
>> -------cut--------
>> #if !defined(NDEBUG)
>> #define BOOST_MULTI_INDEX_ENABLE_INVARIANT_CHECKING
>> #define BOOST_MULTI_INDEX_ENABLE_SAFE_MODE
>> #endif
>> -------cut--------
>> was before all my #include's up to now. And if the "SAFE_MODE" was
>> enabled i was getting SEGFAULTs. But now i dropped this blocked
>> after all my includes in the header file or even in the main
>> implemantation file and dont get the errors anymore!
>> 
>> I hope this helps/makes any sense to you?
> 
> It does make sense, but it doesn't provide much info:
> those macros have to be defined before inclusion of B.MI
> headers to take effect, so that having them after the includes
> is entirely equivalent to not defining them :(

okay, it has clearly no effect to defines variables/flags, hwich are not
used... 

> 
> Can you (I mean, are you allowed to) provide a sketch of
> the class where the offending queue class is being used?
> I remember you said it's a member of a more complex class,
> or something like that.

Hm, i think i can do that, nobody could use a part of the code anyway (or
would like..); But which part of it do you like? The Header? Or a part of
the header? Or parts of code where the first error with the scripts are
raising?

Greetings,
Manu

Ps.: I have some question about the mem_fun. You cannot have a index which
is automaticly recalculated within a member function, can you? Im thinking
about a function checking, if a try_mutex is locked. But i think, this is a
gainst the natured of a index. Am i right?

> 
> Thank you,
> 
> Joaquín M López Muñoz
> Telefónica, Investigación y Desarrollo