On Mar 22, 2010, at 3:15 PM, Helmut Eller wrote:
- Terje Norderhaug [2010-03-22 22:52+0100] writes:
If so, frame-restartable-p should return NIL only when the frame definitely cannot be restarted and T when the frame possibly can be restarted. This makes the current default frame-restartable-p, which always returns NIL, unusable as default for back-ends that have a working restart-frame.
Either back-ends that implement restart-frame also need to implement frame-restartable-p, or the default frame-restartable-p should return T for back-ends that provide a restart-frame function.
With this in place, the backtrace slimefun will more consistently inform the client whether a frame potentially can be restarted or definitely can't be restarted.
Whatever frame-restartable-p returns it doesn't stop the client from calling restart-frame.
No, but a proper value for frame-restartable-p will allow swank clients to consistently inform users about whether a backtrace frame might be restartable or definitely isn't. That's a usability improvement.
BTW, now we're on the topic, let me throw this idea in from a previous related discussion:
On Sat Jul 18 04:37:10 EDT 2009, Tobias C. Rittweiler wrote:
I think FRAME-RESTARTABLE-P should become a FRAME-ATTRIBUTES function with :RESTARTABLE, and :NON-RESTARTABLE being the currently specified values.
-- Terje Norderhaug terje@in-progress.com