00:04:26 -!- tcr [~tcr@92.58.136.182] has quit [Quit: Leaving.] 00:22:34 gor[e] [~svr@gw1.masterhost.ru] has joined #sbcl 00:25:13 -!- gor[e] [~svr@gw1.masterhost.ru] has quit [Client Quit] 00:26:35 gor[e] [~svr@gw1.masterhost.ru] has joined #sbcl 00:26:52 -!- gor[e] [~svr@gw1.masterhost.ru] has quit [Client Quit] 00:27:31 gor[e] [~svr@gw1.masterhost.ru] has joined #sbcl 01:30:14 -!- homie [~levgue@xdsl-78-35-183-65.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 01:33:06 homie [~levgue@xdsl-78-35-183-65.netcologne.de] has joined #sbcl 01:41:07 echo-area [~user@114.251.86.0] has joined #sbcl 01:53:06 -!- hargettp [~hargettp@96.237.121.111] has quit [Quit: Leaving...] 02:19:27 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 252 seconds] 02:21:29 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Ping timeout: 240 seconds] 03:06:33 -!- |3b| [foobar@cpe-72-179-19-4.austin.res.rr.com] has quit [*.net *.split] 03:06:33 -!- nyef [~nyef@pool-64-222-180-202.man.east.myfairpoint.net] has quit [*.net *.split] 03:06:33 -!- fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has quit [*.net *.split] 03:06:33 -!- jsnell [~jsnell@ash.snellman.net] has quit [*.net *.split] 03:06:34 -!- deepfire [~deepfire@80.92.100.69] has quit [*.net *.split] 03:06:35 -!- Neronus [christian@heraklit.ayous.org] has quit [*.net *.split] 03:06:35 -!- Krystof [~csr21@csrhodes.plus.com] has quit [*.net *.split] 03:06:36 -!- minion [~minion@common-lisp.net] has quit [*.net *.split] 03:06:37 -!- foom [~jknight@ita4fw1.itasoftware.com] has quit [*.net *.split] 03:06:38 -!- echo-area [~user@114.251.86.0] has quit [*.net *.split] 03:06:38 -!- cmm [~cmm@109.65.203.181] has quit [*.net *.split] 03:06:38 -!- lnostdal [~Lars@213.80-202-59.nextgentel.com] has quit [*.net *.split] 03:06:38 -!- scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has quit [*.net *.split] 03:06:39 -!- ASau` [~user@93-80-218-22.broadband.corbina.ru] has quit [*.net *.split] 03:06:39 -!- antoszka [~antoszka@unaffiliated/antoszka] has quit [*.net *.split] 03:06:39 -!- gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has quit [*.net *.split] 03:06:40 -!- derrotebaron [johannes@static.7.69.47.78.clients.your-server.de] has quit [*.net *.split] 03:06:40 -!- homie [~levgue@xdsl-78-35-183-65.netcologne.de] has quit [*.net *.split] 03:06:42 -!- lisppaste2 [~lisppaste@common-lisp.net] has quit [*.net *.split] 03:06:42 -!- loke [~elias@bb119-74-213-61.singnet.com.sg] has quit [*.net *.split] 03:06:42 -!- hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has quit [*.net *.split] 03:06:42 -!- peddie [~peddie@XVM-107.MIT.EDU] has quit [*.net *.split] 03:06:42 -!- joshe [~joshe@opal.elsasser.org] has quit [*.net *.split] 03:06:43 -!- clop [~jared@moat3.centtech.com] has quit [*.net *.split] 03:06:44 -!- specbot [~specbot@common-lisp.net] has quit [*.net *.split] 03:06:44 -!- pkhuong [~pkhuong@gravelga.xen.prgmr.com] has quit [*.net *.split] 03:10:50 -!- danlentz [~textual@c-68-46-98-23.hsd1.nj.comcast.net] has quit [Ping timeout: 255 seconds] 03:11:47 danlentz [~danlentz@c-68-46-98-23.hsd1.nj.comcast.net] has joined #sbcl 03:14:24 foom [~jknight@ita4fw1.itasoftware.com] has joined #sbcl 03:18:19 lisppaste2 [~lisppaste@common-lisp.net] has joined #sbcl 03:18:19 loke [~elias@bb119-74-213-61.singnet.com.sg] has joined #sbcl 03:18:46 echo-area [~user@114.251.86.0] has joined #sbcl 03:18:46 antoszka [~antoszka@unaffiliated/antoszka] has joined #sbcl 03:18:46 cmm [~cmm@109.65.203.181] has joined #sbcl 03:18:46 lnostdal [~Lars@213.80-202-59.nextgentel.com] has joined #sbcl 03:18:46 scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has joined #sbcl 03:18:46 ASau` [~user@93-80-218-22.broadband.corbina.ru] has joined #sbcl 03:19:03 hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has joined #sbcl 03:19:03 peddie [~peddie@XVM-107.MIT.EDU] has joined #sbcl 03:19:03 joshe [~joshe@opal.elsasser.org] has joined #sbcl 03:19:03 clop [~jared@moat3.centtech.com] has joined #sbcl 03:19:10 homie [~levgue@xdsl-78-35-183-65.netcologne.de] has joined #sbcl 03:19:18 Neronus [christian@heraklit.ayous.org] has joined #sbcl 03:20:09 |3b| [foobar@cpe-72-179-19-4.austin.res.rr.com] has joined #sbcl 03:20:09 nyef [~nyef@pool-64-222-180-202.man.east.myfairpoint.net] has joined #sbcl 03:20:09 fe[nl]ix [~lacedaemo@pdpc/supporter/professional/fenlix] has joined #sbcl 03:20:09 jsnell [~jsnell@ash.snellman.net] has joined #sbcl 03:20:09 deepfire [~deepfire@80.92.100.69] has joined #sbcl 03:20:13 gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has joined #sbcl 03:20:19 Krystof [~csr21@csrhodes.plus.com] has joined #sbcl 03:20:19 minion [~minion@common-lisp.net] has joined #sbcl 03:20:19 -!- niven.freenode.net has set mode +o Krystof 03:20:25 specbot [~specbot@common-lisp.net] has joined #sbcl 03:20:25 pkhuong [~pkhuong@gravelga.xen.prgmr.com] has joined #sbcl 03:24:48 Fare [~Fare@64.119.159.126] has joined #sbcl 03:26:19 derrotebaron [johannes@static.7.69.47.78.clients.your-server.de] has joined #sbcl 03:37:06 -!- Fare [~Fare@64.119.159.126] has quit [Quit: Leaving] 03:47:07 mtd [~martin@chop.xades.com] has joined #sbcl 03:57:20 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 03:57:20 -!- ChanServ has set mode +o nikodemus 05:13:15 -!- gor[e] [~svr@gw1.masterhost.ru] has quit [Ping timeout: 252 seconds] 05:49:54 *slyrus* saw the headline on news.bbc.com "Tepco bailout speculation grows " and thought they were talking about teclo for a very brief moment... 06:02:20 -!- scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has quit [Remote host closed the connection] 06:09:01 -!- echo-area [~user@114.251.86.0] has quit [Remote host closed the connection] 06:18:33 flip214 [~marek@unaffiliated/flip214] has joined #sbcl 06:23:53 echo-area [~user@114.251.86.0] has joined #sbcl 06:25:42 nixie` [~user@222.92.128.242] has joined #sbcl 06:28:37 slyrus: no, the other notoriously secretive organization (: 06:30:28 it's OK, mobile broadband is not critical infrastructure 06:30:35 if we break it, no-one will complain, right? 06:35:29 tcr [~tcr@92.58.136.182] has joined #sbcl 06:38:00 -!- tcr [~tcr@92.58.136.182] has quit [Client Quit] 06:40:25 onixie [~user@222.92.128.242] has joined #sbcl 06:50:18 -!- onixie [~user@222.92.128.242] has quit [Remote host closed the connection] 07:06:11 -!- gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has quit [Ping timeout: 252 seconds] 07:06:40 gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has joined #sbcl 07:32:06 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 07:36:59 -!- flip214 [~marek@unaffiliated/flip214] has quit [Remote host closed the connection] 07:38:08 flip214 [~marek@2001:858:107:1:7a2b:cbff:fed0:c11c] has joined #sbcl 07:38:08 -!- flip214 [~marek@2001:858:107:1:7a2b:cbff:fed0:c11c] has quit [Changing host] 07:38:08 flip214 [~marek@unaffiliated/flip214] has joined #sbcl 07:50:33 onixie [~user@222.92.128.242] has joined #sbcl 07:52:58 tcr [~tcr@92.58.136.182] has joined #sbcl 07:54:19 -!- onixie [~user@222.92.128.242] has quit [Remote host closed the connection] 07:56:11 -!- homie [~levgue@xdsl-78-35-183-65.netcologne.de] has quit [Read error: Connection reset by peer] 07:59:23 -!- Krystof [~csr21@csrhodes.plus.com] has quit [Ping timeout: 240 seconds] 07:59:52 homie [~levgue@xdsl-78-35-183-65.netcologne.de] has joined #sbcl 08:00:00 -!- tcr [~tcr@92.58.136.182] has quit [Quit: Leaving.] 08:32:08 homie` [~levgue@xdsl-78-35-143-164.netcologne.de] has joined #sbcl 08:33:02 tcr [~tcr@92.58.136.182] has joined #sbcl 08:35:20 -!- homie [~levgue@xdsl-78-35-183-65.netcologne.de] has quit [Ping timeout: 260 seconds] 08:39:55 stassats [~stassats@wikipedia/stassats] has joined #sbcl 08:58:47 -!- nixie` [~user@222.92.128.242] has quit [Read error: Connection reset by peer] 08:59:18 nixie` [~user@222.92.128.242] has joined #sbcl 09:00:32 onixie [~user@222.92.128.242] has joined #sbcl 09:03:23 -!- onixie [~user@222.92.128.242] has quit [Client Quit] 09:14:49 -!- nixie` [~user@222.92.128.242] has quit [Remote host closed the connection] 09:15:12 nixie` [~user@222.92.128.242] has joined #sbcl 09:27:24 pchrist [~spirit@gentoo/developer/pchrist] has joined #sbcl 09:28:12 Hi guys. I didn't know this channel exists :) I'm excited 09:31:24 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has left #sbcl 09:33:45 -!- echo-area [~user@114.251.86.0] has quit [Remote host closed the connection] 09:37:42 Goodday: http://paste.lisp.org/display/120983  any hints how to deal with that miscalculation? 09:38:14 -!- nixie` [~user@222.92.128.242] has quit [Remote host closed the connection] 09:38:34 (on x86) 10:08:58 Krystof [~csr21@cpc2-dals3-0-0-cust1263.hari.cable.virginmedia.com] has joined #sbcl 10:08:58 -!- ChanServ has set mode +o Krystof 10:14:18 hargettp [~hargettp@96.237.121.111] has joined #sbcl 10:23:17 -!- homie` [~levgue@xdsl-78-35-143-164.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 10:27:02 homie [~levgue@xdsl-78-35-143-164.netcologne.de] has joined #sbcl 10:30:04 reducing to what operation introduces an error would help 10:33:55 Not sure how to trace that. 10:34:05 the only problem is that it returns the same results on x86 and on x86-64 for me 10:34:22 oh wait 10:34:27 maybe i didn't use the optimized variant 10:34:35 yeah, that's where the problems shows 10:34:48 works fine non-optimized 10:34:57 indeed, sorry for the fuss 10:36:27 11:55 < antoszka> Update: x86: single, double, long  wrong answer; x64 single  correct answer; double, long  wrong answer (same as x86). 10:36:44 So only optimising as single-float on x64 doesn't break the calculation. 10:37:51 single float is immediate on x86_64 10:38:14 and there's no long floats in sbcl 10:38:33 Well, that code ran :) 10:38:46 that were double floats 10:38:50 ok. 10:39:41 what code are we discussing? 10:39:55 minion: paste 120983? 10:39:55 Paste number 120983: "x86 bug?" by antoszka in None. http://paste.lisp.org/display/120983 10:39:58 Krystof: http://paste.lisp.org/display/120983  optimization problem described here 10:40:09 minion was faster... 10:42:00 I'm not sure that there's any reason to believe that those should give the same answer 10:42:41 optimization can change the level of precision in intermediate calculations, which can change the point at which addition of a small number to a big number becomes the identity operation 10:42:59 this is the same effect as -ffloat-store 10:43:53 do you have reason to believe that that floating point calculation should always give a particular answer under a particular precision and rounding regime? 10:44:51 Well, it's 0,33% off ... that seems to be too much for double-float rounding issues 10:45:40 and especially the optimizing variant should not needlessly throw values into memory - so it should be _more_ exact, by keeping the intermediates in the '87 in 80 bit 10:47:12 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 10:48:11 I'd agree normally except that you've got a z in the denominator of the term being added 10:49:02 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 276 seconds] 10:49:06 i'd just use rationals and wouldn't bother with floats 10:49:52 so there's feedback 10:50:01 stassats`: doesn't work here. a (dotimes (c 33)) ran over a minute, so you won't see a result of (dotimes (c 9999999)) 10:50:25 not in this universe, that is ;-) 10:50:29 is this actually a real calculation anyway, or is it from someone who doesn't understand floating point? 10:50:55 in what sense is this a well-conditioned addition? (adding progressively smaller-and-smaller terms to bigger-and-bigger ones)? 10:51:02 Krystof: That's from a friend who's experimenting with some AI code. Perhaps he's missing some floating point fundamentals :) 10:51:47 But yes, it comes from some working code that gave sensible results without optimisations. 10:52:18 what does "sensible" mean? 10:52:34 in what sense is the first answer the right answer? 10:52:55 Krystof: In the sense, that it is the same on all architectures. 10:53:07 what if it's wrong on all architectures? 10:53:13 That's possible too. 10:54:28 csr21@omega:/tmp$ gcc -O3 -std=gnu99 -o foo-O3 foo.c 10:54:28 csr21@omega:/tmp$ gcc -O0 -std=gnu99 -o foo-O0 foo.c 10:54:33 csr21@omega:/tmp$ ./foo-O0 2.5 4.5 10:54:33 3546.254883 10:54:33 csr21@omega:/tmp$ ./foo-O3 2.5 4.5 10:54:33 3333.338330 10:54:51 Please encourage your friend to read "What every computer scientist should know about floating point" 10:55:05 Krystof: OK, thanks :) 10:56:18 (source code is at the paste, not that it's anything difficult) 10:56:31 thanks. 11:01:02 for some reason this code, but with rationals, is very unwilling to be interrupted in slime 11:01:09 Same here. 11:01:17 I just slime-restarted-inferior-lisp. 11:01:36 it finally succumbed, while at SB-BIGNUM:BIGNUM-GCD 11:02:06 -!- hargettp [~hargettp@96.237.121.111] has quit [Quit: Leaving...] 11:02:32 it might be trying to print implausibly large numerators and denominators 11:02:41 anyway, i'd wait until computers are so powerful that you can use rationals for this problem 11:04:12 hargettp [~hargettp@96.237.121.111] has joined #sbcl 11:05:12 swank truncates big numbers for transmission to slime, but yeah, the truncation itself might take some time 11:07:08 I think the algorithm should be rethinked, anyway. 11:07:57 udzinari [~user@nat/ibm/x-egtakeemkdlksslo] has joined #sbcl 11:24:40 -!- tcr [~tcr@92.58.136.182] has quit [Quit: Leaving.] 11:35:23 A side note: on Allegro CL the optimised version runs 35 times faster and gives same result (whether it's good or not is a different discussion :)) 11:36:01 (same result as in optimized/non-optimized rather than allegro vs sbcl) 11:36:16 tcr [~tcr@92.58.136.182] has joined #sbcl 11:38:34 (and slower than sbcl in either case) 11:44:07 -!- tcr [~tcr@92.58.136.182] has quit [Ping timeout: 252 seconds] 11:52:54 *Krystof* gets working M-- M-x slime RET R 11:53:11 not sbcl, but... it's almost lisp work 11:54:09 -!- hargettp [~hargettp@96.237.121.111] has quit [Quit: Leaving...] 11:54:13 tcr [~tcr@92.58.136.182] has joined #sbcl 11:54:55 Krystof: do you use R for academic work or for the work you can't tell us anything about? 11:55:38 yes! 11:58:11 -!- tcr [~tcr@92.58.136.182] has quit [Client Quit] 11:59:23 my colleagues at University do a fair amount of psychological dataset experimentation where being able to generate tables of p-values is very useful 11:59:45 I use R in s0uper s3kr1t work to quickly summarize numerical results 12:00:14 i see 12:01:28 I don't use R so much in my own academic work because, well, it's been a while since I've done any academic work of my own directly :-( 12:01:59 I used the super sekrit work as an excuse to learn something new 12:02:33 they don't pay much? 12:02:45 who? 12:02:58 sekrit employers 12:03:04 academic pay is fine. super sekrit terms of employment are probably super sekrit 12:04:07 (at the stage when I started doing sekrit work, it's more fair to describe it as "moonlighting" than "highly formalized employment with contracts and stuff" 12:08:46 -!- udzinari [~user@nat/ibm/x-egtakeemkdlksslo] has quit [Remote host closed the connection] 12:36:11 woudshoo [~user@ipleiden.intellimagic.net] has joined #sbcl 13:08:01 -!- hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has quit [Remote host closed the connection] 13:08:25 hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has joined #sbcl 13:13:40 antgreen [~user@CPE00222d6c4710-CM00222d6c470d.cpe.net.cable.rogers.com] has joined #sbcl 13:18:41 -!- hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has quit [Remote host closed the connection] 13:18:43 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 13:38:55 dlowe [~dlowe@ita4fw1.itasoftware.com] has joined #sbcl 13:44:50 -!- flip214 [~marek@unaffiliated/flip214] has quit [Remote host closed the connection] 14:15:37 -!- redline6561 [~user@c-66-56-55-169.hsd1.ga.comcast.net] has quit [Ping timeout: 246 seconds] 14:29:17 -!- antgreen [~user@CPE00222d6c4710-CM00222d6c470d.cpe.net.cable.rogers.com] has quit [Remote host closed the connection] 14:55:52 -!- Krystof [~csr21@cpc2-dals3-0-0-cust1263.hari.cable.virginmedia.com] has quit [Ping timeout: 246 seconds] 15:17:14 tsuru [~charlie@adsl-87-47-213.bna.bellsouth.net] has joined #sbcl 15:28:46 -!- woudshoo [~user@ipleiden.intellimagic.net] has quit [Ping timeout: 246 seconds] 15:38:17 stassats [~stassats@wikipedia/stassats] has joined #sbcl 15:57:23 ignotus [~ignotus@unaffiliated/ignotus] has joined #sbcl 15:57:26 rmarynch [~roman@88.135.194.233] has joined #sbcl 15:57:40 Hello people 15:57:48 hi, with which :external-format can I open unicode files? It seems like "utf16" 15:58:35 <|3b|> 'unicode' is a character set, not a description of a file, you need to figure out how the unicode characters are encoded 15:59:15 |3b|: it is an excel file saved as "Text (unicode) 15:59:33 I don't know what possible keywords I can use as an external-format, I would like to experiment 15:59:59 <|3b|> windows stuff might use utf16 16:00:38 The value NIL is not of type FUNCTION. I get this:-/ like it is a non-existent external format 16:02:08 Here are the definitons for the encodings (just grepped for :utf-8) /usr/share/sbcl-source/src/code/octets.lisp, I don't see utf-16 here 16:04:11 (apropos "utf-16" :keyword) 16:04:21 *|3b|* fails to remember how to get th elist of supported external formats out of sbcl, maybe try :utf16le ? 16:04:50 and this NIL is not of type FUNCTION should be really easy to fix, shouldn't it? 16:04:55 there should be a nicer way to look for external formats, yeah 16:05:28 nikodemus |3b| utf-16le works, thanks! 16:08:16 slime should have auto-completion for :external-format argument 16:08:58 I have another excel file, it is saved as CSV (MSDOS) or something like that, I don't recognise the encoding: http://89.133.33.26:6882/csv maybe something like cp1250 (I don't remember these old stuff well), I can't get SBCL to read them correctly 16:08:59 that'd be right after fuzzy completion has context-sensitivity support, though it can be added to c-p-c already 16:10:29 <|3b|> :cp1250 ? 16:11:26 |3b|: ah, by grepping for that I found /usr/share/sbcl-source/src/code/win32.lisp, there are some things to try there, thanks:D 16:12:21 (maphash (lambda (key value) (print key)) sb-impl::*external-formats*) 16:12:35 <|3b|> sb-impl::*external-formats* seems to have them 16:12:46 *|3b|* is slow :( 16:13:01 thanks 16:15:07 (maphash (lambda (key value) (format t "~%----------~%~S~%~a" key (ignore-errors (m:slurp-text-file "/tmp/2/csv" key)))) sb-impl::*external-formats*) huh, by looking through the results none of them is right, I may have to drop this csv support thing:-/ 16:18:03 (describe '*) says that * is a primitive type specifier. But it is not listed as a _type specifier_ here: http://www.lispworks.com/documentation/HyperSpec/Body/04_bc.htm 16:18:46 I am a bit confused by (the * 4) working fine. Is that right? 16:21:33 As I understood from that page, * can only be a member of a compound type specifier, right? 16:25:28 rmarynch: * can also refer to the last value printed on the REPL (: 16:25:41 oh, wait, wrong 16:25:48 and it can be a function, too 16:25:51 sorry, thought THE evaluates 16:26:20 Yes, I learned this from DESCRIBE :) 16:26:39 why is there no (a TYPE x)? 16:26:39 But what about types? 16:36:10 rmarynch: * isn't a real common lisp type in general, no 16:36:58 it's a wild type parameter, but sbcl parses it into *wild-type* and pretends its a real type 16:37:41 do we want (the * 4) to be fixed too? 16:37:53 theoretically yes 16:38:56 (subtype '* 't) is confused as well, report this? 16:39:12 (as in, i'm not losing any sleep over it, but if someone feels up to untangling *wild-type* related assumptions from sbcl, that'd be great -- just be prepared to be in for a confusing ride) 16:39:22 don't report separately 16:39:44 add to the same bug -- since it's really the same issue with * being treated as an actual type 16:40:26 fine, I will try to fix (the (values *) ...) first, and then we can look further 16:55:26 mon_key [~user@unaffiliated/monkey/x-267253] has joined #sbcl 16:57:44 any idea why the sb-posix:cc slot of sb-posix:termios doesn't get its slot documentation? 16:58:42 The first plausibility to come to mind is there not being any slot documentation defined. 16:58:53 its there on 1.0.47 16:59:07 in interface.lisp 16:59:10 By "there", you mean defined, or in the system? 16:59:25 nyef: Sorry for the define-protocol-class form 17:00:14 its present as: :documentation "Control characters" 17:00:37 Yeah, there's also an :array-length option unique in the class to that slot. 17:00:52 Yes, likely this is the culprit. 17:01:03 Which may be significant, as it apparently has special handling in the macro. 17:02:39 Ah, the macro DEFINE-PROTOCOL-CLASS has special cases for :array-length 17:02:40 What does LDIFF do? 17:03:14 It might be that :array-length has to be the last slot option in order to not screw anything else up. 17:03:24 ...? "Return a new list, whose elements are those of LIST that appear before OBJECT" 17:04:24 Okay, so LDIFF is a hidden gotcha in the macro. 17:04:33 Arguably, the macro should be fixed and not its uses. 17:04:42 (Otherwise this could trip someone else up again later.) 17:05:43 maybe: if array-length ... (getf (cdr slotd) :documentation) ... ? 17:06:18 Umm... doesn't seem right, somehow. 17:07:46 What happens if someone puts array-length before initarg? 17:07:46 :) there are at least three special cases around :array-length in that macro and it would seem that :documentation should (must?) appear after :array-length should that occur 17:09:08 The entire interface pivots on :array-length... Likely much other stuff breaks if it appears out of line? 17:09:33 Arguably, the first mistake is using loop for slotd in slots instead of for (slot-name . slot-options) in slots. 17:10:16 Krystof [~csr21@csrhodes.plus.com] has joined #sbcl 17:10:16 -!- ChanServ has set mode +o Krystof 17:11:01 And beyond that, no, looks like :array-length is only read by GETF, which is position-invariant, and the rest of the slots are pasted directly as a defclass slot description. 17:12:07 I see. 17:12:22 What about the call to MEMBER? 17:12:52 NM 17:12:52 Ah! How about (loop for slotd in slots for array-length-option = (member :array-length slotd) collect (append (ldiff slotd array-length-option) (cddr array-length-option))) ? 17:13:15 A touch ugly, but should splice out the array-length option well enough. 17:14:26 will the member see eql for :array-length? 17:14:52 How do you mean? 17:15:28 Prob. a stupid question... Just wondering if maybe the identity is somehow goofy b/c of the macro semantics? 17:15:46 I don't believe so, no. 17:15:47 e.g. defmacro over loop 17:16:09 Hum. Is termios the only use of array-length? 17:16:29 Looks like it. 17:16:39 AFAICT 17:16:55 define-termios-protocol-class? 17:18:05 Hang on, doing a make-target-contrib build. 17:19:05 How can I tell if the slot has documentation or not? 17:19:21 for which? 17:19:32 termios-cc or whatever it's called. 17:19:46 (describe 'sb-posix:termios t)? 17:20:12 Okay. I'll try that once I get a prompt back. 17:20:32 rgrep contrib/ returns define-protocol-class only in /sb-posix/interface.lisp 17:20:55 Wasn't expecting any hits outside of sb-posix anyway. 17:21:15 Was curious if maybe anyone had piggybacked on it. 17:21:32 FWIW its used for flock, passwd, group, timeval, stat, and termios 17:22:45 It's an internal symbol, so anyone piggybacking has no excuse. 17:24:04 Indeed as the manual states explicitly 17:24:21 "The functionality contained in the package `SB-UNIX' is for SBCL internal use only; its contents are likely to change from version to version." 17:25:56 This is SB-POSIX, not SB-UNIX. 17:26:02 But still, unexported symbol. 17:27:44 thats from the sb-posix node of the manual :) 17:28:49 *nyef* would probably have written "while its contents are unlikely to change drastically from version to version, the developers will be unsympathetic if they do and it breaks anything not internal to SBCL itself." 17:29:32 Less "these are volatile", more "we don't care if we break other peoples stuff that uses this". 17:29:35 Also, the define-protocol-class glitch has the additional unintended consequence of preventing the SBCL documentation facility from generating _any_ reference to sb-posix:cc slot in the manual 17:30:57 e.g. (info "(sbcl)sb-posix") 17:31:27 nyef pasted "For mon_key: this should fix things." at http://paste.lisp.org/display/120989 17:32:24 it was the period?*!#@! 17:33:27 Not really, but I noticed the period when I tested the other bit. :-P 17:33:31 Ah, NM just a note about the diff 17:33:54 There's also a protocol-class in there with no documentation. 17:34:11 group / alien-group. 17:34:31 yea 17:35:50 And committed. 17:37:02 FWIW sb-posix:size slot of stat formats poorly in manual and with describe. 17:37:18 -!- slyrus [~chatzilla@adsl-76-254-45-145.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 250 seconds] 17:38:04 Yeah, looks like it would. 17:38:28 Umm... the same possibly applies to the structure as a whole. 17:39:18 Feel free to open a bug on that one. :-p 17:39:38 NP 17:39:51 I have fixed (VALUES *) by explicitly scanning the list members to find * during IR1, in VALUES-SPECIFIER-TYPE. Not the best way, but it works, and seems to be safe. 17:46:40 (safe means 'tests pass') 17:46:43 christop` [~user@oteiza.siccegge.de] has joined #sbcl 17:48:07 hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has joined #sbcl 18:02:48 -!- rmarynch [~roman@88.135.194.233] has quit [Quit: Leaving] 18:06:12 rmarynch [~roman@88.135.194.233] has joined #sbcl 18:16:11 gor[e] [~svr@79.165.187.105] has joined #sbcl 18:21:13 -!- homie [~levgue@xdsl-78-35-143-164.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 18:26:50 homie [~levgue@xdsl-78-35-143-164.netcologne.de] has joined #sbcl 18:26:57 -!- rmarynch [~roman@88.135.194.233] has quit [Quit: Leaving] 19:05:29 slyrus [~chatzilla@173-228-44-88.dsl.static.sonic.net] has joined #sbcl 20:26:04 nyef: where should I look at to change the fasl format ? 20:26:49 fe[nl]ix: SYS:SRC;CODE;FOP.LISP, SYS:SRC;CODE;LOAD.LISP, SYS:SRC;COMPILER;DUMP.LISP, SYS:SRC;COMPILER;GENERIC;GENESIS.LISP. 20:26:50 nyef: I'd like to have compile-file compute a hash of the source file's contents and put it into the fasl 20:27:24 Hrm. Timestamps no longer good enough? 20:27:54 I'm not even sure it's a good idea, but I want to play with it 20:29:52 Header might be easier to play with than the full format. 20:30:04 Might not even need to tweak GENESIS, as I think that uses LOAD-FASL-GROUP. 20:30:50 And at least there's no longer the mind-twisting FOP-NORMAL-LOAD / FOP-MAYBE-COLD-LOAD stuff. 20:47:05 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Ping timeout: 240 seconds] 21:05:36 -!- homie [~levgue@xdsl-78-35-143-164.netcologne.de] has quit [Read error: Connection reset by peer] 21:06:47 homie [~levgue@xdsl-78-35-170-253.netcologne.de] has joined #sbcl 21:07:16 -!- hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has quit [Remote host closed the connection] 21:07:42 hargettp_ [~hargettp_@dhcp-162.mirrorimage.net] has joined #sbcl 21:21:34 -!- ignotus [~ignotus@unaffiliated/ignotus] has quit [Ping timeout: 246 seconds] 21:24:01 ignotus [~ignotus@catv-89-133-33-26.catv.broadband.hu] has joined #sbcl 21:24:01 -!- ignotus [~ignotus@catv-89-133-33-26.catv.broadband.hu] has quit [Changing host] 21:24:01 ignotus [~ignotus@unaffiliated/ignotus] has joined #sbcl 22:16:03 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 22:32:31 hargettp [~hargettp@96.237.121.111] has joined #sbcl 23:33:44 -!- hargettp [~hargettp@96.237.121.111] has quit [Quit: Leaving...] 23:42:43 tcr [~tcr@217-162-131-235.dynamic.hispeed.ch] has joined #sbcl 23:53:00 antgreen [~user@CPE00222d6c4710-CM00222d6c470d.cpe.net.cable.rogers.com] has joined #sbcl