RE: I-D ACTION:draft-ietf-webdav-quota-01.txt

Any error can be "resolved" by changing the request to something
else that doesn't fail, so that can't be the criterion used to
distinguish a 4xx from a 5xx.  All that sections 10.4 and 10.5 say
are: 

"The 4xx class of status code is intended for cases in which the
 client seems to have erred."

"Response status codes beginning with the digit "5" indicate cases in
 which the server is aware that it has erred or is incapable of
 performing the request."

So requests that are malformed are clearly 4xx errors, but
most other error cases are ambiguous, because looking at the
error one way, the client erred in asking the server to do something
it couldn't do (e.g. create a larger file than is permitted), but
looking at them another way, the server erred by not (or by being
incapable of) performing the specified request (e.g. create the large
file).

Cheers,
Geoff


-----Original Message-----
From: Jim Luther [mailto:luther.j@apple.com]
Sent: Tuesday, March 25, 2003 12:20 PM
To: WebDAV
Subject: Re: I-D ACTION:draft-ietf-webdav-quota-01.txt



I'd refer back to rfc2616's language in sections 10.4 "Client Error 
4xx" and 10.5 "Server Error 5xx". 4xx is for errors the server thinks 
can be resolved by changes in the client's request, and 5xx is for 
errors the server thinks are caused by something server-side.

- Jim

On Tuesday, March 25, 2003, at 08:35  AM, Clemm, Geoff wrote:

>
> That wasn't the question I meant to ask.
>
> Clearly there are clients that handle a specific code
> (such as 507) differently from another specific code (such as 413).
> But in those cases it doesn't matter whether the standard
> defined that specific code to be in the 4xx or 5xx range.
>
> What I was asking was whether there was a client that
> generically handled 4xx codes (i.e. a 4xx code that it had no
> special handling for) in a significantly different way than
> it handles a 5xx code (i.e. a 5xx code that it had no special
> handling for).  I.e., when we are deciding whether to put a
> specific code in the 4xx or 5xx range, does it matter which
> one we pick?
>
> Cheers,
> Geoff
>
> -----Original Message-----
> From: Jim Luther [mailto:luther.j@apple.com]
> Sent: Tuesday, March 25, 2003 10:46 AM
> To: WebDAV
> Subject: Re: I-D ACTION:draft-ietf-webdav-quota-01.txt
>
>
>
> On Monday, March 24, 2003, at 07:53  PM, Clemm, Geoff wrote:
>
>> Just for interests sake, is there any client that acts significantly
>> differently if it were to receive a 4xx response instead of a 5xx
>> response?  If not, this question is merely an aesthetic quibble (:-).
>
> Yes there is a client that handles those responses quite differently.
>
> The Mac OS X WebDAV file system client translates 507 to ENOSPC (No
> space left on device) which is interpreted by most Macintosh
> applications to mean the device is full; the WebDAV file system
> translates 413 as a generic 4xx response to EINVAL (Invalid argument)
> which is interpreted by most Macintosh applications to mean "something
> wasn't right - who knows what?"
>
> The Mac OS X WebDAV file system client is one of the few clients
> actually using quotas today and has been for over 1-1/2 years now.
>
> Jim Luther
> Apple Computer, Inc.
>

Received on Tuesday, 25 March 2003 15:51:09 UTC