[ros-users] Errors compiling gmapping

Bhaskara Marthi bhaskara at willowgarage.com
Sun May 6 18:56:19 UTC 2012


On Sun, May 6, 2012 at 9:07 AM, Kevin Walchko <kevin.walchko at gmail.com> wrote:
> Sorry …
>
> Yes a patch would fix the first issue.
>
> The second issue of path errors ... why is my compiler not finding the header file? It exists, but which Makefile or CMakeLists.txt or whatever else should I add the path too? I assume Linux finds all of the headers, but I don't see anything wrong when looking at the files. Was hoping someone who had helped write this may be able to point me to the correct place and suggest a simple fix to make work.
>
Perhaps you commented out something in the CMakeLists.txt?  I'd
suggest going the first route of modifying the patch file stored in
the gmapping package to include your fix in addition, and not
modifying any makefiles.  If that isn't working, try VERBOSE=1 make to
verify the path being used for compilation.
- b

>
> On May 6, 2012, at 12:01 PM, Jack O'Quin wrote:
>
>> On Sun, May 6, 2012 at 10:54 AM, Kevin Walchko <kevin.walchko at gmail.com> wrote:
>>> OSX 10.7.2
>>> Fuerte
>>> XCode 4.3.2
>>>
>>>
>>> I found a strange error in gmapping under build/gmapping_export/gridfastslam/gfs2rec.cpp, for some reason i gets redefined (line 148)?
>>>
>>> struct ResampleRecord: public Record{
>>>        vector<unsigned int> indexes;
>>>        virtual void read(istream& is){
>>>                is >> dim;
>>>                for (unsigned int i=0; i< dim; i++){
>>>                        unsigned int i; <<<<<<< why redefined???
>>>                        is >> i;
>>>                        indexes.push_back(i);
>>>                }
>>>        }
>>> };
>>>
>>> Unfortunately all of this code appears to be in a tar ball so when I fix the code and do a make clean, it all gets erased and downloaded again.
>>>
>>> I tried not doing the make clean, but it still fails because the makefile tries to reapply patches to code it already applied patches too. I commented out the code in cmakelists.txt so it can move on (it failed because of the failed patch re-applies), but then the next errors are the paths are not setup correctly.
>>>
>>> /Users/kevin/ros_sandbox/stacks/geometry/tf/include/tf/tf.h:91:23: warning:
>>>      in-class initializer for static data member of type 'const double' is a
>>>      GNU extension [-Wgnu]
>>>  static const double DEFAULT_CACHE_TIME = 10.0;  //!< The default ...
>>>                      ^                    ~~~~
>>> In file included from /Users/kevin/ros_sandbox/stacks/slam_gmapping/gmapping/src/slam_gmapping.cpp:107:
>>> /Users/kevin/ros_sandbox/stacks/slam_gmapping/gmapping/src/slam_gmapping.h:28:10: fatal error:
>>>      'gmapping/gridfastslam/gridslamprocessor.h' file not found
>>> #include "gmapping/gridfastslam/gridslamprocessor.h"
>>>         ^
>>> 1 warning and 1 error generated.
>>>
>>> This code is hosed … I assume it runs on Linux right? How can I fix the include path?
>>> Thanks!
>>
>> Sounds like you need to fix it by adding a patch.
>> --
>>  joq
>> _______________________________________________
>> ros-users mailing list
>> ros-users at code.ros.org
>> https://code.ros.org/mailman/listinfo/ros-users
>
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users



-- 
Bhaskara Marthi
Research Scientist
Willow Garage Inc.
650-475-2856



More information about the ros-users mailing list