Boost logo

Boost :

From: John Maddock (John_Maddock_at_[hidden])
Date: 2001-03-10 06:56:17


>Just out of curiosity, are these deficiencies in the Sun compiler (i.e.
not
>behaving in compliance with the C++ standard)? Or are they just "quirks"
of
>the compiler that haven't been adequately hacked around yet? And if the
>former, has anyone submitted bug reports to the Sun guys? I had several
>standards-compliance issues fixed in the early days of the 5.0 compiler by
>submitting bug reports (well I don't know that they fixed them because of
>ME, but I do know that several of the bugs, as reported, were fixed in a
>patch).

I don't have access to sunpro - I've just been looking at Jen's regression
test compiler logs for sunpro 6.1.

It looks as though a thrown exception is getting caught by the wrong catch
statement - I don't know what the issue is here - probably it needs someone
to debug the test program to find out (I can provide guidance if you want
to volanteer).

There is also an issue with the type-traits code: sunpro seems to only
partially support partial specialisation, it seems to be binding to the
correct template but the deduced type is wrong:

template <class T>
struct remove_reference
{ typedef T type; };
template <class T>
struct remove_reference<T&>
{ typedef T type; };

On sunpro "remove_reference<int&>::type" is "int&" and not "int" as it
should be, within the body of the template partial specialisation the "&"
doesn't get stripped off the template parameter T (i'm not sure if I have
explained this very well - look at the type_traits test programs if you
want to see the problem in action).

- John Maddock
http://ourworld.compuserve.com/homepages/john_maddock/


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk