Re: Throwing std::ios_base::failure on formatted input with gcc 6.2

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 10/26/2016 1:04 PM, Jędrzej Dudkiewicz wrote:
I believe that it is the opposite: library is throwing the old one and
you are catching new one. This is my understanding of it.

So catching 'std::ios_base::failure' is actually catching 'std::ios_base::failure[abi:cxx11]' ?


On Wed, Oct 26, 2016 at 6:46 PM, Edward Diener
<eldlistmailingz@xxxxxxxxxxxxxx> wrote:
On 10/26/2016 10:57 AM, Jonathan Wakely wrote:

On 26 October 2016 at 14:42, Edward Diener wrote:

I am missing the syntactical difference between the two. There is
'std::ios_base::failure' and how do you specify the other one ? Your
notation of 'std::ios_base::[abi:__cxx11]failure' I do not understand.


It comes from the abi_tag attribute documented at
https://gcc.gnu.org/onlinedocs/gcc/C_002b_002b-Attributes.html

namespace std {
  struct ios_base {
    struct __attribute__((abi__tag__("cxx11"))) failure { };
  };
}

I got the demangled name slightly wrong, the attribute goes at the end:

#include <iostream>
int main()
{
  throw std::ios_base::failure("");
}
terminate called after throwing an instance of
'std::ios_base::failure[abi:cxx11]'
  what():  : iostream error
Aborted (core dumped)


I am syntactically catching 'std::ios::failure'. Are you saying that the
exception being thrown in my OP is actually
'std::ios_base::failure[abi:cxx11]' ? I am still a bit confused when you
said in a previous reply:


"You're trying to catch the new one, but the library throws the old
one."











[Index of Archives]     [Linux C Programming]     [Linux Kernel]     [eCos]     [Fedora Development]     [Fedora Announce]     [Autoconf]     [The DWARVES Debugging Tools]     [Yosemite Campsites]     [Yosemite News]     [Linux GCC]

  Powered by Linux