[Insight-developers] Sun-CC changes heads up

[Tom Vercauteren]

Hi Luis,

That's great!

For the continuous build, one potential option would be to have only a
"partial" continuous build on this machine. What I mean is that it
could only compile but not run the tests in this continuous mode. Also
it doesn't need to start with a clean empty build tree every morning.

Does that seem doable?

Tom

On Tue, Apr 7, 2009 at 22:28, Luis Ibanez <luis.ibanez at kitware.com> wrote:
>
> Bill, Tom,
>
>
> The "Sun CC compiler + -stlport4 option" Saga seems
> to be finally over.
>
>
> After retouching several hundreds of files, it seems
> that the build in dashsun1 is green again (as far as
> compilation errors go, note that we still have a bunch
> of warnings to deal with...)
>
>
> As Bill pointed out the major change is from now on,
> we should use the vcl_ prefix for (almost) all math
> functions (cos, sin, atan, pow, ceil, floor,...).
>
>
>   Luis
>
>
>
> -----------------------
> Bill Lorensen wrote:
>>
>> Consistent use of the vcl functions versus the native functions is
>> probably a good side effect of this process.
>>
>> On Sun, Apr 5, 2009 at 10:34 PM, Luis Ibanez <luis.ibanez at kitware.com>
>> wrote:
>>
>>> Hi Bill,
>>>
>>> You are right, I should be adding the <string.h> header
>>> instead of replacing the existing <string>.
>>>
>>> What started as a quick fix is turning out to be a major
>>> revisiting of all the files with string and math functions.
>>>
>>> I have been replacing:
>>>
>>>  sin, cos, atan, acos, asin, pow, exp, log, sqrt
>>>
>>> with their vcl_ versions: vcl_sin, vcl_cos... etc.
>>>
>>> and adding header for the functions
>>>
>>> atoi
>>> atof
>>> memset
>>> memcpy
>>>
>>>
>>> I have been doing this by compiling simultaneously
>>> in Linux with gcc 4.2 and with CC under Sun.
>>>
>>> We still have to see how this goes with our usual friends:
>>> Borland and Visual Studio 6.0.
>>>
>>> I'll continue with this tomorrow...
>>>
>>>
>>>  Luis
>>>
>>>
>>> ---------------------
>>> Bill Lorensen wrote:
>>>
>>>> Luis,
>>>>
>>>> I noticed that you are replacing <string> with <string.h>. I
>>>> understand why you need to add string.h, but I thought <string> was
>>>> the include file for the stl string class. Shouldn't you just be
>>>> adding <string.h> and not removing <string>. I see that the continuous
>>>> compiler seem to accept your changes, so maybe I'm off base here.
>>>>
>>>> Bill
>>>>
>>>>
>>>> On Sun, Apr 5, 2009 at 7:35 AM, Luis Ibanez <luis.ibanez at kitware.com>
>>>> wrote:
>>>>
>>>>
>>>>> Hi Mathieu,
>>>>>
>>>>> Thanks for tracking the problem with file writing in the Sun compiler.
>>>>>
>>>>> Just to give you a heads up: Related to the changes that Tom made
>>>>> in order to support derived ImportImageContainer, we introduced a
>>>>> compiler flag: -stlport4 for the SunCC build, because the
>>>>> std::allocator
>>>>> in this compiler was missing the "destroy" method.
>>>>>
>>>>> This new flag triggered a chain of additional changes (and still more
>>>>> to come). Hopefully they will not interfere with your exploration,
>>>>> but...
>>>>> just in case, if you see some dramatic (good or bad) change, it may
>>>>> be related to these other ongoing modifications.
>>>>>
>>>>>
>>>>>  Luis
>>>>> _______________________________________________
>>>>> Powered by www.kitware.com
>>>>>
>>>>> Visit other Kitware open-source projects at
>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>
>>>>> Please keep messages on-topic and check the ITK FAQ at:
>>>>> http://www.itk.org/Wiki/ITK_FAQ
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.itk.org/mailman/listinfo/insight-developers
>>>>>
>>>>
>>>>
>>
>