Hello,

I did ln -s /usr/local/lib/libboost-xx-mt-1_35.dylib /usr/local/lib/ 
libboost-xx-mt.dylib

which did not work.

So, I scrubbed around in the main SConstruct file and found this line:

['system', 'boost/system/system_error.hpp', True], # uncomment this on  
Darwin + boost_1_35

which I uncommented, and still could not find the boost libraries.  
Then, I saw this:

thread_suffix = '-mt'

and thought, to just try adding the version number in there like so:

thread_suffix = '-mt-1_35'

that worked (as sketchy as it is)! At least part way.. g++ is printing  
out a lot of stuff and then:

ld: library not found for -lboost_python-mt
collect2: ld returned 1 exit status
scons: *** [bindings/python/_mapnik.so] Error 1
scons: building terminated because of errors.

Well at least I got past the first part, that was driving me nuts.




> Hi Colin,
>
> Are you positive your symlinks are correct?
>
> I am using boost built from macports, so about all I can offer is  
> that the symlinking in version 1_34 was a tricky problem.  You might  
> benefit from taking a look a the boost port file to see how they  
> managed the correct symlinking:
>
> http://trac.macosforge.org/projects/macports/browser/trunk/dports/devel/boost/Portfile
>
> Overall, keep your notes on building boost for mapnik: posting to  
> mapnik trac about how to get 10.5 running with boost would be a  
> really useful thing (when you get there).
>
> Dane

>> Thank you for your response.
>> I knew I should have added that to my steps in the last email..  
>> Sorry.
>> I thought that the version numbers might be the issue as well and
>> tried sym linking them out but that doesn't help. Same issue.

>>> The issue is that the library is named libboost_filesystem-
>>> mt-1_35.dylib while scons is looking for a library named:
>>> libboost_filesystem-mt.dylib
>>>
>>> So, just link the libboost-XX-mt-1_35.dylib files in /usr/local/lib
>>> to libboost-XX-mt.dylib (without the -1_35) and it should be fine.
>>>
>>> Claire

>>>> Hi,
>>>>
>>>> I am new to this list and Mapnik. I've been attempting to install
>>>> on OS X Leopard but without success and was hoping someone else has
>>>> been through this and would have some pointers.
>>>> First, I followed the instructions on the Mapnik site and installed
>>>> all the necessary libraries into /usr/local/lib (and /usr/local/
>>>> include for the boost headers).
>>>> Then I tried running scons/scons.py without any options and got  
>>>> this:
>>>>
>>>> scons: Reading SConscript files ...
>>>> Building on Darwin ...
>>>> Checking for C library m... yes
>>>> Checking for C library ltdl... yes
>>>> Checking for C library png... yes
>>>> Checking for C library tiff... yes
>>>> Checking for C library z... yes
>>>> Checking for C library jpeg... yes
>>>> Checking for C library proj... yes
>>>> Checking for C library iconv... yes
>>>> Checking for C library pq... no
>>>> Checking for C++ library gdal... no
>>>> Checking for C++ library boost_filesystem-mt... no
>>>> Could not find header or shared library for boost filesystem,
>>>> exiting!
>>>>
>>>> Ok, so there is an issue finding the boost libraries I figured. I
>>>> can see libboost_filesystem-mt-1_35.dylib sitting there in /usr/
>>>> local/lib
>>>>
>>>> Next I removed the scons caches and ran scons.py again but with
>>>> these options: BOOST_LIBS=/usr/local/lib BOOST_INCLUDES=/usr/local/
>>>> include/boost-1_35
>>>> which produced the same output. Always failing on the filesystem
>>>> library.
>>>>
>>>> After searching the web for help I came across some references to
>>>> Boost not being properly built as universal binaries. I originally
>>>> downloaded and built Boost with regular:
>>>> ./configure --prefix=/usr/local
>>>> make
>>>> sudo make install
>>>>
>>>> But maybe there are more hoops for OS X and Intel Macs?
>>>>
>>>> At this point I am at a loss. Thank you in advance for any help,
>>>> and my apologies if this is common knowledge.
_______________________________________________
Mapnik-users mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/mapnik-users

Reply via email to