New post, deriving quickcheck

.gitignore

@@ -1,3 +1,5 @@
 _site/
 _cache/
 dist/
+dist-newstyle/
+result*

default.nix

@@ -1,3 +1,8 @@
-{ haskellPackages, haskell }:
+{ pkgs ? import <nixpkgs> {} }:
 
-haskell.lib.disableSharedExecutables (haskellPackages.callCabal2nix "site" ./. {})
+with pkgs;
+
+rec {
+  site = haskellPackages.callPackage ./site {};
+  rauhala-info = callPackage ./rauhala.info { site = site; };
+}

nixpkgs.json

@@ -0,0 +1,7 @@
+{
+  "url": "https://github.com/NixOS/nixpkgs",
+  "rev": "7e1f60dfbba67b975d1a77d710a6f1437fd9709c",
+  "date": "2020-02-01T09:20:37-05:00",
+  "sha256": "0vk55459iljr5dzwnr5661l44b0wdc15952lk2rjcmxr1620yr5v",
+  "fetchSubmodules": false
+}

posts/guides/demobot.md

@@ -1,148 +0,0 @@
----
-title: Functional architecture Pt. 1
-date: 2018-12-25
----
-
-
-I'm lucky enough to work with Haskell professionally which gives me some view
-to good and maintainable real world architecture. In my opinion, one of the
-biggest contributing factors to how your general architecture is defined, is
-determined by the base application monad stack you are using.
-
-Our actual product is mostly in the regular `LoggingT (ReaderT app IO)` base
-monad with whatever style you would imagine with that base monad in place. It's
-not entirely consistent, but close enough.
-
-With all the talk about just having `IO`, `ReaderT app IO`, free monads or
-tagless final monads, I thought of trying different styles. For this post I'm
-focusing on the tagless final since it's most interesting for me right now.
-
-`IO`
-
-:   The most basic style. This is pretty much only suitable for the most basic
-of needs.
-
-`ReaderT app IO`
-
-:   How we mostly define the base monad. This is a really good way of doing
-things, it gives you a lot of leeway on how you can define the rest of your
-application.
-
-`Free monads`
-
-:   Free monads are a way of having a small constrained DSL or monad stack for
-defining your application. By constraining the user, you are also reducing the
-area for bugs. There is also some possibility for introspection, but usually
-this isn't a usable feature. Also since free monad applications need the full
-AST, they're quite a bit slower than the other solutions.
-
-`Tagless final`
-
-:   This is something I'm the least familiar with. If I have understood
-correctly, free monads and tagless final are more or less equivalent solutions
-in their power, but in tagless final you aren't creating the AST anywhere,
-which also means that you aren't paying for it either.
-
-That out of the way, I had a small project idea for a bot that's easy to
-contribute to, difficult to make errors and easy to reason about. The project
-is at most a proof-of-concept and most definitely not production quality.
-Still, I hope it's complex enough to showcase the architecture.
-
-The full source code is available [at my git repository](https://git.rauhala.info/MasseR/demobot).
-
-For the architecture to make sense, let me introduce two different actors: a
-*core contributor* that's familiar with Haskell and a *external contributor*
-that's familiar with programming, not necessarily with Haskell.
-
-The repository is split into two parts, the library and the application.
-
-The library
-
-:   Provides the restricted monad classes (tagless final), extension points and
-the core bot main loop.
-
-The application
-
-:   Provides the implementation for the tagless final type classes, meaning
-that the application defines how the networking stack is handled, how database
-connectivity is done and so on. It also collects all the extensions for that
-specific application.
-
-The *core contributor* is responsible for maintaining the library as well as
-the type class instances for the application type. The *external contributor*
-is responsible for maintaining one or multiple extensions that are restricted
-in their capability and complexity.
-
-I'm restricting the capabilities of the monad in the library and extensions,
-meaning that I'm not allowing any IO. For example the networking is handled by
-a single `MonadNetwork` type class. This is the most complex type class in the
-library right now, using type families for defining a specific extension point
-for the messages. This could be something like 'event type' for Flowdock
-messages or 'source channel' for IRC messages.
-
-~~~haskell
-data Request meta = Request { content :: Text
-                            , meta    :: meta }
-data Response meta = Response { content :: Text
-                              , meta    :: meta }
-
-class Monad m => MonadNetwork m where
-  type Meta m :: *
-  recvMsg :: m (Request (Meta m))
-  putMsg :: Response (Meta m) -> m ()
-~~~
-
-Then we have the extension point which is more or less just a `Request -> m (Maybe Response)`. I'm using rank n types here for qualifying the `Meta`
-extension point and forcing the allowed type classes to be a subset of the
-application monad stack, I don't want extension writers to be able to write
-messages to the bot network by themselves.
-
-~~~haskell
-data Extension meta =
-  Extension { act :: forall m. (meta ~ Meta m, MonadExtension m) => Request meta -> m (Maybe (Response meta))
-            , name   :: String }
-~~~
-
-Last part of the library is the main loop, which is basically a free monad
-(tagless final) waiting for an interpreter. At least in this POC I find this
-style to be really good, it's really simplified, easy to read and hides a lot
-of the complexity, while bringing forth the core algorithm.
-
-~~~haskell
-mainLoop :: forall m. (MonadCatch m, MonadBot m) => [Extension (Meta m)] -> m ()
-mainLoop extensions = forever $ catch go handleFail
-  where
-    handleFail :: SomeException -> m ()
-    handleFail e = logError $ tshow e
-    go :: m ()
-    go = do
-      msg <- recvMsg
-      responses <- catMaybes <$> mapM (`act` msg) extensions
-      mapM_ putMsg responses
-~~~
-
-Then comes the actual application where we write the effectful interpreters. In
-this POC the interpreter is just a `LoggingT IO a` with the semantics of
-stdin/stdout. This is the only file where we're actually interacting with the
-outside world, everything else is just pure code.
-
-~~~haskell
-instance MonadNetwork AppM where
-  type Meta AppM = ()
-  recvMsg = Request <$> liftIO T.getLine <*> pure ()
-  putMsg Response{..} = liftIO . T.putStrLn $ content
-~~~
-
-Writing the extensions was the responsibility of *external contributors* and we
-already saw how the actual extension point was defined above. Using these
-extension points is really simple and here we see how the implementation is
-just a simple `Request -> m (Maybe Response)`.
-
-~~~haskell
-extension :: Extension ()
-extension = Extension{..}
-  where
-    name = "hello world"
-    act Request{..} | "hello" `T.isPrefixOf` content = return $ Just $ Response "Hello to you" ()
-                    | otherwise = return Nothing
-~~~

rauhala.info/.envrc

@@ -0,0 +1 @@
+use nix

rauhala.info/.gitignore

@@ -0,0 +1,2 @@
+_site
+_cache

rauhala.info/contact.markdown

@@ -8,3 +8,4 @@ I live in Espoo Finland. You can contact me on any of the following services.
 - **Slack**: masser@functionalprogramming.slack.com
 - **IRC**: MasseR@freenode
 - **Mastodon**: MasseR@mastodon.social
+- **Keybase**: [https://keybase.io/MasseR](https://keybase.io/MasseR)

rauhala.info/css/default.css

@@ -39,6 +39,18 @@ h2 {
   font-size: 2rem;
 }
 
+.projectlist {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+}
+
+article.blog {
+  /* display: flex; */
+  /* align-items: center; */
+}
+
+
 article .header {
   font-size: 1.4rem;
   font-style: italic;
@@ -51,6 +63,7 @@ article .header {
   text-decoration: none;
 }
 
+
 @media (max-width: 319px) {
   body {
     width: 90%;
@@ -107,14 +120,40 @@ article .header {
     display: inline;
     margin: 0 0.6rem;
   }
+  div.sidebar-container {
+    /* display: grid; */
+    /* grid-template-columns: minmax(60%, 85%) 1fr; */
+  }
+  div.sidebar-container > div.sidebar {
+    padding-top: 2rem;
+    border-top: 0.2rem solid #000;
+  }
 }
 
+
 @media (min-width: 640px) {
   body {
-    width: 60rem;
+    width: 100rem;
     margin: 0 auto;
     padding: 0;
   }
+  div.sidebar-container {
+    display: grid;
+    grid-template-columns: minmax(60%, 85%) 1fr;
+  }
+  div.sidebar-container > div.sidebar {
+    padding-top: 0px;
+    padding-left: 2rem;
+    border-left: 0.2rem solid #000;
+    border-top: none;
+  }
+  article {
+    width: 60rem;
+  }
+  main {
+    display: flex;
+    justify-content: center;
+  }
   header {
     margin: 0 0 3rem;
     padding: 1.2rem 0;

rauhala.info/default.nix

@@ -0,0 +1,24 @@
+{ stdenv, glibcLocales, site }:
+
+stdenv.mkDerivation {
+  pname = "rauhala.info";
+  version = "0.1.0";
+
+  src = builtins.filterSource (path: type: baseNameOf path != "_cache" && baseNameOf path != "_site") ./.;
+
+  phases = ["buildPhase" "installPhase"];
+
+
+  buildPhase = ''
+    export LOCALE_ARCHIVE="${glibcLocales}/lib/locale/locale-archive"
+    export LANG=en_US.UTF-8
+
+    cp -r $src/* .
+    ${site}/bin/site build
+  '';
+
+  installPhase = ''
+    mkdir -p $out/share/
+    cp -r _site/* $out/share/
+  '';
+}

rauhala.info/index.markdown

@@ -15,12 +15,15 @@ My [GPG key](./resources/2104943D6033C.txt)
 
 ### Notable experience and interests
 
-- **Haskell**
-  - Working on an internal Haskell based tool at Relex Oy
+- **Relex Oy**
+  - Working as a Senior Software Developer at Relex Oy
+  - Three distinct products, all of which written in Haskell
+    - Internal integration tool
+    - Relex Deploy 1, a tool for deploying the core product to hosts
+    - Relex Deploy 2, a tool for deploying the new and shiny core product to hosts
+- **Open source**
   - Some contributions to [darcshub](https://hub.darcs.net/)
   - Some contributions to [xmonad](https://github.com/xmonad)
-  - Dozens of personal projects, most of which ended up being just tests for
-    different libraries, techniques or algorithms.
 - **Systems integrations**
   - Working as a consultant from Avoltus Oy to different companies using
     [Mulesoft](https://developer.mulesoft.com/). Systems include webshops,
@@ -28,4 +31,4 @@ My [GPG key](./resources/2104943D6033C.txt)
 - **Java**
   - Bunch of different smaller projects while working at Avoltus Oy.
 - **Other**
-  - I was involved in creating an email solicit platform.
+  - I was involved in creating an email advertising platform.

rauhala.info/posts/2021-01-26-Tests-with-Deriving.md

@@ -0,0 +1,254 @@
+---
+title: Tests with Deriving Via
+tags: haskell, testing, pbt
+---
+
+I have been using both `hedgehog` and `QuickCheck` based property-based testing
+frameworks, I'm fairly comfortable in writing tests and generators in both.
+Theoretical aspects aside, for a user, I feel like `hedgehog` is more
+ergonomic as it does automatic shrinking *and* does away with typeclasses. The
+former is important as writing good shrinkers is hard, remembering to write
+shrinkers is even harder. The latter is important when you need to modify your
+generation for some tests.
+
+In this post, I'll show that using `DerivingVia` extension and generic
+coercions can help you write almost as ergonomic `Arbitrary` definitions for
+`QuickCheck`. The initial idea is taken from the
+[Deriving Via](https://www.kosmikus.org/DerivingVia/deriving-via-paper.pdf) paper,
+but taken a little bit further. This post assumes some level of understanding
+of type level programming.
+
+For the examples, we're using a `Person` as shown in the examples below. The
+test we'll implement will be the `tripping` property. For the *expected*
+values, the `name` is something name-like and `age` is a range between 1-99.
+
+I'll use `hedgehog` to write the ideal case. The generator is light-weight, but
+has been customized for the business case. I'm using the `hedgehog-corpus`
+package for the name-like generation.
+
+``` haskell
+import GHC.Generics (Generic)
+
+import Data.Text (Text)
+
+import qualified Data.Aeson as A
+
+import Hedgehog
+import qualified Hedgehog.Gen as Gen
+import qualified Hedgehog.Range as Range
+import qualified Hedgehog.Corpus as Corpus
+
+data Person
+  = Person { name :: Text
+           , age :: Int
+           }
+  deriving stock (Show, Eq, Generic)
+  deriving anyclass (A.ToJSON, A.FromJSON)
+
+genValidPerson :: Gen Person
+genValidPerson =
+  Person <$> Gen.element Corpus.simpsons
+         <*> Gen.integral (Range.linear 0 99)
+
+prop_encoding :: Property
+prop_encoding = property $ do
+  p <- forAll genValidPerson
+  pure p === A.eitherDecode (A.encode p)
+```
+
+For comparison, this is what I would write with QuickCheck without any helpers.
+There's quite a bit of added complexity, especially in the shrinker, and only
+with two fields.
+
+``` haskell
+import GHC.Generics (Generic)
+
+import Data.Text (Text)
+import qualified Data.Aeson as A
+
+import Test.QuickCheck
+
+data Person
+  = Person { name :: Text
+           , age :: Int
+           }
+  deriving stock (Show, Eq, Generic)
+  deriving anyclass (A.ToJSON, A.FromJSON)
+
+instance Arbitrary Person where
+  arbitrary = Person <$> elements simpsons <*> choose (1,99)
+    where
+      simpsons = ["bart", "marge", "homer", "lisa", "ned"]
+  shrink Person{name,age} =
+    [Person name' age'
+    | name' <- [name]
+    , age' <- shrinkIntegral age
+    , age' >= 1
+    , age' <= 99
+    ]
+
+prop_encoding :: Person -> Property
+prop_encoding p = pure p === A.eitherDecode (A.encode p)
+```
+
+Good, now that the base is done, let's see what we can do about making
+`QuickCheck` more ergonomic. The solution I'm outlining here relies on these
+features.
+
+- `DerivingVia` extension which can automatically generate instances for you if two types are `Coercible`
+- Isomorphism between the `Generic` representation of two types. For example `(a,b)` has a `Generic` representation that is the same as `data Foo = Foo a b`
+- `QuickCheck` modifiers, for example `PrintableString` which modify the arbitrary generation
+
+The paper defines this piece of code for deriving `Arbitrary` instances for
+anything that is generically isomorphic to something that is already an
+instance.
+
+``` haskell
+newtype SameRepAs a b = SameRepAs a
+
+instance
+  ( Generic a
+  , Generic b
+  , Arbitrary b
+  , Coercible (Rep a ()) (Rep b ())
+  )
+  => Arbitrary (a `SameRepAs` b) where
+  arbitrary = SameRepAs . coerceViaRep <$> arbitrary
+    where
+      coerceViaRep :: b -> a
+      coerceViaRep =
+        to . (coerce :: Rep b () -> Rep a ()) . from
+```
+
+For my implementation, I'll be cleaning the code from the paper. I'm swapping
+the type parameters of the newtype and extract the coercion function to
+top-level so that I can define the `shrink` as well.
+
+``` haskell
+newtype Isomorphic a b = Isomorphic b
+
+type GenericCoercible a b =
+  ( Generic a
+  , Generic b
+  , Coercible (Rep a ()) (Rep b ())
+  )
+
+genericCoerce :: forall a b. GenericCoercible a b => a -> b
+genericCoerce =
+  to . (coerce @(Rep a ()) @(Rep b ())) . from
+
+instance
+  ( Arbitrary a
+  , GenericCoercible a b
+  )
+  => Arbitrary (a `Isomorphic` b) where
+  arbitrary = Isomorphic . genericCoerce @a @b <$> arbitrary
+  shrink (Isomorphic b) =
+    Isomorphic . genericCoerce @a @b
+      <$> shrink (genericCoerce @b @a b)
+```
+
+With this, we can now write `Arbitrary` instances using the tuple
+representation as an intermediary. At least as long as the child types have
+their instances properly set.
+
+``` haskell
+data Person
+  = Person { name :: Text
+           , age :: Int
+           }
+  deriving stock (Show, Eq, Generic)
+  deriving anyclass (A.ToJSON, A.FromJSON)
+  deriving (Arbitrary) via ((Text, Int) `Isomorphic` Person)
+```
+
+This is already a marked improvement to the original `Arbitrary` instance we
+wrote, but this does not yet satisfy our original requirement of generating
+only 'valid' persons. I would like to modify the instance generation on a more
+ad-hoc fashion. For this to happen, I would need some *modifiers* that control
+the arbitrary generation.  I would like to write something like the instance
+below.
+
+``` haskell
+type Simpsons = '["marge", "bart", "homer", "lisa", "ned"]
+data Person
+  = Person { name :: Text
+           , age :: Int
+           }
+  deriving stock (Show, Eq, Generic)
+  deriving anyclass (A.ToJSON, A.FromJSON)
+  deriving (Arbitrary)
+    via ((Corpus Simpsons Text, Range 1 99 Int) `Isomorphic` Person)
+```
+
+
+Let's start by defining the `Range` as it's more straightforward. This is just
+a `newtype` with a couple of phantom type variables, which is used in choosing
+the range of the generator. Shrinking is already quite complex (and probably
+not optimal!), I wouldn't want to write this multiple times.
+
+``` haskell
+newtype Range (from :: Nat) (to :: Nat) a = Range a
+
+instance
+  ( KnownNat from
+  , KnownNat to
+  , Num a
+  , Ord a
+  , Integral a
+  ) => Arbitrary (Range from to a) where
+  arbitrary = Range . fromInteger <$> choose (natVal $ Proxy @from, natVal $ Proxy @to)
+  shrink (Range x) = Range <$> shrunk
+    where
+      shrunk =
+        [ x'
+        | x' <- shrinkIntegral x
+        , x >= fromInteger (natVal $ Proxy @from)
+        , x <= fromInteger (natVal $ Proxy @to)
+        ]
+```
+
+Then the corpus. Just like the `Range` it's a `newtype` with a phantom
+variable, providing the input for the random generation. There's an extra
+typeclass involved to act as a typelevel function.
+
+``` haskell
+newtype Corpus (corpus :: [Symbol]) a = Corpus a
+
+class FromCorpus (corpus :: [Symbol]) where
+  fromCorpus :: [String]
+
+instance FromCorpus '[] where
+  fromCorpus = []
+
+instance (KnownSymbol x, FromCorpus xs) => FromCorpus (x ': xs) where
+  fromCorpus = symbolVal (Proxy @x) : fromCorpus @xs
+
+instance (FromCorpus corpus, IsString x) => Arbitrary (Corpus corpus x) where
+  arbitrary = Corpus . fromString <$> elements (fromCorpus @corpus)
+```
+
+With these instances out of the way, we can redo our original test with
+automatic instances.
+
+``` haskell
+import GHC.Generics (Generic)
+
+import Data.Text (Text)
+import qualified Data.Aeson as A
+
+import Test.QuickCheck
+
+import Isomorphic
+
+data Person
+  = Person { name :: Text
+           , age :: Int
+           }
+  deriving stock (Show, Eq, Generic)
+  deriving anyclass (A.ToJSON, A.FromJSON)
+  deriving Arbitrary via ((Corpus Simpsons Text, Range 1 99 Int) `Isomorphic` Person)
+
+prop_encoding :: Person -> Property
+prop_encoding p = pure p === A.eitherDecode (A.encode p)
+```

rauhala.info/projects/2021-01-21-bidirectional.md

@@ -0,0 +1,37 @@
+---
+title: bidirectional
+github: https://github.com/MasseR/bidirectional
+issues: https://github.com/MasseR/bidirectional/issues
+badge: https://github.com/MasseR/bidirectional/workflows/Test/badge.svg
+---
+
+Bidirectional serialization based on Lysxia's post on [Monadic profunctors for bidirectional programming](https://blog.poisson.chat/posts/2017-01-01-monadic-profunctors.html).
+
+
+Let's assume we have a parser like the following
+
+``` haskell
+int :: Parser (ReaderT String Maybe) (Writer [Int]) Int Int
+int = parser (ReaderT readMaybe) (\x -> x <$ tell [show x])
+```
+
+Then you can use the parser for parsing:
+
+```
+> runReaderT (decode int) "3"
+Just 3
+```
+
+Or for encoding:
+
+```
+> execWriter (encode int 3)
+["3"]
+```
+
+Or combine both of them
+
+```
+> runReaderT (decode int) $ head $ execWriter $ encode int 3
+Just 3
+```

rauhala.info/projects/2021-01-21-phrase.md

@@ -0,0 +1,23 @@
+---
+title: phrase
+github: https://github.com/MasseR/phrase
+issues: https://github.com/MasseR/phrase/issues
+badge: https://github.com/MasseR/phrase/workflows/Run nix build/badge.svg
+---
+
+This project is a command line tool for generating password phrases using the
+[diceware method](https://diceware.dmuth.org/). The passwords are stored in a
+folder structure that is compatible with the
+[pass](https://www.passwordstore.org/) password store manager.
+
+
+
+```
+$ phrase asd
+inlet area crux
+
+$ pass show asd
+inlet area crux
+
+$ file ~/.password-store/asd.gpg
+```

rauhala.info/projects/2021-01-21-zettelkast.md

@@ -0,0 +1,17 @@
+---
+title: zettelkast
+github: https://github.com/MasseR/zettelkast
+issues: https://github.com/MasseR/zettelkast/issues
+badge: https://github.com/MasseR/zettelkast/workflows/Test/badge.svg
+---
+
+Command-line tool for managing zettelkast documents. The tool primarily focuses
+on providing unique ids and showing a graph of document connections. It tries
+to be as unintrusive as possible.
+
+```
+$ zettelkast help
+$ zettelkast new
+$ zettelkast list
+$ zettelkast open <id>
+```

rauhala.info/shell.nix

@@ -0,0 +1,10 @@
+with (import <nixpkgs> {});
+
+let site = haskellPackages.callPackage ../site {};
+
+in
+
+mkShell {
+  buildInputs = [ site ];
+}
+

rauhala.info/templates/default.html

@@ -10,20 +10,25 @@
     </head>
     <body>
         <header>
-            <div class="logo">
-                <a href="/">rauhala.info</a>
-            </div>
+            <!-- <div class="logo"> -->
+            <!--     <a href="/">rauhala.info</a> -->
+            <!-- </div> -->
             <nav>
                 <!-- Git logo from https://git-scm.com/downloads/logos -->
                 <!-- Logo by Jason Long -->
+                <a href="/">Home</a>
+                <a href="/posts.html">Posts</a>
+                <a href="/projects.html">Projects</a>
                 <a href="https://git.rauhala.info"><img src="/images/git_16.png" alt="git" /></a>
                 <a href="/contact.html">Contact</a>
             </nav>
         </header>
 
         <main role="main">
+          <div>
             <h1>$title$</h1>
             $body$
+          </div>
         </main>
 
         <footer>

rauhala.info/templates/post-list.html

@@ -0,0 +1,7 @@
+<ul>
+    $for(items)$
+        <li>
+            <a href="$url$">$title$</a> - $date$
+        </li>
+    $endfor$
+</ul>

rauhala.info/templates/post.html

@@ -0,0 +1,8 @@
+<article>
+  <section class="header">
+    Posted on $date$
+  </section>
+  <section>
+    $body$
+  </section>
+</article>

rauhala.info/templates/project-list.html

@@ -0,0 +1,11 @@
+<ul>
+    $for(items)$
+        <li>
+          <div class="projectlist">
+            <a href="$url$">$title$</a>
+            <a href="$github$">$if(badge)$<img src="$badge$" />$else$ Github $endif$</a>
+          </div>
+        </li>
+    $endfor$
+</ul>
+

rauhala.info/templates/project.html

@@ -0,0 +1,12 @@
+<div class="sidebar-container">
+  <article>
+    <section>
+      $body$
+    </section>
+  </article>
+  <div class="sidebar">
+    <a href="$github$">Github</a>
+    <a href="$issues$">Issues</a>
+    $if(badge)$<img src="$badge$" />$endif$
+  </div>
+</div>

rauhala.info/well-known/keybase.txt

@@ -0,0 +1,111 @@
+==================================================================
+https://keybase.io/masser
+--------------------------------------------------------------------
+
+I hereby claim:
+
+  * I am an admin of https://masser.keybase.pub
+  * I am masser (https://keybase.io/masser) on keybase.
+  * I have a public key ASD7WFMicMJhoArgnvPuR7Dc92WStIUWWXcyngnsTE_IaQo
+
+To do so, I am signing this object:
+
+{
+  "body": {
+    "key": {
+      "eldest_kid": "01202eb5d3b8d2fc63e4bc9c6edefb3a38f1043dfdc44a7268393951fe0ea5214cf00a",
+      "host": "keybase.io",
+      "kid": "0120fb58532270c261a00ae09ef3ee47b0dcf76592b485165977329e09ec4c4fc8690a",
+      "uid": "b01557931e13482e9e646593a64e9119",
+      "username": "masser"
+    },
+    "merkle_root": {
+      "ctime": 1580587604,
+      "hash": "890506579950548b2aa6fd0f35850b9fdb1d064786714f18515d0e3089cdc2458c5fe847d40ca71c2e021ca11677de055ac97804c53d1bef89432f88b57599d4",
+      "hash_meta": "104419595abc52918530f85495417faf35c7a5a608d6730bf53152dfa7288970",
+      "seqno": 14451686
+    },
+    "service": {
+      "entropy": "QIvRp0qNOIZFplyzXERiwp75",
+      "hostname": "masser.keybase.pub",
+      "protocol": "https:"
+    },
+    "type": "web_service_binding",
+    "version": 2
+  },
+  "client": {
+    "name": "keybase.io go client",
+    "version": "4.3.1"
+  },
+  "ctime": 1580587641,
+  "expire_in": 504576000,
+  "prev": "c99f66ba0bfa7ff6be736e24691c8007b0db41b6036b7f1e5f9335ab7f3550f4",
+  "seqno": 34,
+  "tag": "signature"
+}
+
+which yields the signature:
+
+hKRib2R5hqhkZXRhY2hlZMOpaGFzaF90eXBlCqNrZXnEIwEg+1hTInDCYaAK4J7z7kew3PdlkrSFFll3Mp4J7ExPyGkKp3BheWxvYWTESpcCIsQgyZ9mugv6f/a+c24kaRyAB7DbQbYDa38eX5M1q381UPTEIH8FtgpzXt/JJxrhAYknW7xL4+K57JBJnbhFpXWAfEH5AgHCo3NpZ8RApEnmCcZApdOTf+XAHRbbgoysAZHfyTHamWC/BEHKuqEf/BlUFSxo6ASa/YC+Y2sMJfW2iRFlP2bs822TyJTBA6hzaWdfdHlwZSCkaGFzaIKkdHlwZQildmFsdWXEINm3yJdAbmbGJd03EnpuKVh2JzkI8WkngbWbAxbbR0MMo3RhZ80CAqd2ZXJzaW9uAQ==
+
+And finally, I am proving ownership of this host by posting or
+appending to this document.
+
+View my publicly-auditable identity here: https://keybase.io/masser
+
+===================================================================
+https://keybase.io/masser
+--------------------------------------------------------------------
+
+I hereby claim:
+
+  * I am an admin of https://rauhala.info
+  * I am masser (https://keybase.io/masser) on keybase.
+  * I have a public key ASD7WFMicMJhoArgnvPuR7Dc92WStIUWWXcyngnsTE_IaQo
+
+To do so, I am signing this object:
+
+{
+  "body": {
+    "key": {
+      "eldest_kid": "01202eb5d3b8d2fc63e4bc9c6edefb3a38f1043dfdc44a7268393951fe0ea5214cf00a",
+      "host": "keybase.io",
+      "kid": "0120fb58532270c261a00ae09ef3ee47b0dcf76592b485165977329e09ec4c4fc8690a",
+      "uid": "b01557931e13482e9e646593a64e9119",
+      "username": "masser"
+    },
+    "merkle_root": {
+      "ctime": 1580591326,
+      "hash": "2b1b8c5e3de98e3e80ec4b4c28270066744d5d0982c23e4c13ec97d9311e32996b0cc39ab07f51a2418f833d3f03c0219c7ebd8cdc9c99ac09d3f88d0b6129d1",
+      "hash_meta": "0e85fd563e1159d6c5b32f5c424d4c97101c3a11346a0dad77bdbf043ca4f287",
+      "seqno": 14452336
+    },
+    "service": {
+      "entropy": "7UNm+4us1fZE3TRssgGatBC2",
+      "hostname": "rauhala.info",
+      "protocol": "https:"
+    },
+    "type": "web_service_binding",
+    "version": 2
+  },
+  "client": {
+    "name": "keybase.io go client",
+    "version": "4.3.1"
+  },
+  "ctime": 1580591336,
+  "expire_in": 504576000,
+  "prev": "ed5dd9b2571746743f37dd55f59235bd46d7f7de01a2bafba14b9c0929751aee",
+  "seqno": 35,
+  "tag": "signature"
+}
+
+which yields the signature:
+
+hKRib2R5hqhkZXRhY2hlZMOpaGFzaF90eXBlCqNrZXnEIwEg+1hTInDCYaAK4J7z7kew3PdlkrSFFll3Mp4J7ExPyGkKp3BheWxvYWTESpcCI8Qg7V3ZslcXRnQ/N91V9ZI1vUbX994Borr7oUucCSl1Gu7EIJld9Pai+YxhSvZUIA+ZyzwEEbpy98uEqp975YKAEdz9AgHCo3NpZ8RAtksa9w0UQrwg8dPGEzahht+hY4QKQthbZ/yWuOSmfA3C5E1/KorbIEi+tcy5lkSJqNbZRSqBgpl9zerYw5bbAahzaWdfdHlwZSCkaGFzaIKkdHlwZQildmFsdWXEICy29OKKJc5LUELFLo8kW15N3zex43ussKLFeGLhTDapo3RhZ80CAqd2ZXJzaW9uAQ==
+
+And finally, I am proving ownership of this host by posting or
+appending to this document.
+
+View my publicly-auditable identity here: https://keybase.io/masser
+
+===================================================================================================================================

release.nix

@@ -1,21 +0,0 @@
-{ pkgs ? import <nixpkgs> {} }:
-
-let
-  haskellPackages = pkgs.haskellPackages;
-  site = pkgs.callPackage ./default.nix {};
-  shell = pkgs.buildEnv {
-    name = "site-shell";
-    paths = [];
-    buildInputs = [
-      haskellPackages.ghcid
-      haskellPackages.hasktags
-      (haskellPackages.ghcWithHoogle (h: site.buildInputs ++ site.propagatedBuildInputs))
-    ];
-  };
-
-in
-
-{
-  site = site;
-  shell = shell;
-}

site.hs

@@ -1,87 +0,0 @@
---------------------------------------------------------------------------------
-{-# LANGUAGE OverloadedStrings #-}
-import           Data.Monoid (mappend)
-import           Hakyll
-import           Data.List (sortBy, sortOn)
-import           Data.Time (formatTime, defaultTimeLocale)
-
-
---------------------------------------------------------------------------------
-main :: IO ()
-main = hakyllWith defaultConfiguration{ deployCommand = "ipfs add -Q -r _site" } $do
-    match "images/*" $ do
-        route   idRoute
-        compile copyFileCompiler
-
-    match "resources/*" $ do
-        route   idRoute
-        compile copyFileCompiler
-
-    match "css/*" $ do
-        route   idRoute
-        compile compressCssCompiler
-
-    match "js/*" $ do
-        route   idRoute
-        compile compressCssCompiler
-
-    match (fromList ["index.markdown", "contact.markdown"]) $ do
-        route   $ setExtension "html"
-        compile $ pandocCompiler
-            >>= loadAndApplyTemplate "templates/default.html" defaultContext
-            >>= relativizeUrls
-
-    match "posts/incomplete/*" $ do
-      route $ setExtension "html"
-      compile $ pandocCompiler
-        >>= loadAndApplyTemplate "templates/post.html"    postCtx
-        >>= loadAndApplyTemplate "templates/default.html" postCtx
-        >>= relativizeUrls
-
-    match "posts/guides/*" $ do
-      route $ setExtension "html"
-      compile $ pandocCompiler
-        >>= loadAndApplyTemplate "templates/post.html"    postCtx
-        >>= loadAndApplyTemplate "templates/default.html" postCtx
-        >>= relativizeUrls
-
-    match "posts/brainstorming/*" $ do
-      route $ setExtension "html"
-      compile $ pandocCompiler
-        >>= loadAndApplyTemplate "templates/post.html"    postCtx
-        >>= loadAndApplyTemplate "templates/default.html" postCtx
-        >>= relativizeUrls
-
-    create ["guides.html"] $ do
-        route idRoute
-        compile $ do
-            posts <- modFirst =<< loadAll "posts/guides/*"
-            let archiveCtx =
-                    listField "posts" postCtx (return posts) `mappend`
-                    constField "title" "Guides"            `mappend`
-                    defaultContext
-
-            makeItem ""
-                >>= loadAndApplyTemplate "templates/guides.html" archiveCtx
-                >>= loadAndApplyTemplate "templates/default.html" archiveCtx
-                >>= relativizeUrls
-
-
-    match "templates/*" $ compile templateBodyCompiler
-
-modFirst :: [Item a] -> Compiler [Item a]
-modFirst = fmap reverse . modified
-  where
-    modified = sortByM (getItemModificationTime . itemIdentifier)
-    sortByM f xs = map fst . sortOn snd <$> mapM (\x -> (,) x <$> f x) xs
-
---------------------------------------------------------------------------------
-postCtx :: Context String
-postCtx =
-    dateField "date" "%B %e, %Y" `mappend`
-    modifiedField "modified" "%B %e, %Y" `mappend`
-    defaultContext
-  where
-    modifiedField key format = field key $ \i -> do
-      time <- getItemModificationTime $ itemIdentifier i
-      return $ formatTime defaultTimeLocale format time

site/.envrc

@@ -0,0 +1 @@
+use nix

site/.stylish-haskell.yaml

@@ -0,0 +1,249 @@
+# stylish-haskell configuration file
+# ==================================
+
+# The stylish-haskell tool is mainly configured by specifying steps. These steps
+# are a list, so they have an order, and one specific step may appear more than
+# once (if needed). Each file is processed by these steps in the given order.
+steps:
+  # Convert some ASCII sequences to their Unicode equivalents. This is disabled
+  # by default.
+  # - unicode_syntax:
+  #     # In order to make this work, we also need to insert the UnicodeSyntax
+  #     # language pragma. If this flag is set to true, we insert it when it's
+  #     # not already present. You may want to disable it if you configure
+  #     # language extensions using some other method than pragmas. Default:
+  #     # true.
+  #     add_language_pragma: true
+
+  # Align the right hand side of some elements.  This is quite conservative
+  # and only applies to statements where each element occupies a single
+  # line.
+  - simple_align:
+      cases: false
+      top_level_patterns: false
+      records: false
+
+  # Import cleanup
+  - imports:
+      # There are different ways we can align names and lists.
+      #
+      # - global: Align the import names and import list throughout the entire
+      #   file.
+      #
+      # - file: Like global, but don't add padding when there are no qualified
+      #   imports in the file.
+      #
+      # - group: Only align the imports per group (a group is formed by adjacent
+      #   import lines).
+      #
+      # - none: Do not perform any alignment.
+      #
+      # Default: global.
+      align: none
+
+      # The following options affect only import list alignment.
+      #
+      # List align has following options:
+      #
+      # - after_alias: Import list is aligned with end of import including
+      #   'as' and 'hiding' keywords.
+      #
+      #   > import qualified Data.List      as List (concat, foldl, foldr, head,
+      #   >                                          init, last, length)
+      #
+      # - with_alias: Import list is aligned with start of alias or hiding.
+      #
+      #   > import qualified Data.List      as List (concat, foldl, foldr, head,
+      #   >                                 init, last, length)
+      #
+      # - new_line: Import list starts always on new line.
+      #
+      #   > import qualified Data.List      as List
+      #   >     (concat, foldl, foldr, head, init, last, length)
+      #
+      # Default: after_alias
+      list_align: new_line
+
+      # Right-pad the module names to align imports in a group:
+      #
+      # - true: a little more readable
+      #
+      #   > import qualified Data.List       as List (concat, foldl, foldr,
+      #   >                                           init, last, length)
+      #   > import qualified Data.List.Extra as List (concat, foldl, foldr,
+      #   >                                           init, last, length)
+      #
+      # - false: diff-safe
+      #
+      #   > import qualified Data.List as List (concat, foldl, foldr, init,
+      #   >                                     last, length)
+      #   > import qualified Data.List.Extra as List (concat, foldl, foldr,
+      #   >                                           init, last, length)
+      #
+      # Default: true
+      pad_module_names: false
+
+      # Long list align style takes effect when import is too long. This is
+      # determined by 'columns' setting.
+      #
+      # - inline: This option will put as much specs on same line as possible.
+      #
+      # - new_line: Import list will start on new line.
+      #
+      # - new_line_multiline: Import list will start on new line when it's
+      #   short enough to fit to single line. Otherwise it'll be multiline.
+      #
+      # - multiline: One line per import list entry.
+      #   Type with constructor list acts like single import.
+      #
+      #   > import qualified Data.Map as M
+      #   >     ( empty
+      #   >     , singleton
+      #   >     , ...
+      #   >     , delete
+      #   >     )
+      #
+      # Default: inline
+      long_list_align: new_line_multiline
+
+      # Align empty list (importing instances)
+      #
+      # Empty list align has following options
+      #
+      # - inherit: inherit list_align setting
+      #
+      # - right_after: () is right after the module name:
+      #
+      #   > import Vector.Instances ()
+      #
+      # Default: inherit
+      empty_list_align: inherit
+
+      # List padding determines indentation of import list on lines after import.
+      # This option affects 'long_list_align'.
+      #
+      # - <integer>: constant value
+      #
+      # - module_name: align under start of module name.
+      #   Useful for 'file' and 'group' align settings.
+      list_padding: 7
+
+      # Separate lists option affects formatting of import list for type
+      # or class. The only difference is single space between type and list
+      # of constructors, selectors and class functions.
+      #
+      # - true: There is single space between Foldable type and list of it's
+      #   functions.
+      #
+      #   > import Data.Foldable (Foldable (fold, foldl, foldMap))
+      #
+      # - false: There is no space between Foldable type and list of it's
+      #   functions.
+      #
+      #   > import Data.Foldable (Foldable(fold, foldl, foldMap))
+      #
+      # Default: true
+      separate_lists: false
+
+      # Space surround option affects formatting of import lists on a single
+      # line. The only difference is single space after the initial
+      # parenthesis and a single space before the terminal parenthesis.
+      #
+      # - true: There is single space associated with the enclosing
+      #   parenthesis.
+      #
+      #   > import Data.Foo ( foo )
+      #
+      # - false: There is no space associated with the enclosing parenthesis
+      #
+      #   > import Data.Foo (foo)
+      #
+      # Default: false
+      space_surround: false
+
+  # Language pragmas
+  - language_pragmas:
+      # We can generate different styles of language pragma lists.
+      #
+      # - vertical: Vertical-spaced language pragmas, one per line.
+      #
+      # - compact: A more compact style.
+      #
+      # - compact_line: Similar to compact, but wrap each line with
+      #   `{-#LANGUAGE #-}'.
+      #
+      # Default: vertical.
+      style: vertical
+
+      # Align affects alignment of closing pragma brackets.
+      #
+      # - true: Brackets are aligned in same column.
+      #
+      # - false: Brackets are not aligned together. There is only one space
+      #   between actual import and closing bracket.
+      #
+      # Default: true
+      align: false
+
+      # stylish-haskell can detect redundancy of some language pragmas. If this
+      # is set to true, it will remove those redundant pragmas. Default: true.
+      remove_redundant: true
+
+  # Replace tabs by spaces. This is disabled by default.
+  # - tabs:
+  #     # Number of spaces to use for each tab. Default: 8, as specified by the
+  #     # Haskell report.
+  #     spaces: 8
+
+  # Remove trailing whitespace
+  - trailing_whitespace: {}
+
+  # Squash multiple spaces between the left and right hand sides of some
+  # elements into single spaces. Basically, this undoes the effect of
+  # simple_align but is a bit less conservative.
+  # - squash: {}
+
+# A common setting is the number of columns (parts of) code will be wrapped
+# to. Different steps take this into account. Default: 80.
+columns: 80
+
+# By default, line endings are converted according to the OS. You can override
+# preferred format here.
+#
+# - native: Native newline format. CRLF on Windows, LF on other OSes.
+#
+# - lf: Convert to LF ("\n").
+#
+# - crlf: Convert to CRLF ("\r\n").
+#
+# Default: native.
+newline: native
+
+# Sometimes, language extensions are specified in a cabal file or from the
+# command line instead of using language pragmas in the file. stylish-haskell
+# needs to be aware of these, so it can parse the file correctly.
+#
+# No language extensions are enabled by default.
+language_extensions:
+  - RecordWildCards
+  - TemplateHaskell
+  - QuasiQuotes
+  - LambdaCase
+  - TupleSections
+  - MultiParamTypeClasses
+  - TypeApplications
+  - DataKinds
+  - TypeFamilies
+  - FlexibleContexts
+  - NamedFieldPuns
+  - MultiWayIf
+  - PolyKinds
+  - ExplicitForAll
+  - FunctionalDependencies
+  - ExplicitNamespaces
+  - ScopedTypeVariables
+  - ExistentialQuantification
+  - InstanceSigs
+  - GeneralizedNewtypeDeriving
+  - BangPatterns
+

site/app/site.hs

@@ -0,0 +1,107 @@
+{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE RecordWildCards #-}
+{-# LANGUAGE StrictData #-}
+import Hakyll
+
+import Data.List
+       (isPrefixOf, isSuffixOf)
+
+import System.FilePath
+       (takeFileName)
+
+
+--------------------------------------------------------------------------------
+main :: IO ()
+main = hakyllWith defaultConfiguration{ignoreFile = ignore} $ do
+  match "well-known/*" $ do
+    route (customRoute (prepend '.'. toFilePath))
+    compile copyFileCompiler
+
+  match "images/*" $ do
+    route   idRoute
+    compile copyFileCompiler
+
+  match "resources/*" $ do
+    route   idRoute
+    compile copyFileCompiler
+
+  match "css/*" $ do
+    route   idRoute
+    compile compressCssCompiler
+
+  match "js/*" $ do
+    route   idRoute
+    compile compressCssCompiler
+
+  match "posts/*" $ do
+    route $ setExtension "html"
+    compile $ pandocCompiler
+      >>= loadAndApplyTemplate "templates/post.html" postContext
+      >>= loadAndApplyTemplate "templates/default.html" defaultContext
+      >>= relativizeUrls
+
+  match "projects/*" $ do
+    route $ setExtension "html"
+    compile $ pandocCompiler
+      >>= loadAndApplyTemplate "templates/project.html" postContext
+      >>= loadAndApplyTemplate "templates/default.html" defaultContext
+      >>= relativizeUrls
+
+
+  match (fromList ["index.markdown", "contact.markdown"]) $ do
+    route   $ setExtension "html"
+    compile $ pandocCompiler
+      >>= loadAndApplyTemplate "templates/default.html" defaultContext
+      >>= relativizeUrls
+
+  archive $ Archive { output = "posts.html"
+                    , input = "posts/*"
+                    , title = "Posts"
+                    , template = "templates/post-list.html"
+                    , context = postContext
+                    }
+
+  archive $ Archive { output = "projects.html"
+                    , input = "projects/*"
+                    , title = "Projects"
+                    , template = "templates/project-list.html"
+                    , context = postContext
+                    }
+
+  match "templates/*" $ compile templateBodyCompiler
+  where
+
+    postContext :: Context String
+    postContext = dateField "date" "%B %e, %Y" <> defaultContext
+    ignore :: FilePath -> Bool
+    ignore path = any ($ takeFileName path)
+      [ ("." `isPrefixOf`)
+      , ("#" `isPrefixOf`)
+      , ("~" `isSuffixOf`)
+      , (".swp" `isSuffixOf`)
+      ]
+
+data Archive
+  = Archive { output :: Identifier
+            , input :: Pattern
+            , title :: String
+            , template :: Identifier
+            , context :: Context String
+            }
+
+archive :: Archive -> Rules ()
+archive Archive{..} = create [output] $ do
+  route idRoute
+  compile $ do
+    let itemsContext =
+             listField "items" context items
+          <> constField "title" title
+          <> defaultContext
+        items = recentFirst =<< loadAll input
+    makeItem ""
+      >>= loadAndApplyTemplate template itemsContext
+      >>= loadAndApplyTemplate "templates/default.html" itemsContext
+      >>= relativizeUrls
+
+prepend :: a -> [a] -> [a]
+prepend = (:)

site/default.nix

@@ -0,0 +1,10 @@
+{ mkDerivation, base, filepath, hakyll, stdenv, time }:
+mkDerivation {
+  pname = "site";
+  version = "0.1.0.0";
+  src = ./.;
+  isLibrary = false;
+  isExecutable = true;
+  executableHaskellDepends = [ base filepath hakyll time ];
+  license = stdenv.lib.licenses.bsd3;
+}

site/shell.nix

@@ -0,0 +1,16 @@
+with (import <nixpkgs> {});
+
+let site = haskellPackages.callPackage ./. {};
+
+in
+
+mkShell {
+  buildInputs = [
+    ghcid
+    stylish-haskell
+    haskellPackages.cabal-install
+    hlint
+    (haskellPackages.ghcWithPackages (_: site.buildInputs ++ site.propagatedBuildInputs))
+  ];
+}
+

site/site.cabal

@@ -9,8 +9,10 @@ maintainer:         mats.rauhala@iki.fi
 
 executable site
   main-is:          site.hs
+  hs-source-dirs:   app
   build-depends:    base == 4.*
                   , hakyll >= 4.10
                   , time
+                  , filepath
   ghc-options:      -threaded
   default-language: Haskell2010

templates/guides.html

@@ -1,3 +0,0 @@
-A list of small and big guides.
-
-$partial("templates/post-list.html")$

templates/post-list.html

@@ -1,7 +0,0 @@
-<ul>
-    $for(posts)$
-        <li>
-            <a href="$url$">$title$</a> - $modified$
-        </li>
-    $endfor$
-</ul>

templates/post.html

@@ -1,11 +0,0 @@
-<article>
-    <section class="header">
-        Posted on $date$, modified on $modified$
-        $if(author)$
-            by $author$
-        $endif$
-    </section>
-    <section>
-        $body$
-    </section>
-</article>