An observation: on GHC 7.6.3, if I remove c2 entirely, then ghci cooperates.
*Main :t \x - c (c x)
\x - c (c x) :: (C a b, C a1 a) = a1 - b
At first blush, I also expected the definition
-- no signature, no ascriptions
c2 x = c (c x)
to type-check. Perhaps GHC adopted a trade-off giving
I also say +1, but I am concerned about always showing all the bindings.
In my experiences over the years, the times when holes seem they would have
been most helpful is when the bindings were numerous and had large and
complicated types. Dumping all of the bindings in that sort of scenario
would
Is the non-injectivity not an issue here because the type family
application gets immediately simplified?
On Sun, Aug 18, 2013 at 12:45 PM, Joachim Breitner m...@joachim-breitner.de
wrote:
Hi,
now that roles are in HEAD, I could play around a bit with it. They were
introduced to solve the
than is required to read it. :P
On Sun, Aug 18, 2013 at 3:37 PM, Joachim Breitner
m...@joachim-breitner.dewrote:
Hi,
not sure – where would injectivity be needed?
Greetings,
Joachim
Am Sonntag, den 18.08.2013, 15:00 -0500 schrieb Nicolas Frisby:
Is the non-injectivity not an issue here
On Wed, Jul 3, 2013 at 5:33 AM, Joachim Breitner
m...@joachim-breitner.dewrote:
[snip]
strange, why did I miss that?
But I can’t get [the GlobalRdrEnv lookup] to work, even looking up an
element that I took from
the GRE itself returns []:
let e' = head (head (occEnvElts env))
This is an exciting effort! Just a quick reaction to Simon's comments on
CoreM.
On Tue, Jul 2, 2013 at 9:11 AM, Simon Peyton-Jones simo...@microsoft.comwrote:
To your questions:
**
**·**To do these kind of things, CoreM will need more reader
stuff. In particular:
**o
I suspect I've found a situation where, to a first approximation, I'd like
SpecConstr to run more than once.
I'm specializing on GADT constructors that contain coercions, so the
constructor-specialized functions have RHSs with refined types. In my
situation, those refined types enable
When wren's email moved this thread to the top of my inbox, I noticed that
I never sent this draft I wrote. It gives some concrete code along the line
of Wren's suggestion.
-
The included code uses a little of this (singleton types) and a little of
that (implicit configurations).
My GHC 7.6.1 (on a Mac) compiles this code without any warnings or errors.
Do you have some other compilation flags in effect?
On Fri, Nov 9, 2012 at 11:09 AM, Roman Cheplyaka r...@ro-che.info wrote:
For this module
module Test where
import System.Random
data RPS = Rock |
And the important observation is: all of them throw A if interpreted in
ghci or compiled without -O, right?
On Thu, Nov 8, 2012 at 11:24 AM, Albert Y. C. Lai tre...@vex.net wrote:
On 12-11-08 07:12 AM, Simon Hengel wrote:
I was just going to say that I can give at least one counterexample
I share an observation/workaround below.
On Tue, Oct 30, 2012 at 8:49 AM, Andres Löh andres.l...@gmail.com wrote:
This one looks strange to me:
-- Stripping a type from all its arguments
type family Strip (t :: *) :: k
I'd be tempted to read this as saying that
Strip :: forall k. * - k
The question of whether the warnings should come by default or not is
a question of how serious the programmer is when they declare a type
as EPHEMERAL. In Pedro's use cases, I would be very serious about it —
as Ryan said, performance tends to tank otherwise.
### Proposal Extensions
I think a
On Wed, Sep 19, 2012 at 1:51 PM, Richard Eisenberg e...@cis.upenn.edu wrote:
As for recovering kind classes and such once Any is made into a type
family: I'm in favor of finding some way to do explicit kind instantiation,
making the Any trick obsolete. I'm happy to leave it to others to
On Thu, Oct 11, 2012 at 10:20 PM, Nicolas Frisby
nicolas.fri...@gmail.com wrote:
The to my trick key is to use the promotion of this data type.
The key to my trick is to use the promotion of this data type.
Wow — I have no idea what happened
On Fri, Oct 5, 2012 at 9:25 PM, Richard Eisenberg e...@cis.upenn.edu wrote:
For similar reasons, GHC does not bring the constraints on an instance into
the context when an instance matches. So, even if GHC did select the instance
you want, it would not bring ('[] ~ ps) into the context.
Ah,
for
the noise/HTH in the future.
On Mon, Oct 8, 2012 at 1:52 AM, Nicolas Frisby nicolas.fri...@gmail.com wrote:
On Fri, Oct 5, 2012 at 9:25 PM, Richard Eisenberg e...@cis.upenn.edu wrote:
For similar reasons, GHC does not bring the constraints on an instance into
the context when an instance
The issue you had with applications of the [] type seems to be more
insidious than my last email made it out to be. This expression
( $(return $ ConE (mkName [])) ::
$(return $ ConT (mkName []) `AppT` ConT ''Char)
)
fails with [] is applied to too many arguments. I'm thinking that
the []
in the file, but I'm not sure.
Thanks for posting!
Richard
On Oct 5, 2012, at 5:49 PM, Nicolas Frisby wrote:
GHC 7.6 is rejecting some programs that I think ought to be well-typed.
Details here https://gist.github.com/3842579
I find this behavior particularly odd because I can get GHC to deduce
On Sat, Oct 6, 2012 at 12:46 PM, Gábor Lehel illiss...@gmail.com wrote:
What do you use NLong for? I.e. where and how are you taking advantage
of the knowledge that the list is N long?
OK, some context. I'm experimenting with an augmentation of the
generic-deriving generic programming approach.
tl;wr Variables and holes should have disparate syntax, so that code
is easy to locally parse.
Simon, your proposal is very crisp from the GHC implementor's
perspective, but I think it might be harmful from the user's
perspective.
My premise is that free variables — which are normally fatal —
On Thu, Oct 4, 2012 at 4:28 PM, Roman Cheplyaka r...@ro-che.info wrote:
I don't see why it is an issue. You should never encounter holes in the
released code. The only source of holes should be stuff that you just
wrote. With this proposal not only you get an error for the unbound
variable (as
I've investigated this behavior a bit, and I have two things worth mentioning.
(1) One of the ingredients in this behavior is the non-injectivity
of type families. If we make the class parameter f accessible in the
signature of emptyAlt, there are no more type errors. For example, if
you can
On Wed, Aug 22, 2012 at 10:02 AM, Nicolas Frisby
nicolas.fri...@gmail.com wrote:
Maybe just try again in a separate thread? Perhaps under a pseudonym! :)
Whoa, just realized once again that email is tone-deaf. I meant that
'pseudonym' thing cheekily: just to help differentiate the proposal
is kind of irrelevant ...
Regarding impredicativity in GHC, we are still unfortunately on the
whiteboard ...
Hope this helps!
d-
-Original Message-
From: glasgow-haskell-users-boun...@haskell.org [mailto:glasgow-haskell-
users-boun...@haskell.org] On Behalf Of Nicolas Frisby
that the
corresponding instances was totally polymorphic in the argument.
That's bogus reasoning because of ~ (and hence fundeps, as you used).
Thanks again.
On Tue, Apr 17, 2012 at 2:14 PM, Nicolas Frisby
nicolas.fri...@gmail.com wrote:
Thanks! I'll analyze what you've done here.
One thing that jumps out
Great! I'll take a whack at it ;)
On Tue, Apr 17, 2012 at 4:07 PM, Edward Kmett ekm...@gmail.com wrote:
On Mon, Apr 16, 2012 at 6:57 PM, Nicolas Frisby nicolas.fri...@gmail.com
wrote:
I'm simulating skolem variables in order to fake universal
quantification in constraints via unsafeCoerce
is definitely trying to subvert
it; so I vote trustworthy.
I'm adopting Data.Constraints.Forall for my local experimentation.
Thanks for pointing it out.
On Tue, Apr 17, 2012 at 4:15 PM, Nicolas Frisby
nicolas.fri...@gmail.com wrote:
Great! I'll take a whack at it ;)
On Tue, Apr 17, 2012 at 4:07
I'm simulating skolem variables in order to fake universal
quantification in constraints via unsafeCoerce.
http://hpaste.org/67121
I'm not familiar with various categories of types from the run-time's
perspective, but I'd be surprised if there were NOT a way to use this
code to create run-time
Here's another alternative.
newtype Comp f g a = Comp {unComp :: f (g a)}
instance Resolvable e = Resolvable (Maybe `Comp` e) where
resolve db = fmap Comp . resolveMaybe db . unComp
One disadvantage of this approach is that it requires you to pepper
your types with explicit compositions of
I suspect I'm tripping on a gap in the PolyKinds support. I'm trying
to package up a type-level generic view. It uses PolyKinds — and
DataKinds, but I think it's the PolyKinds that matter. If I compile
everything locally in the same build, it works fine. If I isolate the
spine view declarations in
Just a note: as section 6 of [1] notes, one way (possibly the only?)
to satisfy a universally quantified constraint would be a suitably
polymorphic instance — i.e. with a type variable in the head. I think
this would mitigate the need for whole program analysis at least in
some cases, including
Disclaimer: this use case for type-level string ops is still
hypothetical, so these are predictions.
Shooting for the moon, I foresee writing a type-level string
similarity metric. In my experience, that would involve nested
traversals, sliding of sequence windows, etc. In that case, I would
very
I'm interested in type-level strings myself. I'm using an
approximation in order to enrich the instant-generics-style reflection
of data type declarations with a sensitivity to constructor names. For
example, this lets me automatically convert between many the
similarly-named constructors of
(Sorry I'm so late to this dialogue.)
In http://www.haskell.org/pipermail/glasgow-haskell-users/2011-July/020593.html,
SPJ asks
The superclasses are recursive but
a) They constrain only type variables
b) The variables in the superclass context are all
mentioned in the head. In
34 matches
Mail list logo