00:25:53 alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 00:36:55 -!- alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [] 00:38:11 alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 00:41:53 mdc_ [n=mdc@ds9.entity.com] has joined #ccl 01:07:54 -!- billstclair [n=billstcl@unaffiliated/billstclair] has quit [Read error: 110 (Connection timed out)] 01:12:24 -!- alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [] 01:15:44 billstclair [n=billstcl@dsl-205-231-25-39.taconic.net] has joined #ccl 01:16:55 -!- mdc_ [n=mdc@ds9.entity.com] has quit [] 01:18:17 alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 02:12:05 -!- mdc [n=mdc@ds9.entity.com] has quit [] 02:59:27 sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has joined #ccl 03:53:39 -!- sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has quit [] 03:59:15 -!- alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [] 04:58:38 -!- milanj [n=milan@93.86.115.240] has quit ["This computer has gone to sleep"] 05:09:33 segv_ [n=mb@p4FC1C64D.dip.t-dialin.net] has joined #ccl 05:09:37 -!- segv [n=mb@p4FC1E1A7.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 05:21:33 alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 05:25:36 mdc [n=mdc@ds9.entity.com] has joined #ccl 07:47:00 -!- rme [n=rme@pool-70-106-128-43.chi01.dsl-w.verizon.net] has quit [] 08:55:28 hsaliak_ [n=hsaliak@cm89.sigma229.maxonline.com.sg] has joined #ccl 08:59:10 -!- hsaliak [n=hsaliak@cm89.sigma229.maxonline.com.sg] has quit [Read error: 145 (Connection timed out)] 09:04:10 H4ns [n=Hans@p57BB9A7C.dip0.t-ipconnect.de] has joined #ccl 12:13:23 sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has joined #ccl 12:18:38 H4ns1 [n=Hans@p57BB9B32.dip0.t-ipconnect.de] has joined #ccl 12:18:55 -!- H4ns [n=Hans@p57BB9A7C.dip0.t-ipconnect.de] has quit [Nick collision from services.] 12:19:00 -!- H4ns1 is now known as H4ns 12:56:56 -!- sellout [n=greg@c-24-128-50-176.hsd1.ma.comcast.net] has quit [] 13:19:20 -!- alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [] 13:34:38 sellout [n=greg@guest-fw.dc4.itasoftware.com] has joined #ccl 13:50:39 anRch [n=markmill@ip-64-7-14-173.bos.megapath.net] has joined #ccl 13:53:00 alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 14:08:56 -!- anRch [n=markmill@ip-64-7-14-173.bos.megapath.net] has quit [] 14:48:37 jauaor [n=araujo@gentoo/developer/araujo] has joined #ccl 15:26:10 milanj [n=milan@79.101.204.40] has joined #ccl 15:39:43 anRch [n=markmill@m2a5a36d0.tmodns.net] has joined #ccl 15:40:55 CCL 1.2 signals a full WARNING when one typecase clause shadows another, which causes the SBCL cross-compile to think it's failed. Before I send mail, is this something that's fixed in SVN? 15:41:50 You mean fixing SBCL ? 15:42:11 I have no idea. 15:44:01 Hm? What I mean is, does current CCL SVN still signal a full WARNING when compiling things like (etypecase foo (base-char ...) (character ...)) ? 15:44:40 (as opposed to a STYLE-WARNING) 15:53:10 It's not at all obvious to me that signaling a STYLE-WARNING is somehow preferable; the severity of that situation (shadowing or duplicate types) depends a lot on context and the types in question. This isn't new behavior in CCL, and it hasn't changed post-1.2. 16:06:36 sorry, I had to step away for a minute 16:06:55 Quoth the standard, "The type style-warning includes those conditions that represent situations involving code that is conforming code but that is nevertheless considered to be faulty or substandard." - I'd argue that this situation falls into that category, and thus should be a STYLE-WARNING 16:07:25 It's still conforming code, even if one of the clauses is dead. 16:07:29 I disagree,and the same passage notes that it's subjective. 16:24:13 Well, I'm no language lawyer :-) I think the intent was that conforming code shouldn't cause FAILURE-P to be set in COMPILE-FILE, and these situations, while odd, have clearly well-defined behavior since TYPECASE clauses are tested in order. 16:27:13 -!- milanj [n=milan@79.101.204.40] has quit ["Leaving"] 16:29:52 Sure. If you say (typecase x (integer ....) (fixnum ...)) what happens is well-defined, but it's very likely that something's pretty seriously wrong. If anyone counts on getting a full warning on something so obviously wrong in a given implementation (this one), it's not clear to me that things should change to violate that expectation. 16:33:38 I think that presumes that the users are ignoring STYLE-WARNINGS, which I *hope* they aren't. That case is indeed bogus, but consider something like: 16:33:45 chandler pasted "etypecase WARNING and CLOS" at http://paste.lisp.org/display/69825 16:34:27 depending on what the user's doing, at runtime bar could even not be a subtype of foo-mixin 16:35:31 -!- anRch [n=markmill@m2a5a36d0.tmodns.net] has quit [] 16:38:01 chandler annotated #69825 with "in fact, I think that's bugged" at http://paste.lisp.org/display/69825#1 16:43:04 Yes. Some cases indicate fairly severe problems, others may not. Changing the behavior of TYPECASE to not signal full WARNINGs might break some existing systems the depend on full warnings being signaled, but would allow some code that's apparently making non-portable assumptions about type disjointness to compile without full warnings. Does that pretty much sum it up ? 16:48:18 wrt "non-portable assumptions", given that portable code is required to assume that BASE-CHAR and CHARACTER may not be identical, I think (etypecase foo (base-char ...) (character ...)) is perfectly reasonable, even if the second leg is dead on implementations where BASE-CHAR and CHARACTER are the same type. 16:48:19 chandler: have you considered adding #+CCL :IGNORE-FAILURE-P for those files in build-order.lisp-expr, as is already done for clisp in some places? that would be enough to solve the problem, wouldn't it? 16:49:25 lichtblau: working around it in SBCL isn't that hard - in the case I'm looking at now, any of these legs will wind up signalling a runtime error in the cross compiler anyway. 16:51:44 okay. I was under the impression that trying to get all implementations to stick to SBCL's idea of what WARNING and FAILURE-P mean is a fight against windmills anyway. For example, ASDF doesn't do this check on non-SBCL either. 16:51:55 aside from the arguments about the seriousness of the situation, I just don't think there's any leeway for a purports-to-conform implementation to signal a full WARNING here 16:53:27 ccl-logbot [n=ccl-logb@master.clozure.com] has joined #ccl 16:53:27 16:53:27 -!- names: ccl-logbot jauaor alms sellout H4ns hsaliak_ mdc segv_ billstclair bfulgham gbyers mdc_mobile lichtblau lisppaste5 jajcloz xristos @ChanServ gz chandler 16:54:07 gbyers: OK. At the very least, did you see the annotation to my paste? That behavior is definitely wrong. 16:58:05 I'm more inclined to believe that it;s wrong to treat the second clause as unreachable when a redefinable standard class is involved, and even if it's somehow legal it;s not likely to be desirable. 17:04:21 milanj [n=milan@79.101.204.40] has joined #ccl 17:07:59 ccl-logbot [n=ccl-logb@master.clozure.com] has joined #ccl 17:07:59 17:07:59 -!- names: ccl-logbot milanj jauaor alms sellout H4ns hsaliak_ mdc segv_ billstclair bfulgham gbyers mdc_mobile lichtblau lisppaste5 jajcloz chandler gz xristos @ChanServ 17:45:57 -!- H4ns [n=Hans@p57BB9B32.dip0.t-ipconnect.de] has quit ["Leaving."] 18:37:47 rme [n=rme@pool-70-106-128-43.chi01.dsl-w.verizon.net] has joined #ccl 18:44:23 H4ns [n=hans@p57A0D1D2.dip.t-dialin.net] has joined #ccl 22:23:15 -!- alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has quit [] 22:33:14 alms [n=alms@209-150-48-250.c3-0.bkl-ubr1.sbo-bkl.ma.cable.rcn.com] has joined #ccl 23:01:38 -!- milanj [n=milan@79.101.204.40] has quit ["Leaving"] 23:11:17 -!- sellout [n=greg@guest-fw.dc4.itasoftware.com] has quit [] 23:57:04 -!- segv_ [n=mb@p4FC1C64D.dip.t-dialin.net] has quit [calvino.freenode.net irc.freenode.net] 23:58:16 segv_ [n=mb@p4FC1C64D.dip.t-dialin.net] has joined #ccl