Subject: Re: Subclassing numpy.ndarray with Pyrex - msg#00020

Previous Message by Date: click to view message preview

Re: More on enums in 0.9.5.1a

Eric Huss wrote: > i = e + 1 # Fails. Hmmm, it seems you're right, C++ allows this without warning, although it doesn't seem very useful. It would be more useful IMO to allow e = e + 1 but that isn't allowed because the result is of type int instead of enum. I'll see about fixing this. -- Greg

Next Message by Date: click to view message preview

Re: Default class attributes

Pierre GM wrote: > I guess that if I really need a class > attribute, I can always wrap it in a property. If you just want an attribute shared between instances, that will work. It won't be quite like a true class attribute, though, because it won't be accessible via the class. -- Greg

Previous Message by Thread: click to view message preview

Re: Subclassing numpy.ndarray with Pyrex

Lenard Lindstrom wrote: > I don't think numpy was well tested with Pyrex. The provided > c_numpy.pxd causes a pyrex compiler error involving an array of objects. > As for the segment faults even stripping away everything but the class > declaration doesn't stop them. I don't have time to look into this deeply now, but if anyone can investigate and find out whether this is because of something Pyrex is doing wrong or something Numpy is doing wrong, that would be helpful. -- Greg

Next Message by Thread: click to view message preview

Re: Subclassing numpy.ndarray with Pyrex

> Urk, this doesn't bode well for being able to subclass > ndarray in Pyrex. If the normal instance creation > mechanisms are bypassed, various things that Pyrex > relies on won't happen, such as initialisation of > Python-valued attributes. Indeed. Well, thanks anyway > Do you really need to subclass ndarray? Could you > get what you want using a different class that > wraps an ndarray instead? Oh, I tried creating objects in Pyrex that use ndarrays as attributes, and that works pretty well indeed. However, numpy does not recognize them as arrays. Functions using the numpy.asanyarray trick would transform them in basic ndarrays, thus ripping off any specific attributes. The more it goes, the less comfortable I am with direct subclassing of ndarrays anyway...