December 2010 - Haskell-Cafe

Digests
by Permjacov Evgeniy 03 Dec '10

03 Dec '10

The data integrity checks is well-known problem. A common soluting is use of 'checksums'. Most of them , however, are built in quite obfuscated manner (like md5) that results in ugly and error-prone implementations (see reference implementation for same md5). So, the question is: is there a checksum, that is easy to implement over stream of bytes and may work as good checksum and is good in sence that creation of messages with same checksum that given message has is very hard problem (at least 2^128 tries) ? The reason is that I wish a good checksum to be implemented im my stream-oriented library.

4 9

In what language...?
by Andrew Coppin 03 Dec '10

03 Dec '10

Yesterday I read a rather interesting paper: http://www.cl.cam.ac.uk/~mb566/papers/tacc-hs09.pdf It's fascinating stuff, and I *think* I understand the gist of what it's saying. However, the paper is utterly festooned with formulas that look so absurdly over-the-top that they might almost be a spoof of a mathematical formula rather than the real thing. A tiny fraction of the notation is explained in the paper, but the rest is simply "taken to be obvious". The paper also uses several ordinary English words in a way that suggests that they are supposed to have a more specific technical meaning - but I have no idea what. Does anybody have any idea which particular dialect of pure math this paper is speaking? (And where I can go read about it...)

12 26

HStringTemplate and syb-with-class usage
by Christopher Lewis 03 Dec '10

03 Dec '10

I've seen questions on using HStringTemplate with syb-with-class on the list before, but haven't seen this issue addressed. I'm hoping it's a noob issue, and, therefore, thought I would check with the list before spending too much time digging around for solutions. I'm attempting to use HStringTemplate with the syb-with-class support. I've made sure HStringTemplate is installed with 'syb-with-class=True'. 'genericToSElem' is always failing on the 'True =' guard with an error of "Unable to serialize the primitive type [...]" with whatever primitive Haskell type I've included in my ADT; for example, "Unable to serialize the primitive type 'Prelude.Int'". I've made sure to include 'Text.StringTemplate.GenericWithClass', 'Data.Generics.SYB.WithClass.Basics', and 'Data.Generics.SYB.WithClass.Derive', and to derive instances of the 'Typeable' and 'Data' classes with '$(derive [''Test])', where 'Test' is my test ADT. This makes me think that the overloaded 'toSElem' functions defined by the instances in 'Text.StringTemplate.Instances' are not being called? A cursory comparison to RJson, which the HStringTemplate syb-with-class solution is based on, didn't lead me to any obvious differences that I was missing. For reference, my current configuration is (Sorry, it's an older configuration, but the one I'm stuck with for now) GHC 6.10.4 (Mac OS X 10.6.4, Intel) HStringTemplate 0.6.3 syb-with-class 0.6

1 0

Emacs mode for Happy?
by Edward Amsden 03 Dec '10

03 Dec '10

I'm planning on using Happy for my compiler construction class this quarter. The only problem so far is that emacs keeps trying to mangle my .y files because it assumes I mean them for yacc. Is there an emacs mode for Happy or is it maybe time to also buckle down to learning elisp and write one myself? ------ Edward Amsden Undergraduate Computer Science Rochester Institute of Technology

1 0

Most images broken on haskellwiki pages
by Eugene Kirpichov 03 Dec '10

03 Dec '10

Hello. I looked at a couple pages of mine http://www.haskell.org/haskellwiki/Splot , http://www.haskell.org/haskellwiki/Timeplot , and at a couple of random pages http://haskell.org/haskellwiki/Special:Random - and looks like the vast majority of images are not displaying. What's wrong? Does it have to do with the wiki redesign? What should, say, I personally do about my images regarding this? -- Eugene Kirpichov Senior Software Engineer, Grid Dynamics http://www.griddynamics.com/

4 6

Re: [Haskell-cafe] [Haskell] haskell.org migration complete
by Simon Marlow 03 Dec '10

03 Dec '10

On 02/12/2010 23:48, Claus Reinke wrote: >> The haskell.org server migration is now complete. >> Please let us know if you have any problems. > > Beginning this week, the majority of mails from haskell.org > lists seem to end up in my ISP's spam filter. That would be > Yahoo! - I wonder whether others here have seen a similar > effect when checking their spam filters? > > I don't know whether this is caused by the migration, but > it made me wonder about something I recalled reading > about whitelisting mailing list senders with ISPs, and a > little search found the following: > > http://en.wikipedia.org/wiki/Whitelist > http://help.yahoo.com/l/us/yahoo/mail/postmaster/basics/ > http://help.yahoo.com/l/us/yahoo/mail/postmaster/forms_index.html > http://help.yahoo.com/l/us/yahoo/mail/postmaster/bulkv2.html > http://www.novablog.info/2008/10/21/5-simple-steps-to-improving-e-mail-deli… > > > So it might be related to the new server. Is whitelisting > something the mailing list hosting service should deal > with, or is that left to the customer? I'm fairly sure GMail at least whitelists haskell.org, because I get spam sent via the <list>-owner(a)haskell.org aliases that invariably ends up in my inbox, apparently bypassing GMail's usually very good spam filtering. Cheers, Simon

1 0

Manatee Video.
by Andy Stewart 03 Dec '10

03 Dec '10

Hi all, Many people ask "What's Manatee?" A video worth a thousand words : here is video (select 720p HD) http://www.youtube.com/watch?v=weS6zys3U8k And i think the correct answer to "What's Manatee?" should be : Depend on you how to use it. :) Other information look : http://hackage.haskell.org/package/manatee Enjoy! -- Andy

4 8

How to cabal fetch mtl==2.0.0.0?
by Iain Alexander 02 Dec '10

02 Dec '10

<log> F:\Util\Haskell>cabal fetch mtl==2.0.0.0 Resolving dependencies... cabal: internal error: could not construct a valid install plan. The proposed (invalid) plan contained the following problems: The following packages are involved in a dependency cycle transformers-0.2.2.0, special-functors-1.0, mtl-2.0.0.0 F:\Util\Haskell>cabal -V cabal-install version 0.8.2 using version 1.8.0.2 of the Cabal library </log> Could someone please remind me of the correct incantation to make this work. (I observe that transformers-0.1.4.0 optionally (by cabal flag) depends on special-functors, and special-functors-1.0 depends on mtl. So given that mtl-2 no doubt depends on transformers, I can see the potential for confusion, but there presumably must be a way of breaking the (apparent) loop.) TIA, Iain. -- Iain Alexander ia(a)stryx.demon.co.uk

9 15

Monadic design patterns for the web Kickstarter project
by Greg Meredith 02 Dec '10

02 Dec '10

Dear Haskellians, Last week i went in to record the 2nd and 3rd installments of the C9 monadic design patterns for the web videos. Charles Torre, the series producer, told me there were over 40K distinct views of the first video. i know that many of them came from this community. So, i really want to say thanks! i also want to mention that i just initiated a Kickstarter project<http://www.kickstarter.com/projects/1499410734/monadic-design-patterns-for-…>to raise enough money to complete the book. i've been finding that i'm at a point where the book needs a certain level of focus that doesn't mix well with juggling a ton of clients. i'm trying to raise just enough that i can clear off my desk and focus on completing the book and the source code that supports it. So, if it's of interest please let your friends and cronies know and donate yourself, if you can. For your amusement, here's a little video about it <http://www.youtube.com/watch?v=qOj0tgQDSeQ>. Best wishes, --greg -- L.G. Meredith Managing Partner Biosimilarity LLC 7329 39th Ave SW Seattle, WA 98136 +1 206.650.3740 http://biosimilarity.blogspot.com -- L.G. Meredith Managing Partner Biosimilarity LLC 1219 NW 83rd St Seattle, WA 98117 +1 206.650.3740 http://biosimilarity.blogspot.com

1 0

Fwd: typehash patch for base >= 4 && < 4.4
by Andrew Miller 02 Dec '10

02 Dec '10

Hi all, Typehash 1.4.0.0, the latest version of typehash in hackage, has a cabal constraint base < 4, making it difficult to use with packages that only support newer base versions. I sent a patch to fix this to the maintainer for the typehash package (Lennart Augustsson) on the 16th of November, but haven't heard anything back - it is possible that he doesn't read the e-mail address I sent it to, or is no longer interested in maintaining typehash. Does anyone who would be willing to know Lennart in person to confirm this? Failing this, does anyone want to be maintainer for the typehash package? If I don't hear any better ideas, I could simply upload a version 1.4.0.1 with my split base patch applied, and the maintainer changed to "haskell-cafe(a)haskell.org". Best wishes, Andrew -------- Original Message -------- Subject: typehash patch for base >= 4 && < 4.4 Date: Tue, 16 Nov 2010 17:55:13 +1300 From: Andrew Miller <andrew(a)amxl.com> To: lennart(a)augustsson.net diff -rbud typehash-1.4.0.0/src/Data/TypeHash.hs typehash-1.4.0.1/src/Data/TypeHash.hs --- typehash-1.4.0.0/src/Data/TypeHash.hs 2009-10-06 03:43:44.000000000 +1300 +++ typehash-1.4.0.1/src/Data/TypeHash.hs 2010-11-16 17:43:51.917333177 +1300 @@ -44,7 +44,7 @@ data Type = Name { typeName :: String } -- Abstract type, or recursive reference - | Data { typeName :: String, constrs :: [Constructor] } + | Data { typeName :: String, dconstrs :: [Constructor] } deriving (Eq, Ord, Show, Typeable, Data) type Constructor = (String, [Field]) type Field = (String, Type) -- a unique number is used for missing field names @@ -54,7 +54,7 @@ let tn = show $ fullTypeOf x in case dataTypeRep $ dataTypeOf x of AlgRep cs | tn `notElem` tns -> - Data { typeName = tn, constrs = map (gConstr (tn:tns) x) cs } + Data { typeName = tn, dconstrs = map (gConstr (tn:tns) x) cs } _ -> Name { typeName = tn } -- Use type name for truly abstract types and recursive types. gConstr :: (Data a) => [String] -> a -> Constr -> (String, [Field]) Only in typehash-1.4.0.1/src/Data: TypeHash.hs~ diff -rbud typehash-1.4.0.0/typehash.cabal typehash-1.4.0.1/typehash.cabal --- typehash-1.4.0.0/typehash.cabal 2009-10-06 03:43:44.000000000 +1300 +++ typehash-1.4.0.1/typehash.cabal 2010-11-16 17:40:55.496953710 +1300 @@ -1,5 +1,6 @@ Name: typehash -Version: 1.4.0.0 +Cabal-Version: >= 1.2 +Version: 1.4.0.1 License: BSD3 Author: Lennart Augustsson Maintainer: Lennart Augustsson @@ -10,6 +11,18 @@ Description: Produce a hash for a type that is unique for that type. The hash takes both actual type names and type structure into account. This is useful for checking the type of persisted values. -Hs-Source-Dirs: src -Build-Depends: base < 4, mtl, bytestring, pureMD5, binary -Exposed-modules: Data.TypeHash + +Flag base4 + Description: Use base-4.* + Default: True + +Library + Hs-Source-Dirs: src + Build-Depends: mtl, bytestring, pureMD5, binary + If flag(base4) + Build-Depends: + base >= 4 && < 4.4, + syb < 0.3 + else + Build-Depends: base < 4 + Exposed-modules: Data.TypeHash

2 1