On Mon, May 20, 2013 at 12:38 PM, Glyph <[email protected]> wrote:
> > On May 12, 2013, at 9:37 PM, C Anthony Risinger <[email protected]> wrote: > > > however, just yesterday, i realized how to achieve 100% out-of-the-box > support for the gobject realtime bindings (!), *despite* the lack of > support in gobject-introspection (use cytpes manually! of course! it's so > simple!) so i'm currently working to push that out, followed by *gasp* a > real release, but these problems you've encountered are definitely a high > priority, albeit as part of a larger architecture issue... > > Have you considered using cffi rather than ctypes? < > https://pypi.python.org/pypi/cffi> This is a bit more maintainable and > easier to debug when upstream makes changes (you get a nice compiler error > rather than a segfault), which is especially important given GTK+/GLib's > somewhat ... cavalier attitude towards backwards compatibility. no i had not -- i actually just heard of cffi for the first time the other day on IRC in the context of developer adding PyPy support to uWSGI... but i've yet to really check it out myself. i was anly really looking at using ctypes because it's in the stdlib, and (IIRC) is what the goject bindings ultimately use. i know the impl is pretty simple, but i havent had the opportunity to complete yet -- are there other reasons to use cffi over ctypes? my experience with ctypes is not extensive either. -- C Anthony -- --- You received this message because you are subscribed to the Google Groups "Pyjs.org Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
