00:19:34 -!- leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has quit [Quit: sleep] 00:20:51 leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has joined #ccl 00:21:25 -!- leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has quit [Client Quit] 01:03:15 so, I've got a background thread that in the course of its working calls some structure constructors 01:03:42 when I recompile/reload the file including the structure def, the background thread throws errors that the constructor is undefined 01:04:50 instead of just using the new one when it's available 01:05:22 of course, there are various race conditions involved in here, but that the function would be undefined in the middle of the reload surprised me 01:11:29 defstruct will fmakunbound the old structure accessors/constructors/predicate etc. as part of redefinition. 01:12:11 But it does sound like you're getting really unlucky. 01:14:43 some sort of delta detection would be nice, where only those which don't exist in the new version would get fmakunbound'ed 01:14:56 at least to avoid this immediate error 01:15:32 the ability to reload into a running threaded image would be nice without having to write in extra hand-holding 01:15:38 not sure how much work that would be internally 01:23:00 It's a pretty tall order generally, I think, especially given that CLOS is basically a giant shared data structure. 01:36:33 hmm, yeah, if structures are implemented via CLOS 01:36:58 it's really nice to see threaded behavior change in realtime when loading new defuns from an edited file 01:37:31 so the structures not doing that was pretty jarring 02:25:22 rme: Heh. I see you're following darwin-dev for some reason. 02:26:10 Yes, I lurk there. One normally doesn't expect to get such comedy in a forum like that. 02:27:52 Oh, it's been an expectation for darwin-dev for a long time now, going back to proclus and the angry GNU/Darwin rants, the famous "badly with the buttocks" comment, and the strange esteban guy. If it weren't for the comedy, I'd have unsubscribed ages ago. 02:29:32 http://lists.apple.com/archives/Darwin-dev/2006/Aug/msg00200.html - and Jordan Hubbard's followup at http://lists.apple.com/archives/Darwin-dev/2006/Aug/msg00206.html 02:34:49 hah. classic. 02:38:40 *gbyers* just reads darwin-dev in digest form, but will wait for it if it's that good. 03:47:15 bfulgham_ [~brent@adsl-69-234-96-102.dsl.irvnca.pacbell.net] has joined #ccl 03:47:25 -!- bfulgham_ [~brent@adsl-69-234-96-102.dsl.irvnca.pacbell.net] has quit [Remote host closed the connection] 03:50:57 bfulgham_ [~brent@adsl-69-234-96-102.dsl.irvnca.pacbell.net] has joined #ccl 04:01:03 -!- bfulgham_ [~brent@adsl-69-234-96-102.dsl.irvnca.pacbell.net] has quit [Read error: Connection reset by peer] 04:01:10 bfulgham_ [~brent@adsl-69-234-96-102.dsl.irvnca.pacbell.net] has joined #ccl 04:29:15 -!- rme [~rme@pool-70-106-132-157.chi01.dsl-w.verizon.net] has quit [Quit: rme] 05:38:09 nunb [~nundan@59.178.217.120] has joined #ccl 06:48:20 -!- bfulgham_ [~brent@adsl-69-234-96-102.dsl.irvnca.pacbell.net] has quit [Quit: bfulgham_] 08:19:52 -!- nunb [~nundan@59.178.217.120] has quit [Read error: Connection reset by peer] 08:20:19 nunb [~nundan@59.178.207.173] has joined #ccl 08:26:31 hlavaty [~user@77-22-102-139-dynip.superkabel.de] has joined #ccl 08:54:58 tfb [~tfb@212.183.140.37] has joined #ccl 09:39:43 -!- tfb [~tfb@212.183.140.37] has quit [Ping timeout: 240 seconds] 10:37:25 tfb [~tfb@212.183.140.55] has joined #ccl 11:04:26 leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has joined #ccl 11:58:14 -!- tfb [~tfb@212.183.140.55] has quit [Ping timeout: 258 seconds] 12:32:55 tfb [~tfb@212.183.140.0] has joined #ccl 12:40:02 -!- billstclair [~billstcla@unaffiliated/billstclair] has quit [Remote host closed the connection] 12:41:13 billstclair [~billstcla@unaffiliated/billstclair] has joined #ccl 12:44:26 palter [~palter@2002:4b44:b1e1:0:5ab0:35ff:fe78:6749] has joined #ccl 12:44:41 -!- palter [~palter@2002:4b44:b1e1:0:5ab0:35ff:fe78:6749] has quit [Client Quit] 13:52:48 -!- alms [~alms@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [Quit: alms] 14:00:05 alms [~alms@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 14:19:53 -!- nunb [~nundan@59.178.207.173] has quit [Read error: Connection reset by peer] 14:20:21 nunb [~nundan@59.178.213.150] has joined #ccl 15:04:26 anRch [~markmilli@64.134.242.124] has joined #ccl 15:14:51 -!- nunb [~nundan@59.178.213.150] has quit [Quit: This computer has gone to sleep] 15:32:35 gz` [~gz@209-6-40-245.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com] has joined #ccl 15:33:30 -!- gz` [~gz@209-6-40-245.c3-0.smr-ubr1.sbo-smr.ma.cable.rcn.com] has quit [Client Quit] 15:44:03 -!- leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has quit [Quit: rcirc on GNU Emacs 23.1.97.1] 16:10:24 Fare [~Fare@ita4fw1.itasoftware.com] has joined #ccl 16:14:53 -!- anRch [~markmilli@64.134.242.124] has quit [Ping timeout: 252 seconds] 17:03:47 rme [~rme@pool-70-106-132-157.chi01.dsl-w.verizon.net] has joined #ccl 17:22:15 roffe [~roffe@89.191.2.214] has joined #ccl 17:23:01 I think perhaps I have a problem with 1.5. 17:24:55 What's the problem? 17:27:01 anRch [~markmilli@64.134.67.77] has joined #ccl 17:30:12 -!- tfb [~tfb@212.183.140.0] has quit [Ping timeout: 258 seconds] 17:37:25 leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has joined #ccl 17:38:58 I keep getting the error message "Error: Unbound variable: DIRECTORY-NAMESTRING" 17:39:23 It seems that it happens when ccl looks for a file and doesn't find it ... 17:39:57 Problem is, (require 'does-not-exist) gives this error message, not nil which I think is the correct answer 17:40:46 platform is MacOS X, Snow Leopard, 64 bit implementation 17:45:18 obvious workaround: don't do it 17:50:02 No problem as far as the code I write is concerned, but lots of contributed code relies on being able to look for files that arent there ... 17:51:32 uh? such as? 17:52:09 (can you get a backtrace?) 17:52:15 swank-ccl that comes with slime, for example ... 17:53:02 lisppaste5: url 17:53:03 To use the lisppaste bot, visit http://paste.lisp.org/new/ccl and enter your paste. 17:54:05 roffe pasted "directory-namestring problem" at http://paste.lisp.org/display/99144 17:55:27 this is just an example of the problem ... are there any debug swiches I should turn on to get a decent backtrace? 17:56:54 swank-ccl.lisp does _not_ contain the symbol DIRECTORY-NAMESTRING 17:59:03 Where is your slime installed? /usr/local/source/slime, I assume. Why is /usr/local/share/emacs/site-lisp/ involved here? 18:02:23 one is a symlink to the other ?  18:06:28 removed symlink, copied, consistently refers to /usr/local/share/emacs/site-lisp/ consistently now, same thing happens. 18:07:39 Everything works if I do 18:07:46 (setq directory-namestring nil) 18:07:48 roffe: can you get a debugger prompt at the error (maybe with (setf *break-on-signals* 'error) and ask a :b backtrace? 18:07:57 sure 18:08:19 (restart the lisp to undo the setq) 18:08:28 (or makunbound the symbol) 18:10:06 roffe pasted "directory-namestring problem" at http://paste.lisp.org/display/99146 18:11:31 haha, might be MY problem, after all... 18:12:13 do tell ... :) 18:12:37 I'm a newbie to ccl (not really to lisp though), feel free to flame me ... 18:12:53 roffe: what does (or #+asdf2 (asdf:asdf-version) #+asdf :old) return ? 18:13:08 in a clean Lisp? 18:13:13 in the erroring one 18:13:42 :old 18:13:46 roffe is using 1.5, which doesn't include asdf2. 18:13:56 I blame old ASDF :) 18:14:15 ¡no me digas! 18:14:21 roffe: does your problem go away if you upgrade ccl's ASDF to the latest from git? 18:14:41 ok, just give me a few hours to find out how to do that ... 18:15:04 it the asdf git checkout, run: ccl --eval '(progn(compile-file "asdf.lisp" :output-file (format nil "~Atools/asdf.lx64fsl" (ccl::ccl-directory)))(quit))' 18:15:22 (if ccl is installed as root you may have to sudo that) 18:15:47 to get an asdf checkout, try: git clone git://common-lisp.net/projects/asdf/asdf.git 18:15:51 i never used git, I've just used asdf-install 18:16:40 problem: I broke asdf-install with the latestest ASDF 2, and the patch that fixes it hasn't made it upstream yet... 18:16:59 roffe: what's the value of asdf:*central-registry* 18:16:59 apt-get install git (or whatever the darwin equivalent is) 18:18:06 rme: ("home:.asdf-install-dir;systems;" "ccl:tools;asdf-install;" DIRECTORY-NAMESTRING *DEFAULT-PATHNAME-DEFAULTS*) 18:18:29 rme: bingo! 18:18:29 but no problem, it'll be a bit of manual install, but darwin comes with git. there's always a first time for everything. 18:19:00 roffe: your *central-registry* is wrong. 18:19:22 rme: congratulations! 18:19:32 all hail, etc. 18:19:50 *rme* bows 18:19:58 what, er, should it be? 18:19:58 roffe: whichever script or program sets up your *central-registry* is borked. 18:21:08 that just might be asdf 18:21:28 try (require :asdf) in a fresh ccl, and look at the variable. 18:22:09 You probably set up asdf:*central-registry* in your init file. If you want to annotate one of your pastes with the contents of your init file, maybe that would show the problem. 18:22:09 the bug is in my init file ... i created a workaround for _another_ problem 18:22:23 arth 18:23:07 anyway 18:23:17 here's my init file (with no init file, things seem to work) 18:23:54 roffe pasted "directory-namestring problem" at http://paste.lisp.org/display/99147 18:24:11 with the init file, lisp breaks on 18:24:23 (pushnew "ccl:tools;asdf-install;" asdf:*central-registry* :test #'string-equal) 18:25:13 peterbb [~peterbb@ves1-1x-dhcp324.uio.no] has joined #ccl 18:25:17 roffe annotated #99147 "directory-namestring problem" at http://paste.lisp.org/display/99147#1 18:25:58 I assume an upgrade to asdf2 might fix this 18:28:38 roffe: might. And then, I recommend using the source-registry instead of *central-registry* (see the doc) 18:29:23 it seems to me that asdf does not upgrade as easily as CPAN :) 18:29:35 roffe: not yet 18:29:52 roffe: but the hope is it soon will, thanks to said source-registry. 18:31:18 my project is to write the code to typeset a norwegian-to-swahili dictionary. the text is a csv file. I used perl in 1997. I don't care if I have to tweak a bit. 18:31:33 I still feel the pain of 1997. 18:31:55 roffe: I hope my fare-csv library can help. 18:32:01 -!- peterbb [~peterbb@ves1-1x-dhcp324.uio.no] has quit [Remote host closed the connection] 18:32:20 roffe: Here are a couple of comments. First, when adding components to asdf:*central-registry*, I recommend using #'equal as the test function. *central-registry* may contain things other than strings. 18:32:46 peterbb [~peterbb@ves1-1x-dhcp324.uio.no] has joined #ccl 18:33:12 (yup, or use ASDF 2 and configure source-registry instead of central-registry :) ) 18:33:53 Fare: I've used Alain Picard's cvs.lisp. Sorry about that. 18:34:10 I hope we can still be friends. 18:34:28 roffe: I hate your guts. 18:34:30 Second, there is no need to push anything onto *module-provider-functions*. CCL already contains code to have require load asdf systems. 18:34:48 roffe: Alain's library is OK. It shamed me into making mine better. 18:35:03 (which it initially wasn't at all) 18:35:17 I havent looked at either for a while. Alain's works for now. 18:36:24 You're doing a great job. Lisp lacks a CPAN. Everybody keeps saying that the real problem with Lisp is the lack of standardized libraries, not the parentheses. 18:36:59 we'll see if source-registry actually allows for improved source-distribution tools. 18:37:00 Everybody as in those I know who have worked with Lisp. 18:37:59 first step was implement it, then debug ASDF enough to release it. Now it's get ASDF2 widely distributed, then get clbuild, asdf-install and such to adopt it. 18:38:55 I have asdf2 now, fresh from git 18:39:09 uninstalling asdf is no problem 18:39:30 but installing asdf2 from source files ... i guess its a matter of reading some docs. 18:40:11 roffe: the ccl trunk includes asdf2 and a patched asdf-install. 18:40:35 oh 18:44:09 it the asdf git checkout, run: ccl --eval '(progn(compile-file "asdf.lisp" :output-file (format nil "~Atools/asdf.lx64fsl" (ccl::ccl-directory)))(quit))' 18:44:22 rme: which version did you check in? 18:44:46 1.719 18:44:52 (there's some noticeable breakage introduced in 1.714 and fixed in 1.719) 18:44:54 good 18:45:54 I removed the top-level (declaim (optimize ...)), but I see that you commented that out in 1.720, so that's good. 18:46:27 in a clean ccl, (require 'asdf) works but not (require 'asdf-install) 18:46:58 rme: I did integrate your feedback in 1.720 indeed. 18:47:12 an 1.722 is coming with the latest CMUCL tweak. 18:47:34 (I'm waiting for more feedback, if I get none today, will release 1.722 with just that) 18:48:19 roffe: right. you have to add #p"ccl:tools;asdf-install;" to asdf:*central-registry* before (require 'asdf-install) will work. 18:49:02 are you sure you don't mean source-registry? 18:49:18 or just use asdf2 and put that path in your source-registry 18:50:14 source-registry is ... ? 18:50:35 rme: do we want (:tree #p"ccl:tools;") at the end of the asdf::wrapping-source-registry ? 18:51:00 reading http://common-lisp.net/project/asdf/asdf/FAQ.html now ... 18:52:41 roffe: mkdir -p ~/.config/common-lisp/source-registry.conf.d/ ; echo > ~/.config/common-lisp/source-registry.conf.d/90-ccl-tools.conf '#+clozure (:tree #p"ccl:tools;")' 18:52:46 roffe: or something 18:54:30 (require 'asdf-install) still doesn't work 18:54:55 roffe: what does (or #+asdf2 (asdf:asdf-version) #+asdf :old) return ? 18:55:20 NIL 18:55:30 so, first (require :asdf) 18:55:41 done 18:55:42 THEN check that you have 1.719 18:55:48 THEN (require :asdf-install) 18:56:31 Fare: I'm not sure I like that idea (adding (:tree #p"ccl:tools;")) myself. 18:57:37 ? (asdf:asdf-version) 18:57:38 > Error: Reader error: No external symbol named "ASDF-VERSION" in package # . 18:57:56 rme: why not? 18:58:05 especially at the END of the source-registry. 18:58:15 milanj [~milanj_@93.87.151.235] has joined #ccl 18:58:24 roffe: you're still using the old asdf, then. 18:58:29 Fare: just conservatism. I don't really feel that strongly about it, though. 18:58:54 rme: the alternative is to require all users of these tools to setup the source registry themselves... 18:59:10 roffe: did you install the new one on top, as I suggested? 18:59:23 Is making asdf-install hard to use such a bad thing? :P 18:59:31 I have ccl 1.5, and the asdf that comes with it 19:00:18 roffe: I recommend using the latest asdf, but I'm biased 19:00:25 sure 19:00:41 but if I'm in the git repositoy 19:00:46 and say (load "asdf") 19:00:55 works too 19:00:57 then (asdf:asdf-version) says 1.721 19:01:02 wonderful 19:01:20 you can replace the ccl asdf with it the asdf git checkout, run: ccl --eval '(progn(compile-file "asdf.lisp" :output-file (format nil "~Atools/asdf.lx64fsl" (ccl::ccl-directory)))(quit))' 19:01:54 but there's a problem again: 19:02:18 roffe annotated #99147 "directory-namestring problem" at http://paste.lisp.org/display/99147#2 19:03:07 asdf2 seems to stumble over asdf1 19:04:57 roffe: if you want to use asdf2 with ccl 1.5, then you will need to patch your asdf-install as in http://trac.clozure.com/ccl/changeset/13689 19:05:11 As I said before, in the trunk, this is already done. 19:05:50 I sort of thought i had the latest svn 19:06:07 And your problem was never about asdf1 vs. asdf2, per se. You were just setting up asdf:*central-registry* incorrectly. 19:06:14 rme: I suppose asdf2 is not going into 1.5, is it? 19:06:53 roffe: there are two main branches of ccl. there is the trunk, which is where bleeding-edge development takes place, and there is the current release branch, which is stable, and gets only important bug fixes. 19:07:17 leo2007: I wasn't planning to put asdf2 into 1.5. 19:07:30 I'm sort of dependant on a production environment here :) 19:08:31 *Fare* notices that that makes yet another divergent version of asdf-install. 19:08:52 how many are there in the wild, with no clear version control? 19:09:26 recompiling ... 19:09:31 rme: btw, when you patch asdf, maybe you should add a .1 to the version, as in "1.719.1" or some such. Yuck. 19:09:34 rme: ccl usually releases every 6 months? 19:10:14 Fare: you are probably right. I will do that in future, but I'm hoping that I won't have to make any further local patches to asdf. 19:10:42 so am I hoping :) 19:11:48 leo2007: Yes, we're trying to do releases every six months or so. 19:13:39 attempting asdf-install after recompile 19:13:50 (I remember when recompiling Emacs took a day) 19:14:00 Error: Server responded 404 for GET http://www.cliki.net/asdf?download 19:15:00 roffe: that's exaggerating. It shouldn't be more than 1 hour unless you are using a machine from 1995. 19:15:15 I'm 43. 19:15:32 ok 19:15:32 we're talking 1989 here. 19:15:37 ok 19:16:06 You have my sympathy. But usually you can grab the precompiled version. 19:19:11 OK, now asdf-install exhibits the usual behavior of asking where I want to store the stuff, but doesn't download anything, gives a 404 error 19:20:29 (asdf-install:install 'csv) => Error: Server responded 404 for GET http://www.cliki.net/CSV?download 19:20:37 So close, yet so far away 19:22:25 is CSV meant to be asdf-installable? 19:22:34 if not, you can fix its cliki page 19:23:10 my bad 19:23:14 fare-csv it is 19:23:21 (asdf-install:install 'fare-csv) 19:23:55 roffe annotated #99147 "directory-namestring problem" at http://paste.lisp.org/display/99147#3 19:26:00 warmer but still no cigar 19:29:36 I have installed split-sequence. 19:29:40 which I actually used. 19:29:42 use. 19:29:49 but not asdf itself. 19:33:08 roffe: :c 2 means you skip the signature 19:33:23 :R for a list of restarts 19:34:25 the signature is not a problem, I've turned it off 19:34:33 but it seems that asdf-install will not install asdf itself 19:35:18 -!- anRch [~markmilli@64.134.67.77] has quit [Quit: anRch] 19:38:57 roffe: I am confused. You already have asdf, right? 19:39:54 Yes, If I'm in the git repository and say (load "asdf") then I can install stuff with asdf2 19:40:49 but I'm not all that familiar with ccl and would sort of like to be able to skip the (load "asdf") part and have it installed properly 19:41:30 since cvs.lisp is not in asdf, I'd like to know how to place it where (require) can find it 19:41:38 but I think I can read myself up on that 19:42:24 I feel like I'm not communicating well. A version of asdf comes with ccl 1.5. You load it with (require 'asdf). 19:43:08 roffe: asdf-install never installs asdf 19:43:29 I've compiled ccl 19:43:33 1.5 19:44:08 There is also a version of asdf-install that comes with ccl 1.5. After you have done (require 'asdf), you load asdf-install by doing (push #p"ccl:tools;asdf-install;" asdf:*central-registy*) and then (require 'asdf-install). 19:45:18 or (asdf:oos 'asdf:load-op :asdf-install), of course. 19:45:30 ? (require 'asdf) 19:45:31 ASDF 19:45:31 ("ASDF") 19:45:31 ? (push #p"ccl:tools;asdf-install;" asdf:*central-registy*) 19:45:31 > Error: Reader error: No external symbol named "*CENTRAL-REGISTY*" in package # . 19:45:52 spelling error, sorry: it's asdf:*central-registry* 19:46:04 it seems my ccl 1.5, fresh from cvs up, may not be what it ought to be 19:46:20 but no matter, it's almost ten in my time zone. I'll call it a day. 19:46:55 thanks so far, I have something that works and that's enough just now. you've been very patient. keep up the good work. 19:47:12 I'm was a schoolmate of Erik Naggum, by the way. 19:48:58 Good luck. I'm often here in #ccl. 19:50:13 thanks, and good night. 19:50:17 -!- roffe [~roffe@89.191.2.214] has quit [Quit: roffe] 20:32:33 so should I include anything from ccl: in the source-registry? 20:37:50 My opinion is that the answer should be no. I don't want to have to worry about asdf possibly altering the behavior of require for anything in the ccl directory. If users want asdf-install, they can add its path to their asdf configuration manually. 20:39:24 ok 20:46:23 roffe [~roffe@cm-84.215.145.23.getinternet.no] has joined #ccl 21:12:11 -!- milanj [~milanj_@93.87.151.235] has quit [Ping timeout: 248 seconds] 21:16:21 milanj [~milanj_@93.87.150.90] has joined #ccl 22:24:43 anRch [~markmilli@64.134.243.163] has joined #ccl 22:25:20 -!- alms [~alms@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [Quit: alms] 22:36:20 alms [~alms@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 22:58:21 -!- milanj [~milanj_@93.87.150.90] has quit [Quit: Leaving] 23:00:22 -!- roffe [~roffe@cm-84.215.145.23.getinternet.no] has quit [Remote host closed the connection] 23:00:39 roffe [~roffe@vpn-client192.uio.no] has joined #ccl 23:09:36 -!- leo2007 [~leo@cpc1-cmbg13-0-0-cust596.cmbg.cable.ntl.com] has left #ccl 23:21:59 -!- anRch [~markmilli@64.134.243.163] has quit [Quit: anRch] 23:30:49 -!- peterbb [~peterbb@ves1-1x-dhcp324.uio.no] has quit [Quit: Leaving]