On Mon, Oct 05, 2015 at 08:06:29AM +1300, Evan Hanson wrote: > On 2015-10-04 14:55, Peter Bex wrote: > > Cool. Have you been able to reproduce the crash (without patch) at all? > > Yeah, by simply running `make check` in a loop until it fails. It > doesn't usually take more than 10 or 12 runs to hit the error (and only > one or two when inside a VM).
Ah, that's interesting. Good to know! > > > Just two minor things: (1) it looks like one too many words is allocated > > > for the C_apply_values argvector > > > > The argvector holds the continuation followed by each item in the > > argument list, which is why I added 1 to it. > > Yes, but that +1 is already done once before the C_demand (line 7304 > after applying the first patch), then done again in the argument to > C_alloc (line 7309). I think only the first one is necessary; that way > we'll be C_alloc'ing the same amount that's C_demand'ed, and that's used > for the eventual C_do_apply call, (+ (length lst) 1). I overlooked that. Thanks for pointing it out! You're right of course, and I'm now convinced that your first patch is fine as well. Cheers, Peter
signature.asc
Description: Digital signature
_______________________________________________ Chicken-hackers mailing list Chicken-hackers@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-hackers