warnings=off is not working

David Abrahams <daveboost-consulting.com> writes:

> Vladimir Prus <ghostcs.msu.su> writes:
>
>> On Friday 27 January 2006 19:36, David Abrahams
wrote:
>>> "Reece Dunn" <msclrhdhotmail.com> writes:
>>> > David Abrahams wrote:
>>> >>I get loads of warnings when compiling
with msvc with warnings=off
>>> >
>>> > warnings=off is using the /W0 flag which
is what the visual studio
>>> > solution files use to turn off warnings,
so this is doing the right
>>> > thing.
>>> >
>>> > NOTE: the warning feature is an incidental
feature, so it is not
>>> > propagated.
>>>
>>> That is bad.  When users build something with
warnings=off from the
>>> command-line and get a pile of warnings, they
will think, like I did,
>>> that something is wrong.  And it is.
>>
>> In fact, incidental features *are* propagated, at
least they are supposed to 
>> be.
>>
>> Dave, do you have a minimal example reproducing the
problem? 
>
> I'm afraid not.  I was building the Boost.Python
tests, though, so it
> should be easy enough to reproduce.

Maybe this gives you a clue.  Notice that the first TU is
/W0 and the
second one is /W4:

  msvc.compile.c++
..\..\..\bin.v2\libs\python\test\select_from_python_t
est.test\msvc-6.0\debug\type_id.obj

      call "C:\Program Files\Microsoft Visual
Studio\VC98\Bin\VCVARS32.BAT"  >nul
  cl /Zm800 -nologo -TP  /Z7 /Od /Ob0 /W0 /GR /MDd /GB /EHs 
"..\..\..\bin.v2\libs\python\test\select
_from_python_test.test\msvc-6.0\debug\type_id.obj.rsp&qu
ot; -c
-Fo"..\..\..\bin.v2\libs\python\test\select_fro
m_python_test.test\msvc-6.0\debug\type_id.obj"  

  type_id.cpp
  ..\..\..\libs\python\src\converter\type_id.cpp(181)
: error C2491: '<<' : definition of dllimport
function not allowed
  ..\..\..\libs\python\src\converter\type_id.cpp(188)
: error C2491: '<<' : definition of dllimport
function not allowed
  msvc.compile.c++
..\..\..\bin.v2\libs\python\build\msvc-6.0\debug\sl
ice.obj

      call "C:\Program Files\Microsoft Visual
Studio\VC98\Bin\VCVARS32.BAT"  >nul
  cl /Zm800 -nologo -TP  /Z7 /Od /Ob0 /W4 /GR /MDd /GB /EHs 
"..\..\..\bin.v2\libs\python\build\msvc-
6.0\debug\slice.obj.rsp" -c
-Fo"..\..\..\bin.v2\libs\python\build\msvc-6.0
debug\slice.obj"  

  slice.cpp

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

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

On Friday 10 February 2006 21:26, David Abrahams wrote:

> >> Dave, do you have a minimal example
reproducing the problem?
> >
> > I'm afraid not.  I was building the Boost.Python
tests, though, so it
> > should be easy enough to reproduce.
>
> Maybe this gives you a clue.  Notice that the first TU
is /W0 and the
> second one is /W4:


>       call "C:\Program Files\Microsoft Visual
Studio\VC98\Bin\VCVARS32.BAT"
>  >nul cl /Zm800 -nologo -TP  /Z7 /Od /Ob0 /W0 /GR
/MDd /GB /EHs 
> "..\..\..\bin.v2\libs\python\test\select
_from_python_test.test\msvc-6.0\de
>bug\type_id.obj.rsp" -c
>
-Fo"..\..\..\bin.v2\libs\python\test\select_fro
m_python_test.test\msvc-6.0\
>debug\type_id.obj"
.......
>       call "C:\Program Files\Microsoft Visual
Studio\VC98\Bin\VCVARS32.BAT"
>  >nul cl /Zm800 -nologo -TP  /Z7 /Od /Ob0 /W4 /GR
/MDd /GB /EHs 
> "..\..\..\bin.v2\libs\python\build\msvc-
6.0\debug\slice.obj.rsp" -c
>
-Fo"..\..\..\bin.v2\libs\python\build\msvc-6.0
debug\slice.obj"
>
>   slice.cpp

Running "bjam" with --debug-building showed me
that the <warnings>off was just 
not propagated to the boost_python targets, and the reason
was the the 
"warnings" feature was no marked as propagated.
Fixed in CVS now, let me know 
if this fixes to problems for you.

- Volodya


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build