How to convert character encoding from CP932 to UTF-8 in nodejs javascript, using the nodejs-iconv module (or other solution)

Question

I'm attempting to convert a string from CP932 (aka Windows-31J) to utf8 in javascript. Basically I'm crawling a site that ignores the utf-8 request in the request header and returns cp932 encoded text (even though the html metatag indicates that the page is shift_jis).

Anyway, I have the entire page stored in a string variable called "html". From there I'm attempting to convert it to utf8 using this code:

var Iconv = require('iconv').Iconv;
var conv = new Iconv('CP932', 'UTF-8//TRANSLIT//IGNORE');

var myBuffer = new Buffer(html.length * 3);
myBuffer.write(html, 0, 'utf8')
var utf8html = (conv.convert(myBuffer)).toString('utf8');

The result is not what it's supposed to be. For example, the string: "投稿者さんの 稚内全日空ホテル のクチコミ (感想・情報)" comes out as "ｿｽｿｽｿｽeｿｽﾒゑｿｽｿｽｿｽｿｽｿｽ ｿｽtｿｽｿｽｿｽSｿｽｿｽｿｽｿｽｿｽzｿｽeｿｽｿｽ ｿｽﾌクｿｽ`ｿｽRｿｽ~ (ｿｽｿｽｿｽzｿｽEｿｽｿｽｿｽｿｽ)"

If I remove //TRANSLIT//IGNORE (Which should cause it to return similar characters for missing characters, and failing that omit non-transcode-able characters), I get this error: Error: EILSEQ, Illegal character sequence.

I'm open to using any solution that can be implemented in nodejs, but my search results haven't yielded many options outside of the nodejs-iconv module.

nodejs-iconv ref: https://github.com/bnoordhuis/node-iconv

Thanks!

Edit 24.06.2011: I've gone ahead and implemented a solution in Java. However I'd still be interested in a javascript solution to this problem if somebody can solve it.

Answer 1

I got same trouble today :)
It depends libiconv. You need libiconv-1.13-ja-1.patch.
Please check followings.

• http://d.hatena.ne.jp/ushiboy/20110422/1303481470
• http://code.xenophy.com/?p=1529

or you can avoid problem using iconv-jp try

npm install iconv-jp

Answer 2

I had same problem, but with CP1250. I was looking for problem everywhere and everything was OK, except call of request – I had to add encoding: 'binary'.

request = require('request')
Iconv  = require('iconv').Iconv

request({uri: url, encoding: 'binary'}, function(err, response, body) {
    body = new Buffer(body, 'binary')
    iconv = new Iconv('CP1250', 'UTF8')
    body = iconv.convert(body).toString()
    // ...
})