One thing that would help immensely is an "infinite" undo-redo on the
debugger. Tracking down certian errors has become more complex than reasonable for using the old system of laying "self halt" traps to catch the cause of an error before it causes enough trouble to create walkback. |
"Kirk W. Fraser" <[hidden email]> wrote in message
news:eWBV7.1341$[hidden email]... > One thing that would help immensely is an "infinite" undo-redo on the > debugger. Tracking down certian errors has become more complex than > reasonable for using the old system of laying "self halt" traps to catch the > cause of an error before it causes enough trouble to create walkback. Hi Kirk, Are you referring to an ability to step back (instead of forward) in the debugger? If so, that is theoretically possible, but generally not in practice. I say 'generally', because I'm pretty sure there is a way to do it usefully under certain controlled circumstances, but I'll let you know about this when I'm sure :-). Regards, Peter van Rooijen |
Kirk
Hmm.... I'm afraid this idea is far too sophisticated for a modest enhancement to the existing debugger. I know of no language that provides this sort of debugging capability (except perhaps Prolog?) and the implementation would generate hundreds of thousands of new objects. In order to step back one would need to keep the entire state at each step, or at least have some sort of backtracking mechanism. It just doesn't seem a feasible way of working to me. However, it sounds like something a third party might want to add (remember the entire debugger is implemented in Smalltalk) so maybe you can garner some interest from others on the Dolphin newsgroup. Given the general adequacy of the current debugger and the likely implementation effort I think it unlikely to rise to a high priority on our list of future enhancements. Best Regards, Andy Bower Dolphin Support http://www.object-arts.com --- Are you trying too hard? http://www.object-arts.com/Relax.htm --- "Peter van Rooijen" <[hidden email]> wrote in message news:a0c8bm$9f6$[hidden email]... > "Kirk W. Fraser" <[hidden email]> wrote in message > news:eWBV7.1341$[hidden email]... > > One thing that would help immensely is an "infinite" undo-redo on the > > debugger. Tracking down certian errors has become more complex than > > reasonable for using the old system of laying "self halt" traps to catch > the > > cause of an error before it causes enough trouble to create walkback. > > Hi Kirk, > > Are you referring to an ability to step back (instead of forward) in the > debugger? If so, that is theoretically possible, but generally not in > practice. I say 'generally', because I'm pretty sure there is a way to do > usefully under certain controlled circumstances, but I'll let you know about > this when I'm sure :-). > > Regards, > > Peter van Rooijen > > |
[This followup was posted to comp.lang.smalltalk.dolphin and a copy was
sent to the cited author.] Folks, The Caml debugger architecture for ML languages looks like it might apply to Smalltalk http://citeseer.nj.nec.com/tolmach93debugger.html Caml Debugger events look similar to the points where the Smalltalk debuggers insert/implement their breakpoints. Looks very interesting - checkout 'reverse execution'. The papers related to http://citeseer.nj.nec.com/booth97walk.html (Walk Backwards to Happiness, Debugging by Time Travel) look interesting too! --Peter Goodall In article <a11kgb$nlkde$[hidden email]>, bower@object- arts.com says... > Kirk > > Hmm.... I'm afraid this idea is far too sophisticated for a modest > enhancement to the existing debugger. I know of no language that provides > this sort of debugging capability (except perhaps Prolog?) and the > implementation would generate hundreds of thousands of new objects. In order > to step back one would need to keep the entire state at each step, or at > least have some sort of backtracking mechanism. It just doesn't seem a > feasible way of working to me. > > However, it sounds like something a third party might want to add (remember > the entire debugger is implemented in Smalltalk) so maybe you can garner > some interest from others on the Dolphin newsgroup. Given the general > adequacy of the current debugger and the likely implementation effort I > think it unlikely to rise to a high priority on our list of future > enhancements. > > Best Regards, > > Andy Bower > Dolphin Support > http://www.object-arts.com > --- > Are you trying too hard? > http://www.object-arts.com/Relax.htm > --- > |
Free forum by Nabble | Edit this page |