«

Dear ,

My C++ program crashes with the following error:
SIGNAL BUS (invalid addresses aligment)
at random pleaces.



t@1 (l@1) signal BUS (invalid address alignment) in
__rwstd::digit_writer<char,std:streambuf_iterato r
<char,std::char_traits<char> > >:ut_digits at 0xfdf09af0
0xfdf09af0: put_digits+0x0298: ld [%o0], %l5
Current function is Analytic::Schedule (optimized)
4838 str << " " << ":R" << ":::"
<< "01/01/1900" << ":" << DF ;
(dbx) quit

Thanks in advance

bye
Tony

ADS7328 wrote:
> Dear ,
>
> My C++ program crashes with the following error:
> SIGNAL BUS (invalid addresses aligment)
> at random pleaces.
>
>
>
> t@1 (l@1) signal BUS (invalid address alignment) in
> __rwstd::digit_writer<char,std:streambuf_iterato r
> <char,std::char_traits<char> > >:ut_digits at 0xfdf09af0
> 0xfdf09af0: put_digits+0x0298: ld [%o0], %l5
> Current function is Analytic::Schedule (optimized)
> 4838 str << " " << ":R" << ":::"
> << "01/01/1900" << ":" << DF ;
> (dbx) quit

Usually a bad cast. Are you casting to values directly from a char *
somewhere ? The compiler usually does not do this.

The resolution to this is to either use some compiler specific features
(like #pragma pack) or read a byte at a time.

The information you provide above is next to useless to us. Is it IRIX?

On May 22, 2:09 pm, Gianni Mariani <gi3nos...@mariani.ws> wrote:
> ADS7328 wrote:
> > Dear ,
>
> > My C++ program crashes with the following error:
> > SIGNAL BUS (invalid addresses aligment)
> > at random pleaces.
>
> > t@1 (l@1) signal BUS (invalidaddressalignment) in
> > __rwstd::digit_writer<char,std:streambuf_iterato r
> > <char,std::char_traits<char> > >:ut_digits at 0xfdf09af0
> > 0xfdf09af0: put_digits+0x0298: ld [%o0], %l5
> > Current function is Analytic::Schedule (optimized)
> > 4838 str << " " << ":R" << ":::"
> > << "01/01/1900" << ":" << DF ;
> > (dbx) quit
>
> Usually a bad cast. Are you casting to values directly from a char *
> somewhere ? The compiler usually does not do this.
>
> The resolution to this is to either use some compiler specific features
> (like #pragma pack) or read a byte at a time.
>
> The information you provide above is next to useless to us. Is it IRIX?

Thanks a lot, I'm working on Solaris 2.8

I changed into :

double DF= 0.0 ;

ostringstream str ;

str << " " << ":R" << ":::" <<
"01/01/1900" << ":" << DF ;

and I deleted the

str.~ostingstream() ;


and it seems to work fine.


But I don't undertsand if the entry " str.~ostringstream() "
was the problem.


bye
Tony

ADS7328 wrote:
> [..]
> I changed into :
>
> double DF= 0.0 ;
>
> ostringstream str ;
>
> str << " " << ":R" << ":::" <<
> "01/01/1900" << ":" << DF ;
>
> and I deleted the
>
> str.~ostingstream() ;
>
>
> and it seems to work fine.
>
>
> But I don't undertsand if the entry " str.~ostringstream() "
> was the problem.

If you were doing

char storage[sizeof(SomeType)];
SomeType *myObject = new (storage) SomeType;

it is not guaranteed to work because the alignment requirements
for a 'char' array and 'SomeType' can be different. To perform
proper placement new, your storage has to be allocated in free
store itself or other measures have to be taken (implementation-
specific, usually) to ensure proper alignment.

Why were you calling the destructor of 'str' yourself, anyway?

V
--
Please remove capital 'A's when replying by e-mail
I do not respond to top-posted replies, please don't ask