[whatwg] asking for UA developer's feedback on Content-Disposition filename I18N issue

Julian Reschke julian.reschke at gmx.de
Sun Jul 20 09:24:58 PDT 2008


I posted the mail below four weeks ago HTML WG's mailing list, hoping 
for feedback from Microsoft and Apple. All I got was some off-list 
feedback that wasn't really helpful, thus I'm reposting over here 
(trying to find out whether people over here are more receptive ;-).

BR, Julian

-------- Original Message --------
From: - Thu Jun 26 14:29:05 2008
X-Mozilla-Status: 1001
X-Mozilla-Status2: 00800000
Message-ID: <48638B90.9050906 at gmx.de>
Date: Thu, 26 Jun 2008 14:29:04 +0200
From: Julian Reschke <julian.reschke at gmx.de>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv: 
Gecko/20060516 Thunderbird/ Mnenhy/
MIME-Version: 1.0
To: public-html at w3.org <public-html at w3.org>
Subject: asking for UA developer's feedback on Content-Disposition 
filename I18N issue
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit


our chair suggested to me that I summarize & repost this issue, in order
to collect feedback from the developers of UAs who may have missed
previous discussions on this mailing list (and the httpbis mailing
list). So...:

- RFC2183 defines the Content-Disposition header

- RFC2616 (HTTP/1.1) defines the Content-Disposition header, as used in
HTTP (<http://tools.ietf.org/html/rfc2616#section-19.5.1>)

- As far as I can tell, all major UAs implement simple cases, such as

   Content-Disposition: attachment; filename="fname.ext"

- HTTP/1.1 does not allow non-ISO8859-1 characters in headers, so the
simple format can not be used for filenames that use non ISO8859-1

- For this case, RFC2231 defines escaping rules
(<http://tools.ietf.org/html/rfc2231>). A simple example would be:

    Content-Disposition: attachment;

   for a filename of "äöü߀.txt".

- This has been implemented for a long time now in Firefox and Opera.

- Other UAs do not seem to provide any reliable way to transport
non-8859-1 characters in the filename.

- Implementing the RFC2231-style escaping would be backwards-compatible;
right now, for instance, IE simply ignores the extended format.

- More discussion: around
<http://www.w3.org/Bugs/Public/show_bug.cgi?id=5773#c7> (HTML WG's
Bugzilla) and (HTTPbis mailing list)

 From those UA implementors who do not implement this, I'd like to find
out what the chances are that this may be supported in future releases.
What's holding you back? Are there any issues with the combo of
RFC2616/RFC2183/RFC2231 that need to be addressed?

BR, Julian

More information about the whatwg mailing list