On Tue, 2008-05-20 at 07:34 +0000, [EMAIL PROTECTED] wrote: > Author: markt > Date: Tue May 20 00:34:01 2008 > New Revision: 658132 > > URL: http://svn.apache.org/viewvc?rev=658132&view=rev > Log: > Add some comments to Remy's -1s > -1: remm (I think it is a good patch, but it throws an error for something > which was working before, > so I would think keeping it for the next release would be better) > + markt How about making it optional with a system property that can be > used > + to restore the current behaviour is an app depends on it?
Ok. > * Fix https://issues.apache.org/bugzilla/show_bug.cgi?id=44994 > Correct BNF grammar so ${0 lt a ? 1 lt a ? "many": "one": "none"} works > @@ -93,6 +95,10 @@ > it may not apply to HTTP/1.1, which may be more strict; passing > full URLs in the > request line is normally an HTTP/1.1 thing, from what I > remember, usable as an alternative > to the mandatory Host header) > + markt Both the HTTP/1.1 and HTTP/1.0 contain broadly the same text > + regarding tolerant applications. (ie allow multiple SP/HT between > + method-url and url-httpVersion). However, it is optional so we > don't > + have to support it. What does httpd do about that BTW ? > * Fix https://issues.apache.org/bugzilla/show_bug.cgi?id=42747 > Use any META-INF/context.xml in a consistent manner and ensure it is used > on > @@ -100,3 +106,6 @@ > http://svn.apache.org/viewvc?rev=657995&view=rev > +1: markt > -1: remm (would need convincing that this is really well tested, otherwise > delay to next release) > + markt Exactly what would convince you that this was well tested above > and > + beyond the local testing that I have done to convince myself that > + this patch does what I want it to. It's a start. Rémy --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
