Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: pass (athena.apache.org: domain of heshiming@gmail.com
 designates 209.85.214.180 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:date:message-id:subject:from:to:content-type;
        b=LrJaQeoeuE2TG3Xr2BvEblF35GAtNOFG4/6nDFaVvjx12DMp5jkpTBL/qMXXNcbv8U
         BK9GdjZVXaAPGxSS31jI+xnBj6X8RhoMOTu9PGRzPmDbKXxlx128nArUtNQtX74fRmf7
         35MftNx/HpVO7d1OfmW+DbxLvsnhR1ld9gujs=
MIME-Version: 1.0
Date: Thu, 3 Feb 2011 20:59:05 +0800
Message-ID: <AANLkTinq1=D2tuTazDw1+TALooyMQZbiy46ayxPYRLxM@mail.gmail.com>
Subject: 400, ('bad_request', 'Attachment name is not UTF-8 encoded')
From: He Shiming <heshiming@gmail.com>
To: user@couchdb.apache.org
Content-Type: text/plain; charset=ISO-8859-1

Hi,

I'm accessing CouchDB 1.0.1 on localhost from python-couchdb. I
discovered this problem while trying to use Chinese characters to name
attachments of the document.

So I'm calling couchdb.Database.put_attachment with some Chinese
characters as name: u'\u7eff\u53f6\u60c5-2753303'. Attachment saving
will cause an HTTP 400 error, with message 'Attachment name is not
UTF-8 encoded'.

I'm not sure if I should encode the string to utf-8 myself. So I tried
to encode the above string by unicode(filename).encode('utf-8'). It
becomes: '\xe7\xbb\xbf\xe5\x8f\xb6\xe6\x83\x85-2753303'. I still got
the same error.

Sometimes it's not even HTTP 400 error. It's "error: [Errno 104]
Connection reset by peer", which means the connection got cut off
unexpectedly. But other functions of CouchDB remain fine after this
error. It's happening to non-alphanumeric character attachment names
in particular.

Is this a bug or should I avoid using such names as attachment name?

-- 
Best regards,
He Shiming