01:49:08 alms_ [~alms_@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 01:49:11 -!- alms_ [~alms_@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [Client Quit] 04:14:09 -!- Adlai [~adlai@unaffiliated/adlai] has quit [Ping timeout: 265 seconds] 06:18:14 leo2007 [~leo@cpc1-cmbg13-0-0-cust596.5-4.cable.virginmedia.com] has joined #ccl 07:42:10 e-user [~akahl@nat/nokia/x-cotnegrfxbkxwbxj] has joined #ccl 08:15:20 e-user: you were looking for me yesterday ? 08:23:23 gbyers: hi gary, yes. three things: congrats and thank you for 1.6; and, still no boolean for that .stabs, no bad though; lastly: we (that means, my reviewer and i) have found a new problem, the x86_64 kernel has an executable heap which easily breaks with selinux and makes ccl unusable for many server systems out there in the world. why is that necessary? 08:24:44 Compiling a lisp function puts that function in memory; functions that can't be executed aren't generally very useful. 08:26:47 hmm, the problem is only reported for ccl x86_64, not for i686 or any other cl implementation with a similar architecture (e.g. sbcl); setting selinux' allow_exemem to false also prevents execution of the kernel reproducibly 08:27:45 Which makes one wonder about whatever's reporting it. 08:34:30 if you have a minute i can rebuild 1.6 on this machine and check 08:34:41 Sure. 08:38:38 first of all, where it was reported: https://bugzilla.redhat.com/show_bug.cgi?id=562226#c17 08:38:48 (comment 17) 08:39:17 if you could comment on this directly in bugzilla, that would be super helpful 08:40:13 I think that selinux is a useless piece of crap; I'd probably just say something inflammatory. 08:41:36 that piece of software, how matter how useless or crappy, is installed on many machines out there and many people rely on it ;) 08:43:01 rpmlint also reports this an a possible error, but gives more information that might be useful here: "One common source for needlessly executable stack cases are object files built from assembler files which don't define a proper .note.GNU-stack section." - well, of course it's an executable heap but maybe the same applies here. 08:43:28 And it gets Fedora/RedHat on GSA schedules in the US, and they make money off of that. Even if they agreed about how craptacular it is, they aren't likely to stop distributing it. 08:43:43 exactly. 08:44:06 This isn't C code. 08:44:09 it has a big, blinking "NSA" sign on it that attracts customers 08:45:25 well i guess the right solution would be to create a custom selinux rule that gets installed with ccl by default on fedora and derived systems (RHEL & co.) 08:45:54 my part again, i guess ;) 08:46:16 for now, i have to take care of $DAYJOB though 08:46:23 see you later! 08:46:51 Sorry. I could say "there are reasons for making some stacks executable, so that closures can be stack-allocated." I'm not sure that that'd settle the issue. 08:49:07 gbyers: any upstream-contributed comment is warmly welcomed 08:50:02 I'll see if I can think of anything non-inflammatory to say. 08:51:54 is it so bad, really? 09:31:59 I added a comment. 09:34:31 I think that it's pretty bad, yes. I'm not sure that it's better or worse than the Windows thing that pops up a dialog whenever you try to do something useful; too close to call. 09:59:09 jdz [~jdz@193.206.22.97] has joined #ccl 10:11:32 jdz_ [~jdz@193.206.22.97] has joined #ccl 10:11:45 -!- jdz_ [~jdz@193.206.22.97] has quit [Read error: Connection reset by peer] 13:00:05 gbyers: thanks for commenting :) this is really helpful for the review process 13:05:15 maybe your statements even reveal flaws in selinux that i can push for fixing thanks to your statements 13:43:48 alms_ [~alms_@146-115-42-237.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 14:07:53 -!- oconnore [~eric@ip68-225-113-244.mc.at.cox.net] has quit [Ping timeout: 255 seconds] 14:13:32 DanLentz [~danlentz@c-68-32-54-29.hsd1.nj.comcast.net] has joined #ccl 14:24:04 -!- DanLentz [~danlentz@c-68-32-54-29.hsd1.nj.comcast.net] has quit [Quit: Colloquy for iPhone - http://colloquy.mobi] 14:30:12 oconnore [~eric@wsip-70-168-242-70.mc.at.cox.net] has joined #ccl 14:35:29 -!- e-user [~akahl@nat/nokia/x-cotnegrfxbkxwbxj] has quit [Ping timeout: 264 seconds] 14:41:41 consolers [heuristics@59.92.48.58] has joined #ccl 14:41:53 anyone familiar with ccl's foreign system? I've been reading the manual for an hour without figuring how to to convert a lisp array of strings into a suitable argv that i can pass to a main(int argc, char **argv) 14:43:38 consolers: Do you just want to run an external program? If so, ccl:run-program will do the job. 14:44:02 so theres no way to do it with ccl primitives? heh 14:44:57 no, run-program wont cut it 14:45:18 cmucl was the last lisp that let me seamlessly treat C's char * with unsigned byte 14:46:12 of course that was before satan incorporated the unicode consortium in menlo park. 14:46:53 -!- oconnore [~eric@wsip-70-168-242-70.mc.at.cox.net] has quit [Ping timeout: 264 seconds] 14:46:55 Right, because all those funny foreigners who use languages that aren't representable in ISO 8859-1 are satanic. 14:46:59 yes, it's so nice to live in the world where all text can be represented using characters from ASCII set 14:47:27 well my problem is i have a function which takes a char**, can i or can i not construct something to pass it 14:47:35 In order to pass an array of strings into something that is suitable for **argv, you'll need to allocate some foreign memory to hold the pointers (unless there's some easier way I'm missing). 14:47:56 right. so the question is do i allocate an ivector of macptrs 14:48:00 that i can pass it 14:48:41 consolers: see ccl:with-enocded-cstrs 14:49:27 e-user [~akahl@nat/nokia/x-zciquxxjumosoamh] has joined #ccl 14:50:05 rme: That solves only half of it, right? 14:50:08 that doesnt cut it 14:50:15 That will take care of copying and encoding the lisp string data into stack-allocated memory. 14:50:21 i can create the cstring with make-cstring etc 14:50:35 i need to create an array of those 14:50:45 where is the char ** 14:51:01 Sorry, I can't type instantaneously. 14:52:18 consolers: `make-heap-ivector' allocates an array on the heap that you can then fill with the pointers to your cstrings. There's a tutorial on doing something similar (passing a pointer to an array of pointers to arrays of ints) at http://ccl.clozure.com/manual/chapter12.11.html#Tutorial--Allocating-Foreign-Data-on-the-Lisp-Heap . 14:52:29 chandler i've been through that 14:52:37 ... and? 14:52:50 i need to do make-heap-ivector 10 'macptr 14:52:53 that doesnt work 14:53:06 Then, use something like (rlet ((argv (:array :address 5))) ...) to allocate your argv. 14:53:08 i need a pointer to an array of char * 14:53:15 not pointer to array of ints 14:54:13 rme: That winds up with everything stack allocated, right? 14:54:24 chandler: that is true. 14:55:37 i havent tried rlet... 14:55:57 is there a long-living version 14:56:04 To set the argv, you'd say (setf (paref argv (:array :address) 0) some-ptr) 14:56:41 If you want to, you can heap-allocate memory with #_malloc 14:56:55 maybe http://trac.clozure.com/ccl/wiki/GradientWindow would be a useful example 14:57:19 I'm on X86-LINUX64, and i'm dealing with a binary only ccl 14:57:52 i'll probably have to go the malloc way 14:59:02 Then, (defparameter *my-argv* (#_malloc (* n target::word-size-in-bytes))) or something. 14:59:21 btw I need to add a css block: code {font-size: 200%;} pre {font-size: 150%;} em.replaceable { font-size: 200%; } , to read the ccl pages with my browser 14:59:40 (for some reason) 15:00:23 -!- e-user [~akahl@nat/nokia/x-zciquxxjumosoamh] has quit [Ping timeout: 265 seconds] 15:00:25 And what browser is that? 15:01:16 i dont have #_malloc do I ? I'll have to define that first 15:01:47 Unless you've got an incomplete installation of ccl, you should have it. 15:02:27 yeah its incomplete. but that should be doable 15:02:40 There's CCL::MALLOC. 15:02:50 can i use (setf %%get-signed-longlong) on the malloced pointer 15:03:17 consolers: yes. 15:04:16 Or (external-call "malloc" :unsigned-long 12 :address) 15:04:36 thanks! 15:07:06 (setf paref) works, i think. 15:13:15 e-user [~akahl@nat/nokia/x-peoqsoesjcnzwidg] has joined #ccl 15:13:20 no, kills the lisp somewhere. 15:14:08 does paref know :address is word-sized ? 15:14:12 I have this: 15:14:15 (defun lisp-string-array-to-c-string-array (a) (let ((r (ccl::malloc (* target::word-size-in-bytes (length a))))) (dotimes (i (length a)) (setf (ccl:paref r (:array :address) i) (ccl::make-cstring (elt a i)))) r)) 15:16:00 oh i was using using basic xulrunner-1.9.2.12 with conkeror. let me check how it looks in opera 15:16:35 -!- e-user [~akahl@nat/nokia/x-peoqsoesjcnzwidg] has quit [Remote host closed the connection] 15:16:38 no problem in opera 15:16:40 e-user [~akahl@nat/nokia/x-hjcvbejvksdaekbs] has joined #ccl 15:18:49 well well with stylesheets, on, in opera, code is still too tiny 15:19:30 the fonts are from a stock opensuse-11.2 distro 15:19:44 consolers: You don't have ccl sources available? 15:19:53 not atm 15:20:43 btw the fonts css was for ccl-documentation.html 15:20:57 which for 1.5 did not link to any stylesheet 15:22:16 consolers: You may find the source for ccl::call-with-string-vector and ccl::with-string-vector helpful. Please see http://trac.clozure.com/ccl/browser/trunk/source/level-1/linux-files.lisp#L985 15:22:51 thx let me take a look 15:23:59 i'm out of time now. but that function i posted doesnt cut it but kills lisp. 15:24:09 -!- consolers [heuristics@59.92.48.58] has quit [Quit: off.] 15:26:50 -!- e-user [~akahl@nat/nokia/x-hjcvbejvksdaekbs] has quit [Ping timeout: 240 seconds] 15:28:16 e-user [~akahl@nat/nokia/x-hurktiqsbrbfkukv] has joined #ccl 15:36:44 -!- e-user [~akahl@nat/nokia/x-hurktiqsbrbfkukv] has quit [Remote host closed the connection] 15:37:00 e-user [~akahl@nat/nokia/x-junkdrydpmfuemdy] has joined #ccl 15:48:37 -!- eslick [~eslick@206.127.32.140] has quit [Quit: This computer has gone to sleep] 15:48:38 -!- eslick_ [~eslick@206.127.32.140] has quit [Quit: This computer has gone to sleep] 16:13:49 -!- e-user [~akahl@nat/nokia/x-junkdrydpmfuemdy] has quit [*.net *.split] 16:13:49 -!- jdz [~jdz@193.206.22.97] has quit [*.net *.split] 16:15:18 oconnore [~eric@wsip-70-168-242-70.mc.at.cox.net] has joined #ccl 16:17:49 e-user [~akahl@nat/nokia/x-junkdrydpmfuemdy] has joined #ccl 16:17:49 jdz [~jdz@193.206.22.97] has joined #ccl 16:33:57 -!- oconnore [~eric@wsip-70-168-242-70.mc.at.cox.net] has quit [Ping timeout: 272 seconds] 16:34:16 anRch [~markmilli@64.134.66.111] has joined #ccl 17:05:48 -!- jdz [~jdz@193.206.22.97] has quit [Quit: Leaving] 17:06:44 -!- anRch [~markmilli@64.134.66.111] has quit [Quit: anRch] 17:10:55 milanj [~milanj_@178-223-146-93.dynamic.isp.telekom.rs] has joined #ccl 17:13:58 what is the branch URL to rebuild 1.6 from source on amd64? 17:14:38 http://svn.clozure.com/publicsvn/openmcl/release/1.6/linuxx86/ccl 17:14:45 thanks! 17:19:18 anRch_ [~markmilli@64.134.66.111] has joined #ccl 17:41:00 -!- e-user [~akahl@nat/nokia/x-junkdrydpmfuemdy] has quit [Quit: Leaving.] 17:49:46 looks like you're using unpinned externals... 17:50:07 the revision of some externals is later than the revision of the url I'm checking out 17:50:40 For the record, we at ITA decided unpinned externals were a bad practice that defeats reproducibility 17:52:39 I think we rather abuse svn externals. 17:55:14 -!- anRch_ [~markmilli@64.134.66.111] has quit [Quit: anRch_] 17:55:15 The externals refer to directories in .../1.6/release/source. 17:55:55 It's possible to check out 1.6/release/source and copy in binaries (and interfaces). It's then much easier to svn up to particular revisions. 17:57:37 We wanted a one-line command for users to check out a complete working ccl, and we have that, at least. 18:07:44 eslick [~eslick@173-13-186-118-sfba.hfc.comcastbusiness.net] has joined #ccl 18:07:44 eslick_ [~eslick@173-13-186-118-sfba.hfc.comcastbusiness.net] has joined #ccl 19:02:10 anRch [~markmilli@64.134.68.90] has joined #ccl 20:09:55 -!- milanj [~milanj_@178-223-146-93.dynamic.isp.telekom.rs] has quit [Ping timeout: 272 seconds] 20:14:06 milanj [~milanj_@77-46-187-235.dynamic.isp.telekom.rs] has joined #ccl 20:17:11 bfulgham [~brent@wsip-72-215-191-226.sb.sd.cox.net] has joined #ccl 20:34:10 -!- eslick [~eslick@173-13-186-118-sfba.hfc.comcastbusiness.net] has quit [Quit: Reverting to analog] 20:34:11 -!- eslick_ [~eslick@173-13-186-118-sfba.hfc.comcastbusiness.net] has quit [Quit: Reverting to analog] 20:56:37 -!- anRch [~markmilli@64.134.68.90] has quit [Quit: anRch] 22:38:48 -!- milanj [~milanj_@77-46-187-235.dynamic.isp.telekom.rs] has quit [Quit: Leaving] 22:44:25 -!- leo2007 [~leo@cpc1-cmbg13-0-0-cust596.5-4.cable.virginmedia.com] has quit [Ping timeout: 245 seconds]