Update of explicit failures markup (more platforms added, that have bad float types conversion implementation)

[SVN r74760]
This commit is contained in:
Antony Polukhin 2011-10-06 16:47:38 +00:00
parent dfd3d0229a
commit e4d2cfbb26

View file

@ -646,6 +646,7 @@
<toolset name="borland-5.8*"/>
<toolset name="borland-5.9*"/>
<toolset name="intel-darwin-11.*"/>
<toolset name="intel-darwin-12.0"/>
<toolset name="gcc-3.4.0*"/>
<toolset name="gcc-4.1*"/>
<toolset name="gcc-4.2.0*"/>
@ -670,8 +671,13 @@
<mark-failure>
<toolset name="gcc-mingw-4.*"/>
<toolset name="vacpp-10.1"/>
<toolset name="vacpp"/>
<toolset name="msvc-9.0~stlport5.2"/>
<toolset name="msvc-9.0~wm5~stlport5.2"/>
<toolset name="intel-darwin-11.*"/>
<toolset name="intel-darwin-12.0"/>
<note author="Antony Polukhin">
Some compilers and STL realizations convert double and long double types with bigger precision loss than minimal. Such failures are not a lexical_cast, but a compiler fault.
Some compilers and STL realizations convert double and long double types with bigger precision loss than minimal (or even round to infinity). Such failures are not a lexical_cast, but a compiler fault.
</note>
</mark-failure>
</test>