On Fri, Jan 15, 2010 at 9:41 AM, Aaron Meurer <[email protected]> wrote: > > On Jan 15, 2010, at 8:18 AM, mindcorrosive wrote: > >> I've reinstalled python with sympy (cleaning up everything in the >> install directory in between), and the issue still occurs. Upgrading >> to a more recent Python version is not an option -- I'm developing a >> plugin for a certain application that has embedded Python scripting >> engine -- and they use Py2.4. Still, I've tested the latest sympy and >> Py2.6 and it works as it should on the same machine. >> >> I did some additional tests and it seems that the crux is in the evalf >> method -- other similar functions, like sin and log, are also affected >> in the same way, and the a.evalf() (see code sample for "a") method >> does not work as it should as well. >> >> Sympy versions 0.6.1 and 0.6.0 work properly, although 0.6.5, 0.6.4 >> and 0.6.2 (and most likely 0.6.3 as well) do not (yes, I've tested >> them all, and yes, I'm that desperate). > What about sympy 0.6.6? > > Also, since it seems to have worked before, could you use git to bisect the > source to find the offending commit. Here is a little tutorial in case you > don't know git. > > You will probably need msysgit or some similar. > Just do: > > git clone git://git.sympy.org/sympy.git # This downloads the repository, > including the whole history. > git bisect start # This starts the bisect process > git checkout sympy-0.6.4 # The first version that doesn't work > git bisect bad # Set it as bad > git checkout sympy-0.6.3 # The latest version that does work > git bisect good # Set it as good > > At this point, git will checkout a commit somewhere in between the two > versions. It should give you a good guide as you go. In each case, run > check if it works by running isympy in the bin directory and type git bisect > good or git bisect bad accordingly until it tells you what the first > offending commit is. > > Then please paste the SHA1 of that offending commit here (git show SHA1 will > show you what changed with that commit). Hopefully this will provide more > insight to help fix this.
Yes, use git bisect to find out which revision broke that. Then let's fix it. I think it's some problem with long vs int in mpmath. Ondrej
-- You received this message because you are subscribed to the Google Groups "sympy" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/sympy?hl=en.
