request-time attributes and xml representations...

[Eduardo Pelegri-Llopart <pe...@eng.sun.com>]

To be double clear.  My intretation of the spec is the same as
Danno's.  I have not yet completed reading all my tomcat-dev digest
messages, but the rest of the 23rd seemed to suggest that Ruby and
Danno clarified the misunderstanding.

Thanks!

	- eduard/o

======

Date: Thu, 23 Dec 1999 13:44:56 -0500
To: tomcat-dev@jakarta.apache.org
From: rubys@us.ibm.com
Subject: Re: [PATCH] Bogus request time attrs accepted
Message-ID: <85...@d54mta04.raleigh.ibm.com>

>Only request time attrs of "<%= foo %>" should be recognized.  Bizare
things
>such as "<jsp:expression foo /> and "<%= foo />" are accepted by existing
>code, which is probobly left over from previous revs of the spec.

-1

I'm all for making a change which ensures that the open and closing
delimiters match.  And if there are any other changes which makes the
Jasper reference implementation more compliant with the current spec, then
I am all for them too.

If someone wants to take the position that the current spec is misguided
and should be changed - this is a subject that I am absolutely neutral to.
But the right response it to fix the spec, not to submit patches that
remove support from the reference implementation of portions that are in
the spec.

------------------------------

Date: Thu, 23 Dec 1999 16:43:03 -0700
To: <to...@jakarta.apache.org>
From: "Danno Ferrin" <sh...@earthlink.net>
Subject: Re: [PATCH] Bogus request time attrs accepted
Message-ID: <00...@DRF26>

Perhaps I was a bit to terse in posting the patch.  This is in the context
of request time parameters in actions (standard ant taglib) and does not
have anything to do with the XML implmentations

Spec says foo='<%= foo %>' only, or '%= foo %' for XML.

Refrence implmentatnion accepts the bogus stuff like
'<jsp:expression foo />' and the mixed cases.

The patch makes jasper accept only the form specified in the specification.

--Danno