[vtk-developers] dashboard failure with Java wrapping and Obj-C++?

[Chris Harris]

David,

Can you give me an example of one of the OUTPUTS that don't match the
actual location?

Thanks,

Chris

On Mon, Apr 16, 2012 at 9:22 AM, Chris Harris <chris.harris at kitware.com> wrote:
> I can take a look at this.
>
> Chris
>
> On Fri, Apr 13, 2012 at 6:36 PM, David Gobbi <david.gobbi at gmail.com> wrote:
>> On Fri, Apr 13, 2012 at 10:48 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>>> On Fri, Apr 13, 2012 at 9:18 AM, Marcus D. Hanwell
>>> <marcus.hanwell at kitware.com> wrote:
>>>
>>>> That looks like the parallel build issues we have been seeing. You
>>>> could confirm by running make in that build tree a second time, this
>>>> was fragile and I think modularization has caused some of the duct
>>>> tape to fray. We would appreciate help if anyone knows the Java
>>>> wrapping/parallel build issues.
>>>
>>> I located this previous fix for the java parallel build issue.  I'm
>>> not 100% certain what the problem was.
>>> http://vtk.org/gitweb?p=VTK.git;a=commit;h=875f8232
>>
>> I suspect that part of the problem is that for several of the java
>> add_custom_command lines in Wrapping/Java/CMakeLists.txt
>> (and perhaps elsewhere), the OUTPUTS does not match the
>> actual location of the output file generated by javac.  Someone
>> should go through the cmake files and verify that each one of
>> these custom commands is correct.
>>
>> This would explain another problem that I've seen with the
>> java compile: sometimes when I do "make", I see java files
>> being compiled that have nothing to do with the changes that
>> I've made.  Pointless re-execution of commands suggests
>> that the OUTPUTS file wasn't actually generated (or was
>> generated, but put in the wrong directory).
>>
>>  - David