On Jul 8, 2010, at 10:43 AM, Chris Withers wrote: > Michael Bayer wrote: >>> http://svn.zope.org/zope.interface/trunk/setup.py?view=auto >> we wanted to annoy buildout/setuptools/Glyph as much as possible, and I am >> happy to say we succeeded. > > I hope you're joking :-S > >> The issue of them not building by default wasn't that they might not build, >> just that they have not been widely tested, and its C code, and we really >> didn't want people complaining that SQLAlchemy makes their application core >> dump. > > But that's what we have a wonderful and shiny set of unit tests for, right? > :-)
unit tests in Python can't really test the wider variety of mishaps that native code can produce, i.e. subtle memory leaks, weird race conditions, etc. Not that our C code is subject to any of that, but again, wanted to be very cautious. We wanted to get 0.6 out without having to worry about the C code having untested problems (and there were a few late in the game fixes too). > >> Not sure if we'd want to change this default midway through 0.6 or wait til >> 0.7. > > Fair enough, but this is sadly one of those things where you're only likely > to find the edge cases when you do make it "on by default"... The C exts also don't improve performance that much for most users. It would be nicer if we could get people to write a lot more C extensions for us, then it would really be something worth flipping on. -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalch...@googlegroups.com. To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en.