feat: allow user to specific encoding

Support for manually specifying the source encoding using iconv-lite to convert to utf-8 from the specified encoding.

cli/index.js

@@ -1,16 +1,19 @@
 #!/usr/bin/env node
 
+var minimist = require('minimist');
 var readFileSync = require('fs').readFileSync;
 
-var argv = require('minimist')(process.argv.slice(2), opts={
-  'boolean': ['V', 'h', 'd', 'v', 'i', 'n',],
-  'alias': {
+var argv = minimist(process.argv.slice(2), opts = {
+  boolean: ['V', 'h', 'd', 'v', 'i', 'n',],
+  string: ['e',],
+  alias: {
     V: 'version',
     h: 'help',
     d: 'debug',
     v: 'verbose',
     i: 'images',
     n: 'nocompress',
+    e: 'encoding',
   },
 });
 
@@ -77,4 +80,4 @@ if (argv.verbose) {
   inliner.on('jobs', function jobs(event) {
     console.error(event);
   });
-}
+}

docs/usage.txt

@@ -3,17 +3,22 @@
 
     $ inliner [options] url-or-filename
 
-  Options:
+  Flags:
 
     -h, --help        output usage information
     -V, --version     output the version number
     -v, --verbose     echo on STDERR the progress of inlining
     -n, --nocompress  don't compress CSS or HTML - useful for debugging
     -i, --images      don't encode images - keeps files size small, but more requests
 
+  Options:
+
+    -e, --encoding    attempt to transcode the page from the specified encoding to utf-8
+
   Examples:
 
     $ inliner -v https://twitter.com > twitter.html
     $ inliner -ni local-file.html > local-file.min.html
+    $ inliner -e windows-1253 http://foofootos.gr > foofootos-utf8.html
 
   For more details see http://github.com/remy/inliner/

lib/get.js

@@ -41,6 +41,7 @@ module.exports = function get(url, options) {
   debug('inliner.get url: %s', url);
 
   var settings = assign({}, options, {
+    encoding: null,
     followRedirect: true,
   });
 
@@ -63,4 +64,4 @@ module.exports = function get(url, options) {
   });
 
   return cache[url];
-};
+};

lib/index.js

@@ -9,6 +9,7 @@ var assign = require('lodash.assign');
 var forEach = require('lodash.foreach');
 var Promise = require('es6-promise').Promise; // jshint ignore:line
 var findAssets = require('./find-assets');
+var iconv = require('iconv-lite');
 
 function Inliner(url, options, callback) {
   var inliner = this;
@@ -129,10 +130,30 @@ function main() {
     inliner.jobs.done.html();
     debug('processing HTML');
 
-    var todo = findAssets(res.body);
+    if (inliner.options.encoding !== undefined) {
+      // force transcoding to utf-8 from specified encoding
+      // todo: check the meta header to automatically pick encoding
+      var body = iconv.encode(
+        iconv.decode(res.body, inliner.options.encoding),
+        'utf-8'
+      );
+    } else {
+      var body = res.body;
+    }
+
+    var todo = findAssets(body);
     var $ = todo.$;
     delete todo.$;
 
+    if (inliner.options.encoding !== undefined) {
+      // when transcoding remove any meta tags setting the charset
+      $('meta').each(function charMeta(index) {
+        if ($(this).attr('content').toLowerCase().indexOf('charset=')) {
+          $(this).remove();
+        }
+      });
+    }
+
     forEach(todo, function forEach(todo, key) {
       if (key === 'images' && !inliner.options.images) {
         // skip images if the user doesn't want them
@@ -346,4 +367,4 @@ function completeJob(type) {
 
   // this allows me to include addJob as part of a promise chain
   return arguments[1];
-}
+}

package.json

@@ -19,6 +19,7 @@
     "cheerio": "^0.19.0",
     "debug": "^2.2.0",
     "es6-promise": "^2.3.0",
+    "iconv-lite": "^0.4.11",
     "lodash.assign": "^3.2.0",
     "lodash.defaults": "^3.1.2",
     "lodash.foreach": "^3.0.3",