[cmake-developers] FindGTest.cmake and CMP0064

[David Chen]

> Date: Wed, 18 Nov 2015 10:44:24 -0500
> From: Brad King <brad.king at kitware.com>
> To: cmake-developers at cmake.org
> Subject: Re: [cmake-developers] FindGTest.cmake and CMP0064
> Message-ID: <564C9CD8.9090900 at kitware.com>
> Content-Type: text/plain; charset=utf-8
> 
> On 11/17/2015 11:37 AM, Rolf Eike Beer wrote:
>> Am 17.11.2015 17:08, schrieb David Chen:
>>> We?ve been getting a warning from FindGTest.cmake about CMP0064 when
>>> building SimpleITK.  The warning occurs at lines 127, 129 and 131 when
>>> the variable ${test_type} is equal to ?TEST?.  With CMP0064 this TEST
>>> could be interpreted as commands in the if() and elseif() statements
>>> unless the policy is set to OLD.
> 
> With CMP0054 and CMP0064 both set to NEW then the original code
> would work as expected.  However, CMake's modules must tolerate
> many policy (non-)setting combinations.
> 
>> The correct fix would probably to just remove the dereference, i.e. 
>> change
>> 
>>   if(${test_type} STREQUAL "TEST_P")
>> 
>> to
>> 
>>   if(test_type STREQUAL "TEST_P")
> 
> The code
> 
>  if(test_type STREQUAL "TEST")
> 
> could still trigger CMP0064 warnings.  We can use the style
> 
>  if("x${test_type}" STREQUAL "xTEST_P")
> 
> to avoid both CMP0054 and CMP0064.  Done here:
> 
> FindGTest: Refactor test type checks to avoid cases triggering CMP0064
> https://cmake.org/gitweb?p=cmake.git;a=commitdiff;h=b5e7b22d
> 
> David, please verify that this approach still works for your case.

Your patch works fine for me.

Thanks,
Dave


David T. Chen, PhD                      [MSC contractor]
mailto:dchen at mail.nih.gov <mailto:dchen at mail.nih.gov>               http://lhncbc.nlm.nih.gov/personnel/david-chen <http://lhncbc.nlm.nih.gov/personnel/david-chen>
phone:301.435.3264 <tel:301.435.3264>                      iphone:301.524.3174 <tel:301.524.3174>
Office of High Performance Computing and Communications
National Library of Medicine


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/cmake-developers/attachments/20151118/ed892892/attachment.html>