- From: Erik Bruchez <ebruchez@orbeon.com>
- Date: Sun, 20 Jun 2010 17:22:33 -0700
- To: Forms WG <public-forms@w3.org>
- Message-ID: <AANLkTil9aIZ1OintLo_-x_y7uPe0rbGpafXr39PLvGR1@mail.gmail.com>
I think that context() with id is useful and a good shortcut.
Note that variables also solve (among other things) this issue, although in
a slightly more verbose way, e.g.:
<xf:repeat id="status-repeat"
nodeset="instance('statuses')/statuses/status">
<xf:variable name="status-repeat" value="."/>
...
<xf:repeat id="set-status-repeat"
nodeset="instance('statuses')/statuses/status[$status-repeat/status != .]">
-Erik
On Fri, Jun 18, 2010 at 2:46 PM, Klotz, Leigh <Leigh.Klotz@xerox.com> wrote:
> OK, I put in a proposal in the wiki with the hotels use case:
>
> http://www.w3.org/MarkUp/Forms/wiki/@context_id
>
> Leigh.
>
> ------------------------------
> *From:* John Boyer [mailto:boyerj@ca.ibm.com]
> *Sent:* Friday, June 18, 2010 2:28 PM
>
> *To:* Klotz, Leigh
> *Cc:* public-forms@w3.org
> *Subject:* RE: @context everywhere; applies to ref and value separately,
> or applies before ref, then ref applies to value?
>
>
> Hi Leigh,
>
> Yes, I agree that context() would need to be more powerful, but frankly I
> thought we eventually would need to do that.
>
> So, once we decide there is a real use case, maybe we should discuss how an
> additional optional parameter could be added to context() to allow it to
> obtain the inscope eval context of an element other than the default of the
> immediate ancestor element.
>
> Cheers,
> John M. Boyer, Ph.D.
> IBM Distinguished Engineer, Interactive Web Applications and Documents
> IBM Lotus Forms Architect
> Workplace, Portal and Collaboration Software
> IBM Canada Software Lab, Victoria
> E-Mail: boyerj@ca.ibm.com
>
> Blog: http://www.ibm.com/developerworks/blogs/page/JohnBoyer
> Blog RSS feed:
> http://www.ibm.com/developerworks/blogs/rss/JohnBoyer?flavor=rssdw
>
>
>
>
> From: "Klotz, Leigh" <Leigh.Klotz@xerox.com> To: John
> Boyer/CanWest/IBM@IBMCA Cc: <public-forms@w3.org> Date: 06/18/2010 01:56
> PM Subject: RE: @context everywhere; applies to ref and value separately,
> or applies before ref, then ref applies to value?
> ------------------------------
>
>
>
> Thanks for clarifying the context on child elements. That's a relief.
>
> I believe you're right about the expressive power of repeat nodesets and
> index() being up to the task of the context() function.
>
> The problem I encountered was when the inner repeat used, as a predicate on
> its nodeset, a value from the outer repeat.
> It does turn out to be solvable with index, but it's not pretty.
>
> In fact, I believe that when you get to triply-nested repeats, you have to
> go back to copying repeat-nodesets and index, and cannot use context(),
> because index() takes an ID argument but context() does not, so there's no
> way to get back to the grandparent.
>
>
> Here's the context example with two nested repeats, using context:
>
> <xf:repeat id="statuses-repeat"
> nodeset="instance('statuses')/statuses/status">
> <xf:repeat id="hotels-repeat"
> nodeset="instance('hotels')/hotel[status=current()]">
> <xf:trigger>
> <xf:label>Mark Vacant</xf:label>
> <xf:action ev:event="DOMActivate">
> <xf:setvalue ref="instance('mark')/status">vacant</xf:setvalue>
> <xf:setvalue ref="instance('mark')/name"
> value="context()/name" />
> <xf:send submission="mark-status" />
> </xf:action>
> </xf:trigger>
> </xf:repeat>
> </xf:repeat>
>
> and here it is without context(), using index(), position(), and
> duplicating the nodesets of the two repeats, including their predicates:
>
> <xf:repeat id="statuses-repeat"
> nodeset="instance('statuses')/statuses/status">
> <xf:repeat id="hotels-repeat"
> nodeset="instance('hotels')/hotel[status=current()]">
> <xf:trigger>
> <xf:label>Mark Vacant</xf:label>
> <xf:action ev:event="DOMActivate">
> <xf:setvalue ref="instance('mark')/status">vacant</xf:setvalue>
> <xf:setvalue ref="instance('mark')/name"
> value="instance('hotels')/hotel[status=instance('statuses')/statuses/status[index('statuses-repeat')]
> and position()=index('hotels-repeat')]/name" />
> <xf:send submission="mark-status" />
> </xf:action>
> </xf:trigger>
> </xf:repeat>
> </xf:repeat>
>
> Now let's add a set of triggers, one for each status, here done with index:
>
> <xf:repeat id="statuses-repeat"
> nodeset="instance('statuses')/statuses/status">
> <xf:repeat id="hotels-repeat"
> nodeset="instance('hotels')/hotel[status=current()]">
> <xf:repeat id="set-status-repeat"
> nodeset="instance('statuses')/statuses/status">
> <xf:trigger>
> <xf:label>Mark <output value="." /></xf:label>
> <xf:action ev:event="DOMActivate">
> <xf:setvalue ref="instance('mark')/status" value="
> instance('status')/status[index('set-status-repeat')]" />
> <xf:setvalue ref="instance('mark')/name"
> value="instance('hotels')/hotel[status=instance('statuses')/statuses/status[index('statuses-repeat')]
> and position()=index('hotels-repeat')]/name" />
> <xf:send submission="mark-status" />
> </xf:action>
> </xf:trigger>
> </xf:repeat>
> </xf:repeat>
> </xf:repeat>
>
> I don't see that we can handle this with context() though. We can do this
> line:
>
> <xf:setvalue ref="instance('mark')/status" value="context()" />
> but I don't see how the next setvalue can profit from context().
> <xf:setvalue ref="instance('mark')/name"
> value="???" />
>
> Leigh.
>
>
>
>
>
>
>
>
>
>
> ------------------------------
> *From:* John Boyer [mailto:boyerj@ca.ibm.com <boyerj@ca.ibm.com>] *
> Sent:* Friday, June 18, 2010 11:49 AM*
> To:* Klotz, Leigh*
> Cc:* public-forms@w3.org*
> Subject:* RE: @context everywhere; applies to ref and value separately, or
> applies before ref, then ref applies to value?
>
>
> Hi Leigh,
>
> Every element has an in-scope evaluation context, including those child
> elements you talked about [1]. See especially the text after the list,
> which indicate that elements that are not binding elements have an in-scope
> evaluation context equivalent to what they would have if they were binding
> elements.
>
> [1] *http://www.w3.org/TR/xforms11/#expr-eval*<http://www.w3.org/TR/xforms11/#expr-eval>
>
> So, yes you actually should be able to use context() in
> output/mediatype/@value.
>
> Although context() is the "right" answer, can you still let me know whether
> or not you see a specific problem with using index() in that case you raised
> earlier, when further amended to be an inner repeat? I'd like to assume it
> is actually not a problem unless you say otherwise.
>
> Cheers,
> John M. Boyer, Ph.D.
> IBM Distinguished Engineer, Interactive Web Applications and Documents
> IBM Lotus Forms Architect
> Workplace, Portal and Collaboration Software
> IBM Canada Software Lab, Victoria
> E-Mail: boyerj@ca.ibm.com
>
> Blog: *http://www.ibm.com/developerworks/blogs/page/JohnBoyer*<http://www.ibm.com/developerworks/blogs/page/JohnBoyer>
> Blog RSS feed: *
> http://www.ibm.com/developerworks/blogs/rss/JohnBoyer?flavor=rssdw*<http://www.ibm.com/developerworks/blogs/rss/JohnBoyer?flavor=rssdw>
>
>
>
> From: "Klotz, Leigh" <Leigh.Klotz@xerox.com> To: John
> Boyer/CanWest/IBM@IBMCA Cc: <public-forms@w3.org> Date: 06/18/2010 11:24
> AM Subject: RE: @context everywhere; applies to ref and value separately,
> or applies before ref, then ref applies to value?
>
> ------------------------------
>
>
>
> Thank you, John. That's great news. I had gotten confused and thought
> context was an XForms 1.2 feature; so @context everywhere is, but context()
> isn't.
> I'll note this in the wikipage on @context everywhere.
>
> I have another question about context(), though, and it would have applied
> equally to my misunderstood @context:
>
> #fn-context says that it applies to "context node of the nearest ancestor
> element". So, that means it isn't useful for child elemements which allow
> evaluation of xpath expressions for what would otherwise be static attribute
> values.
> For example, *output/mediatype/@value* <output/mediatype/@value> cannot
> use context().
>
> Leigh.
>
>
> ------------------------------
> *
> From:* John Boyer [*mailto:boyerj@ca.ibm.com* <boyerj@ca.ibm.com>] *
> Sent:* Friday, June 18, 2010 7:19 AM*
> To:* Klotz, Leigh*
> Cc:* public-forms@w3.org*
> Subject:* Re: @context everywhere; applies to ref and value separately, or
> applies before ref, then ref applies to value?
>
>
> Hi Leigh,
>
> First, to clarify, you're trying to get the value attr in setvalue to
> evaluate "name" as the name of the hotel that is both good and cheap, i.e.
> the value expression evaluates relative to the repeat item node rather than
> the result of the ref in setvalue.
>
> While the ref sets the starting node of the value, the repeat item node
> happens to be the "inscope evaluation context node" of the setvalue action,
> which sets the starting node of the ref.
>
> Therefore, no, the context attribute will not help you with resetting the
> evaluation node for the value attr. The context attribute allows override
> of the inscope evaluation context of the element containing the context
> attribute, so it would only help you to set a context for the ref of the
> setvalue.
>
> Unfortunately, Mark's blog post is also incorrect, and you are correct,
> that using the current() function does not work. Later in Mark's post, he
> does acknowledge that.
>
> The actual fix to the problem is the context() function [1], and you're
> hitting on the exact reason it was added.
>
> [1] *http://www.w3.org/TR/xforms11/#fn-context*<http://www.w3.org/TR/xforms11/#fn-context>
>
> In your markup, use this:
> <setvalue ev:event="DOMActivate"
> ref="instance('home')/hang/your/hat" value="*context()/*name" />
>
> If Mark's post contained context() where it currently contains current(),
> then his blog post would work without having to resort to the index()
> function:
>
> <xf:setvalue bind="testBind" value="context()/VO_ID" />
>
> Finally, note that context() is one of the harder functions to implement,
> so you may have to file an issue with your flavourite XForms implementer if
> the above doesn't work. And meanwhile, could you explain further what
> breaks when you use the alternative method of the index() function to solve
> this problem in an inner repeat? Offhand, the only reason I can think of
> why it wouldn't work is that the harder part of implementing the index()
> function has not been done properly, which would mean another issue for your
> flavourite XForms implementer.
>
> Cheers,
> John M. Boyer, Ph.D.
> IBM Distinguished Engineer, Interactive Web Applications and Documents
> IBM Lotus Forms Architect
> Workplace, Portal and Collaboration Software
> IBM Canada Software Lab, Victoria
> E-Mail: boyerj@ca.ibm.com
>
> Blog: *http://www.ibm.com/developerworks/blogs/page/JohnBoyer*<http://www.ibm.com/developerworks/blogs/page/JohnBoyer>
> Blog RSS feed: *
> http://www.ibm.com/developerworks/blogs/rss/JohnBoyer?flavor=rssdw*<http://www.ibm.com/developerworks/blogs/rss/JohnBoyer?flavor=rssdw>
>
>
> From: "Leigh L. Klotz, Jr." <Leigh.Klotz@Xerox.com> To:
> public-forms@w3.org Date: 06/17/2010 11:37 PM Subject: @context
> everywhere; applies to ref and value separately, or applies before ref,
> then ref applies to value?
>
>
> ------------------------------
>
>
>
> When we add @context to actions [1], would this work to deal with the
> lost repeat context in setvalue problem?
>
> <repeat nodeset="instance('directory')/hotels/hotel[@good='yes' and
> @cheap='yes']">
> <trigger>
> <label>Book <output ref="name" /></label>
> <setvalue ev:event="DOMActivate" context="."
> ref="instance('home')/hang/your/hat" value="name" />
> </trigger>
> </repeat>
>
> (You can work around this loss of context with index(), but not in the
> case of nested repeats.)
>
> Interestingly, Mark Birbeck suggests [2] that current() solves this
> problem, but my understanding was that
> <setvalue ref="instance('home')/hang/your/hat" value="current()/name" />
> was the same as
> <setvalue ref="instance('home')/hang/your/hat" value="./name" />
> and current() would be useful only inside predicates.
>
> Am I wrong on this count?
>
> [1] *http://www.w3.org/MarkUp/Forms/wiki/@context_everywhere*<http://www.w3.org/MarkUp/Forms/wiki/@context_everywhere>
> [2] *http://www.formsplayer.com/node/200*<http://www.formsplayer.com/node/200>
>
> Leigh.
>
>
>
>
>
>
>
Received on Monday, 21 June 2010 00:23:26 UTC