libera/#commonlisp - IRC Chatlog

13:27:37 Demosthenex i'd love to have seen minecraft in lisp :P i know there was an abandoned POC version somewhere

13:30:10 scymtym https://github.com/gregcman/sucle

15:12:57 kagev splittist: what is (was?) ILC?

15:13:30 beach International Lisp Conference.

15:13:33 jackdaniel International Lisp Conference

15:13:50 jackdaniel ACTION is too slow, blames the weather

15:13:56 kagev I see. Was it also held in Europe?

15:14:10 beach Not usually, no.

15:14:11 kagev well ... I guess I can search for that ansewr ...

15:14:20 beach kagev: It used to be organized by ALU, the Association of Lisp Users. But I think that organization is dormant.

15:14:53 beach I attended it in San Francisco and in Montreal. Perhaps in some other place as well, I can't remember.

15:15:15 kagev I see ... found a site announcing it to be held in Kyoto Japan (back in 2012)

15:37:18 splittist kagev: what they said

15:48:26 NotThatRPG ILC was very fun: I went to it in Cambridge, MA and Reno, NV. There was one more after that, but none sinse

15:48:32 NotThatRPG s/sinse/since/

15:49:37 NotThatRPG Shinmera: ping?

16:32:52 Shinmera ?

16:34:02 jackdaniel minion: what is ping-pong?

16:34:03 minion maybe you need to ask my master, chandler - he knows a lot

17:06:13 NotThatRPG Shinmera: I've been finding some problems with repeatability in random-state...

17:06:41 NotThatRPG Posted a PR and an issue. Wondering if you could have a peek; I have a proposed fix, but I'm not sure it's legit

17:07:07 Shinmera You don't need to ping me for that stuff, I get emails and respond when I can.

17:10:26 NotThatRPG Shinmera: OK.

17:13:45 NotThatRPG One question, though: I have been running a relatively modest number of tests on each of the generators (which involves copying the generators for repeatability). I'm seeing SBCL give heap exhausted errors when I do this. Is there any reason this should be memory leaking?

17:14:22 NotThatRPG num-generators * 3 * 2 as far as I can tell

17:14:51 NotThatRPG Do generators need to be explicitly deleted?

17:21:37 NotThatRPG Trying to test random-state, I go from Dynamic space usage is: 91,911,680 bytes. to Dynamic space usage is: 522,143,632 bytes. and error out.

17:22:41 dstein64- ** NICK dstein64

17:35:07 Shinmera It doesn't do any CFFI, so, not to my knowledge.

17:39:56 NotThatRPG It looks like the resources are being sucked up by KISS11. When I drop that from my test suite, the test-suite runs to completion (but repeatability tests all fail).

17:40:36 NotThatRPG I'll verify and then update my notes.

17:41:38 NotThatRPG Yes, it's KISS11

17:46:41 Bike i just did a quick test locally, and it looks like some of the generators do copy enough - the dumb lcg and middle-square. mersenne 64 and 32 do seem to have a problem, though. are you seeing it fail on all generators?

17:47:56 Bike squirrel and tt800 also seem ok, but not rc4

17:48:32 NotThatRPG Bike: Now only MERSENNE-TWISTER (32 and 64) and RC4 are failing.

17:48:49 NotThatRPG That's good -- I can replace my use of :MERSENNE-TWISTER for now.

17:49:05 NotThatRPG I will stick to SQUIRREL for now.

17:49:30 NotThatRPG ACTION wishes FIVEAM had an XFAIL...

17:49:47 Bike it looks like the copy method on mersenne only does a shallow copy of the generator structure rather than actually copying the underlying arrays

17:50:55 NotThatRPG Using CL:*RANDOM-STATE* caused errors for me in RANDOM-STATE on SBCL. I had to replace it with a new global as default value for RANDOM's GENERATOR argument

18:02:23 NotThatRPG I have a fork with these fixes and PR for upstream