On May 12, 2009, at 1:30 PM, Michael Abshoff wrote: > Dag Sverre Seljebotn wrote: >> Michael Abshoff wrote: >>> Dag Sverre Seljebotn wrote: >>>> Robert Bradshaw wrote: >>> <SNIP> >>> >>> Hi Dag, >>> >>>> Status on Sage is that these fail: >>>> >>>> sage -t "devel/sage/sage/libs/ntl/ntl_ZZ_pX.pyx" >>>> sage -t "devel/sage/sage/rings/integer.pyx" >>>> sage -t "devel/sage/sage/plot/plot.py" >>>> sage -t "devel/sage/sage/interfaces/maxima.py" >>>> >>>> The last two seem to be because of fragile doctests. >>> Can you post the first doctest failure from each failing file? >> >> This is on Sage 3.4.1 (my own build). > > Ok. Which compiler and host? 3.4.2 is out by the way and 4.0.alpha0 > should drop tonight :) > > >> sage -t "devel/sage/sage/libs/ntl/ntl_ZZ_pX.pyx" >> ********************************************************************* >> * > > <SNIP> > >> OverflowError: can't convert negative value to unsigned int >> ********************************************************************* >> * > > Ok, changing semantics like that makes me queasy, but I assume you > guys > know what you are doing. It seems like a rather deep fundamental > change > for 0.11.1->0.11.2.
I consider it a bug that we used to allow this--what it's doing now is converting -1 to 2^64-1. The patch for this is already in Sage: http://trac.sagemath.org/ sage_trac/attachment/ticket/5803/5803-cython-0.11.1.patch - Robert _______________________________________________ Cython-dev mailing list [email protected] http://codespeak.net/mailman/listinfo/cython-dev
