From novikov.leon at gmail.com Tue Dec 1 14:11:42 2009 From: novikov.leon at gmail.com (Leonid Novikov) Date: Tue, 1 Dec 2009 17:11:42 +0300 Subject: [Cl-perec-devel] mapping for set Message-ID: <11527e190912010611u155163fcn9205cb8665d381be@mail.gmail.com> I worked with hu.dwim.perec (live repo from 04.11.2009) and collided with the following behavior: HOME> (defpclass* test1 () ((c1 :type string))) # HOME> (hu.dwim.perec::mapping-of (nth 4 (class-slots (find-class 'test1)))) # HOME> (defpclass* test2 () ((c2 :type (set test1)))) # HOME> (hu.dwim.perec::mapping-of (nth 4 (class-slots (find-class 'test2)))) NIL This is correct? From levente.meszaros at gmail.com Tue Dec 1 17:26:38 2009 From: levente.meszaros at gmail.com (=?ISO-8859-1?Q?Levente_M=E9sz=E1ros?=) Date: Tue, 1 Dec 2009 18:26:38 +0100 Subject: [Cl-perec-devel] mapping for set In-Reply-To: <11527e190912010611u155163fcn9205cb8665d381be@mail.gmail.com> References: <11527e190912010611u155163fcn9205cb8665d381be@mail.gmail.com> Message-ID: On Tue, Dec 1, 2009 at 3:11 PM, Leonid Novikov wrote: > I worked with hu.dwim.perec (live repo from 04.11.2009) and collided > with the following behavior: > > HOME> (defpclass* test1 () ((c1 :type string))) > # > HOME> (hu.dwim.perec::mapping-of (nth 4 (class-slots (find-class 'test1)))) > # > HOME> (defpclass* test2 () ((c2 :type (set test1)))) > # > HOME> (hu.dwim.perec::mapping-of (nth 4 (class-slots (find-class 'test2)))) > NIL > > This is correct? This is not correct, but this is the expected behavior. There are a couple of failed test cases when a slot's type is a set. We actually never use this feature, because it does not provide navigation from the other direction. You supposed to use associations instead, btw do you have any particular reason not to? Cheers, levy -- There's no perfectoin From novikov.leon at gmail.com Wed Dec 2 06:26:13 2009 From: novikov.leon at gmail.com (Leonid Novikov) Date: Wed, 2 Dec 2009 09:26:13 +0300 Subject: [Cl-perec-devel] mapping for set In-Reply-To: References: <11527e190912010611u155163fcn9205cb8665d381be@mail.gmail.com> Message-ID: <11527e190912012226l725be646x1a0c6646af2b8c71@mail.gmail.com> Indeed there is no need to use set instead of associations. I accidentally stumbled upon it when executed (export-model) with the loaded hu.dwim.perec.test package Thank you for your response.