- From: Henrik Nordstrom <henrik@henriknordstrom.net>
- Date: Tue, 28 Jul 2009 18:07:48 +0200
- To: Yves Lafon <ylafon@w3.org>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
tis 2009-07-28 klockan 05:55 -0400 skrev Yves Lafon: > Implementation-wise it is clear that if the SHOULD NOT is not respected > then the MUST applies, so I don't think that rewording is really necessary. I read the text as the MUST include only applying if the request is not the result of an If-Range conditional request. I don't read it as specifying which exact headers to include in an If-Range based response when not respecting the SHOULD NOT, that case to me is implementation defined. But that's my reading.. My point of bringing this up in the content of Content-MD5 is only to show that sending Content-MD5 is NOT RECOMMENDED in most HTTP/1.1 206 responses, with the use of ETag and If-Range strongly recommended. This kind of rules out Content-MD5 to be used as a message transport validation method, even ignoring the other problems demonstrated earlier. Regards Henrik
Received on Tuesday, 28 July 2009 16:08:26 UTC