Idris Programming Language Google Group

Matrix multiplication

2013-05-25T03:26:39Z

Hi,
I'm using the latest Idris 5b2f0910a9<[link]>
I'm trying to solve the matrix multiplication problem from this document:
[link]
Here's the code that does work:
module Lecture1
total
repeat : (n : Nat) -> a -> Vect a n

Removing "return" from class Monad

2013-05-23T12:43:10Z

Hello fellow Idris folks,

I've been thinking lately it's somewhat inelegant that all of the
following are true:

1. Every Monad is an Applicative, and Idris' library enfoces this by
having Applicative be a superclass of Monad
2. The "pure" operation on Applicative should always do the same thing

How to output the Java code?

2013-05-07T02:08:40Z

How do I compile a Java file?
I compiled the JAR file as you said, I copied it to
/Users/orsa/.cabal/share/idris -0.9.7/ path. Then I copied the entire java
directory (from this project) to that path. Then I changed the file name
from executable_pom_template.xml to executable_pom.xml.
I get the following:

exposing functions through FFI

2013-05-04T23:20:10Z

Hi all,
I'm trying to use a c library from idris.
This library [1] provides threaded coroutines, useful for non-blocking IO
and lightweight signaling, which is very much like the RTS that GHC uses.
As a consequence to this architecture, the library drives the main (epoll
based) event loop and expects you to plug your own code into that.

Idris and ASMJS?

2013-04-28T04:09:31Z

Hi,
Could Idris make use of ASMJS to make it faster within Javascript execution
environment?
It says "For complex applications (such as the above games) performance is
only around 2x slower than normally-compiled C++"
[link]
Thanks,
Philip

Verified Software Competition - second announcement - http://vscomp.org

2013-04-18T08:02:50Z

(first announcement available at
[link])

Second Announcement
VSTTE Competition 2013
20-22 April 2013

Organizers: Joseph Kiniry, Hannes Mehnert, Dan Zimmerman

This edition of the VSTTE programming contest is an experiment of a

Problem with dsl specification

2013-04-17T15:22:56Z

Hi.

I am trying to implement dsl specification
for the simple imperative language defined in Idrid excercises
(Lecture 2. problems 9 and 10).
[link]

The specification of dsl is the following:

dsl imp
variable = Var
index_first = stop

Extending the code from Lecture 2, programming with Idris.

2013-04-11T17:13:42Z

I am trying to extend the code from lecture 2,
programming with Idris (the code in at the end of the post).
I would like to do the following:
1. Create an environment with one integer variable.
Something like this:
envT: Env G
envT = [Var TyInt 5]
And use it like:
testFac2 : Int
testFac2 = interp envT fact 4

Introductory articles on Idris? (for general programmers)

2013-04-11T04:42:08Z

Hi,
I came across this good article.
[link]
Are there any other articles for people who come from non-Haskell,
non-mathematical background's such as myself where I come from a general
programming background (but I have functional programming knowledge with

Simplifying pattern matching with do notation

2013-04-10T15:06:08Z

I am learning Idris and I implemented the following function:
total
elemInTree : DecEq a => (x : a) -> (t : Tree a) -> Maybe (ElemTree x t)
elemInTree _ Leaf = Nothing
elemInTree x (Node l y r) with (decEq x y)
elemInTree x (Node l x r) | Yes _ = Just hereET
--- I want to simplify the following line ---

Pattern-matching on Types

2013-04-05T21:31:27Z

I'm trying to compare Lists of Types element-wise but I'm getting some
bizare behaviour. Is pattern-matching Types possible? Should it be?

I've narrowed down the problem I'm hitting to a simple test case:

This behaves as expected:

*test> :r
Type checking ./test.idr
*test> eq Int Int
True : Bool

Mac OS users - what arch?

2013-04-09T17:02:07Z

Hi folks,

I've just gotten ahold of a Mac to test my Idris work on. This should
cut down on cross-platform build issues, and I'll get Windows set up
too.

But it seems that some of you run 32-bit, and some 64-bit Haskell.
This has a bit impact on what I need to test for various reasons.
Could Mac users please let me know (in a mail to me, rather than the

Simplifying function

2013-04-09T14:03:49Z

Hi.
How can I write following function in 'pure idris' (without the proof)?
--- code ---
total
plus_nSm : (n : Nat) -> (m : Nat) -> n + S m = S (n + m)
plus_nSm O y = refl
plus_nSm (S x) y = let ih = plus_nSm x y in ?plus_nSm_case
-- I want to eliminate this proof.
Main.plus_nSm_case = proof {
compute;

Idris programming languages exercises

2013-04-08T12:31:25Z

Hi.
I am trying to learn programming language Idris
and I am no familiar with dependent types.
I am doing the exercises posted on:
[link]
Currently, I am stuck at the proglem: Lecture 1. problem 2. a)
where I am supposed to implement matrix transpose function.

building idris libs fails with LANG=C

2013-04-03T10:41:23Z

Hi,

While trying to prepare an RPM package again of Idris
for Fedora I noticed that buildings the libs fails if LANG=C.
(For better or worse rpmbuild defaults to LANG=C when building.)

Anyway it is not such a big deal - I can workaround it easily enough
but I thought I would just report it here since it took me

Unfortunate dependency on unix package

2013-03-28T18:03:32Z

Hi,
today I grabbed current sources of Idris-dev from github just to find that
idris.cabal now lists the "unix" package among its dependencies.
That's very unfortunate because "The package is not supported under Windows
(except under Cygwin)" (quote from Hackage).
This dependency forbids building Idris under Windows. And no, cygwin is not

REPL holds file in Windows if parsing failed

2013-03-28T09:09:22Z

Hi,
there is a problem with Idris REPL which manifests in Windows but not in
Linux (and OS X). If one loads a source file and there is a syntax error,
then after reporting this error REPL keeps the file open and forbids any
editing, every text editor will complain it cannot save the file. Only
after quitting REPL one can save the file again. However if the error is

dependent pairs in 0.9.7

2013-03-28T08:57:02Z

Hi,
Here's some very basic code which compiled fine in versions 0.9.5 and 0.9.6
but fails to type check in 0.9.7:
using (A : Type, B : A->Type, C : Type)
uncurry : ((x:A) -> B x -> C) -> ((x:A ** B x) -> C)
uncurry f = \p => f (getWitness p) (getProof p)
Version 0.9.7 says "Can't convert (x ** B x) with Exists A B". Can anyone

Build of Idris fails because of `libffi'

2013-03-23T10:49:36Z

Hello,

when I try to install the package `libffi' (by `cabal install') now
required by Idris, I always get the error:

Resolving dependencies...
Configuring libffi-0.1...
cabal: The pkg-config package libffi is required but it could not be found.
Failed to install libffi-0.1
cabal: Error: some packages failed to install:

How to compile to C source code?

2013-03-19T10:38:12Z

Hi there,
So I was trying out the tutorial here
[link]
It can compile to Javascript which is great, but if I want to compile to C
I am not getting what I expected which would be C source code.
idris --target C test.idr -o test.c
So - its outputting a binary rather than C source code is that correct?

Request for comments on the elemList exercise

2013-03-17T22:05:37Z

I'm following the videos from the workshop that Edwin Brady is giving.
In the first video ([link] around 1:10:10 mark), Edwin
mentions a more correct version of elemList , one that returns Dec (Elem x
xs).
Below is a solution (it typechecks with the latest version).
Since I'm just beginning to learn dependently typed programming (mostly on

Some questions about totality checking and impossible matches

2013-03-16T19:05:34Z

Hi Idris hackers,
I just tried to do my first steps in Idris and encountered some problems
when trying to write a simple proof:
1. Why do the following functions type-check? I expected them to be
rejected by the totality checker:
total test : Nat -> Nat
total test2 : Nat -> Nat
test2 x with (x,x) | (O,O) = O

Idris integration with Javascript?

2013-03-14T14:06:46Z

Hi,
How does Idris integrate with Javascript, can I pass javascript data
structures into Idris code? can I call javascript from Idris code via a ffi?
Thanks, Philip

FFI extensions for globals and aggregate types

2013-03-13T19:07:21Z

I'd like to extend Idris' FFI to C with support for reading/writing
globals, and for handling values of aggregate type (i.e. structs, arrays,
and unions). The backend implementation for accessing globals is
straightforward enough: just as the output IRs have (L/D/S)Foreign
constructors for FFI calls, additional constructors can be added to

New release, Idris 0.9.7

2013-03-10T21:01:48Z

Dear all,
I have released a new version of Idris onto hackage. Details are here:

[link]

As usual, you can install with 'cabal update; cabal install idris'. There are few user visible changes, but there has been a large change to the implementation of type checking and unification, so please let me know if anything unexpected happens (either in a good or bad way!).

Questions regarding the "Design and Implementation" paper

2013-03-03T13:31:47Z

Hi,
I am reading "Design and Implementation" paper. I have two questions.
1. Is it necessary to have `let' in the core language? It seems to me in
such a rich type system we can introduce it as usual a syntactic sugar of
function application. Of course, it may cause efficiency problems. Is
this the main reason to have it built in? Or are there other reasons I

Idris compiler Haddock

2013-02-26T15:22:30Z

Hi folks,

I just convinced Haddock to build reasonable docs for Idris, and I
translated some of my own notes as well as existing comments into
Haddock documentation. You can see the result at:

[link]

Please let me know if it's useful, or if something should be changed.

Question about applicative functor syntax

2013-02-14T00:13:01Z

Is there any particular reason why Idris uses the <$> symbol for
application in applicative functors, rather than the <*> from Haskell? I
find it slightly confusing because <$> is a synonym for fmap in both
Haskell and the original McBride and Patterson paper. I apologize if this
has been asked before, but I couldn't find anything.

codata, productivity question

2013-02-12T11:09:36Z

I tried to play with codata. The last method in the listing is rejected
with the message:
"Idris03.drop is not productive", without any further explanation.
Can you give a hint what I'm doing wrong? Thank you, Falko
to get some squares, we now can go
take 5 (map (\x => x*x) natStream)

Segmentation Fault

2013-02-09T16:19:41Z

Hello All
I wrote this simple code to compute the factorial.
module Main

product : List Integer -> Integer
product = foldl (*) 1

fac : Integer -> Integer
fac n = product [1..n]

main : IO ()
main = do
arg <- getLine
print . fac . cast $ arg

Mukeshs-MacBook-Pro:Idris mukeshtiwari$ idris test.idr -o test

Bug: function with ?= returns argument instead of result

2013-02-08T18:26:06Z

Hi,
this program:
module Main
rotLeft : Vect a (S n) -> Vect a (S n)
rotLeft v ?= tail v ++ [head v]
rotLeft_lemma_1 = proof { intros; trivial; }
main : IO ()
main = print $ rotLeft [1,2,3]
outputs "[1,2,3]" while expected output is [2,3,1].
When I just evaluate tail v ++ [head v] in REPL it works properly but when

Advice on Idris style; possible stdlib contribution

2013-02-07T14:50:44Z

Hello folks,

I just threw together the start of a decidable equality module for
Idris. Basically, it's like the Eq type class, but it gives evidence
instead of Booleans. It's quite similar to Agda's stdlib, but not
quite as well factored yet (for example, it doesn't have the
abstractions over binary relations and such).

New unification

2013-02-01T10:46:59Z

Hi all,
I have just pushed a patch to github with a major change to unification. Specifically, instead of failing early (and backtracking), we now maintain a list of problems and refine as we go. (I can hear some people saying "I told you you should have done it that way in the first place!" :)).

F#-style type providers through I/O in the type system

2013-01-29T18:44:54Z

Dear Idris list,

I've been kicking around an idea for the last little while that I
think would be worth implementing and getting some milage out of, but
I wanted to run it by the list for feedback first.

F# has a feature called "type providers". These are libraries that
are loaded by the compiler that generate new types at compile time.

Possible bug, term not reducing.

2013-01-27T03:43:37Z

In the following code, I am defining the co-natural numbers, addition on them, and a bisimilarity relation.
My goal is to prove the dreaded fact: 0 + x = x. However, I get the following message
%default total
codata CoNat = cO | cS (CoNat)
cnplus : CoNat -> CoNat -> CoNat
cnplus cO y = y
cnplus (cS x) y = cS (cnplus x y)

StrLen missing?

2013-01-24T11:20:58Z

Hi,
What is the right way of getting the length of a String in Idris?
In the RTS source code I see function idris_strlen which should be invoked
by a primitive %StrLen but I don't know how to call it from Idris code. It
seems to be missing from primitives mentioned in Primitives.hs and
Builtins.idr.

Idris 0.9.6 released

2013-01-23T17:45:32Z

Hi all,
An updated version of Idris, 0.9.6, has been released onto Hackage ([link]). This has lots of updates and improvements over the previous released version. You can find full details here:

[link]

Have fun!

Edwin

Idris speed

2013-01-17T09:02:56Z

Hi,
In the ICFP'10 paper (Scrapping your Inefficient Engine) I saw some
benchmarks mentioned where Idris code performed close to C and Java.
Recently I installed fresh version of Idris (0.9.5.2 from Idris-dev at
github) on Windows (with GHC 7.4.1 and GCC 4.6.1 from MinGW) and tried some
experiments involving simple math with Ints. In my tests Idris programs

Building Idris (Git master) with GHC 7.0, too

2012-12-29T23:35:51Z

Hello,

with the trivial patch appended you can build Idris from Git master
with all GHC 7.x versions. I've tested it with GHC 7.0.4, GHC 7.2.2,
GHC 7.4.2 and GHC 7.6.1 (together with upd-to-date versions of the
additional packages from Hackage needed).

Dirk

How to compile Idris from source?

2012-12-26T19:41:46Z

Hi.
As Idris seems to develop fast, I'd like to use the most up to date version.
Until now I used "cabal install idris" to download the latest release, and
I simply don't know enough of the Haskell infra structure to be sure not to
kill my running version without getting a new one.
Can you give / point me to information what to do after git clone

Partiality monad

2012-12-23T15:34:02Z

Instead of having a special keyword for partial functions, why not use the
partiality monad? Basically, the partiality monad is defined like this:
codata Partial a = Delay (Partial a) | Now a
A partial function from a to b would then be a function a -> Partial b.

Blobs: Proposal, discussion topic, code for preview

2012-12-17T22:27:34Z

TLDR: Blobs as a primitive datatype?

Now for the rambling story-telling mode of discourse that I so enjoy and am
probably already known for!

So I thought to myself, it would be nice if I could start porting some data
structures or something over from Haskell, possibly in semi-automated
fashion. So, I thought, what do we need most? Ah! There's a quantitative

Get Idris lazy?

2012-12-13T16:02:16Z

Hi,

It is really nice to see the turn from Set to Type for the type of types.
So following that, how about change idris to use lazy evaluation? Here
are the reasons:

1. If I remember correctly, Martin-Löf's type theory features lazy
evaluation. I did not quite get and now also forget why he did that.

Type of types

2012-12-11T23:07:50Z

Hi all,
Following last week's discussion, and because it's good to make these changes early, the type of types is now 'Type', rather than 'Set'.

I'm not changing it again :). Or back.

Edwin.

RTS for LLVM vs. C backends

2012-12-10T01:17:54Z

The current LLVM backend attempts to use the same RTS as the C backend to
reduce duplication of effort, but it's looking more and more likely that
its needs significantly diverge from those of the C backend. Specifically,
the LLVM backend attempts to store references (and hopefully, in the
future, values as well) on the machine stack, which must then be scanned

Help with function images / inverses

2012-12-09T17:06:13Z

I've been working my way through dependent types at work [DTAW], which is
written with agda, but seems like a good place to gain some familiarity
with dependent types and gives me some practice with trying to work with
them in idris.
Fair warning: I'm no agda expert, which might be making some of the

Renaming 'Set' to 'Type'

2012-12-05T21:32:48Z

Pros:
- More intuitive to people new to dependent typing
- Makes implementation and use of something like Haskell's Data.Set much
less awkward and confusing
Cons:
- Breaks pretty much all existing code
A straw poll on IRC has most people vaguely supportive of the idea. Any
dissenters?

Local type inference in 'where'

2012-11-28T22:14:32Z

I've just implemented something I've wanted for a while now, spurred on by someone complained to me (in person) recently that it was annoying to have to gives types for everything in a where clause.

You can now omit a type declaration for a function f in a where block, provided that:

* 'f' appears in the right hand side of the top level definition

filter for vector, a question

2012-11-28T20:37:12Z

Hello,
probably I get somewhat demanding, but if anybody has the time to explain
in simple words, why filter1 (a copy from prelude) works, and
filter2 doesn't?
How can p be understood by the type checker in one case, and not in the
other?
Is there a way to make filter2 work?
Thank you, and greetings, Falko

A wish

2012-11-28T19:39:05Z

Hello,
just in case it is easy, I'd be happy about a configuration, an run time
argument or so, to limit the memory usage to a certain value.
I kind of get emotional, when my computer doesn't respond anymore ;-)
Kind regards & happy hacking, Falko