What about using a function that throws an exception to quit? The nice thing about using a function to quit is that then you can return values.
I've done so here: http://gist.github.com/255883 In addition I've made it so the debug-repl call can wrap a form. If you call quit-dr with no parameters, the debug-repl will just return the value of the form wrapped. But you can also call quit with another form, to replace the original wrapped value. A couple of examples: user=> (let [a 10] (debug-repl (* a a))) dr-1-1006 => (quit-dr) 100 user=> (let [a 10] (debug-repl (* a a))) dr-1-1007 => (quit-dr (* 9 a)) 90 Finally, I've added a level and counter to the dr prompt so it is easier to tell if you are in nested repls or if you quit one debug-repl and enter another. g On Dec 12, 4:06 am, Konrad Hinsen <konrad.hin...@fastmail.net> wrote: > On 11 Dec 2009, at 23:26, Dan Larkin wrote: > > > On Dec 11, 2009, at 3:04 PM, Sean Devlin wrote: > > >> Wouldn't ::quit do the same thing? > > > It wouldn't, because the repl is evaluating in the context of > > wherever you put the (debug-repl) call, so its namespace won't be > > "dr". > > > What about instead of using keywords for commands, we use functions > > for commands: > > > (debug-repl-quit) > > I don't see any straightforward way to write a function that, when > executed inside a debug REPL, quits the REPL. The only way to get out > of a REPL is to have the its reader function return an object created > specifically for this purpose (and passed to the reader function). > Therefore any straightfoward method for quitting the REPL must be > based on objects returned by the reader, before evaluation. It is > probably possible to get around this by having a function that > modifies a global var which in turn is read by the REPL to decide when > to quit, but that's not what I'd call straightforward. > > What *is* doable easily is quit whenever the reader sees the form > (debug-repl-quit). But it's the literal form and not its value. I > guess that would be good enough for the typical usage scenarios of a > debug REPL, but still it's not really a function call. Perhaps a > keyword would be a better choice to mark the difference. > > Konrad. -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
