But, again, I insist that native code stepper would be much, much better in architectural sence even despite less detail it would give. _______________________________________________ Lisp Hug - the mailing list for LispWorks users lisp-hug@lispworks.com http://www.lispworks.com/support/lisp-hug.html
But, again, I insist that native code stepper would be much, much
better in architectural sence even despite less detail it would give.
_______________________________________________
Lisp Hug - the mailing list for LispWorks users
lisp-hug@lispworks.com
http://www.lispworks.com/support/lisp-hug.html
Lisp machine… just sayin’ On Mar 6, 2014, at 6:41 AM, Denis Budyak <budden73@gmail.com> wrote: > > But, again, I insist that native code stepper would be much, much > better in architectural sence even despite less detail it would give. > > _______________________________________________ > Lisp Hug - the mailing list for LispWorks users > lisp-hug@lispworks.com > http://www.lispworks.com/support/lisp-hug.html > _______________________________________________ Lisp Hug - the mailing list for LispWorks users lisp-hug@lispworks.com http://www.lispworks.com/support/lisp-hug.html
On 6 Mar 2014, at 11:41, Denis Budyak <budden73@gmail.com> wrote: > But, again, I insist that native code stepper would be much, much > better in architectural sence even despite less detail it would give. I think that native-code stepping is reasonably hard to do in a useful way in a language with macros. I've never actually found stepping Lisp useful (including on LispMs). _______________________________________________ Lisp Hug - the mailing list for LispWorks users lisp-hug@lispworks.com http://www.lispworks.com/support/lisp-hug.html
I think that native-code stepping is reasonably hard to do in a useful way in a language with macros. I've never actually found stepping Lisp useful (including on LispMs).
On 6 Mar 2014, at 11:41, Denis Budyak <budden73@gmail.com> wrote:
> But, again, I insist that native code stepper would be much, much
> better in architectural sence even despite less detail it would give.
_______________________________________________
Lisp Hug - the mailing list for LispWorks users
lisp-hug@lispworks.com
http://www.lispworks.com/support/lisp-hug.html
> II think that native-code stepping is reasonably hard to do in a useful way in a language with macros. Indeed this is hard, but there are more or less successful examples of such implementation. E.g. #line directive in C allows to control a stepper when code transformations are used. Some time ago I suggested to implement similar things for Lispworks, and I even posted some prototype code for particular cases (e.g., for iterate). Lispworks developers promised to work on it. > The moment you stop single-stepping the code, you will quite likely be able to re-compile the code that you were stepping through. Is I have mentioned above, there are problems: i) compilation context might differ in some tricky cases and there is no predicate to check if we are in tricky case or not. So, setting a breakpoint potentially converts function in a way which differs its behaviour ii) When you return from stepped function to a function which was not compiled with stepper enabled, you're unable to step through it. I guess there could be possible to recompile caller function just before actual return and to redirect execution path from non-stepped function into stepped, but this is not implemented in current version of Lispworks stepper and I'm not sure that is easier than implement native-code stepper. iii) doing "restart frame stepping" works well for a code without side effects only. Again, there is no "pure-functional-p" predicate to check if it is safe to do restart frame. _______________________________________________ Lisp Hug - the mailing list for LispWorks users lisp-hug@lispworks.com http://www.lispworks.com/support/lisp-hug.html
Bradford Miller wrote: > Lisp machine... > just sayin' Yeah ... I wish I still had one :( Like the Symbolics 3600 I used 30 years ago. On Mar 6, 2014, at 6:41 AM, Denis Budyak <budden73@gmail.com> wrote: > But, again, I insist that native code stepper would be much, much better in architectural sence even despite less detail it would give. _______________________________________________ Lisp Hug - the mailing list for LispWorks users lisp-hug@lispworks.com http://www.lispworks.com/support/lisp-hug.html