- From: Willy Tarreau <w@1wt.eu>
- Date: Thu, 30 Jun 2011 15:11:29 +0200
- To: Julian Reschke <julian.reschke@gmx.de>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
On Thu, Jun 30, 2011 at 02:46:37PM +0200, Julian Reschke wrote: > Hi, > > I quickly ran some tests, and the results are (with what I suppose are > the default settings): > > Apache-Coyote/1.1 (Tomcat): > > Limit for a single header field: ~8000 > Limit for all fields: ~8000 > Status Code: 400 > > Apache/2.2.14: > > Limit for a single header field: ~8180 > Limit for all fields: > 16000 > Status Code: 400 > > microsoft.com: > > Limit for a single header field: ~16000 > Limit for all fields: ~16000 > Status Code: 400 > > google.com: > > Limit for a single header field: ~15400 > Limit for all fields: > 15400 > Status Code: 400 (413 for the larger values) > > > So 413 doesn't seem to be used in general for this case. > > Should it? In that case we should clarify the spec... I think we should target 413 for this. Every time I got reports of 400 due to too large requests, it wasted several persons' time trying to figure what was wrong. Anyway if a UA doesn't know about 413, its interpretation falls back to 400 as for any unknown 4xx. Just my 2 cents, Willy
Received on Thursday, 30 June 2011 13:12:05 UTC