[Insight-developers] A massive number of new warnings

[Hans Johnson]

I was toying with the idea of adding something like is shown at the bottom
of this message.

We could safely build up a larger set of "recommended" compiler flags for
each of the release types rather than just "-g" and "-O3".

I think that -Wall should be the default flag, and that developers have to
explicitly turn it off.  That way it's not just the 4 of us triaging all
these complier warnings.

NOTE:  The code below is not working, but I think it gets the point across.

Hans




#To create a portable build system, it is best to not test for platforms,
but to test for features.
#
#Instead of testing "if Windows then do this", test for "if the
-Wno-invalid-offsetof flag works then use it". You can do that with the
CheckCCompilerFlag module, for example:

include(CheckCCompilerFlag)
include(CheckCXXCompilerFlag)

set(C_RELEASE_DESIRED_FLAGS "")
set(CXX_RELEASE_DESIRED_FLAGS "")

# -Wall
check_c_compiler_flag(-Wall HAS_ALL_WARNINGS_C)
if (HAS_ALL_WARNINGS_C)
    set(CMAKE_C_FLAGS  "${CMAKE_C_FLAGS} -Wall")
endif()

check_cxx_compiler_flag(-Wall HAS_ALL_WARNINGS_CXX)
if (HAS_ALL_WARNINGS_CXX)
    set(CMAKE_CXX_FLAGS  "${CMAKE_C_FLAGS} -Wall")
endif()

## -pedantic
check_c_compiler_flag(-pedantic HAS_PEDANTIC_C)
if (HAS_PEDANTIC_C)
    set(CMAKE_C_FLAGS  "${CMAKE_C_FLAGS} -pedantic")
endif()

check_cxx_compiler_flag(-pedantic HAS_PEDANTIC_CXX)
if (HAS_PEDANTIC_CXX)
    set(CMAKE_CXX_FLAGS  "${CMAKE_C_FLAGS} -pedantic")
endif()

## -Wextra
check_c_compiler_flag(-Wextra HAS_EXTRA_WARNINGS_C)
if (HAS_EXTRA_WARNINGS_C)
    set(CMAKE_C_FLAGS  "${CMAKE_C_FLAGS} -Wextra")
endif()

check_cxx_compiler_flag(-Wextra HAS_EXTRA_WARNINGS_CXX)
if (HAS_EXTRA_WARNINGS_CXX)
    set(CMAKE_CXX_FLAGS  "${CMAKE_C_FLAGS} -Wextra")
endif()

## -Wshadow
check_c_compiler_flag(-Wshadow HAS_SHADOW_WARNINGS_C)
if (HAS_SHADOW_WARNINGS_C)
    set(CMAKE_C_FLAGS  "${CMAKE_C_FLAGS} -Wshadow")
endif()

check_cxx_compiler_flag(-Wshadow HAS_SHADOW_WARNINGS_CXX)
if (HAS_SHADOW_WARNINGS_CXX)
    set(CMAKE_CXX_FLAGS "${CMAKE_C_FLAGS} -Wshadow")
endif()

## -Wlong-long  We should not use the long-long type in ITK
check_c_compiler_flag(-Wlong-long HAS_LONG_LONG_WARNINGS_C)
if (HAS_LONG_LONG_WARNINGS_C)
    set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wlong-long")
endif()

check_cxx_compiler_flag(-Wlong-long HAS_LONG_LONG_WARNINGS_CXX)
if (HAS_LONG_LONG_WARNINGS_CXX)
    set(CMAKE_CXX_FLAGS "${CMAKE_C_FLAGS} -Wlong-long")
endif()


check_cxx_compiler_flag(-Wno-invalid-offsetof HAS_NO_INVALID_OFFSETOF)
if (HAS_NO_INVALID_OFFSETOF)
    set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wno-invalid-offsetof")
endif()

message(${CMAKE_C_FLAGS})
message(${CMAKE_CXX_FLAGS})



On 1/6/11 2:55 PM, "Luis Ibanez" <luis.ibanez at kitware.com> wrote:

> I just added -Wall and -Wextra to the builds in
> zion, thurmite and eldorado.
> 
> It's time to follow Bill L. in the zero-warning campaign      :-)
> 
> 
>      Luis
> 
> 
> ----------------------------------
> On Wed, Jan 5, 2011 at 6:32 PM, Sean McBride <sean at rogue-research.com> wrote:
>> That build is only using -Wall and -Wextra, which my gcc dashboards also
>> use, and yet mine show a much smaller number of warnings.  Strange that
>> a newer gcc would not warn where an older gcc does, one reason could be
>> that they were false positives in the older version.  In any case, there
>> are less warnings than it seems since some are caused by headers and so
>> repeated in many source files.
>> 
>> On Wed, 5 Jan 2011 17:01:46 -0500, Bill Lorensen said:
>> 
>>> I think they are valid warnings.
>>> 
>>> On Wed, Jan 5, 2011 at 4:20 PM, Hans Johnson <hans-johnson at uiowa.edu> wrote:
>>>> This looks like the Insight-gcc-3.3. Has a definitive decision been made
>>>> regarding compilers less than 4?  I thought that less than 3.4 was excluded
>>>> because of incompatibility issues.
>> 
>> 
>> 

-- 
Hans J. Johnson, Ph.D.
Assistant Professor
200 Hawkins Drive
T205 BT, The University of Iowa
Iowa City, IA 52242

hans-johnson at uiowa.edu
PHONE: 319 353 8587