Re: [boost] Re: Next revision of boost::thread & OS error code.

From: Beman Dawes (bdawes_at_[hidden])
Date: 2003-01-14 16:50:21

At 10:36 AM 1/14/2003, David Abrahams wrote:

 >David Abrahams <dave_at_[hidden]> writes:
 >
 >> Beman Dawes <bdawes_at_[hidden]> writes:
 >>
 >>> At 03:29 PM 1/13/2003, David Abrahams wrote:
 >>>
 >>>  >Remember that it's a bad idea to carry dynamically-allocated
 >>>  >state in an exception object.
 >>>
 >>> I wrestled with that a long time with the Filesystem Library, and
 >>> finally ignored it. The advice is good in general, but there just
 >>> didn't seem to be any way to meet the needs of users without
 >>> carrying several paths (which contain std::strings.)
 >>
 >> Then as a last resort you ought to be holding them via shared_ptr.
 >> It's critical that the exception object doesn't throw when it's
 >> copied.
 >
 >Having looked at filesystem_error, I see that it does not adopt this
 >practice.  This is a very serious design error, IMO, which can be
 >easily avoided.  Just to clarify, if an exception object's copy
 >constructor throws, the program goes directly to terminate().

Yes, the code needs to be changed.

Hopefully, the interface is now close to stable. I probably should change 
who() to return a const char * however.

--Beman