On Jun 28, 2:16 pm, mabshoff <[EMAIL PROTECTED]> wrote: > On Jun 28, 2:10 pm, Carl Witty <[EMAIL PROTECTED]> wrote: > > Another question about the rewrite: does anybody mind if I break > > unpickling of old pickled fast_float objects? I can make them work, > > but it would require keeping around hundreds of lines of code that I'd > > rather get rid of. > > I believe the current policy is to preserve the old pickles since we > now have the pickle jar to do regression testing. And I agree with > that policy since we must avoid breaking backward compatibility at all > (reasonable) cost. Since we have versioned pickles would that cause > massive pain for you?
Well, I'm planning to basically rewrite it from scratch; supporting the old pickles means either writing a lot of new code to translate old objects into the new system, or keeping around the entire old implementation only to support the old pickles. It seems quite possible that nobody has any old pickled fast_float objects that they care about. If we think people really do have old fast_float objects laying around, then I'll do the work to support unpickling them. > > Carl > > By the way: What happened to looking into LLVM in the context of > fast_float? That's phase 2 :) > Cheers, > > Michael Carl --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
