Re: [boost] sunpro 5_8 test at OSL (was static templates in BOOST)

From: Steve Clamage (Stephen.Clamage_at_[hidden])
Date: 2005-07-08 21:44:26

The macro BOOST_NO_IS_ABSTRACT should be defined.

Also, be sure you are using the option -library=stlport4 on every CC command 
line. The configuration files for Sun should set up that option already.

---
Steve Clamage, stephen.clamage_at_[hidden]
Jonathan Turkanis wrote:
> Jeff Garland wrote:
> 
>>On Thu, 7 Jul 2005 21:29:09 -0500, Douglas Gregor wrote
>>
>>
>>
>>>Just a follow-up: I've installed the Sun compilers on one of our OSL
>>>systems. It shows up as "OSL4" (sunpro-5_8u1-sunos) on the
>>>regression tests. It's far too late in the Boost release cycle to
>>>consider it a release platform, but we'll be running the tests
>>>nightly regardless.
>>
>>This is interesting.  Looks like Sun has made some serious forward
>>progress. By my eye it looks like about 1/2 of the date-time tests
>>now pass.  And that's a bit deceiving as it looks like a handful of
>>problems are preventing most of the rest from compiling. As usual,
>>it's I/O that's at the root of a big chunk of the problems. As I
>>recall when Caleb was running SunPro almost none of date-time would
>>compile...
> 
> 
> Only one Iiostreams tests passes. All the errors I've checked are either
> internal errors or problems with is_abstract. is_abstract_test, from type
> traits, is also failing.
> 
> Config defines BOOST_NO_IS_ABSTRACT for __SUNPRO_CC < 0x570; I wonder if this
> needs to be defined for  0x580?
> Jonathan
>