allow use of multipart/form-data in draft-reschke-webdav-post

[Arnaud Quillaud]

Hello,

Late comment on draft-reschke-webdav-post:

Browser based applications are not allowed to upload files from the
client filesystem by any mean other than using an HTML form containing a
file param and an enctype of multipart/form-data. As a consequence,
those applications have no way to upload local files to a WebDAV
repository using PUT (my knowledge of browser based technologies is
rather limited but I have found at least a few references to this
limitation).

http://tools.ietf.org/html/draft-reschke-webdav-post could partially
solve this issue by simply allowing a multipart/form-data containing a
single subpart to be POSTed.

Modifying a resource by uploading it again would still not be possible
of course but that can be somehow workarounded.

Arnaud Quillaud

[Petr Tomasek]

On Mon, Nov 02, 2009 at 11:15:02AM +0100, Arnaud Quillaud wrote:

> Hello,
>
> Late comment on draft-reschke-webdav-post:
>
> Browser based applications are not allowed to upload files from the
> client filesystem by any mean other than using an HTML form containing a
> file param and an enctype of multipart/form-data. As a consequence,
> those applications have no way to upload local files to a WebDAV
> repository using PUT (my knowledge of browser based technologies is
> rather limited but I have found at least a few references to this
> limitation).
>
> http://tools.ietf.org/html/draft-reschke-webdav-post could partially
> solve this issue by simply allowing a multipart/form-data containing a
> single subpart to be POSTed.
>
> Modifying a resource by uploading it again would still not be possible
> of course but that can be somehow workarounded.
>
> Arnaud Quillaud

Wouldn't it be easier (and a cleaner solution) if the browsers just supported PUT/WebDAV?

P.T.

--
Petr Tomasek <http://www.etf.cuni.cz/~tomasek>
Jabber: butrus@...
SIP: butrus@...

[Arnaud Quillaud]

On 11/2/09 8:36 PM, Petr Tomasek wrote:

> On Mon, Nov 02, 2009 at 11:15:02AM +0100, Arnaud Quillaud wrote:
>    
>> Hello,
>>
>> Late comment on draft-reschke-webdav-post:
>>
>> Browser based applications are not allowed to upload files from the
>> client filesystem by any mean other than using an HTML form containing a
>> file param and an enctype of multipart/form-data. As a consequence,
>> those applications have no way to upload local files to a WebDAV
>> repository using PUT (my knowledge of browser based technologies is
>> rather limited but I have found at least a few references to this
>> limitation).
>>
>> http://tools.ietf.org/html/draft-reschke-webdav-post could partially
>> solve this issue by simply allowing a multipart/form-data containing a
>> single subpart to be POSTed.
>>
>> Modifying a resource by uploading it again would still not be possible
>> of course but that can be somehow workarounded.
>>
>> Arnaud Quillaud
>>      
> Wouldn't it be easier (and a cleaner solution) if the browsers just supported PUT/WebDAV?
>    

It might be cleaner but definitely not easier to get this standardized
on all browsers (and have all legacy browsers disappear).

Found a page describing the behavior of different browsers at
http://pseudo-flaw.net/content/web-browsers/form-data-encoding-roundup/

Arnaud Quillaud
> P.T.
>
>    

[Julian Reschke]

Arnaud Quillaud wrote:

> Hello,
>
> Late comment on draft-reschke-webdav-post:
>
> Browser based applications are not allowed to upload files from the
> client filesystem by any mean other than using an HTML form containing
> a file param and an enctype of multipart/form-data. As a consequence,
> those applications have no way to upload local files to a WebDAV
> repository using PUT (my knowledge of browser based technologies is
> rather limited but I have found at least a few references to this
> limitation).
>
> http://tools.ietf.org/html/draft-reschke-webdav-post could partially
> solve this issue by simply allowing a multipart/form-data containing a
> single subpart to be POSTed.

It could. However I'm not entirely happe with special casing a single mime type.

An alternative would be to define a *second* URI that would accept POST requests from HTML forms.

> Modifying a resource by uploading it again would still not be possible > of course but that can be somehow workarounded.

Orthogonal problem...

Best regards, Julian

--
DSL-Preisknaller: DSL Komplettpakete schon für 16,99 Euro mtl.!*
http://portal.gmx.net/de/go/dsl02