tesseract: 3.02.02 -> 3.04.01

From the upstream changelog:

 * Tesseract development is now done with Git and hosted at github.com
   (Previously we used Subversion as a VCS and code.google.com for
   hosting).

So let's move over to the GitHub repository, where the organisation also
includes a full repository for tessdata, so we no longer need to fetch
it one-by-one.

The build also got significantly simpler, because we no longer need to
run autoconf, neither do we need to patch the configure script for
Leptonica headers.

This also has the advantage that we don't need to use the
enableLanguages attribute for the test runner anymore.

Full upstream changelog can be found at:

https://github.com/tesseract-ocr/tesseract/blob/c4d273d33cc36e/ChangeLog

Tested against all NixOS tests with enabled OCR (chromium, emacs-daemon,
installer.luksroot and lightdm).

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @viric

nixos/lib/testing.nix

@@ -93,7 +93,7 @@ rec {
 
       vms = map (m: m.config.system.build.vm) (lib.attrValues nodes);
 
-      ocrProg = tesseract.override { enableLanguages = [ "eng" ]; };
+      ocrProg = tesseract;
 
       # Generate onvenience wrappers for running the test driver
       # interactively with the specified network, and for starting the

pkgs/applications/graphics/tesseract/default.nix

@@ -1,53 +1,31 @@
-{ stdenv, fetchurl, autoconf, automake, libtool, leptonica, libpng, libtiff
-, enableLanguages ? null
+{ stdenv, fetchFromGitHub, pkgconfig, leptonica, libpng, libtiff
+, icu, pango, opencl-headers
 }:
 
-with stdenv.lib;
-
-let
-  majVersion = "3.02";
-  version = "${majVersion}.02";
-
-  mkLang = lang: sha256: let
-    src = fetchurl {
-      url = "http://tesseract-ocr.googlecode.com/files/tesseract-ocr-${majVersion}.${lang}.tar.gz";
-      inherit sha256;
-    };
-  in "tar xfvz ${src} -C $out/share/ --strip=1";
-
-  wantLang = name: const (enableLanguages == null || elem name enableLanguages);
-
-  extraLanguages = mapAttrsToList mkLang (filterAttrs wantLang {
-    cat = "0d1smiv1b3k9ay2s05sl7q08mb3ln4w5iiiymv2cs8g8333z8jl9";
-    rus = "059336mkhsj9m3hwfb818xjlxkcdpy7wfgr62qwz65cx914xl709";
-    spa = "1c9iza5mbahd9pa7znnq8yv09v5kz3gbd2sarcgcgc1ps1jc437l";
-    nld = "162acxp1yb6gyki2is3ay2msalmfcsnrlsd9wml2ja05k94m6bjy";
-    eng = "1y5xf794n832s3lymzlsdm2s9nlrd2v27jjjp0fd9xp7c2ah4461";
-    slv = "0rqng43435cly32idxm1lvxkcippvc3xpxbfizwq5j0155ym00dr";
-    jpn = "07v8pymd0iwyzh946lxylybda20gsw7p4fsb09jw147955x49gq9";
-  });
-in
-
 stdenv.mkDerivation rec {
   name = "tesseract-${version}";
+  version = "3.04.01";
 
-  src = fetchurl {
-    url = "http://tesseract-ocr.googlecode.com/files/tesseract-ocr-${version}.tar.gz";
-    sha256 = "0g81m9y4iydp7kgr56mlkvjdwpp3mb01q385yhdnyvra7z5kkk96";
+  src = fetchFromGitHub {
+    owner = "tesseract-ocr";
+    repo = "tesseract";
+    rev = version;
+    sha256 = "0h1x4z1h86n2gwknd0wck6gykkp99bmm02lg4a47a698g4az6ybv";
   };
 
-  buildInputs = [ autoconf automake libtool leptonica libpng libtiff ];
+  tessdata = fetchFromGitHub {
+    owner = "tesseract-ocr";
+    repo = "tessdata";
+    rev = "3cf1e2df1fe1d1da29295c9ef0983796c7958b7d";
+    sha256 = "1v4b63v5nzcxr2y3635r19l7lj5smjmc9vfk0wmxlryxncb4vpg7";
+  };
 
-  hardeningDisable = [ "format" ];
+  nativeBuildInputs = [ pkgconfig ];
+  buildInputs = [ leptonica libpng libtiff icu pango opencl-headers ];
 
-  preConfigure = ''
-      ./autogen.sh
-      substituteInPlace "configure" \
-        --replace 'LIBLEPT_HEADERSDIR="/usr/local/include /usr/include"' \
-                  'LIBLEPT_HEADERSDIR=${leptonica}/include'
-  '';
+  LIBLEPT_HEADERSDIR = "${leptonica}/include";
 
-  postInstall = concatStringsSep "; " extraLanguages;
+  postInstall = "cp -Rt \"$out/share/tessdata\" \"$tessdata/\"*";
 
   meta = {
     description = "OCR engine";