#4370: Bring back monad comprehensions
---------------------------------+------------------------------------------
Reporter: simonpj | Owner: nsch
Type: feature request | Status: new
Priority: normal | Milestone: 7.2.1
Component: Compiler | Version: 6.12.3
Keywords: | Testcase:
Blockedby: | Difficulty:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: None/Unknown
---------------------------------+------------------------------------------
Comment(by nsch):
Replying to [comment:14 simonpj]:
> I'm afraid that the `PostTcTable` in `MDoExpr` is misleading. As you'll
see, it's the ''only'' use of `PostTcTable` and it shouldn't really be
there.
I'll get ridd of it. I didn't like it anyway. :)
> Here you mean "`guard` is looked up '''only''' for monad
comprehensions", I assume?
Yes.
> In other cases, the `guard` field stays as bottom?
Yes.
> Actually, on reflection, consider this.
> {{{
> do notation: do { e ; Q } --> e >> do { Q }
> monad comp: [ e | g; Q ] --> guard g >> [ e | Q ]
> }}}
> Which suggests that you can typecheck the `ExprStmt` of a monad
comprhension
> in the above way, and then attach a `SyntaxExpr` of ( (>>) . guard ) to
the `ExprStmt`.
> Then you'd only need the one field. The `(>>)` and `guard` would be
looked up
> (they are rebindable) but the compose operation `(.)` is the real built-
in one,
> not rebindable. This would be much neater than having two fields, one
of which
> is usually bottom.
I actually tried that, but I cannot typecheck that `((>>) . guard)`
function in the typechecker using `tcSyntaxOp` since it results in some
GHC panic error message on runtime. It's a bit difficult to reproduce
right now since it requires quite a few changes - but if you think the
exact error message would be helpful I can get it.
However, I'm not sure if this is the right way to typecheck this
expression, but using `tcMonoExpr` is obvisouly wrong and I didn't see any
other usefull functions.
> Do you think you could write the documentation first?
I'll work on that. Thanks for the examples.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4370#comment:15>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
