- Haskell-Cafe - Haskell.org

Hackage upload log
by Michael Snoyman 11 Apr '14

11 Apr '14

There used to be a log of Hackage uploads available at [1], but "it's just not here" anymore. Has that content moved to a different URL? Or is there some new way to get that information? I'd like to know the upload dates of every package. I could definitely screenscrape fro the date on the package listing pages, but I was hoping for something a bit more elegant. [1] http://hackage.haskell.org/packages/archive/log

2 4

Re: [Haskell-cafe] package 'permutation' no longer compiles and the maintainer is uncontactable
by Patrick Perry 11 Apr '14

11 Apr '14

It's fine with me if Sophie takes over maintenance of the "permutation" package. On Friday, April 11, 2014 at 11:03 AM, Patrick Perry wrote: > Hi Sophie, > > It's fine with me if you take over maintenance of the package. > > > Patrick > > On Friday, April 11, 2014 at 10:01 AM, Carter Schonwald wrote: > > > Ccing him now > > Get his ok and I'll give you uploaded acls > > > > On Friday, April 11, 2014, Sophie Taylor <sophie(a)traumapony.org (mailto:sophie@traumapony.org)> wrote: > > > Hi, > > > > > > The permutation package no longer compiles due to the change in comparison primops in 7.8.1. Unfortunately, the maintainer Patrick Perry no longer appears to be contactable (the email address bounces) and so cannot submit a patch. > > > > > > I have a patch I wish to upload and according to https://www.haskell.org/haskellwiki/Taking_over_a_package > > > > > > I have to state my intention to take over the maintainence of the package, so yeah, this is me doing that. > > > > > > Thanks, > > > Sophie Taylor (spacekitteh) > > > > > > > > > > > > > > > > > >

1 0

Re: [Haskell-cafe] Seems a wrong implementation of Graham Scan in Haskell Wiki
by Kim-Ee Yeoh 11 Apr '14

11 Apr '14

On Fri, Apr 11, 2014 at 6:16 PM, Yang Leo <jeremyrobturtle(a)gmail.com> wrote: > I don't have the account on Haskell wiki. How can I get an account on it > so I can contribute? Look here: http://www.haskell.org/haskellwiki/HaskellWiki:Contributing -- Kim-Ee

1 0

package 'permutation' no longer compiles and the maintainer is uncontactable
by Sophie Taylor 11 Apr '14

11 Apr '14

Hi, The permutation package no longer compiles due to the change in comparison primops in 7.8.1. Unfortunately, the maintainer Patrick Perry no longer appears to be contactable (the email address bounces) and so cannot submit a patch. I have a patch I wish to upload and according to https://www.haskell.org/haskellwiki/Taking_over_a_package I have to state my intention to take over the maintainence of the package, so yeah, this is me doing that. Thanks, Sophie Taylor (spacekitteh)

3 2

Seems a wrong implementation of Graham Scan in Haskell Wiki
by Yang Leo 11 Apr '14

11 Apr '14

The implementation in the page: http://www.haskell.org/haskellwiki/Graham_Scan_Implementation seems to be wrong. Given a Point list as follows: pts5 = [Point (0,0), Point (0,4), Point (1,4), Point (2,3), Point (6,5), Point (6,0)] The gscan will output a wrong answer which included a inner point Point (1,4). That is because the implemented algorithm cannot handle the situation where multiple consecutive points need to be removed. Here is my implementation: ``` module GrahamScan where import Data.List (sortBy) import Data.Function (on) -- 9. define data `Direction` data Direction = GoLeft | GoRight | GoStraight deriving (Show, Eq) -- 10. determine direct change via point a->b->c -- -- define 2D point data Pt = Pt (Double, Double) deriving (Show, Eq, Ord) isTurned :: Pt -> Pt -> Pt -> Direction isTurned (Pt (ax, ay)) (Pt (bx, by)) (Pt (cx, cy)) = case sign of EQ -> GoStraight LT -> GoRight GT -> GoLeft where sign = compare ((bx - ax) * (cy - ay)) ((cx - ax) * (by - ay)) -- 12. implement Graham scan algorithm for convex -- -- Helper functions -- -- Find the most button left point buttonLeft :: [Pt] -> Pt buttonLeft [] = Pt (1/0, 1/0) buttonLeft [pt] = pt buttonLeft (pt:pts) = minY pt (buttonLeft pts) where minY (Pt (ax, ay)) (Pt (bx, by)) | ay > by = Pt (bx, by) | ay < by = Pt (ax, ay) | ax < bx = Pt (ax, ay) | otherwise = Pt (bx, by) -- -- Main convex :: [Pt] -> [Pt] convex [] = [] convex [pt] = [pt] convex [pt0, pt1] = [pt0, pt1] convex pts = scan [pt0] spts where -- Find the most buttonleft point pt0 pt0 = buttonLeft pts -- Sort other points `ptx` based on angle <pt0->ptx> spts = tail (sortBy (compare `on` compkey pt0) pts) where compkey (Pt (ax, ay)) (Pt (bx, by)) = (atan2 (by - ay) (bx - ax), {-the secondary key make sure collinear points in order-} abs (bx - ax)) -- Scan the points to find out convex -- -- handle the case that all points are collinear scan [p0] (p1:ps) | isTurned pz p0 p1 == GoStraight = [pz, p0] where pz = last ps scan (x:xs) (y:z:rsts) = case isTurned x y z of GoRight -> scan xs (x:z:rsts) GoStraight -> scan (x:xs) (z:rsts) -- I choose to skip the collinear points GoLeft -> scan (y:x:xs) (z:rsts) scan xs [z] = z : xs ``` The source file is at https://github.com/robturtle/Haskell/blob/master/ch03/GrahamScan.hs Maybe not elegant for I just finished the first 3 chapters in Real World Haskell. However it at least correctly computes the example `pts5`

2 1

How to find consumer packages on Hackage?
by Han Joosten 11 Apr '14

11 Apr '14

Hi, I wonder if there is a (simple?) way to find out which packages use some package foo? The reason I would like this is that I am interested in using configFile in a project, and I would like to look at some example usage of it. So in my case I am interested in knowing what packages use configFile. I presume there is a generic way to find out.... Cheers Han Joosten

2 2

Confused about type seen in the wild
by Alexander Solla 10 Apr '14

10 Apr '14

I'm using the yesod-markdown package as a "template" for making my own yesod-mathjax package. I'm pretty much done, and it does compile now, but I'm just really confused by a type that looks like bad syntax. In fact, I ended up trying to "fix" it, which lead to errors: mathJaxField :: Monad m => RenderMessage (HandlerSite m) FormMessage => Field m MathJax Notice that there are two type arrows (=>). Interestingly, mathJaxField :: ( Monad m , RenderMessage (HandlerSite m) FormMessage ) => Field m MathJax compiles too. So it looks like currying/uncurrying, at the type level. But when did this start? I'm using GHC 7.6.3

8 18

ANN.: TreeViz v0.0.6 release.
by David Banas 10 Apr '14

10 Apr '14

I’ve just posted version `v0.0.6’ of the TreeViz package to both Hackage and GitHub. This release fixes the “mixed decimation style failure” bug, in which a DIF stage followed by a DIT stage, followed by anything was giving incorrect numerical results for the FFT. Cheers, -db

1 0

ANN.: TreeViz v0.0.6 release.
by David Banas 10 Apr '14

10 Apr '14

I’ve just posted version `v0.0.6’ of the TreeViz package to both Hackage and GitHub. This release fixes the “mixed decimation style failure” bug, in which a DIF stage followed by a DIT stage, followed by anything was giving incorrect numerical results for the FFT. Cheers, -db

1 0

Stackage with GHC 7.8: build results
by Michael Snoyman 10 Apr '14

10 Apr '14

I'm happy to report that after about six months, enough upper bounds have been corrected on Hackage that a Stackage build with GHC 7.8 can now proceed. I've run a build with 7.8.0rc2, and have uploaded the build.log[1] and the complete cabal output for each package[2]. The good news is that most packages are building correctly, though there is still at least one package deep in the dependency tree which does not build (pcre-light). Here's a list of packages which failed to build, ignoring those which couldn't be built due to broken dependencies: * fpco-api (yes, I'm working on it) * gitlib * groundhog * monad-peel * parsestar * pcre-light * recursion-schemes * statistics * syb-extras * thyme * uniqueid [1] http://download.fpcomplete.com/ghc-7.8-builds/2014-04-09/build.log [2] http://download.fpcomplete.com/ghc-7.8-builds/2014-04-09/logs.zip

3 4