01:25:09 LiamH [n=nobody@pool-72-75-101-81.washdc.east.verizon.net] has joined #ccl 01:26:58 I'm noticing for some of my functions CCL returns the wrong answer unless the function is traced. Does anyone recognize this problem? 01:27:55 Doesn't sound familiar right off hand. Do you have a simple example? 01:28:12 Unfortunately, not simple. 01:28:58 It happens when I make the call inside a let form. As an alternative to tracing it, if I take it out of the let and use defparameter, it works. 01:39:14 You could try compiling with (DEBUG 3) to see if that works. If it does, then that would make me suspect a bug involving register assignment. 01:46:10 Your let form (that you replace with defparameter in the test), does it bind a special variable? 02:01:27 gz: no 02:04:28 But defparameter does. 02:04:40 rme: setting (declaim (optimize (debug 3))) and recompiling everything does not help 02:04:50 gz: yes, usually 02:05:17 I have put together an email for the mailing list. 02:29:55 -!- mdc_mobile [n=mdc_mobi@ds9.entity.com] has quit [] 02:49:04 mdc_mobile [n=mdc_mobi@ds9.entity.com] has joined #ccl 03:41:20 LiamH: still there? 03:41:54 yes 03:45:00 The type of data in a gsl_permutation is a size_t. That's a 64 bit type. It looks like you'd be copying over 32-bit words. 03:46:26 scratch that please. what am i thinking. 03:46:41 OK 03:46:41 erase that completely 03:46:52 you had me going there for a sec 03:48:42 what's (element-type (make-permutation 4 t)) in a 64-bit lisp? 03:49:38 (UNSIGNED-BYTE 64) 03:50:24 for both CCL and SBCL, the only 64 bit lisps I have working (CLISP has a bug on 64 bit platforms) 03:51:11 I was looking at your code in 32-bit CCL. Bad idea, probably. 03:52:25 No, I'd like to get GSLL running on as many platforms as I can. 03:52:41 I guess you're on PPC? 03:52:52 x86, actually. 03:53:09 Didn't think CCL existed for 32 bit x86. 03:53:42 The trunk version runs on 32-bit x86. 03:54:25 OK, maybe I should try that (I have a 32 bit emulator chroot). 03:55:51 CLISP on 32 bits falls over on 64 bit tests, I'm wondering if that's CLISP, CFFI, or something about the 32 bit glibc/kernel. 04:17:39 rme pasted "gsl permutations" at http://paste.lisp.org/display/72863 04:19:01 LiamH: the act of printing a permutation object before calling permutation-reverse is what seems to be making it work. 04:19:27 yes, gbyers got me thinking that too 04:19:59 there is a C to CL synchronization step which is triggered by printing 04:20:34 If I break on gsl_permutation_reverse in gdb, and call your permutation-reverse without printing the permutation object first, the data is all zeros on entry to gsl_permutation_reverse. 04:20:59 Hmm 04:21:17 -!- mdc_mobile [n=mdc_mobi@ds9.entity.com] has quit [] 04:21:54 weird, because I did a macroexpansion and confirmed that (COPY-CL-TO-C P) is in there 04:23:58 This requires some thought, for which I'm to tired to conduct at the moment. 04:24:23 I'm going to think about this tomorrow. Thanks for your help. 04:24:32 Good luck and good night. 04:26:48 Thanks, and thanks again for your help. 04:26:59 -!- LiamH [n=nobody@pool-72-75-101-81.washdc.east.verizon.net] has quit ["Leaving."] 08:11:38 -!- rme [n=rme@pool-70-105-87-66.chi.dsl-w.verizon.net] has quit [] 12:05:31 H4ns [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has joined #ccl 12:22:04 segv__ [n=mb@p4FC1F343.dip.t-dialin.net] has joined #ccl 12:22:33 -!- segv_ [n=mb@p4FC1CCE3.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 13:05:48 H4ns1 [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has joined #ccl 13:06:28 mdc_mobile [n=mdc_mobi@ds9.entity.com] has joined #ccl 13:29:19 -!- H4ns [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has quit [Read error: 110 (Connection timed out)] 13:31:36 -!- H4ns1 is now known as H4ns 14:03:39 LiamH [n=none@common-lisp.net] has joined #ccl 14:29:50 milanj [n=milan@77.46.248.69] has joined #ccl 14:58:20 -!- H4ns [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has quit [Read error: 104 (Connection reset by peer)] 15:02:07 H4ns [n=hans@dialin-145-254-047-187.pools.arcor-ip.net] has joined #ccl 15:21:11 -!- segv__ is now known as segv 16:09:58 anRch [n=markmill@nmd.sbx09269.readima.wayport.net] has joined #ccl 16:41:58 sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has joined #ccl 16:49:05 -!- anRch [n=markmill@nmd.sbx09269.readima.wayport.net] has quit [] 16:59:24 H4ns1 [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has joined #ccl 17:07:15 -!- H4ns [n=hans@dialin-145-254-047-187.pools.arcor-ip.net] has quit [Read error: 113 (No route to host)] 19:00:06 H4ns [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has joined #ccl 19:06:05 -!- H4ns1 [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has quit [Read error: 60 (Operation timed out)] 20:34:31 -!- milanj [n=milan@77.46.248.69] has quit ["This computer has gone to sleep"] 21:01:12 -!- H4ns [n=hans@dialin-145-254-062-083.pools.arcor-ip.net] has quit [Read error: 131 (Connection reset by peer)] 21:01:24 H4ns [n=hans@dialin-145-254-065-240.pools.arcor-ip.net] has joined #ccl 21:25:52 mdc_ [n=mdc_mobi@ds9.entity.com] has joined #ccl 21:25:52 -!- mdc_mobile [n=mdc_mobi@ds9.entity.com] has quit [Read error: 104 (Connection reset by peer)] 21:43:53 -!- LiamH [n=none@common-lisp.net] has quit ["Leaving."] 21:49:39 rme [n=rme@pool-70-105-87-66.chi.dsl-w.verizon.net] has joined #ccl 21:51:21 -!- H4ns [n=hans@dialin-145-254-065-240.pools.arcor-ip.net] has quit [Read error: 110 (Connection timed out)]