00:23:18 mega1 [~quassel@2001:470:1f05:548:215:58ff:fe7d:773b] has joined #sbcl 00:32:24 -!- hargettp [~hargettp@96.237.127.144] has quit [Quit: Linkinus - http://linkinus.com] 02:03:30 hargettp [~hargettp@96.237.127.144] has joined #sbcl 02:30:16 slyrus [~chatzilla@adsl-75-36-217-249.dsl.pltn13.sbcglobal.net] has joined #sbcl 03:56:36 -!- nyef [~nyef@pool-70-109-148-200.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 04:15:19 -!- hargettp [~hargettp@96.237.127.144] has quit [Quit: Leaving...] 04:36:05 -!- mega1 [~quassel@2001:470:1f05:548:215:58ff:fe7d:773b] has quit [Ping timeout: 260 seconds] 06:29:42 tcr [~tcr@217-162-131-235.dynamic.hispeed.ch] has joined #sbcl 07:19:29 flip214 [~marek@2001:858:107:1:baac:6fff:fe6b:9183] has joined #sbcl 07:19:29 -!- flip214 [~marek@2001:858:107:1:baac:6fff:fe6b:9183] has quit [Changing host] 07:19:29 flip214 [~marek@unaffiliated/flip214] has joined #sbcl 07:35:39 ugh... sf.net was hacked and their CVS servers are down. 07:35:52 are we in a freeze yet? 07:36:27 might this be a good time to consider moving to git somewhere? 09:58:07 homie` [~levgue@xdsl-78-35-139-245.netcologne.de] has joined #sbcl 10:00:27 -!- homie [~levgue@xdsl-78-35-131-23.netcologne.de] has quit [Ping timeout: 246 seconds] 10:04:42 -!- homie` [~levgue@xdsl-78-35-139-245.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 10:10:21 homie [~levgue@xdsl-78-35-139-245.netcologne.de] has joined #sbcl 10:16:40 -!- homie [~levgue@xdsl-78-35-139-245.netcologne.de] has quit [Remote host closed the connection] 10:44:12 stassats [~stassats@wikipedia/stassats] has joined #sbcl 11:08:48 hargettp [~hargettp@96.237.127.144] has joined #sbcl 11:24:21 homie [~levgue@xdsl-78-35-139-245.netcologne.de] has joined #sbcl 12:02:48 -!- hargettp [~hargettp@96.237.127.144] has quit [Quit: Leaving...] 12:04:47 hargettp [~hargettp@96.237.127.144] has joined #sbcl 12:38:33 -!- hargettp [~hargettp@96.237.127.144] has quit [Quit: Leaving...] 12:43:09 Blkt [~user@93-33-140-157.ip44.fastwebnet.it] has joined #sbcl 13:00:50 -!- Blkt [~user@93-33-140-157.ip44.fastwebnet.it] has quit [Ping timeout: 240 seconds] 13:49:31 -!- flip214 [~marek@unaffiliated/flip214] has quit [Remote host closed the connection] 14:06:07 nyef [~nyef@pool-70-109-148-200.cncdnh.east.myfairpoint.net] has joined #sbcl 14:06:24 G'morning all. 14:07:28 Does anyone know if the commentary at the top of SYS:SRC;COMPILER;CONSTRAINT.LISP is still accurate, particularly the pieces about TODO and known problems? 14:43:14 attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has joined #sbcl 15:05:25 hargettp [~hargettp@dhcp-162.mirrorimage.net] has joined #sbcl 15:17:40 nyef: I believe so. 15:25:03 -!- tcr [~tcr@217-162-131-235.dynamic.hispeed.ch] has quit [Quit: Leaving.] 15:26:06 tcr [~tcr@217-162-131-235.dynamic.hispeed.ch] has joined #sbcl 15:27:07 Blkt [~user@93-33-132-79.ip44.fastwebnet.it] has joined #sbcl 15:47:56 Joy. 16:59:20 rpg [~rpg@216.243.156.16.real-time.com] has joined #sbcl 17:23:22 -!- attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has left #sbcl 17:31:16 -!- slyrus [~chatzilla@adsl-75-36-217-249.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 250 seconds] 17:32:05 mega1 [~quassel@2001:470:1f05:548:215:58ff:fe7d:773b] has joined #sbcl 17:45:07 Interesting bug: (lambda (x) (declare (type undefined x)) (aref (the simple-string x) 0)) 17:50:43 Ouch. 17:51:07 mon_key [~user@unaffiliated/monkey/x-267253] has joined #sbcl 17:52:06 hello sbcl! still wondering if anyone can give some insight wrt http://paste.lisp.org/+2JX0 17:52:20 Need to know if i should report a bug. 17:52:29 And, of course, it errors during IR2TRAN, which is nowhere near it fails. 17:53:26 ... Is this the same bug? 17:53:43 nyef: is that for me? 17:54:32 More for pkhuong, since he just gave a very minimal test case with a related error message. 17:55:02 (The difference is that it mentions DATA-VECTOR-REF, not DATA-VECTOR-REF-WITH-OFFSET.) 17:55:39 Mmm. What's a SIMPLE-ISO-LATIN-1-STRING in your example, other than an undefined type? 17:56:04 nyef: I left it out in the first paste. Its included in the first annotation. 17:56:15 Ah. 18:00:15 the function is an adaptation of ironclad/src/util.lisp `ascii-string-to-byte-array' v 0.28 from quicklisp 18:02:35 which AFAICT doesn't appear to have any callers and which has a funny notion of what constitutes an ASCII string... 18:07:14 Cute. The array TN is being assigned a primitive-type of T. 18:07:42 Probably because it's a "hairy" type. 18:08:20 right. 18:08:38 might be good to map primitive types with csubtypep? 18:09:04 That's plausible, yes. 18:10:37 Gah! 18:11:08 The magic mapping occurs in PRIMITIVE-TYPE-AUX in SYS:SRC;COMPILER;GENERIC;PRIMTYPE. 18:11:15 My -eyes-! 18:12:30 reduced even further. 18:12:46 mm.. 18:14:48 Can we get away with replacing this mess with a priority-list of primitive-types and CSUBTYPEP? 18:15:00 ok. So arguably, there is some dumbness in primitive-type-aux. 18:16:00 however, the transform for hairy-data-vector-ref of string in generic/vm-tran.lisp isn't necessarily right. 18:17:10 it's probably better to transform to aref instead of attempting to bypass the aref -> data-vector-ref conversion chain 18:19:09 Be careful messing with this, the last time I dug into this part of the system was when I sorted out that non-unicode build failure. 18:20:32 the only badness I can imagine so far is that more transforms have to trigger, so either slower compilations, or suboptimal codegen in a couple complex cases because of some limit. 18:21:28 Mmm... I guess as long as the KLUDGE in ARRAY-TYPE-UPGRADED-ELEMENT-TYPE still applies... 18:24:05 in any case, converting to aref is more robust 18:24:42 This is "just" optimization, and unrelated to the bug we started with, right? 18:24:50 no. 18:24:54 it's the bug we started with. 18:25:03 Oh? 18:25:09 Ah, right. 18:25:27 Still, there's arguably a bug in PRIMITIVE-TYPE here. 18:26:03 Since the types are all recognizably subtypes of a specialized array type. 18:27:05 We're converting hairy-data-vector-ref to data-vector-ref, seemingly because we assume that the only reason they're hairy is because they're an array type. 18:28:07 but a subtype of (simple-array character 1) can still be hairy enough to need some additional transforms. 18:29:35 ... Didn't it /start/ as an AREF, though? Couldn't this end up in an infinite loop? 18:30:01 no, the type is simpler. 18:31:36 I guess we should check that the type is wide enough to need the etypecase. 18:49:58 pkhuong: Trying to understand the fix. so the checks csubptypep ar for arrays with dimension 1 b/c of the implicit #\null for C? 18:50:15 attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has joined #sbcl 18:50:40 mon_key: no. There's a difference between (simple-array * 1) and (simple-array * (1)). 18:52:41 -!- hargettp [~hargettp@dhcp-162.mirrorimage.net] has quit [Remote host closed the connection] 18:53:08 hargettp [~hargettp@dhcp-162.mirrorimage.net] has joined #sbcl 18:54:31 pkhuong: OK. specifically for `specifier-type' or for CL in general? 18:55:16 pkhuong: Not being pedantic, just tyring to understand. 18:56:56 mon_key: the spec will describe it all in detail. 18:57:36 -!- deepfire [~deepfire@80.92.100.69] has quit [Ping timeout: 265 seconds] 18:57:40 deepfire [~deepfire@80.92.100.69] has joined #sbcl 19:18:11 slyrus [~chatzilla@173-228-44-88.dsl.static.sonic.net] has joined #sbcl 19:18:52 pkhuong: I don't find it in the spec. but I guess these are things of type: 19:18:53 (typep (make-array 0 :element-type 'base-char) '(simple-array base-char 1)) 19:19:59 does that more or less embody the diff? 19:22:36 clhs array 19:22:37 http://www.lispworks.com/reference/HyperSpec/Body/t_array.htm 19:22:38 see dimension-spec 19:39:27 stassats: thanks. So, specifier-type is matching the restricted set of simple-array's of rank 1 specialized to hold . 19:41:04 that's mumbo-jumbo to me, '(simple-array base-char (1)) means a simple-base-string of length 1 19:42:00 antifuchs [~foobar@ec2-184-73-244-158.compute-1.amazonaws.com] has joined #sbcl 19:42:03 we should (: 19:42:04 hi slyrus 19:42:48 Abezethibou [~user@92.44.9.151] has joined #sbcl 19:42:49 stassats: I was concerned as to the difference between (simple-array base-char 1) vs (simple-array base-char (1)) which pkhuong informs (mumbo jumbo aside) are not the same things. 19:42:59 just for the sake of argument, github vs. gitorius? which is better? 19:43:07 dlowe [~dlowe@ita4fw1.itasoftware.com] has joined #sbcl 19:43:20 slyrus: Is the concern uptime, UI or freedom? :) 19:43:29 yes 19:43:29 *nyef* isn't quite a fan of github, and has never looked at gitorius. 19:43:31 mon_key: they are not indeed, the former is a simple-base-string of arbitrary length, and the latter is of length 1 19:43:49 I definitely prefer github, for its social features, and because it has some very contributor- (and user-)friendly features 19:43:56 like rendering the README, and stuff. 19:44:10 common-lisp.net does git as well, right? 19:44:14 I think I'd be happy enough with git.boinkor as the master repository. :-p 19:44:20 Yes, clnet does git. 19:44:29 -!- Abezethibou [~user@92.44.9.151] has quit [Read error: Connection reset by peer] 19:44:30 anytime I see a gitorious-hosted project, I click around a little, and then forget about it, because there's nowhere that tells me what the project is about 19:44:30 That'd work as well. 19:44:31 -!- Blkt [~user@93-33-132-79.ip44.fastwebnet.it] has quit [Quit: Error: do not makunbound t please!] 19:45:08 I'm not sure having git.boinkor as the master would work that well 19:45:15 for one, I can't really make reliability guarantees 19:45:23 stassats: that is where i am confused. Just to be clear (simple-array base-char 1) is can be any length? 19:45:27 what with me being one substantial land mass plus one ocean away from the machine 19:45:27 Fair enough. clnet, then? 19:45:35 well, github.com/sbcl is there already (: 19:46:12 mon_key: n means it has n dimensions, and with (...) it describes each dimension 19:46:22 While I do have a github account, I find their interface to be annoying at best, and completely bloody useless at worst. 19:46:43 hm. 19:46:57 The other question, though, is what happens with version.lisp-expr? 19:47:08 the visualization of branches on github is kinda nice 19:47:11 I suppose that will be the first git-only commit (: 19:47:55 Did anyone actually review the scripts and whatnot I put together for automatically updating it with each push? 19:48:29 I skimmed them, but I don't remember anything about them anymore /-: 19:49:29 I know that every time the topic gets raised, it devolves into an argument about git-specific version-naming schemes, which probably does little for the more conservative committers beyond annoy them. 19:51:41 stassats: So (simple-array * (1)) is an array of rank 1 with valid indices 0,1 whereas (simple-array * 1) is an array of rank 1 with valid indices 0 ... 19:52:33 not quite 19:52:45 but this really isn't in scope of #sbcl 19:56:32 nyef: if you want to be conservative you can wait for sf.net to come up again, I guess (; 19:56:33 nyef: what was the reason git desribe is not good enough? 20:01:29 attila_lendvai: It's more that we currently have a linear series of versions, not some odd web of stuff. Essentially, the point is to be minimally disruptive to any procedure not directly connected with the mechanics of checkin / checkout. 20:02:50 dunno... I'd just set up a call to 'git describe' in the build process and forget cvs... but then I'm a reckless guy by reputation already, so... :) 20:03:39 seriously, yeah. 20:04:03 I'm all for that. screw the patch numbers. :) 20:07:22 Fare [~Fare@ita4fw1.itasoftware.com] has joined #sbcl 20:16:54 How well does 'git describe' work on a source distro (not a full checkout)? 20:19:31 it doesn't, that's why you encode it into a source distro when you export. 20:20:01 you'd put @@GIT_VERSION@@ into a file, have a build rule that substitutes git describe for that. 20:20:13 and also run that inplace before making a source distrib 20:25:10 -!- cmm [~cmm@109.66.206.178] has quit [Ping timeout: 260 seconds] 20:25:48 cmm [~cmm@109.66.206.178] has joined #sbcl 20:32:25 kurtsmith [~chatzilla@ksmith.physics.wisc.edu] has joined #sbcl 20:56:54 -!- hargettp [~hargettp@dhcp-162.mirrorimage.net] has quit [Remote host closed the connection] 21:02:16 or ship a flattened git tree that has only the current object, and the release tag. 21:02:40 I think that's possible, and would be pretty neat. have an "origin" remote pointing to the real thing, let users run diff (: 21:02:42 nah: you don't want the build process for someone downloading a tarball to include "compile git" 21:02:48 that's true 21:03:06 Still, either way you're talking about restructuring some existing procedures. 21:07:01 Sounds like a plan 21:07:09 nyef: in a fairly simple manner 21:19:46 you could look at GIT-VERSION-GEN in git.git. 21:56:13 -!- tcr [~tcr@217-162-131-235.dynamic.hispeed.ch] has quit [Quit: Leaving.] 22:22:17 -!- cmm [~cmm@109.66.206.178] has quit [Ping timeout: 264 seconds] 22:23:11 cmm [~cmm@bzq-109-66-206-178.red.bezeqint.net] has joined #sbcl 22:26:58 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 22:28:29 -!- dlowe [~dlowe@ita4fw1.itasoftware.com] has quit [Quit: *poof*] 22:53:44 homie` [~levgue@xdsl-78-35-132-152.netcologne.de] has joined #sbcl 22:54:41 -!- homie` [~levgue@xdsl-78-35-132-152.netcologne.de] has quit [Remote host closed the connection] 22:56:15 -!- homie [~levgue@xdsl-78-35-139-245.netcologne.de] has quit [Ping timeout: 240 seconds] 22:56:17 homie` [~levgue@xdsl-78-35-132-152.netcologne.de] has joined #sbcl 22:56:19 -!- homie` [~levgue@xdsl-78-35-132-152.netcologne.de] has quit [Remote host closed the connection] 22:59:17 homie [~levgue@xdsl-78-35-132-152.netcologne.de] has joined #sbcl 23:20:05 hargettp [~hargettp@96.237.121.38] has joined #sbcl