How does CouchDB handle UTF-8?

I'm quite puzzled by CouchDB: if I send a PUT request with some JSON string fields encoded as UTF-8, the non 7 bit ASCII characters get converted to the "\uXXXX" escape sequence. Is there any way to tell it not to esc开发者_JAVA百科ape UNICODE?

Those \uXXXX are the correct way of encoding UTF-8 characters in Javascript.

Considering CouchDB is accessed using JSON (i.e. Javascript data), those sequences should be interepreted when using the data, and this should not be a problem.

CouchDB use mochiweb to handle JSON encoding/decoding.

There is an argument do encoding routine witch tells to output without those \uXXXX.

Simple way to apply patch is:

1. get CouchDB source
2. edit src/mochiweb/mochijson2.erl
3. Find -record(encoder, {handler=null, utf8=false}). around line 45.
4. Change to utf8=true
5. make clean; make; make install

I found the discussion with Chris Anderson http://erlangine.feautec.pp.ru/?p=232 and it tells me there is a chance to get this behavior out of box if someone care to make proper patch to CouchDB.