------- Comment #23 from gdr at cs dot tamu dot edu 2007-01-30 02:11 ------- Subject: Re: std::bad_alloc::what() does not explain what happened
"pcarlini at suse dot de" <[EMAIL PROTECTED]> writes: | ------- Comment #22 from pcarlini at suse dot de 2007-01-30 01:42 ------- | (In reply to comment #21) | > I suspect Andrew Pinski's point might be that what() could return a | > string that represents the name of the most derived type of the | > exception. But, nothing so far forces to do that. A reasonable | > definition is to what Paolo suggest, with clear documentation (that | > mentions this). | | Agreed. Gaby, do you have any strong opinion about std::exception | itself? In my current patch draft I'm leaving it alone, but in | principle we could change also its what() to return "std::exception" | instead of typeid(*this).name(). >From consistency point of view I would say that the change should also be done for std::exception. However, the use of typeid is very convenient in the sense that we have to defined what() only once. Now, if we change that definition in std::exception, it means that we should revisit all other exception classes, such as std::runtime_error, etc. I have no strong opinion. Probably just leave it as is, and document the choices we face so that people see the rationale behind the implementations. Thanks, -- Gaby -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14493
