00:05:50 -!- foom [jknight@nat/google/x-txpygwjytydkbeik] has quit [Ping timeout: 245 seconds] 00:09:22 foom [jknight@nat/google/x-nfwyiltaivbteiti] has joined #sbcl 01:07:14 prxq_ [~mommer@mnhm-590c312a.pool.mediaWays.net] has joined #sbcl 01:09:57 -!- prxq [~mommer@mnhm-590c173a.pool.mediaWays.net] has quit [Ping timeout: 248 seconds] 01:51:47 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 02:11:24 -!- Bike [~Glossina@71-222-63-186.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 02:13:54 Bike [~Glossina@75-164-166-65.ptld.qwest.net] has joined #sbcl 02:28:49 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Quit: Leaving.] 02:33:58 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 02:38:33 -!- christoph_debian [~christoph@ppp-188-174-85-72.dynamic.mnet-online.de] has quit [Ping timeout: 264 seconds] 02:41:26 -!- oleo [~oleo@xdsl-78-35-177-161.netcologne.de] has quit [Ping timeout: 240 seconds] 02:42:02 oleo [~oleo@xdsl-78-35-144-164.netcologne.de] has joined #sbcl 02:47:28 -!- Bike [~Glossina@75-164-166-65.ptld.qwest.net] has quit [Ping timeout: 246 seconds] 02:49:38 Bike [~Glossina@174-25-36-8.ptld.qwest.net] has joined #sbcl 02:51:57 christoph_debian [~christoph@ppp-188-174-59-77.dynamic.mnet-online.de] has joined #sbcl 03:04:09 -!- Tribal [tribal@rcfreak0.com] has quit [Read error: Connection reset by peer] 03:23:54 Tribal [tribal@rcfreak0.com] has joined #sbcl 03:54:07 pranavrc [~pranavrc@122.164.14.107] has joined #sbcl 03:54:07 -!- pranavrc [~pranavrc@122.164.14.107] has quit [Changing host] 03:54:07 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 04:19:26 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Ping timeout: 240 seconds] 04:21:02 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 04:31:36 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 04:32:19 momo-reina [~user@173.244.199.112] has joined #sbcl 04:45:49 pnpuff [~ln@unaffiliated/pnpuff] has joined #sbcl 05:31:37 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 05:33:50 teggi [~teggi@123.25.130.142] has joined #sbcl 05:34:08 sdemarre [~serge@38.166-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 06:01:46 -!- vi1 [~vi1@93.92.216.186] has quit [Remote host closed the connection] 06:32:37 -!- oleo [~oleo@xdsl-78-35-144-164.netcologne.de] has quit [Ping timeout: 248 seconds] 06:33:16 oleo [~oleo@xdsl-78-35-177-70.netcologne.de] has joined #sbcl 07:04:37 -!- momo-reina [~user@173.244.199.112] has left #sbcl 07:08:20 vi1 [~vi1@93.92.216.186] has joined #sbcl 07:09:48 -!- vi1 [~vi1@93.92.216.186] has left #sbcl 07:11:31 vi1 [~vi1@93.92.216.186] has joined #sbcl 07:12:14 -!- sdemarre [~serge@38.166-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 240 seconds] 07:16:25 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 07:19:02 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 07:50:24 -!- oleo [~oleo@xdsl-78-35-177-70.netcologne.de] has quit [Remote host closed the connection] 07:53:04 stassats [~stassats@wikipedia/stassats] has joined #sbcl 07:53:36 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Remote host closed the connection] 07:55:35 sdemarre [~serge@38.166-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 08:00:05 -!- sdemarre [~serge@38.166-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 248 seconds] 08:01:03 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 08:03:17 -!- Bike [~Glossina@174-25-36-8.ptld.qwest.net] has quit [Ping timeout: 248 seconds] 08:50:36 -!- pnpuff [~ln@unaffiliated/pnpuff] has quit [] 09:05:39 milosn [~milosn@user-5af50bca.broadband.tesco.net] has joined #sbcl 09:40:08 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Remote host closed the connection] 10:53:47 pnpuff [~ln@unaffiliated/pnpuff] has joined #sbcl 11:16:33 oleo [~oleo@xdsl-78-35-177-70.netcologne.de] has joined #sbcl 12:10:09 pnpuff- [~ln@unaffiliated/pnpuff] has joined #sbcl 12:13:21 -!- pnpuff [~ln@unaffiliated/pnpuff] has quit [Ping timeout: 264 seconds] 12:16:30 -!- pnpuff- is now known as pnpuff 12:19:06 robgssp` [~user@cpe-24-93-28-218.rochester.res.rr.com] has joined #sbcl 12:21:43 -!- robgssp [~user@cpe-24-93-28-218.rochester.res.rr.com] has quit [Ping timeout: 240 seconds] 12:48:28 -!- pnpuff [~ln@unaffiliated/pnpuff] has quit [] 13:00:59 pnpuff [~l@unaffiliated/pnpuff] has joined #sbcl 13:21:45 -!- pnpuff [~l@unaffiliated/pnpuff] has quit [Ping timeout: 264 seconds] 13:36:12 sdemarre [~serge@50.124-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 13:43:18 -!- sdemarre [~serge@50.124-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 264 seconds] 14:17:09 dto [~user@pool-96-252-62-13.bstnma.fios.verizon.net] has joined #sbcl 14:21:03 tcr [~tcr@77-56-40-229.dclient.hispeed.ch] has joined #sbcl 14:26:10 davazp [~user@204.Red-79-153-96.dynamicIP.rima-tde.net] has joined #sbcl 14:59:30 -!- davazp [~user@204.Red-79-153-96.dynamicIP.rima-tde.net] has quit [Ping timeout: 264 seconds] 15:22:19 attila_lendvai [~attila_le@95.56.68.98] has joined #sbcl 15:22:19 -!- attila_lendvai [~attila_le@95.56.68.98] has quit [Changing host] 15:22:19 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 16:07:09 pkhuong: I put out the new versions of the game today. Windows, Linux and Mac ---- all three OS's with SBCL :) i don't know how you guys do it, but it's fantastic! 16:08:16 it's all thanks to the students from the 80s 16:14:34 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 16:15:25 stassats: no kidding eh? 16:16:58 a couple students from the naughties helped as well. 16:17:17 hmm :) 16:18:01 pkhuong: thus far the only incompatibility i've encountered was that oddball 64 bit Windows XP Professional x64 Edition 16:30:03 sdemarre [~serge@50.124-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 16:34:37 I'm pretty sure I can't care about that ;) 16:36:15 pkhuong: definitely :) 16:36:32 they're in the wild now, i'll keep you apprised of any incompatibilty reports. 16:43:09 *stassats* found where type propagation for tests happens, in add-test-constraints 16:43:23 it just looks at *backend-predicate-types* 16:43:48 yes. 16:46:15 so, if type-identity is added, that's where it would be handled, but how to make it an identity, if, i assume, constraint propagation happens after transforms are applied 16:46:38 at which stage will this type-identity wrapper be removed? 16:46:49 ir2? 16:48:38 -!- teggi [~teggi@123.25.130.142] has quit [Remote host closed the connection] 16:52:08 it'd have to be a mix of ir1 (to detect when predicates already are in an If context), if-if optimisation and ir2. 16:52:18 so I'm considering other approaches now. 16:53:00 the obvious and easy way is jut to modify add-test-constraints 16:53:01 Something like a dummy combination to encode the information that [leaf X is of type Y after this program point]. 16:53:20 stassats: that only works for type tests that desugar into a single combination. 16:53:56 that is the easy fix I suggester earlier, but it will fail to catch a lot of cases. 16:54:31 and it means that you'll have to write specialised type-based transforms to elide redundant tests. 16:56:01 A special node to assert flow-sensitive type information would be more general. And another special node to denote branches inferred to be impossible would then let IF nodes naturally eliminate redundant type tests (unless they're side-effectful). 16:56:29 that sounds complicated 16:56:40 This seems more robust (it even works for union or intersection types), and is less complicated than getting type-identity right. 16:57:45 complicated, as in over my head, if you're planning to do it, then fine 16:58:05 The special nodes can just be combinations. I'll see about sketching something later today, if the GSoC midterm report doesn't eat up too much time. 16:58:44 are combinations "available" before source transforms happen? 16:59:16 what do you mean? 16:59:33 i just thought that it was a define-source-transform for typep, but it's deftransform 17:00:10 it's both. 17:00:51 right 17:00:56 the source transforms fires early in the common case. That way, we avoid most round trips through IR1 translation to a lambda, let recognition, etc. 17:02:08 so, would that be a problem? 17:02:19 no. Combinations are just function calls. 17:03:06 can we extend that to other things then, not just type tests 17:03:20 (typep x 'foo) -> (cond ([type-test-form ...] (really-derive-type x 'type) t) (t (really-derive-type x '(not type)) nil)) 17:04:38 and that pseudo derive type thing might be useful to speed up IR1 transforms: a big issue is that when we perform a type-based transform, the lambda that's spliced in "forgets" the tight type that was derived, until let conversion and constraints happen. 17:05:17 in general, we'll want both that (to handle general TYPEP forms), and a new kind of optimiser to hook in constraints. 17:05:25 Bike_ [~Glossina@174-25-36-8.ptld.qwest.net] has joined #sbcl 17:06:59 -!- Bike_ is now known as Bike 17:07:08 will this (mod x) pessimization be considered a regression and warrant at least a simple add-test-constraints work around? 17:07:48 if I can't get the general case right before the next release, I'll at least do something in constraints. 17:08:13 well, it's already freezed, so something complex might not be appropriate 17:08:56 i considered reverting typep changes, but then it'll leave type-checks with the same situation 17:10:00 before the next release. I don't think it's important enough to commit anything during the freeze. 17:11:01 if someone complains of a big regression, we can add an ad-hoc two-line patch. 17:11:10 (unsigned-byte x) is a common declaration 17:11:13 davazp [~user@204.Red-79-153-96.dynamicIP.rima-tde.net] has joined #sbcl 17:11:49 what would Krystof do? 17:11:49 declarations still work fine. 17:12:07 it's type-based dispatch that'll suffer. 17:12:57 it works indeed, casts are handled specially then 17:13:14 stassats: what do you mean by "how important swap-bytes-16 on 64-bit is" ? 17:13:40 fe[nl]ix: how often is it done to warrant some further optimizations 17:15:12 not very often 17:15:44 it's required reading TCP/UDP ports, but that doesn't happen often 17:16:55 pnpuff [~ln@unaffiliated/pnpuff] has joined #sbcl 17:17:19 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 17:17:24 -!- slyrus [~chatzilla@107.200.11.156] has quit [Ping timeout: 240 seconds] 17:21:57 well. The GSoC questionnaire is very short; much better than what granting bodies foist upon us ;) 17:25:40 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 17:28:04 -!- davazp [~user@204.Red-79-153-96.dynamicIP.rima-tde.net] has quit [Ping timeout: 246 seconds] 17:31:28 pnpuff- [~ln@unaffiliated/pnpuff] has joined #sbcl 17:34:31 -!- pnpuff [~ln@unaffiliated/pnpuff] has quit [Ping timeout: 260 seconds] 17:35:33 some type tests are so huge, they probably could be put into functions 17:37:10 or rewritten if constraints get cleverer about them. 17:37:13 especially when testing more than one argument 17:37:30 (simple-array * 1) is fugly. 17:37:38 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Remote host closed the connection] 17:38:27 (unsigned-byte 64) is worse 17:39:22 it could be very simple though 17:40:34 i got it (unsigned-byte 64) slightly improved, by 5 bytes 17:41:13 I am not bothered by some minor performance regressions in type dispatch 17:41:49 Krystof: well, it's more that the pattern of (etypecase x ((ub 64) ...)) doesn't work. 17:42:44 wait, doesn't work at all? 17:43:10 or doesn't usefully infer x as (ub 64) within ...)? 17:43:16 type declaration isn't propagated to the branch after the successful check 17:43:17 the latter. 17:44:14 ok. That is slightly sucky but only slightly and I can write in my release note about it. If someone is motivated to do something about it before I release in the next day or so, a test marked as failing would be a fine thing 17:44:38 it's easy to fix in an ad-hocy way 17:45:19 I would rather not take ad-hoc now. 17:46:08 and not for ub 64, but for ub less than 64 17:48:50 to be exact, for any (mod X) where is fixnum and not equivalent to (unsigned-byte #.n-word-bits) 17:53:34 pkhuong: do you have anything specific in mind how (unsigned-byte 64) can be made cheaper? 17:55:24 maybe if unsigned bignums were marked as such not with zero digits, but by some tagging 17:55:28 stassats: seems like it should be possible to do a case analysis on: fixnum, bignum of length 1, bignum of length 2. 17:56:01 so, when combined with other test, not just for a single (unsigned-byte 64) test 18:01:55 -!- pnpuff- is now known as pnpuff 18:09:04 Bike_ [~Glossina@67-5-251-178.ptld.qwest.net] has joined #sbcl 18:09:15 -!- Bike_ [~Glossina@67-5-251-178.ptld.qwest.net] has quit [Client Quit] 18:09:32 Bike_ [~Glossina@67-5-251-178.ptld.qwest.net] has joined #sbcl 18:11:53 -!- Bike [~Glossina@174-25-36-8.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 18:13:23 -!- Bike_ is now known as Bike 18:19:49 pkhuong: now this is a curious circumstance. I have a friend on 64-bit Windows Vista, and neither the SBCL-compiled EXE nor the CCL-compiled one work. it's strange because the SBCL one is now working on systems that the older version failed on before. and weirder, I hadn't seen ANY windows users report the CCL versions failing to run. 18:19:54 and i had many downloads/installs 18:20:11 could be something wrong specifically with his system, 18:20:16 or, nobody uses vista? 18:21:14 here, not many people use any kind of windows 18:21:24 vista, even less so 18:21:45 there doesn't seem to be any output from the crash in the terminal, not even the VirtualAlloc error that used to be shown. 18:21:48 hello stassats . 18:23:34 <|3b|> could be some dll taking up address space, not sure how to check for that though 18:23:49 like a third-party application? yeah 18:24:43 i suspect it's something like that, because with many downloads and much feedback, i've never had any complaints about the CCL versions crashing on startup. except from this one guy. 18:25:03 ok well, just keeping you updated. I'm making a list of all the test results 18:26:51 pnpuff- [~ln@unaffiliated/pnpuff] has joined #sbcl 18:27:07 -!- pnpuff- [~ln@unaffiliated/pnpuff] has quit [Client Quit] 18:30:04 -!- pnpuff [~ln@unaffiliated/pnpuff] has quit [Ping timeout: 260 seconds] 18:35:29 http://stackoverflow.com/questions/3407066/error-starting-sbcl-under-slime-on-vista hmm. this might be relevant 18:35:52 but it's two years old... 18:37:23 <|3b|> looks more like broken code than broken sbcl there 18:37:46 <|3b|> something tried to load sb-posix, and that wasn't available 18:38:05 yeah, i don't think that's happening here, my friend said there was no output. 18:38:31 <|3b|> also package problems shouldn't depend on platform for a dumped image 18:38:41 yes true |3b| 18:38:50 ok so that one link isn't relevant. 18:38:55 what about DEP Data Execution Prevention? 18:39:01 friend says it's turned on 18:39:50 could explain why neither SBCL nor CCL worked 19:01:31 -!- oleo [~oleo@xdsl-78-35-177-70.netcologne.de] has quit [Ping timeout: 246 seconds] 19:02:02 oleo [~oleo@xdsl-78-35-164-53.netcologne.de] has joined #sbcl 19:02:53 the good news about that is that it won't newly break for anyone who got it to work before :) 19:11:24 *stassats* thought of a too-clever way to do unsigned-byte-64-test 19:12:19 basically, load the left most bignum digit, then jump to the test that does either compares it to zero or greater-than-zero 19:20:18 which isn't any better once it works for all cases 19:20:58 maybe if all bignums had a most-significant digit which reflected its sign, all ones for negative, all zeros for positive 19:25:59 or i can save one jump if just let it fall through after the test for zero do a test for positiveness too, which will succeed 19:46:48 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 19:46:55 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 20:08:01 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 20:24:50 -!- prxq_ [~mommer@mnhm-590c312a.pool.mediaWays.net] has quit [Remote host closed the connection] 20:25:24 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 20:26:54 -!- dto [~user@pool-96-252-62-13.bstnma.fios.verizon.net] has quit [Remote host closed the connection] 20:31:59 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 20:42:48 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 20:52:19 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 21:00:06 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Read error: Connection reset by peer] 21:02:57 pnpuff [~ln@unaffiliated/pnpuff] has joined #sbcl 21:03:44 -!- pnpuff [~ln@unaffiliated/pnpuff] has left #sbcl 21:04:22 -!- sdemarre [~serge@50.124-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 246 seconds] 21:07:45 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 21:17:13 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 21:24:03 pnpuff [~ln@unaffiliated/pnpuff] has joined #sbcl 21:24:25 -!- pnpuff [~ln@unaffiliated/pnpuff] has left #sbcl 21:32:17 victory. (lambda (x) (and (typep x '(member a b)) (typep x '(member c d)))) compiles to a straight NIL 21:53:56 I hope not. 21:54:11 (functionp nil) -> NIL ; (functionp (lambda (x) (and (typep x '(member a b)) (typep x '(member c d))))) -> T 21:58:09 <|3b|> pjb: who says it is wrapped in a call to functionp? 21:58:21 *|3b|* assumes that isn't what was meant anyway though 22:01:47 |3b|: pkhuong said that (lambda (x) (and (typep x '(member a b)) (typep x '(member c d)))) was replaced by NIL, I exhibited a function for which they return different results, therefore for which they cannot be substituted. Therefore his compiler, as described, is wrong. 22:02:12 <|3b|> only if you assume he is an idiot, and meant that exactly literally 22:02:26 No, not AND, but OR. 22:03:22 <|3b|> without any code context, that code is dead code with no side effects, so could just as legally compile to /noting/ 22:03:26 <|3b|> *nothing 22:04:20 <|3b|> with irc context, it seems pretty obvious that it meant it compiles to a function that returns NIL without any runtime type checks 22:18:09 You're a fun drain. 22:19:09 <|3b|> ah, if it was intended as humor, i take all that back :) 22:19:30 <|3b|> (except the part about it being dead code, since that fits in that context too) 22:20:27 :-) 22:25:03 minion: memo for stassats: something like https://github.com/pkhuong/sbcl/tree/mnesic-type-tests 22:25:03 Remembered. I'll tell stassats when he/she/it next speaks. 22:34:54 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 22:41:46 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 22:51:33 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 22:53:26 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 22:54:14 still not sure if I want to add an annotation node type, or continue hacking it with combinations... 23:02:02 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Read error: Connection reset by peer] 23:13:31 benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has joined #sbcl 23:56:36 -!- benkard [~benkard@mnch-4d04dcfd.pool.mediaWays.net] has quit [Ping timeout: 256 seconds]