[09:22:42] *** Quits: wmacmil (~wmacmil@c83-252-138-144.bredband.comhem.se) (Ping timeout: 256 seconds)
[10:54:20] *** Joins: wmacmil (~wmacmil@c83-252-138-144.bredband.comhem.se)
[11:17:36] <wmacmil> how do you understand the `no trees found` response
[11:17:38] <wmacmil> PlType> gr
[11:17:38] <wmacmil> Eval (TypSig (MkName "Foo") (ExtendCls (BaseClas Num) Xx EmptCls) (TPair (TVar Yy) (TVar Zz)))
[11:17:38] <wmacmil> 1 msec
[11:17:38] <wmacmil> PlType> i PlTypeHs.gf
[11:17:38] <wmacmil> linking ... OK
[11:17:39] <wmacmil> Languages: PlTypeHs
[11:17:41] <wmacmil> 8 msec
[11:17:43] <wmacmil> PlType> gr
[11:17:45] <wmacmil> no trees found
[11:17:47] <wmacmil> 0 msec
[11:18:14] <wmacmil> i can generate trees with just the abstract file but then when I load the concrete I get nothing
[11:18:30] <wmacmil> why is this?  can share code if needed!
[11:37:38] <daherb> wmacmil: wrong start category?
[11:38:01] <daherb> inariksit: i just discovered this frankenstein phd thesis cover: http://www.cse.chalmers.se/temporary/nytt-omslag-disp.pdf
[11:38:59] <daherb> wmacmil: having a look at the code would help
[11:39:09] <daherb> and forget my previous question
[11:39:58] <daherb> i would guess some function is missing that is necessary to create a term of the start category
[11:40:10] <daherb> what happens if you linearize the abstract tree?
[11:41:27] <wmacmil> it works
[11:41:27] <wmacmil> PlType> l (MkName "Foo")
[11:41:27] <wmacmil> Foo
[11:41:27] <wmacmil> 1 msec
[11:42:28] <daherb> what about "l Eval (TypSig (MkName "Foo") (ExtendCls (BaseClas Num) Xx EmptCls) (TPair (TVar Yy) (TVar Zz)))"
[11:42:31] <daherb> ?
[11:43:37] <wmacmil> i haven't yet completed the lincats
[11:44:21] <wmacmil> https://gist.github.com/wmacmil/5c2e8e85cf1a8e6684ceaf1f393af4ce
[11:44:24] <daherb> then that's probably the problem
[11:45:15] <wmacmil> so you can't generate trees even though you aren't linearizing them? with incomplete lincats
[11:45:42] <daherb> what is the resulting category of MkName? if it is Name, you can try if "gr -cat=Name" works
[11:46:09] <daherb> i am not sure but that seems like an acceptable guess
[11:46:38] <wmacmil> https://gist.github.com/wmacmil/40b862f4e2ae849171323c788f306cef
[11:47:15] <wmacmil> PlType> gr -cat=Name
[11:47:15] <wmacmil> MkName "Foo"
[11:48:42] <wmacmil> ok i'll complete the lincats
[11:48:48] <wmacmil> and let you know if any problems persis
[11:48:50] <wmacmil> t
[11:48:51] <wmacmil> thanks
[11:53:46] *** Quits: wmacmil (~wmacmil@c83-252-138-144.bredband.comhem.se) (Ping timeout: 272 seconds)
[12:54:48] *** Joins: wmacmil (~wmacmil@212.85.68.186)
[13:04:26] <wmacmil> i'm getting the follow error for a function
[13:04:28] <wmacmil> PlTypeHs.gf:
[13:04:28] <wmacmil>    PlTypeHs.gf:49:
[13:04:28] <wmacmil>      Happened in operation addClass
[13:04:28] <wmacmil>       {s = str;
[13:04:28] <wmacmil>        empt = False} is not in the lincat of Clas; try wrapping it with lin Clas
[13:04:45] <wmacmil> oper
[13:04:46] <wmacmil> 	addClass : Str -> Clas ;
[13:04:46] <wmacmil> 	addClass str = { s = str ; empt = False } ;
[13:04:57] <wmacmil>   Clas =  { s : Str } ** { empt : Bool } ;
[13:05:12] <wmacmil> lincat   Clas =  { s : Str } ** { empt : Bool } ;
[13:05:54] <wmacmil> Will anyone help me decipher what it means/is wrong
[13:08:25] <daherb> wmacmil: just write addClass str = lin Clas { s = str ; empt = False } ;
[13:08:40] <wmacmil> tack!
[13:09:02] <daherb> lincats have hidden so-called lock fields that are missing in the plain record
[13:09:25] <daherb> by putting lin XXX in front of it you add the lock field for category XXX
[13:17:41] <wmacmil> @daherb, how's it going teaching gf via zoom btw
[13:17:57] <wmacmil> or whatever means you're using
[13:26:14] <daherb> wmacmil: unfortunately it is not as much interaction as in a computer lab
[13:26:31] <daherb> but we are doing our best
[15:20:15] *** Quits: wmacmil (~wmacmil@212.85.68.186) (Ping timeout: 260 seconds)
[15:49:17] *** Joins: wmacmil (~wmacmil@c83-252-138-144.bredband.comhem.se)
[16:19:58] <wmacmil> following up from yesterday, regarding list categories
[16:21:49] <wmacmil> is there any way to either customize the ConsC function to take an additional arguement, or to somehow do it explicitly
[16:26:03] <wmacmil> If i Have a function defined as follows, how can i intersperse the TypVar variable into the [Clas] linearlization.
[16:26:04] <wmacmil> fun
[16:26:04] <wmacmil>   ExtendCls : [Clas] -> TypVar -> TypCls -> TypCls ;
[16:26:04] <wmacmil> lincat
[16:26:04] <wmacmil>   Clas =  Str ;
[16:26:05] <wmacmil> lin
[16:26:07] <wmacmil>   BaseClas cl = cl ;
[16:26:09] <wmacmil>   ConsClas cl lcl = cl ++ "," ++ lcl ; -- : C -> ListC -> ListC
[16:26:11] <wmacmil> I want
[16:26:15] <wmacmil> lin
[16:26:17] <wmacmil>   ConsClas cl lcl = \x (cl ++ x ++ "," ++ lcl); -- where x here represents the Typvar
[16:29:51] <wmacmil> because the only way to define a linearization category (Clas, in this case) is via a table, right?  But the table parameters can't be an arbitrary string? and it doesn't seem like a correct approach to solving this problem
[17:03:41] <daherb> wmacmil: i am afraid but i don't think it is possible
[17:04:13] <daherb> i tried something similar when i implemented coordination in latin
[17:04:24] <daherb> what strings would you expect for x?
[17:13:44] <wmacmil> just names
[17:14:48] <wmacmil> the goal is just for proper typeclass constraints in a haskell typing judgement
[17:15:30] <wmacmil> for instance
[17:15:31] <wmacmil> idBb :: (Num b, Ord b, Monad asdf ) => b -> asdf Int ->  b
[17:16:48] <wmacmil> so the Num and Ord would constitute the list, and we'd just have one abstract node for the b (in the typeclass part of the signature) with seperate b nodes for the actual typing part
[17:17:43] <wmacmil> so then if it got linearized to English we could say "where b has typeclasses Num and Ord, and asdf is a  Monad)
[17:18:14] <wmacmil> i'm going to run but if you post an answer i'll check the logs when i get back.  or you can just wait til later