transitioning from boost pyste and py++ python wrapping of a C++ template library

142 views
Skip to first unread message

tki...@gmail.com

unread,
Aug 14, 2013, 6:40:39 PM8/14/13
to xdr...@googlegroups.com
Hi,
I previously wrapped a lot of C++ classes and functions using boost pyste and py++ (both of which seem to be no longer updated). My C++ library in turn uses some basic boost header files. It seems like I should be able to move to Xdress rather easily based on my research so far (incl SciPy talk), but before I make the effort to try this I was wondering if anyone knew of potential issues I should be aware of in attempting to do this.
Thanks!

Anthony Scopatz

unread,
Aug 18, 2013, 3:01:29 AM8/18/13
to tki...@gmail.com, xdress
Hey!  

Sorry for taking so long to actually get to a response for this.  You happened to write while I was on vacation and have a huge backlog to get through =).

First off, awesome news that you are considering transitioning to xdress wrappers.  I have never used pyste or py++ so I am not really sure what issues there are in transitioning from these tools to xdress.  However, if I were to take a stab at what the major issue might be it is prabably going to be boost.  Right now, it may be more difficult if more of the API has boost constructs in it than not.  For example, if you are returning a boost::shared_ptr in all of your functions, it is not entirely clear what this means in Python...yet.  

However, I'd love / have a personal need for us to more explicitly support boost.  It would be great if your were willing to be the guinea pig here and try things out, report what doesn't work, and hopefully help us to address these issues.  I am very willing to help out here in whatever way I can.

Be Well
Anthony

tki...@gmail.com

unread,
Aug 19, 2013, 1:42:53 PM8/19/13
to xdr...@googlegroups.com, tki...@gmail.com
Hi Anthony,
Thanks for the reply. I think the only thing I'm really using from boost is the shared_ptr or shared_array and that may be bypassed if necessary. I basically want to be able to pass numpy arrays in and out of the python/C++ interface. I will fill you in and give some feedback once I get around to playing with this. It may take me a while as it's a side project and I need to work through the examples in xdress, etc.  Thanks also for offer to help. By the way the documentation says to use option "--dump-desc" but it seems like it's actually "--dumpdesc" in git repo
Tony

Anthony Scopatz

unread,
Aug 19, 2013, 2:10:01 PM8/19/13
to tki...@gmail.com, xdress
On Mon, Aug 19, 2013 at 10:42 AM, <tki...@gmail.com> wrote:
Hi Anthony,
Thanks for the reply. I think the only thing I'm really using from boost is the shared_ptr or shared_array and that may be bypassed if necessary. I basically want to be able to pass numpy arrays in and out of the python/C++ interface.

Great! We have a really nice way of doing this through std::vector on the C++
 
I will fill you in and give some feedback once I get around to playing with this. It may take me a while as it's a side project and I need to work through the examples in xdress, etc.  Thanks also for offer to help.

No problem ;).  Any feedback is really greatly appreciated!
 
By the way the documentation says to use option "--dump-desc" but it seems like it's actually "--dumpdesc" in git repo

Thanks for reporting!

Be Well
Anthony

Thinh Nguyen

unread,
Apr 4, 2014, 5:18:48 PM4/4/14
to xdr...@googlegroups.com, tki...@gmail.com
Hi all,

I am new to Xdress and I am wondering if there was any work done on shared_ptr's or if there's any workarounds for it.

Regards,
Thinh

Anthony Scopatz

unread,
Apr 4, 2014, 5:30:30 PM4/4/14
to Thinh Nguyen, xdress, tki...@gmail.com
Hi Thinh!

No one has done any explicit work on / with boost to the best of my knowledge.  (Please pipe up if you have.)  However, we now have much, much, much better template support than when this was first opened.  I'd encourage you to try it out.  If something doesn't work, please let us know.  I also *highly* encourage pull requests that support people's use cases.

Be Well
Anthony


--
You received this message because you are subscribed to the Google Groups "xdress" group.
To view this discussion on the web visit https://groups.google.com/d/msgid/xdress/259109d4-7eb8-413f-8024-f9b9116ed7a4%40googlegroups.com.

Reply all
Reply to author
Forward
0 new messages