Jason Grout wrote: > I'm just finishing making RDF and CDF vectors use numpy under the hood. > The last problem I'm having is trying to use the Cython buffer > interface in sage/rings/polynomial/real_roots.pyx to access the > underlying numpy array. >
FYI, Dag and I corresponded about this (see below). So now the patch converting RDF/CDF vectors to use numpy is ready for review; see http://trac.sagemath.org/sage_trac/ticket/4206 Dag Sverre Seljebotn wrote: > Seems like an issue with circular cimports? Which means those aren't tested in the test framework. Anyway I was unable to provoke the behaviour with a testcase myself, but "coding in blind" then I assume that the following patch should fix it. > > Jason: To answer the question in the wiki about what to do for SAGE, I assume that we can quite quickly release a Cython 0.10.2 that incorporates this patch. Though Robert would be the one to give a real answer. > > diff -r 04e83ffd8ea2 Cython/Compiler/Buffer.py > --- a/Cython/Compiler/Buffer.py Fri Nov 07 06:55:37 2008 +0100 > +++ b/Cython/Compiler/Buffer.py Sun Nov 23 16:58:15 2008 +0100 > @@ -710,7 +710,11 @@ def use_py2_buffer_functions(env): > > # Search all types for __getbuffer__ overloads > types = [] > + visited_scopes = set() > def find_buffer_types(scope): > + if scope in visited_scopes: > + return > + visited_scopes.add(scope) > for m in scope.cimported_modules: > find_buffer_types(m) > for e in scope.type_entries: > This patch made it compile and all the doctests pass. Can we get this patch (or an equivalent one) into Sage as soon as possible? Thanks, Jason --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---