00:17:47 stassats [~stassats@wikipedia/stassats] has joined #sbcl 00:22:20 -!- tcr1 [~tcr@81-233-246-97-no37.tbcn.telia.com] has quit [Quit: Leaving.] 00:53:26 tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has joined #sbcl 01:15:04 rbarraud [~rbarraud@118-92-11-130.dsl.dyn.ihug.co.nz] has joined #sbcl 01:23:09 -!- tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has quit [Quit: Leaving.] 01:32:46 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 265 seconds] 06:21:54 tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has joined #sbcl 06:36:55 rbarraud_ [~rbarraud@118-92-18-14.dsl.dyn.ihug.co.nz] has joined #sbcl 06:37:08 -!- rbarraud [~rbarraud@118-92-11-130.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 258 seconds] 07:40:47 RUN-PROGRAM-ED fails for me on 1.0.42.32 on linux x86/64 07:48:18 "SBCL_HOME environment variable is set, and conflicts with INSTALL_ROOT." 07:48:25 Is this a new check? What does it mean? 08:03:28 -!- tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has quit [Ping timeout: 245 seconds] 08:06:42 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 08:06:42 -!- ChanServ has set mode +o nikodemus 08:20:03 I have seen intermittent deadlocks with that test on openbsd 08:21:46 Or rather, with (:RUN-PROGRAM :PTY-STREAM) 08:24:32 `micro [~micro@www.bway.net] has joined #sbcl 08:51:04 attila_lendvai [~attila_le@adsl-89-134-26-40.monradsl.monornet.hu] has joined #sbcl 08:59:06 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 240 seconds] 09:00:21 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 09:25:26 -!- attila_lendvai [~attila_le@adsl-89-134-26-40.monradsl.monornet.hu] has quit [Quit: Leaving.] 10:05:56 -!- froydnj [~froydnj@gateway.codesourcery.com] has quit [Ping timeout: 265 seconds] 10:09:08 stassats [~stassats@wikipedia/stassats] has joined #sbcl 10:26:59 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Remote host closed the connection] 10:27:18 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 10:55:38 tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has joined #sbcl 11:32:13 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 12:00:07 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 265 seconds] 12:17:22 So why is an error in a finalizer caught and degraded to a warning? 12:19:04 -!- rbarraud_ [~rbarraud@118-92-18-14.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 258 seconds] 12:19:12 one obvious reason in the current implementation is that otherwise some finalizers would be lost 12:19:19 I'm thinking of the gc-disposes-unclosed-stream issue and thinking about signaling an error in that case; I mean yeah it can happen any time, in any thread, but isn't the same also true for, say, an out-of-memory condition? 12:21:51 that's a serious-condition though, not an error 12:22:02 tcr: OOM gives you a place where something was allocated though 12:22:58 with error for zombie stream, the error will be uncorrelated to anything done to that stream, and without any hint as to where that stream could have been created or even manipulated 12:24:21 pkhuong_: OOM will give you the last thread that tried to cons, right? That might not necessarily be the one consuming most of it; I guess there's some likelihood involved that it is 12:25:23 depending on the root of the memory scarcity (inner loop consing lot, or gradual memory leak?) 12:25:56 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:27:35 Another issue, disable-debugger prints the backtrace which is sometimes what you want, in case of scripts less so, though 12:28:30 A compromise might be to log the backtrace into a file; though I haven't yet had a good idea where to put that file or how to name it in an application-independent way 12:29:26 i don't like when something pollute my fs 12:30:00 many (most?) scripting languages print a backtrace for an unhandled exception, I think it's appropriate for --script to do likewise 12:30:55 related issue, are there thoughts of global condition handlers? I know clisp has them, I'm not sure how they integrated them with threads 12:31:37 jsnell: sbcl's backtraces are long, so I actually have scroll up rather a bit to see the actual error message 12:32:14 in my ideal world, the terminal would be smart, so applications can hide the backtrace by default, and on click it would appear :-) 12:34:36 I know we've talked about global condition handlers in the past, and never reached any agreement on pretty much any part of how it should work 12:35:32 what issues came up? 12:41:36 I don't remember exactly, and not having a lot of luck finding the discussions on sbcl-devel. but it was at least about whether they were even useful, syntax for defining them, order of applying them, how to remove them 12:42:35 whether to block GC globally while defining them? :) 12:42:44 I think every time it was discussed it was in the narrow context of muffling load-time style-warnings, which might have had an effect on how people thought about it :-) 12:46:27 also how to deal with collisions 12:46:49 I mean the order they're executed relatively to each other 12:49:31 right, that's what I meant by order of applying them 12:52:39 I'd say by inheritance and most-specific-first, and provide a CALL-NEXT-HANDLER (which should also be provided in handler-bind, and handler-case because that functionality is sorely lacking 12:56:00 that implies not being able to define multiple global handlers for the same condition 13:10:00 jsnell: since you were the last one to fiddle with hash tables: what are your feelings on keeping a per-generation table of address -> hash value mapping, and use that to for eq hashing? 13:14:07 hmm... wouldn't you need the same amount of gc synchronization magic in the hash tables as right now? 13:15:01 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 260 seconds] 13:15:27 only table to rehash per GC, and I can write a dedicated function for the batch inserts 13:15:30 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 13:16:26 and since the only update to the address -> hash table is by the GC, with everyone locked out, we can look at a version counter before and after. 13:17:58 ah, I think I misunderstood what you meant. so this would be basically be like an out-of-band and lazily computed constant hash code for each object 13:18:04 right. 13:18:59 that sounds neat 13:19:40 I think I discussed the issue with Feeley ~3-4 years ago, and that's what he was doing in gambit. I felt it was awful to introduce an additional uncorrelated access, but now, I'm not so sure: rehashes are pretty expensive, and one update with O(1) consing is enough to rehash an arbitrarily-large table. 13:20:30 yeah, we kind of moved into exactly the opposite direction a few years ago 13:20:47 from the gc doing the rehash for each table to doing it lazily once the table was accessed 13:22:55 but of course that change was driven by other implementation considerations, we never had any real world numbers about how often the greedy rehashing was wasted 13:28:41 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 276 seconds] 13:28:45 so, downside: hashing is a hash lookup; advantage: sxhash doesn't suck as much, GC doesn't cause rehashes (corollary: users can define other hashed data structures based on pointer identity). 13:38:47 it'd be nice for sxhash to not suck on sbcl, even if it's still useless for the general case of wanting to write portable code 13:41:42 Isn't sxhash used internally for hashing? 13:41:58 tcr: only for equal(p?) hashing 13:42:16 bah, it's lazy. I can implement constant-eq-hash, and using it in hash-tables is an easy choice for users to make. 13:42:23 so it sucks because it does not take the address into account? 13:42:46 because it can't. 13:43:47 I'm missing your point of sb-ext:auto-close; I would not have expect obviousness to play any role 13:44:37 if we're going to extend the standard, we should at least make sure our users can easily tell what's standard and what's an extension. 13:46:12 Bah screw newbies :-) 13:47:13 if the mechanism is going to get implemented anyway, it might as well replace the existing eq-hashes. the only good code is deleted code 13:49:53 upward [~upward@modemcable004.209-80-70.mc.videotron.ca] has joined #sbcl 13:50:24 attila_lendvai [~attila_le@adsl-89-134-26-40.monradsl.monornet.hu] has joined #sbcl 13:50:44 If I download and build sbcl from source on darwin, are there any changes I need to make to enable threading support before running sh make.sh ? 13:51:21 upward: yes. see INSTALL, section 2.2 13:55:13 pkhuong_: thanks! 14:15:11 hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has joined #sbcl 14:30:31 -!- `micro [~micro@www.bway.net] has left #sbcl 14:34:50 -!- lnostdal [~quassel@56.84-48-233.nextgentel.com] has quit [Read error: Operation timed out] 14:36:31 lnostdal [~quassel@56.84-48-233.nextgentel.com] has joined #sbcl 14:48:03 -!- upward [~upward@modemcable004.209-80-70.mc.videotron.ca] has left #sbcl 15:02:22 -!- tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has quit [Quit: Leaving.] 15:24:00 -!- Krystof [~csr21@84-51-132-95.christ977.adsl.metronet.co.uk] has quit [Read error: Connection reset by peer] 15:26:13 -!- attila_lendvai [~attila_le@adsl-89-134-26-40.monradsl.monornet.hu] has quit [Quit: Leaving.] 15:53:21 Krystof [~csr21@84-51-132-95.christ977.adsl.metronet.co.uk] has joined #sbcl 15:53:21 -!- ChanServ has set mode +o Krystof 15:56:02 -!- hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has quit [Quit: hargettp] 16:10:24 stassats [~stassats@wikipedia/stassats] has joined #sbcl 17:03:51 rmarynch [~roman@bras-4-ge-62.122.200.230.utm.if.ua] has joined #sbcl 18:43:14 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 18:43:14 -!- ChanServ has set mode +o nikodemus 19:40:37 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 265 seconds] 19:41:19 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 19:57:12 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 252 seconds] 20:12:50 -!- rmarynch [~roman@bras-4-ge-62.122.200.230.utm.if.ua] has quit [Remote host closed the connection] 20:57:42 hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has joined #sbcl 21:04:59 rbarraud_ [~rbarraud@118-92-18-14.dsl.dyn.ihug.co.nz] has joined #sbcl 21:16:47 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 21:18:38 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 264 seconds] 21:19:24 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 22:16:16 attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has joined #sbcl 22:48:32 -!- angavrilov [~angavrilo@217.71.227.181] has quit [Read error: Operation timed out] 22:48:58 angavrilov [~angavrilo@217.71.227.181] has joined #sbcl