[Distutils] setup.py build_ext --rpath behaviour

[Floris Bruynooghe]

On Thu, Apr 16, 2009 at 11:42:28AM +0100, Floris Bruynooghe wrote:
> Now when using the --rpath option to the build_ext command
> distutils.unixcompiler.UnixCCompiler.runtime_library_dir_option() will
> use some heuristics to figure out which option to pass to compiler to
> get the runpath in (i.e. "-R" or "-Wl,-R" etc).  I'm going to argue
> that this needs to be extened pass in -Wl,--enable-new-dtags,-R if the
> GNU linker is used so that the newer and better RUNPATH gets put into
> the shared objects all the time.  (I don't yet know how to detect the
> GNU linker but would like a consensus on the desired behaviour before
> looking into this).

The patch for this is attached to http://bugs.python.org/issue5900
Feel free to give feedback/criticism if you care about this.


> The second issue with build_ext --rpath is on AIX.  Again some
> background on AIX shared objects, AIX is not SysV-like and uses the
> XCOFF binary format instead of ELF.  Therefore they don't have a RPATH
> or RUNPATH, but they do have a think called LIBPATH which does
> something similar.  The difference between XCOFF's LIBPATH and ELF's
> RUNPATH is that AIX's runtime linker does not have a default search
> path, hence the full search path needs to be encoded into the LIBPATH
> of the shared object.
> 
> Now I would propose for build_ext --rpath to encode the LIBPATH when
> used on AIX since that is the correct thing to do IMHO.

I was kind of hoping someone would pick up on this and at least argue
about how it is not a good thing to try and convert an option for one
concept (RPATH/RUNPATH) onto another one that isn't quite the same
(LIBPATH).  The side-effects are just not as clearly defined and open
to interpretation.

Anyway, having given this some more tought I now think it's not a good
thing.  Instead I'm going to make a patch that raises and exception if
the --rpath option is used on AIX as it's behaviour is undefined[0].  If
a user does want -blibpath:/some/path to be passed on to the linker
they can still use LDFLAGS.


Other opinions would be appreciated.

Regards
Floris


[0] The linker will actually simply ignore it lurring the user into
    false security.  Only when -bsvr4 is passed to the linker too will
    it be treated identical as -blibpath, but this is again the same
    fuzzy mapping as a SysV user will not be used to passing in the
    default search path to -R/--rpath and hence probably not end up
    with what they wanted too.

-- 
Debian GNU/Linux -- The Power of Freedom
www.debian.org | www.gnu.org | www.kernel.org