00:19:28 palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has joined #ccl 00:42:34 -!- ayrnieu [n=_ayrnieu@69.171.164.152] has quit [Read error: 110 (Connection timed out)] 01:05:46 lars_t_h [n=lars_t_h@84.238.90.144] has joined #ccl 01:06:19 -!- lars_t_h [n=lars_t_h@84.238.90.144] has left #ccl 01:19:21 -!- leo2007 [n=leo@smaug.linux.pwf.cam.ac.uk] has quit [Remote closed the connection] 02:00:49 -!- palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has quit [] 02:01:07 palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has joined #ccl 02:01:20 -!- palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has quit [Client Quit] 02:07:08 psismondi [n=user@d24-150-10-209.home.cgocable.net] has joined #ccl 02:08:03 -!- psismondi [n=user@d24-150-10-209.home.cgocable.net] has left #ccl 02:16:33 ayrnieu [n=_ayrnieu@69.171.164.41] has joined #ccl 02:27:04 -!- pem [n=pem@159.226.35.246] has quit [Remote closed the connection] 02:27:22 pem [n=pem@159.226.35.246] has joined #ccl 02:28:54 palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has joined #ccl 03:14:14 -!- Adlai is now known as crickets 03:14:21 -!- crickets is now known as Adlai 03:29:34 bfulgham_ [n=brent@adsl-69-234-114-78.dsl.irvnca.pacbell.net] has joined #ccl 04:04:29 -!- mdc_mobile [n=mdc_mobi@ds9.entity.com] has quit [Read error: 104 (Connection reset by peer)] 04:04:42 mdc_mobile [n=mdc_mobi@ds9.entity.com] has joined #ccl 04:28:48 palter_ [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has joined #ccl 04:28:48 -!- palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has quit [Read error: 104 (Connection reset by peer)] 04:28:51 -!- palter_ is now known as palter 05:20:57 -!- ayrnieu [n=_ayrnieu@69.171.164.41] has quit [Read error: 110 (Connection timed out)] 05:25:02 -!- rme [n=rme@pool-70-105-85-54.chi.dsl-w.verizon.net] has quit [] 05:56:45 -!- bfulgham_ [n=brent@adsl-69-234-114-78.dsl.irvnca.pacbell.net] has quit [] 11:09:54 -!- pem [n=pem@159.226.35.246] has quit [Read error: 54 (Connection reset by peer)] 11:10:08 pem [n=pem@159.226.35.246] has joined #ccl 12:04:01 rme [n=rme@pool-70-105-85-54.chi.dsl-w.verizon.net] has joined #ccl 12:47:44 -!- xristos [n=x@research.suspicious.org] has quit [Read error: 110 (Connection timed out)] 12:47:48 Adlai: your compare-and-swap was using an immediate register to hold a pointer, that won't work if a gc happens while it's running. 13:02:24 -!- Adlai [n=adlai@unaffiliated/adlai] has quit [Remote closed the connection] 13:02:37 Adlai [n=adlai@unaffiliated/adlai] has joined #ccl 14:23:38 gz: ah, thank you for clearing that up 14:25:12 but the pointer has to be in that register (rax), because of the semantics of cmpxchg -- is there some form of without-gc macro? 14:27:38 if a GC happens at the wrong time, the cmpxchg will fail. 14:29:28 is this why ccl::%ptr-store-conditional and ccl::%ptr-store-fixnum-conditional have that extra control flow? 14:30:12 I'm not looking at the code at the moment, but if the cmpxchg fails they should load rax again. 14:30:35 (to retry.) 14:31:36 gbyers: I'll look into this later, but thank you for the advice. Thank you too, gz. 14:32:07 the latest version of my code is at http://github.com/adlai/chanl/tree/devel/src/trivial-cas.lisp 14:33:17 milanj [n=milan@77.46.202.106] has joined #ccl 15:00:12 There's a primitive that does a conditional store of a node (a tagged lisp value), and something similar used by the lock-free hash table code. Both are implemented in the CCL kernel and there's magic in the code that suspends and interrupts threads to handle the case where the thread is interrupted while executing those routines. 15:25:55 -!- sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has quit [Read error: 110 (Connection timed out)] 15:33:22 sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has joined #ccl 15:45:02 anRch [n=markmill@98.98.163.38] has joined #ccl 15:57:44 -!- sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has quit [Read error: 110 (Connection timed out)] 16:41:45 -!- anRch [n=markmill@98.98.163.38] has quit [] 17:28:46 anRch [n=markmill@nmd.sbx07279.burlima.wayport.net] has joined #ccl 18:56:35 -!- anRch [n=markmill@nmd.sbx07279.burlima.wayport.net] has quit [] 19:24:14 nyef [n=nyef@pool-71-161-71-17.cncdnh.east.myfairpoint.net] has joined #ccl 20:24:54 -!- rme [n=rme@pool-70-105-85-54.chi.dsl-w.verizon.net] has quit [verne.freenode.net irc.freenode.net] 20:24:54 -!- pem [n=pem@159.226.35.246] has quit [verne.freenode.net irc.freenode.net] 20:24:54 -!- palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has quit [verne.freenode.net irc.freenode.net] 20:24:54 -!- mdc [n=mdc@ds9.entity.com] has quit [verne.freenode.net irc.freenode.net] 20:34:57 mdc [n=mdc@ds9.entity.com] has joined #ccl 20:34:57 palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has joined #ccl 20:34:57 pem [n=pem@159.226.35.246] has joined #ccl 20:34:57 rme [n=rme@pool-70-105-85-54.chi.dsl-w.verizon.net] has joined #ccl 20:34:57 sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has joined #ccl 20:36:06 -!- sellout is now known as Guest42964 20:44:18 -!- Guest42964 is now known as sellout 21:08:55 -!- milanj [n=milan@77.46.202.106] has quit ["Leaving"] 22:00:50 sykopomp [n=root@unaffiliated/sykopomp] has joined #ccl 23:09:21 -!- palter [n=palter@c-24-128-76-188.hsd1.ma.comcast.net] has quit [] 23:15:42 I'm running ccl + slime + emacs on Windows. Any "gotchas" I should watch out for??? Many thanks in advance, oh CCL wizards :-) 23:19:33 Be sure you're using an up-to-date ccl. Other than that, if you run into bugs, please report them by making a ticket at http://trac.clozure.com/ccl/ 23:24:57 Does "up to date" mean SVN? I'm new to SVN (and old fashioned CVS user) and have "Turle SVN" if I have the name right. Is there any configuration I have to do after an SVN update? 23:25:33 Whoops, that's "Turtle" - the simple interactive SVN client. 23:25:50 You might want to review http://trac.clozure.com/openmcl/wiki/WindowsNotes 23:25:59 Thanks, will do 23:26:36 So far seems ccl is more stable and "nicely packaged" than SBCL - Yay! 23:27:19 And kudos to the development team and active community 23:27:42 At any given time, there are typically two branches being maintained: a release branch (currently at version 1.3), and the trunk (which is where development takes place). 23:29:31 And does the trunk have more bug fixes? 23:29:54 Possibly. It may also have more bugs... 23:30:10 So what does SVN pull down? 23:31:07 Quick Q to avoid me spelunking all the docs and libs: Does ccl support Streams and Generators from the appendix to CLTL2 ??? 23:31:37 It depends what URL you use. http://svn.clozure.com/publicsvn/openmcl/release/1.3/windows/ccl is the 1.3 release branch. http://svn.clozure.com/publicsvn/openmcl/trunk/windows/ccl is the trunk. 23:31:46 Summermute: There's a Series library on cliki. 23:33:12 Summermute: Weren't you already warned about the horrible consequences of using up the supply of question marks like that? 23:35:39 Aha and aha - so I should be "safe" pulling down the release branch? Pity the Series library isn't nicely integrated. 23:36:17 Summermute: (asdf-install:install :series) seems nicely-integrated enough for my tastes. 23:36:58 Any efforts to port the GUI to Windows and/or Linux - something like wxWindows or the now nicely portable GTK+ might go along way there. 23:37:58 I'll sheepishly admint I'm an ASDF virgin ;-) My CL experience is roughly 20 years old :-( 23:38:01 saw this recently: http://clgtk2.wordpress.com/2009/10/04/progress-update/ 23:38:01 (Or the now nicely licensed Qt - but pity the C++) 23:38:39 Summermute: since you're running Windows, I suggest that you run the trunk. When the 1.4 release comes out (real soon now), you can switch to that. 23:38:57 With all the controversy regarding Qt over the last decade, i just don't see why not stick to wxWindows or better yet, GTK+. No muss no fuss. 23:39:16 Trunk? Is this not the stable version? 23:39:47 Qt tends to work a lot better on OS X than the alternatives, and also has a lot more integrated functionality (such as a HTML rendering widget built on WebKit). Much of the "controversy" had to do with licensing, and that's essentially a non-issue now that it's LGPL'ed. 23:40:23 Alice ML used GTK+ plus to good effect. Really snappy response time and UI delivered "principle of least surprise." 23:41:04 Sure, I'll admit I don't know much about OS X, althouth I do entertain buying a cheap max box. 23:41:08 I doubt it would've been least surprise on a Mac, or even on Windows. I just installed Pidgin on my Windows system, and it still doesn't quite feel native (as well as plonking a Gtk theme selector thing in my start menu that I wasn't expecting). 23:41:49 Worked very nicely on Windoze. My linux boxen are down and way out of date, so I couldn't "test" those. 23:42:19 I may very well be less critical, though. 23:43:28 In any case, it sure would be nice to have a slick Lisp IDE on Windoze and Linux - any plans, efforts underway? 23:44:57 chandler: Re wxWindows, or at least the ancient version in DrScheme, the response time just sucks, especially in light of today's CPU and GPU prowess. 23:45:54 PLT diverged from wx a long time ago, but my experiences with another wx application (Audacity) have not been all that positive either. 23:46:31 Oh, hey, speaking of ccl on windows, I tried looking in the kernel source for any indication of proper interaction between lisp unwind-protect and non-local exits and win32 SEH but didn't see anything. Am I missing something, or is it just not there? And if it's just not there, what happens if you throw from a window procedure to toplevel? 23:49:28 No unwind-protect - that really sucks, and pretty much eliminates ccl from commercial consideration. DRAT! 23:49:50 What? 23:49:53 Summermute: I think you're leaping to unwarranted conclusions... 23:50:14 Sure, am I missing nyef's missive incorrectly? 23:50:19 Absolutely. 23:50:47 nyef is asking about a technical detail of the interaction between UNWIND-PROTECT and Win32 Structured Exception Handling (SEH). 23:50:48 What I'm asking is if unwind-protect and the various non-local exit forms behave well with respect to non-lisp stack frames. 23:52:12 Well, am I correct tht (block here (block here (unwind-protect (cleanup-stuff)) won't work? 23:52:12 It's an important detail if you like to be able to debug window procedures, though. 23:52:31 Summermute: Your syntax is all wrong, so of course it won't work. 23:52:35 I don't understand the question beyond that. 23:52:38 Whoops, stick a (throw here) in there :-) 23:52:50 I'm not trying to suggest that purely lisp-side unwind-protect is broken. 23:52:51 That's wrong too; you mean (return-from here) presumably. 23:53:06 Sorry, to be frank I'm kind of medicated (legally) 23:53:28 Yes, (return-from here) sorry 23:53:43 Well, then assume that everything is peachy keen with CCL for now, and holler if you run into a bug. 23:54:14 So are you honing in only to the interaction between Windoze structured excepton handling and Lisp condition system? 23:54:18 Yes. 23:54:32 I can probably live with that for now. 23:54:39 gbyers might your question, nyef. I don't know enough about Windows to be sure. 23:54:45 Not the Lisp condition system per se, but the interaction between native unwinds and Lisp-side UNWIND-PROTECT. 23:55:04 chandler: And vice-versa. Lisp unwinds and native UNWIND-PROTECT. 23:55:18 The bugs is even more isolated - good. Again, I can live with that, I think 23:55:29 There might not be a bug. 23:55:32 I actually find that the native unwind-protect thing is the more critical issue, or at least the one I ran into first. 23:55:35 ?? 23:55:58 I think nyef is just looking for implementation source, at least as I parsed his question. 23:56:04 Summermute: I'm not saying there's a bug. I'm saying "I can't see that this is handled, is it?" 23:56:21 chandler: are you saying the condition system should trap these and make them resumable, hence no unwind-protect? 23:56:59 No. Probably best if you ignore this discussion for now. It's neck deep in implementor talk. 23:57:15 Again, my ignorace - what sorts of things trigger Windoze SEH ? 23:57:29 Summermute: Window procedure calls, for starters. 23:57:48 Which means that normal native UI handling does so. 23:57:57 chandler: it's how i learn. I have a real interet in runtime systems and really need to bruch up on modern OS features. 23:58:16 Whaaaaaaaaaat? 23:58:30 But it's only a problem in this specific case if you take an unwind from the window procedure (a callback) to outside that callback. 23:58:49 A simple system call to the GUI returns triggers and SEH? Ridiculous and a PITA. 23:58:52 Learning about this will teach you about Windows. That may or may not have anything to do with "modern OS features". This is Windows-specific stuff only. 23:59:15 i'm Windoze/Linux agnostic :-) 23:59:50 Summermute: No. The window-procedure invocation establishes a native unwind-protect. Windows gets nasty if its native unwind-protect stuff isn't treated with respect.