[pgrouting-dev] pgRouting 2.0 using OSX homebrew

[Stephen Woodbridge]

On 5/8/2013 10:01 PM, Alec Gosse wrote:
> Steve,
>
> The pull request I submitted puts these files in the lib directory:
> librouting.so            librouting_dd.so         librouting_tsp.a
> librouting_bd.so         librouting_ksp.a
>
> Cmake clearly thinks that tsp and ksp should be built as static libraries, while it calls the others shared modules, but I couldn't figure out what gave it this idea. I just checked on my linux machine, and it's doing the same thing there, so it doesn't seem to be mac-specific.

Yeah, I'm getting the same thing here. I take a look at this. it should 
be easy to fix.

Thanks,
   -Steve

> Best,
> Alec
>
>
>
>
> On May 8, 2013, at 8:14 PM, Stephen Woodbridge <woodbri at swoodbridge.com> wrote:
>
>> On 5/8/2013 7:49 PM, sanak wrote:
>>> Hi Alec,
>>>
>>> Thanks for your pull request!
>>
>> I'm merging the pull request and will push it shortly.
>>
>> I have one question, does anyone know what the commented out code is supposed to do? IE: why is there in the first place? Commenting it out seems to be OK on Linux.
>>
>> I google about but could not find an answer regarding why its there and I don't know C++. I suppose I could ask on the boost list if no one knows.
>>
>>> 2013/5/9 Alec Gosse <alec at thegosses.com <mailto:alec at thegosses.com>>
>>>
>>>     Thank you all for your guidance. I've submitted a pull request for
>>>     my changes. As I noted in the text of the request, librouting_tsp
>>>     and librouting_ksp are being built as static rather than dynamic
>>>     libraries, so make install fails since they have the extension .a
>>>     rather than .so. Any thoughts on this would be most appreciated.
>>>
>>>     The only other trouble I've run across is a very long list of
>>>     warnings like this while compiling:
>>>     /usr/local/src/pgrouting/src/driving_distance/src/drivedist.c:220:34: warning:
>>>     use of GNU old-style field designator extension [-Wgnu-designator]
>>>        edge_columns_t edge_columns = {id: -1, source: -1, target: -1,
>>
>> Before we #ifdef about this, it seems that there might be a better way to do this initialization, ie: NOT using the GNU old-style field designator extension but rather an ANSI C way of doing this.
>>
>> I'll take a stab at it but if anyone knows how to change these please do a pull request.
>>
>> Hi Alec,
>>
>> So your pull request is generating *.a and not *.so, is this correct?
>>
>> Thanks,
>>   -Steve
>>
>>> I guess that above warning cause is Clang(LLVM)'s g++ compatibility,
>>> so, more custom CMake settings may be necessary.
>>> I will check about it in this weekend (or later).
>>>
>>> Thanks,
>>> Sanak
>>>
>>>
>>> _______________________________________________
>>> pgrouting-dev mailing list
>>> pgrouting-dev at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/pgrouting-dev
>>>
>>
>> _______________________________________________
>> pgrouting-dev mailing list
>> pgrouting-dev at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/pgrouting-dev
>
> _______________________________________________
> pgrouting-dev mailing list
> pgrouting-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/pgrouting-dev
>