15:42:51heisigAny ideas how a programmer can avoid the "No more immobile pages available" error in sbcl when allocating too many funcallable instances?
15:59:49stassatsbut then, how come there's so many funcallable instances? they shouldn't be in immutable space
15:59:56stassatsunless they are also generic functions
16:15:54beachThey are instances of funcallable-standard-object.
16:17:08beachI was lazy. Since funcallable-standard-class is more general than standard-class, I didn't make a distinction and used funcallable-standard-class for everything.
16:17:35beachI might have done it differently if I had known there was a limit like this.
16:21:06beachI guess I'll check the sources and see if I can find something.
16:21:22stassatssubclassing funcallable-standard-class does allocate in immobile space
16:21:49beachThat is very likely what is going on then.
16:22:09beachI could try to avoid that, but that would be a lot of work.
16:22:23beachSo if it is easier to bump the limit, I would do that instead.
16:22:49stassatsdo you not intend distributing your code?
16:23:23stassatsimmobile or not, funcallable instances are more wasteful
16:26:21beachIt is the bootstrapping code for SICL.
16:26:31beachSo yes, I intend to distribute it at some point.
16:27:02beachIt doesn't matter if it is somewhat wasteful.
16:27:50beachAnd distribution is a long time in the future, so right now I would just rather continue my work on making progress with the bootstrapping procedure.
16:28:09beachIf I have to, I will of course try to decrease the number of immobile objects.
17:47:59beachstassats: I need to vanish in order to have dinner with my (admittedly small) family. I'll be back later and read any suggestions you might have.
17:53:05stassatsok, my suggestion would be to build the latest SBCL
17:53:24stassatsi made only generic functions go immobile
19:02:22beachI compiled it, but I need to figure out how to purge some fasls. I'll do that tomorrow when I have more time and I am more alert. Thank you for your help.