[Insight-developers] MAC Compile errors
Bradley Lowekamp
blowekamp at mail.nih.gov
Mon Jul 14 15:24:33 EDT 2008
Hi all,
This is looking more like a cmake issue now.
I have looked at Insight/Utilities/vxl/core/vxl_config.h.in under a
variety of different conditions.
Usually something like this is produced:
#define VXL_HAS_INT_32 1
#define VXL_INT_32_STRING "int"
#if 1
typedef int vxl_int_32;
typedef signed int vxl_sint_32;
typedef unsigned int vxl_uint_32;
#else
# define vxl_int_32 void
# define vxl_sint_32 void
# define vxl_uint_32 void
#endif
The "unix make files" are always producing this. I just started a
experimental build with cmake 2.4.6, OSX10.5.4 and it's also producing
this correct file.
However with OSX 10.5.4 and cmake 2.6, the following is produced:
#define VXL_HAS_INT_32 0
#define VXL_INT_32_STRING "void"
#if 0
typedef void vxl_int_32;
typedef signed void vxl_sint_32;
typedef unsigned void vxl_uint_32;
#else
# define vxl_int_32 void
# define vxl_sint_32 void
# define vxl_uint_32 void
#endif
Also all "#define VXL_HAS_??" are 0 and "void"
I am looking at how these variables are generated now...
Brad
On Jul 14, 2008, at 11:08 AM, Luis Ibanez wrote:
>
> Hi Brad,
>
>
> The submissions from
>
>
> "mini1.nlm"
> "mini2.nlm"
>
> http://www.cdash.org/CDash/viewBuildError.php?buildid=123006
> http://www.cdash.org/CDash/viewBuildError.php?buildid=122976
>
>
> are still reporting the type "A" compilation errors of the form:
>
>
> /.../Insight/Utilities/vxl/core/vnl/vnl_matlab_header.h:16: error:
> variable or field 'type' declared void
>
> (same for lines 17,18,19,20)
>
> The problem seem to be that these platforms do not have defined
> the type "vxl_int_32".
>
>
> 16: vxl_int_32 type;
> 17: vxl_int_32 rows;
> 18: vxl_int_32 cols;
> 19: vxl_int_32 imag;
> 20: vxl_int_32 namlen;
>
>
> The default definition of this type is made in
>
> Insight/Utilities/vxl/core/vxl_config.h.in
>
>
> in line 67-77:
>
> #define VXL_HAS_INT_32 @VXL_HAS_INT_32@
> #define VXL_INT_32_STRING "@VXL_INT_32@"
> #if @VXL_HAS_INT_32@
> typedef @VXL_INT_32@ vxl_int_32;
> typedef signed @VXL_INT_32@ vxl_sint_32;
> typedef unsigned @VXL_INT_32@ vxl_uint_32;
> #else
> # define vxl_int_32 void
> # define vxl_sint_32 void
> # define vxl_uint_32 void
> #endif
>
>
> In VXL, this type is "manually" defined for the VisualStudio 6.0
> and 7.0 compiler
>
> cd Insight/Utilities/vxl/vcl
> $ grep -r VXL_HAS_INT_32 *
> config.stlport.win32-vc60/vxl_config.h:#define VXL_HAS_INT_32 1
> config.win32/vc60/vxl_config.h:#define VXL_HAS_INT_32 1
> config.win32/vc70/vxl_config.h:#define VXL_HAS_INT_32 1
>
>
> Overall, it seems that with this upgrade, Apple have succeded
> to make GCC as bad as Visual Studio 6.0.
>
> They may have a strange conception of what "progress" means :-/
>
>
> Is there an expectation that APPLE will fix this anytime soon ?
>
>
> If not, could you please convert the submissions fromm these two
> machines from Nightly to Experimentals ?
>
>
>
> Thanks
>
>
> Luis
>
>
> -------------------------
> Bradley Lowekamp wrote:
>> You are right. I thought "mini3" (which is a osx make build) was a
>> type "A" failure. But it's the type "B" link problem like the rest
>> of them.
>> Interestingly enough, on July they 9th, "mini1" and "mini2" also
>> had type "B" link errors, when they were running cmake 2.4.6 and
>> 10.5.4:
>> http://www.cdash.org/CDash/viewBuildError.php?buildid=119307
>> Brad
>> On Jul 11, 2008, at 10:48 AM, Sean McBride wrote:
>>> On 7/11/08 10:11 AM, Luis Ibanez said:
>>>
>>>> It seems that we are dealing with two different problems.
>>>
>>>>
>>>> A) Compilation errors in "mini1" and "mini2"
>>>
>>>>
>>>> (BTW: could you lease update the BUILDNAME of these
>>>
>>>> machines to indicate their institutional affiliation ?
>>>
>>>> something like mini1.nlm.gov, for example).
>>>
>>>>
>>>> http://www.cdash.org/CDash/viewBuildError.php?buildid=120638
>>>
>>>>
>>>> with errors of type:
>>>
>>>>
>>>> /.../Insight/Utilities/vxl/core/vnl/vnl_matlab_header.h:16: error:
>>>
>>>> variable or field 'type' declared void
>>>
>>>
>>> Those 2 builds are using CMake's Xcode generator. I believe they
>>> are
>>> they only ITK builds that do. That may be why only they are
>>> failing.
>>>
>>> --
>>> ____________________________________________________________
>>> Sean McBride, B. Eng sean at rogue-research.com <mailto:sean at rogue-research.com
>>> >
>>> Rogue Research www.rogue-research.com <http://www.rogue-research.com
>>> >
>>> Mac Software Developer Montréal, Québec, Canada
>>>
>>>
>> ========================================================
>> Bradley Lowekamp Lockheed Martin Contractor for
>> Office of High Performance Computing and Communications
>> National Library of Medicine blowekamp at mail.nih.gov <mailto:blowekamp at mail.nih.gov
>> >
========================================================
Bradley Lowekamp
Lockheed Martin Contractor for
Office of High Performance Computing and Communications
National Library of Medicine
blowekamp at mail.nih.gov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/mailman/private/insight-developers/attachments/20080714/8f72b138/attachment.htm>
More information about the Insight-developers
mailing list