From edward.burns at oracle.com Fri Nov 5 14:32:23 2010 Content-Type: multipart/mixed; boundary="===============0916570468809466601==" MIME-Version: 1.0 From: Ed Burns To: jsr-314-open-mirror at lists.jboss.org Subject: Re: [jsr-314-open-mirror] [jsr-314-open] ANNOUNCE: JSF 2.1 Spec, Release Candidate Five Date: Fri, 05 Nov 2010 11:32:01 -0700 Message-ID: <19668.19873.81969.743842@gargle.gargle.HOWL> In-Reply-To: Re: [jsr-314-open] ANNOUNCE: JSF 2.1 Spec, Release Candidate Five on 4 November 2010 --===============0916570468809466601== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable >>>>> On Thu, 04 Nov 2010 23:33:29 -0700, Ed Burns said: >>>>> On Wed, 3 Nov 2010 13:09:30 -0700, Ed Burns said: EB> Hello JSF Expert Community, EB> This is the fifth, and most likely final, iteration of the JSF 2.1 EB> spec. Please let's get this done so we can move on to JSF 2.2 and EB> resolve all these nice composite component issues. EB> Well, here's the sixth iteration. It has the improvements suggested by EB> Andy. EB> https://javaserverfaces-spec-public.dev.java.net/files/documents/1936/1= 53911/jsf-spec-2.1-SNAPSHOT-20101105.zip Imre Osswald, another excellent community member, grabbed me on IM this morning and pointed out another little hole in the namespace solution. This prompted me to make an edit to the PDF, which I have updated in the above zip. If you downloaded it, download it again. It has a new pdf but the rest is the same. Here is the commit message. M integrationWithFacelets.fm M preface.fm M title.fm - update date - 10.3.2 add spec language that asserts that the root namespace must be passed through unmodified. The unprefixed namespace, also known as the root namespace, must be passed through without modification or check for validity. The passing through of the root namespace must occur on any non-prefixed element in a facelet page. For example, the following markup declaration: a + b 27 would be rendered as a + b 27 I added an automated test to Mojarra to make sure this works and I do indeed see the square root of a + b to the 27th when viewing the page in Firefox with the necessary MathML fonts installed. Ed -- = | edward.burns(a)oracle.com | office: +1 407 458 0017 | homepage: | http://ridingthecrest.com/ | 3 work days until German Oracle User's Group Conference | 10 work days until GlassFish 3.1 Hard Code Freeze --===============0916570468809466601==-- From andy.schwartz at oracle.com Fri Nov 5 14:56:16 2010 Content-Type: multipart/mixed; boundary="===============3812196750539280602==" MIME-Version: 1.0 From: Andy Schwartz To: jsr-314-open-mirror at lists.jboss.org Subject: Re: [jsr-314-open-mirror] [jsr-314-open] ANNOUNCE: JSF 2.1 Spec, Release Candidate Five Date: Fri, 05 Nov 2010 14:53:48 -0400 Message-ID: <4CD452BC.2050800@oracle.com> In-Reply-To: 19668.19873.81969.743842@gargle.gargle.HOWL --===============3812196750539280602== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On 11/5/10 2:32 PM, Ed Burns wrote: >>>>>> On Thu, 04 Nov 2010 23:33:29 -0700, Ed Burns said: >>>>>> = > > The unprefixed namespace, also known as the root namespace, must be > passed through without modification or check for validity. The > passing through of the root namespace must occur on any non-prefixed > element in a facelet page. Sounds good. (Sounds like we are just specifying behavior that Faclets = has always implemented, right?) BTW, I see that we still have and components in the = latest spec. I haven't seen an explanation for why this are = necessary/useful. When I asked about this, you mentioned that = was useful as a resource target. However, it is not. From our earlier = thread: > I see. Unfortunately this approach is flawed. The element may = > only contain and elements. As such, "html" is not a = > valid target for resource relocation. We need to remove this new = > requirement from the 2.1 specification. > > If cannot serve as a resource target and if is a valid = > element in XML processing mode, I am still unclear on the value of = > and wonder whether this should be included in the 2.1 = > specification. > > On a related note, what's up with ? :-) Can we kill this = > off? Or is there some reason why folks would need to use this instead = > of ? = I don't remember seeing a response to this, though with all of the = emails lately, I may have missed it. I still do not understand why our users would use <h:html> and <h:title> = over <html> and <title>. If we don't have a strong reason for why these = components are useful, we should remove these from the specification in = order to avoid spec/component bloat. Andy --===============3812196750539280602==-- From blake.sullivan at oracle.com Fri Nov 5 19:38:59 2010 Content-Type: multipart/mixed; boundary="===============5382568104137529319==" MIME-Version: 1.0 From: Blake Sullivan <blake.sullivan at oracle.com> To: jsr-314-open-mirror at lists.jboss.org Subject: Re: [jsr-314-open-mirror] [jsr-314-open] ANNOUNCE: JSF 2.1 Spec, Release Candidate Five Date: Fri, 05 Nov 2010 16:37:55 -0700 Message-ID: <4CD49553.6010103@oracle.com> In-Reply-To: 4CD452BC.2050800@oracle.com --===============5382568104137529319== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Just to chime in further on the continued presence of <h:html> and = <h:title>, as Andy said, in addition to just pointlessly bloating the = specification, having <h:html> act as a resource relocation target seems = just plain wrong. -- Blake Sullivan On 11/5/10 11:53 AM, Andy Schwartz wrote: > On 11/5/10 2:32 PM, Ed Burns wrote: >>>>>>> On Thu, 04 Nov 2010 23:33:29 -0700, Ed Burns = >>>>>>> <edward.burns(a)oracle.com> said: >> >> The unprefixed namespace, also known as the root namespace, must be >> passed through without modification or check for validity. The >> passing through of the root namespace must occur on any non-prefixed >> element in a facelet page. > > Sounds good. (Sounds like we are just specifying behavior that = > Faclets has always implemented, right?) > > BTW, I see that we still have <h:html> and <h:title> components in the = > latest spec. I haven't seen an explanation for why this are = > necessary/useful. When I asked about this, you mentioned that = > <h:html> was useful as a resource target. However, it is not. From = > our earlier thread: > >> I see. Unfortunately this approach is flawed. The <html> element = >> may only contain <head> and <body> elements. As such, "html" is not = >> a valid target for resource relocation. We need to remove this new = >> requirement from the 2.1 specification. >> >> If <h:html> cannot serve as a resource target and if <html> is a = >> valid element in XML processing mode, I am still unclear on the value = >> of <h:html> and wonder whether this should be included in the 2.1 = >> specification. >> >> On a related note, what's up with <h:title>? :-) Can we kill this = >> off? Or is there some reason why folks would need to use this = >> instead of <title>? = > > I don't remember seeing a response to this, though with all of the = > emails lately, I may have missed it. > > I still do not understand why our users would use <h:html> and = > <h:title> over <html> and <title>. If we don't have a strong reason = > for why these components are useful, we should remove these from the = > specification in order to avoid spec/component bloat. > > Andy > > --===============5382568104137529319==--