Boost Testing :
From: Eric Niebler (eric_at_[hidden])
Date: 2007-03-01 13:22:52
Sean Huang wrote:
> There are quite a few intel-vc8-win-9.1 errors but it seems they are all
> caused by a single problem shown below. If we can find a workaround for this
> problem then intel 9.1 + VC8 can be added back to the release platforms. See
> Huang-WinXP-x86_32 test results in the regression reports at
> .\boost/xpressive/detail/utility/boyer_moore.hpp(64): error: more than one
> instance of overloaded function "std::fill_n" matches the argument list:
> function template "void __cdecl std::fill_n(_OutElem (&)[_Size],
> _Diff, const _Ty &)"
> function template "std::_Enable_if<<expression>, void>::_Result
> __cdecl std::fill_n(_OutIt, _Diff, const _Ty &)"
> argument types are: (unsigned char , int, unsigned char)
> std::fill_n(this->offsets_, uchar_max + 1, this->length_);
I just changed this to:
std::fill_n(static_cast<unsigned char *>(this->offsets_), uchar_max
+ 1, this->length_);
That should remove the (bogus) ambiguity. Change committed to HEAD and RC.
-- Eric Niebler Boost Consulting www.boost-consulting.com