12:06:26stassatsi'll remove it when i add a way to turn it back on optionally
12:06:39stassats(at an undefined time in the future)
12:08:02stassatsfor this, and the defmethod-no-defgeneric warning
12:08:18luisSounds good, thanks. :-) Generally being able to turn warnings on and off would be useful to incrementally tame the tens of thousands of warnings SBCL complains about in our software. Maybe, I'm not sure there's any hope.
12:09:01stassatsthe &optional and &key has always been nothing more than a useless nuisance
12:09:07luissetq warning at the repl is also quite annoying, while you're at it (though that might not have other complications, not sure)
12:09:18stassatsi have never looked at it and said "it's right, i'll change my ways"
13:27:07teegheeI typed (time (mod (expt 777121 500000) 9981037)) into slime sbcl repl, and it took about 14 seconds to return the result, but time output said the real time passed was 0.000 seconds.
13:27:42teegheeI don't know what's happening inside time macro, anyone has some idea?
13:28:10_deathlikely it computes the number at compile-time
13:28:31_deathif you define a function with the form, you can see that it does
13:28:55teegheeaha so it measures the running time not compile time
15:04:27clothespinnot from dmalloc, I would have to hack it further to report what objects it is freeing and what objects not, plus currently you cant tell if the c++ library is calling alloc or free internally or from an API
15:04:57clothespinit should make for a serious tool in the future when things get really hairy
15:06:06clothespinbut I'm finding leaks the tedious old-fashioned way, update the code/longevity test
15:11:58_deathI first learned about dmalloc from https://blog.regehr.org/archives/896 (he uses it in the code repository)