html-tidy: unify with its drop-in replacement tidy-html5

- the original project has been unmaintained for years
- some dependants needed to be patched due to renamed headers
  https://github.com/htacg/tidy-html5/issues/326#issuecomment-160329114
- separate tidy-html5 package was removed, as since the 5.0.0 version
  it's meant as a successor to both, and library name got back
  from libtidy5.so to libtidy.so
  https://github.com/htacg/tidy-html5/issues/326#issuecomment-160314666

/cc committers to tidy-html5: @edwjto and @zimbatm.

pkgs/tools/text/html-tidy/default.nix

@@ -1,41 +1,31 @@
-{ fetchcvs, stdenv, autoconf, automake, libtool }:
+{ stdenv, fetchurl, cmake, libxslt }:
 
-let date = "2009-07-04"; in
-  stdenv.mkDerivation rec {
-    name = "html-tidy-20090704";
+let
+  version = "5.0.0";
+in
+stdenv.mkDerivation rec {
+  name = "html-tidy-${version}";
 
-    # According to http://tidy.sourceforge.net/, there are no new
-    # release tarballs, so one has to either get the code from CVS or
-    # use a decade-old tarball.
+  src = fetchurl {
+    url = "https://github.com/htacg/tidy-html5/archive/${version}.tar.gz";
+    sha256 = "1qz7hgk482496agngp9grz4jqkyxrp29r2ywbccc9i5198yspca4";
+  };
 
-    src = fetchcvs {
-      inherit date;
-      cvsRoot = ":pserver:anonymous@tidy.cvs.sourceforge.net:/cvsroot/tidy";
-      module = "tidy";
-      sha256 = "d2e68b4335ebfde65ef66d5684f7693675c98bdd50b7a63c0b04f61db673aa6d";
-    };
+  nativeBuildInputs = [ cmake libxslt/*manpage*/ ];
 
-    buildInputs = [ autoconf automake libtool ];
+  # ATM bin/tidy is statically linked, as upstream provides no other option yet.
+  # https://github.com/htacg/tidy-html5/issues/326#issuecomment-160322107
 
-    preConfigure = ''
-      cp -rv build/gnuauto/* .
-      AUTOMAKE="automake --foreign" autoreconf -vfi
+  meta = with stdenv.lib; {
+    description = "A HTML validator and `tidier'";
+    longDescription = ''
+      HTML Tidy is a command-line tool and C library that can be
+      used to validate and fix HTML data.
     '';
+    license = licenses.libpng; # very close to it - the 3 clauses are identical
+    homepage = http://html-tidy.org;
+    platforms = platforms.all;
+    maintainers = with maintainers; [ edwtjo ];
+  };
+}
 
-    doCheck = true;
-
-    meta = {
-      description = "HTML Tidy, an HTML validator and `tidier'";
-
-      longDescription = ''
-        HTML Tidy is a command-line tool and C library that can be
-        used to validate and fix HTML data.
-      '';
-
-      license = stdenv.lib.licenses.mit;
-
-      homepage = http://tidy.sourceforge.net/;
-
-      maintainers = [ ];
-    };
-  }