On Thu, Aug 28, 2008 at 5:13 PM, mabshoff <[EMAIL PROTECTED]> wrote:

> On Aug 28, 2:54 pm, Jason Grout <[EMAIL PROTECTED]> wrote:
> > > The Python extension is linked against nauty object files, and gcc
> > > complains these should be compiled with the flag "-fPIC" to make
> > > relocatable code.  nauty does not have this flag normally, so I
> modified
> > > the makefile to add this flag using sed.  Is the right way to handle
> > > this problem?  Is it okay for the executables that Jason is installing
> > > in bin to be compiled with -fPIC, or should two separate compilings be
> done?
> >
> > Others (mabshoff?) can comment on what to do about -fPIC.  However, my
> > guess is that using sed at spkg install time is probably not a good
> > idea, or at least not in line with official practices.  It seems that
> > the official way to change the makefile is to store the new makefile in
> > the patches directory and copy it over to the src directory at spkg
> > install time.
>
> Using sed in general is fine since various packages in Sage do depend
> on it at build time. But Jason is right that it should be changed in
> the makefile, put into patches and then copied over in spkg-install at
> build time. If someone opens a ticket I will review the spkg and
> hopefully someone else will look at the Python bindings. Depending on
> how things are we might also upgrade to the latest upstream release at
> the same time.
>

I'll try to rework the change to the Makefile in this way.  Is the Makefile
ever changed when running ./configure?  If so, then I'll need to make a
patch that's applied to the Makefile as opposed to just copying over a
modified Makefile.

Jason's spkg has the latest upstream release in it.

There's a few changes that need to be made in the Python bindings (mainly
the way that edge-orbits are calculated; should that be done in a separate
function call as I have it now, or done in one single function call with
options?), so I don't anticipate this release getting put into the next
release.  Mainly I needed to know how to handle this -fPIC problem, which is
why I sent the mail to the list today. :)  However, I think the (C code for
the Python bindings) is pretty useful at this point for people who wanted to
use it (and they may have helpful comments about how to structure the code).

Thanks!
Stephen

--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to