[Insight-developers] RE: SGI Flag set

[William A. Hoffman]

Slow dashboard rollups for ITK.  It takes a very long time to roll up a dashboard
in ITK.  I think a majority of the time is spent building the TestOverview.html.
ITK has the highest number of tests for any of the projects at Kitware:

ITK 50,000 tests
VTK 13,000 tests. 
ParaView has about 6,000 
CMake  3,000 tests run.


I don't think anyone is really looking at this page anyway because it takes
too long to load.   For smaller numbers of tests like CMake it comes up
very quick.  But, if I click on the tests overview for ITK it takes too long
to load anyway.  By the end of the day as more tests are added it takes longer
and longer for the rollup.  

I have been running a rollup that has taken over an hour, and is still going.
Perhaps we should just remove the testing overview for ITK.  Is that possible 
with Dart?


-Bill


At 02:40 PM 8/16/2004, Lorensen, William E (Research) wrote:
>Bill,
>
>That sounds like a great solution. We could add similar pragmas to other
>isolated itk files. Let's see how the brain dead compilers at GE handle the
>pragmas tonight.
>
>Bill 
>
>-----Original Message-----
>From: William A. Hoffman [mailto:billlist at nycap.rr.com]
>Sent: Monday, August 16, 2004 1:57 PM
>To: Lorensen, William E (Research);
>insight-developers at public.kitware.com
>Subject: RE: [Insight-developers] RE: SGI Flag set
>
>
>I think I found another way to do this.  
>The SGI compiler supports pragma to turn off warnings.
>
>I am adding the following to vcl_compiler.h:
>
>#   define VCL_SGI_CC_730
>// 1116  Non-void function "std::__malloc_alloc_template<0>::_S_oom_malloc"
>(declared
>//    at line 163) should return a value.
>// 3505  setjmp not marked as unknown_control_flow because it is not
>declared as a
>// 3438 ceil not marked as no_side_effects because it is not declared as a
>function
>#   pragma set woff 3438
>#   pragma set woff 1116
>#   pragma set woff 3505
>
>So, if code includes vcl code, then it will turn off the warnings.
>
>
>-Bill
>
>At 11:02 AM 8/16/2004, Lorensen, William E (Research) wrote:
>>Looks like vxl has most of the trouble. We could turn it on in that tree.
>If
>>the warnings are isolated in other parts of itk we could also do that
>there.
>>
>>-----Original Message-----
>>From: William A. Hoffman [mailto:billlist at nycap.rr.com]
>>Sent: Monday, August 16, 2004 10:08 AM
>>To: Lorensen, William E (Research);
>>insight-developers at public.kitware.com
>>Subject: Re: [Insight-developers] RE: SGI Flag set
>>
>>
>>How would people feel about excluding this warning on the SGI by default?
>>The problem is that users will see these warnings when they build ITK and
>>think something is wrong.
>>
>>-woff 1116 
>>
>>I know it is a good warning, but other compilers that do not have bad code
>>in STL headers will catch the warnings in the dashboards.  It really should
>>only affect folks developing on the SGI, and I don't think there are many
>>of them.
>>
>>-Bill
>>
>>
>>At 08:14 AM 8/16/2004, Lorensen, William E (Research) wrote:
>>>Dart is also blocking:
>>>
>>>cc-3505 CC: WARNING File = /usr/include/internal/setjmp_core.h, Line = 74
>>>
>>>
>>>-----Original Message-----
>>>From: William A. Hoffman [mailto:billlist at nycap.rr.com]
>>>Sent: Sunday, August 15, 2004 7:41 PM
>>>To: Lorensen, William E (Research); Miller, James V (Research); 'Luis
>>>Ibanez'; Insight Developers List
>>>Subject: SGI Flag set
>>>
>>>
>>>I am down to these flags:
>>>
>>>    SET(ITK_REQUIRED_CXX_FLAGS "${ITK_REQUIRED_CXX_FLAGS} -woff
>>>15,84,1110,1209,1424,1468,3102,3210,3439 -ptused -no_prelink")
>>>    SET(ITK_REQUIRED_LINK_FLAGS "${ITK_REQUIRED_LINK_FLAGS}  -Wl,-woff15")
>>>
>>>http://www.itk.org/Testing/Dashboard/20040815-0500-Nightly/Dashboard.html
>>>
>>>However, the rapture build still has 400+ warnings in the STL directory.
>>>What is being done at IOWA and CRD to block these warnings?  Is that
>>>a Dart block?
>>>
>>>-Bill
>>>_______________________________________________
>>>Insight-developers mailing list
>>>Insight-developers at itk.org
>>>http://www.itk.org/mailman/listinfo/insight-developers 
>_______________________________________________
>Insight-developers mailing list
>Insight-developers at itk.org
>http://www.itk.org/mailman/listinfo/insight-developers