[boost] Re: [MultiArray] Regression failures on Borland

From: David Abrahams (dave_at_[hidden])
Date: 2003-12-11 10:36:26

garcia <garcia_at_[hidden]> writes:

> Douglas Paul Gregor wrote:
>
>>I tried getting the MultiArray regression tests to run well on Borland,
>>but failed miserably. The problem goes beyond the typical refactoring to
>>make BOOST_STATIC_CONSTANT's visible in the right places, and touches on
>>some of the overloading/template deduction bugs. I'd be willing to just
>>mark the compiler broken w.r.t. the library.
>>
>>  
>>
> Thanks for the effort, Doug.  Quite some time ago, John Maddock had
> suggested that it might be possible to get multi_array to work with
> Borland, but I don't remember the details of what he thought was
> involved.  I thought it had something to do with the numeric template
> parameters, but I may very well be mistaken.

Getting multi-array to work with Borland would probably not be too
hard if you replaced all the non-type template parameters with MPL
integral constant wrappers.  I didn't bother because it didn't seem
to actually be a regression from 1.30.2.  The meta-comm result pages
seem to indicate that whatever's failing on Borland never worked
there.

http://www.meta-comm.com/engineering/resources/cs-win32_metacomm/developer_result_page.html#multi_array

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com